Microeconomics, Competition and Strategic Behaviour
Strategy and Decision-Making in Markets
0905
2022
978-3-8385-5908-7
978-3-8252-5908-2
UTB
Markus Thomas Münter
10.36198/9783838559087
Microeconomics with case studies and applications for strategic management and consulting
Microeconomics is not applied math - this textbook explains microeconomic frameworks used by managers in strategic decision-making. Numerous case studies cover price discrimination, economies of scale, digital business models, game theory, dealing with uncertainty, entry barriers or sunk costs - all of which are crucial for understanding market dynamics and competitive behaviour. Frameworks and concepts are highlighted that have strong connections to strategic management.
Content:
1 Microeconomics, competition and strategic behaviour
2 Customer behaviour, products and network effects
3 Decisions under risk and behavioural economics
4 Firms, competition and innovation
5 Firm size, technology and decisions on production
6 Costs, restructuring and M&A
7 Perfect competition, monopoly and competition policy
8 Pricing strategies and price discrimination
9 Strategic decision-making and game theory
10 Strategic competition in oligopoly
<?page no="0"?> Markus Thomas Münter Microeconomics, Competition and Strategic Behaviour <?page no="1"?> utb 5908 Eine Arbeitsgemeinschaft der Verlage Brill | Schöningh - Fink · Paderborn Brill | Vandenhoeck & Ruprecht · Göttingen - Böhlau · Wien · Köln Verlag Barbara Budrich · Opladen · Toronto facultas · Wien Haupt Verlag · Bern Verlag Julius Klinkhardt · Bad Heilbrunn Mohr Siebeck · Tübingen Narr Francke Attempto Verlag - expert verlag · Tübingen Psychiatrie Verlag · Köln Ernst Reinhardt Verlag · München transcript Verlag · Bielefeld Verlag Eugen Ulmer · Stuttgart UVK Verlag · München Waxmann · Münster · New York wbv Publikation · Bielefeld Wochenschau Verlag · Frankfurt am Main <?page no="2"?> Markus Thomas Münter is Professor of Economics at htw saar, Germany. He has a more than 15-years background in management consulting and banking. Markus was a Visiting Assistant Professor at London Business School and a Visiting Professor at La Trobe University in Melbourne. His research is on dynamics of markets, technology and corporate strategy. <?page no="3"?> Markus Thomas Münter Microeconomics, Competition and Strategic Behaviour Strategy and Decision-Making in Markets UVK Verlag · München <?page no="4"?> Umschlagabbildung: © Wavebreakmedia · iStock Autorenfoto (S. 2): © privat Bibliografische Information der Deutschen Nationalbibliothek Die Deutsche Nationalbibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet über http: / / dnb.dnb.de abrufbar. 1. Auflage 2022 DOI: https: / / doi.org/ 10.36198/ 9783838559087 © UVK Verlag 2022 - ein Unternehmen der Narr Francke Attempto Verlag GmbH + Co. KG Dischingerweg 5 · D-72070 Tübingen Das Werk einschließlich aller seiner Teile ist urheberrechtlich geschützt. Jede Verwertung außerhalb der engen Grenzen des Urheberrechtsgesetzes ist ohne Zustimmung des Verlags unzulässig und strafbar. Das gilt insbesondere für Vervielfältigungen, Übersetzungen, Mikroverfilmungen und die Einspeicherung und Verarbeitung in elektronischen Systemen. Alle Informationen in diesem Buch wurden mit großer Sorgfalt erstellt. Fehler können dennoch nicht völlig ausgeschlossen werden. Weder Verlag noch Autor: innen oder Herausgeber: innen übernehmen deshalb eine Gewährleistung für die Korrektheit des Inhaltes und haften nicht für fehlerhafte Angaben und deren Folgen. Diese Publikation enthält gegebenenfalls Links zu externen Inhalten Dritter, auf die weder Verlag noch Autor: innen oder Herausgeber: innen Einfluss haben. Für die Inhalte der verlinkten Seiten sind stets die jeweiligen Anbieter oder Betreibenden der Seiten verantwortlich. Internet: www.narr.de eMail: info@narr.de Einbandgestaltung: Atelier Reichert, Stuttgart CPI books GmbH, Leck utb-Nr. 5908 ISBN 978-3-8252-5908-2 (Print) ISBN 978-3-8385-5908-7 (ePDF) <?page no="5"?> 5 Preface When I am talking to C-level managers, with a background in economics or business, about what they remember from their microeconomics classes at university, mostly I get the response “hmm … not much … some maths, maybe”. If I ask them, what their most pressing business issues are, they typically mention the competitive behaviour of rivals, whether markets go up or down, and how to find a profitable new business model in the context of digitalisation and platforms. Well - it’s exactly these topics that microeconomics is about. Why is there such a strong mis-match? Complaints about unrealistic teaching, lack of applicability of methods, and the allegedly obvious failure of economists in connection with the financial and debt crisis of the past decade, are almost commonplace. Many students also complain about a too high level of abstraction, unrealistic assumptions regarding rationality of decision-makers, and an excessive use of mathematics. Sometimes there are even requests that professorships in economics be reallocated to more applied subjects. However, applied microeconomics, as it is taught in international business schools today, has little in common with the microeconomics of the 1980s or 1990s, which was sometimes reduced to applied mathematics. Essentially, the changes can be summarised around three pillars for successful teaching: a sound matching of empirical evidence and theory, integration of findings from behavioural sciences, focus on managerial methods for strategy development in organisations. In this respect, the always overused and appropriately scolded homo oeconomicus now resembles Schrödinger's cat: don't open the box, or it might be dead. In the meantime, we have, also rewarded by Nobel Prizes to Herbert Simon, Reinhard Selten, Daniel Kahneman and Richard H. Thaler, some robust findings about situations in which people decide almost rationally, and when and especially why they deviate in other situations. Lecturers at many universities already teach applied microeconomics explaining key empirical observations in industryor firm-specific case studies against the backdrop of digitalisation and global competition. Microeconomics also provides significant insights for competitive strategy and the development of new business models. Already in 2012, the Economist titled this development "A Golden Age of Micro" highlighting that Amazon, Google, Facebook and ebay all employ leading academic microeconomists in their strategy departments. It is in this context that Microeconomics, Competition and Strategic Behaviour emerged from my lectures in Saarbrücken. Students are not afraid of models, if these are derived from their perceived reality, coherent and direct. Empirical research, the use of real firm data and case studies based on current developments create a sound basis for this. To make this point even clearer: two-sided markets in digital business models appear somewhat more important than a supposed Giffen-good somewhere in Scotland in the 18th century, if microeconomics is to lay a foundation for students’ future as decision-makers in firms. For microeconomics to have a raison d'être in the education of busi- <?page no="6"?> Preface 6 ness students at universities of applied sciences, it is not enough to impart an abstract view of economic contexts or merely to create the foundations for business courses such as marketing, management or finance. Rather, teaching must also show which microeconomic methods are actually used in everyday business practice. If these concepts are understood as generally applicable tools, then microeconomics, which is often feared and presumably far from reality, becomes a Swiss army knife that students can use immediately - especially when entering strategy departments or management consultancies. Stuff that is in this book - and things that did not make it There are truly fantastic textbooks on microeconomics available - why is this book different? I do have, and this might be an exception, two hearts beating as one: I am an economist, and I have more than 15 years background in management and consulting, mainly in the European financial services industry. I had been working on strategy and transformation projects, on M&A deals, business development programmes, and restructuring cases - and of course doing daily business with colleagues from operations, finance, IT, HR, product management or even sales and marketing. From this experience I have some idea, what people in firms should know about microeconomics, and I know what they (or at least some of them) actually do not know. This book tries to address this gap. The target audience of this book are students in their first semesters of Bachelor's degree programmes in business administration whose possible careers are management roles in firms or an entry into management consulting. Hence, the focus of this textbook is on microeconomic concepts relevant from a management perspective that are used everyday in strategic decisionmaking. Concepts include price discrimination, economies of scale, game theory, dealing with uncertainty, entry barriers or sunk costs - all of which are necessary for understanding market structures, innovations and competitive behaviour. The focus is on concepts that have strong connections to business administration modules and are further deepened in Masters programmes in modules such as industrial organisation, managerial economics or competition policy. Of course, mathematics is used - one of the crucial questions 'in real life' is almost always: “does it pay off? ”. In addition, with ever more data available in firms, models provide some guidance on how to use this data to make intelligent decisions. Conversely, many areas of microeconomics are left out of consideration, which I have never seen or ran into them in more than 15 years of strategy consulting and management or which have no close connection to business administration and management decisions. Moreover, some obviously essential topics - management of state-owned firms, natural resources or labour markets - are left out. In most curricula, this content is covered in courses on economic policy. In addition, topics are left out that I may find exciting, but which, when viewed in the light of breadth and depth, are only essential for students of economics on their way to an academic career. Excellent additional textbooks are listed in the appendix to ► Chapter 1. Thank you I would like to thank my students. Through discussions in the lectures, but especially during collaboration in projects and theses, I have hopefully also been able to gain a little more clarity. My special thanks go to Franziska Müller of Verbraucherzentrale Brandenburg, Georg Sobbe of <?page no="7"?> Preface 7 Bundesverband der Deutschen Musikindustrie, Steffen Häfele of Bundeskartellamt, and Rainer Berger for once again providing great support and supervision. However, this book would not have been possible without the help of some friends: Fabiane Mihut-Albeck did an enormous job putting together bits and pieces from hundreds of slides for the first German edition of this book and creating the index; Anahita Amirsadari commented on a first draft and made my Aussie-style English a lot more readable and understandable; and Caroline Wilson and Nika Hein reviewed the whole thing in great detail. Thank you so much for being companions in this journey. To my wife Cecile for making this life possible. Karlsruhe and Saarbrücken, Summer 2022 Markus Thomas Münter <?page no="9"?> 9 Content Preface .............................................................................................................................................5 Content ............................................................................................................................................9 List of abbreviations .................................................................................................................... 12 List of mathematical and economic variables ......................................................................... 13 1 Microeconomics, competition and strategic behaviour ....................................... 17 1.1 Microeconomics between empirical research, theory and experiments ........................ 18 1.2 Markets, demand and supply ........................................................................................................... 25 1.3 Price elasticity and marginal revenues ......................................................................................... 38 1.4 Summary and key learnings ............................................................................................................. 48 Recommendations for further reading ....................................................................................................... 48 Questions for review .......................................................................................................................................... 49 Literature ................................................................................................................................................................. 49 2 Customer behaviour, products and network effects ............................................. 53 2.1 Customer behaviour and decisions on demand ....................................................................... 54 2.2 Market delineation and product categories ............................................................................... 66 2.3 Network effects and multisided markets ..................................................................................... 83 2.4 Summary and key learnings ............................................................................................................. 94 Recommendations for further reading ....................................................................................................... 94 Questions for review .......................................................................................................................................... 95 Literature ................................................................................................................................................................. 95 3 Decisions under risk and behavioural economics .................................................101 3.1 Decisions under risk and uncertainty ..........................................................................................102 3.2 Bounded rationality and behavioural economics ..................................................................116 3.3 Summary and key learnings ...........................................................................................................131 Recommendations for further reading .....................................................................................................131 Questions for review ........................................................................................................................................132 Literature ...............................................................................................................................................................133 4 Firms, competition and innovation ........................................................................ 139 4.1 Firms, objectives and strategies ....................................................................................................140 4.2 Competitive advantage, market structure and firm-specific capabilities .....................151 4.3 Competition and innovation ..........................................................................................................165 4.4 Summary and key learnings ...........................................................................................................179 <?page no="10"?> Content 10 Recommendations for further reading ..................................................................................................... 180 Questions for review ........................................................................................................................................ 180 Literature ............................................................................................................................................................... 181 5 Firm size, technology and decisions on production .............................................187 5.1 Production function and technology .......................................................................................... 189 5.2 Short run decisions: diminishing marginal product and productivity ........................... 194 5.3 Long run decisions: technical progress and returns to scale ............................................ 201 5.4 Summary and key learnings ........................................................................................................... 211 Recommendations for further reading ..................................................................................................... 212 Questions for review ........................................................................................................................................ 212 Literature ............................................................................................................................................................... 213 6 Costs, restructuring and M&A .................................................................................215 6.1 Cost functions, decisions and competitiveness ...................................................................... 216 6.2 Short run decisions: Fixed costs and marginal costs ............................................................ 220 6.3 Long run decisions: adjusting cost structures ......................................................................... 225 6.4 Cost-based competitive advantage and M&A ....................................................................... 235 6.5 Summary and key learnings ........................................................................................................... 245 Recommendations for further reading ..................................................................................................... 246 Questions for review ........................................................................................................................................ 246 Literature ............................................................................................................................................................... 247 7 Perfect competition, monopoly and competition policy ................................... 249 7.1 Decisions of a firm under perfect competition ....................................................................... 250 7.2 Producer and consumer surplus as a measure of economic welfare ............................ 258 7.3 Monopoly and dominant firms ..................................................................................................... 263 7.4 Restraints of competition, competition policy and competition authorities .............. 271 7.5 Summary and key learnings ........................................................................................................... 287 Recommendations for further reading ..................................................................................................... 287 Questions for review ........................................................................................................................................ 288 Literature ............................................................................................................................................................... 288 8 Pricing strategies and price discrimination .......................................................... 293 8.1 Forms and requirements of price discrimination ................................................................... 294 8.2 Direct price discrimination and market segmentation ........................................................ 296 8.3 Indirect price discrimination and two-part tariffs .................................................................. 302 8.4 Bundling ................................................................................................................................................. 310 8.5 Auctions .................................................................................................................................................. 316 <?page no="11"?> Content 11 8.6 Summary and key learnings ...........................................................................................................321 Recommendations for further reading .....................................................................................................322 Questions for review ........................................................................................................................................322 Literature ................................................................................................................................................................323 9 Strategic decision-making and game theory ........................................................ 327 9.1 Nash equilibria in simultaneous games .....................................................................................330 9.2 Risk aversion and mixed strategies ..............................................................................................342 9.3 Sequential decisions and commitment ......................................................................................348 9.4 Summary and key learnings ...........................................................................................................354 Recommendations for further reading .....................................................................................................355 Questions for review ........................................................................................................................................355 Literature ...............................................................................................................................................................356 10 Strategic competition in oligopoly ......................................................................... 359 10.1 Capacity decisions and strategies in Cournot competition ...............................................362 10.2 Sequential decisions and strategies under Stackelberg competition ............................374 10.3 Pricing decisions and strategies in Bertrand competition ..................................................378 10.4 Strategic competition with product differentiation ..............................................................385 10.5 Relevance for competitive strategies ..........................................................................................394 10.6 Summary and key learnings ...........................................................................................................397 Recommendations for further reading .....................................................................................................398 Questions for review ........................................................................................................................................398 Literature ...............................................................................................................................................................400 Index ............................................................................................................................................ 403 <?page no="12"?> 12 List of abbreviations B2B business to business B2C business to consumer BPO business process outsourcing CAGR compound annual growth rate e.g. for example EBIT earnings before interest and taxes FTE full time equivalents GWB Gesetz gegen Wettbewerbsbeschränkung i.e. that is to say M&A mergers and acquisitions P&L profit and loss statement p.a. per annum PEST political-legal, economic, social and technological environment R&D research and development RoE return on equity ROIC return on invested capital SCP structure-conduct-performance SIEC significant impediment to effective competition SSNIP small but significant non-transitory increase in price SWOT strength-weaknesses and opportunities-threats analysis TK transaction costs <?page no="13"?> 13 List of mathematical and economic variables 𝟏𝟏/ 𝒃𝒃 size of a market 𝒂𝒂 maximum willigness to pay 𝑨𝑨 technological efficiency 𝑨𝑨𝑨𝑨 productivity 𝑨𝑨𝑨𝑨𝑨𝑨 average total costs 𝑨𝑨𝑨𝑨𝑨𝑨 average variable costs 𝒃𝒃 slope of demand function 𝑨𝑨𝑪𝑪 consumer surplus 𝒅𝒅 total differential 𝑫𝑫 demand 𝑫𝑫𝑫𝑫𝑫𝑫 deadweight loss 𝒆𝒆 𝑨𝑨𝑨𝑨 elasticity of total costs 𝑬𝑬 price elasticity of one market side given network effects 𝑬𝑬𝑬𝑬 equity 𝑬𝑬𝑬𝑬 expected utility 𝑬𝑬𝑨𝑨 expected value 𝑭𝑭𝑨𝑨 fixed costs 𝑭𝑭𝑬𝑬 debt 𝑴𝑴𝑴𝑴𝑪𝑪 marginal rate of substitution 𝑴𝑴𝑴𝑴𝑨𝑨 marginal rate of technological substitution 𝑰𝑰 income 𝑬𝑬 capital, amount of capital input 𝑫𝑫 labour, number of workers 𝑫𝑫 𝒊𝒊 Lerner-index 𝑴𝑴𝑨𝑨 marginal costs 𝑴𝑴𝑬𝑬𝑪𝑪 minimimum efficient size 𝑴𝑴𝑨𝑨 marginal product 𝑴𝑴𝑴𝑴 marginal revenues <?page no="14"?> 14 𝒏𝒏 number of firms 𝒑𝒑 price 𝑨𝑨𝑨𝑨𝑴𝑴 price-cost-margin 𝒑𝒑𝒓𝒓 probability of an event 𝑨𝑨𝑪𝑪 producer surplus 𝒒𝒒 quantity, production capacity of a firm 𝒒𝒒 𝑫𝑫 quantity demanded 𝒒𝒒 𝑪𝑪 quantity supplied 𝑸𝑸 total production, capacity of all firms 𝒓𝒓 interest rate, discount rate 𝒓𝒓 𝑫𝑫 interest rate of debt, borrowing rate 𝒓𝒓 𝑪𝑪𝑺𝑺 interest rate of equity, profit expectations of shareholders 𝑴𝑴 revenues 𝑴𝑴 𝟐𝟐 coefficient of determination 𝒔𝒔 𝒊𝒊 market share of firm i 𝑪𝑪 supply 𝑪𝑪𝑨𝑨 sunk costs 𝒕𝒕 time 𝑨𝑨 technology, technological path 𝑨𝑨𝑨𝑨 total costs 𝒖𝒖 utility 𝒗𝒗 value 𝑨𝑨 firm value 𝑨𝑨𝑨𝑨 variable costs 𝒘𝒘 wage rate (hourly-, monthlyor annual-salary) −𝒘𝒘/ 𝒓𝒓 wage-/ interest-rate relation, slope of isocost line 𝑫𝑫 wealth 𝑫𝑫 𝑴𝑴 risk-weighted expected value of wealth 𝑫𝑫 𝑪𝑪 certainty-equivalent of wealth 𝑫𝑫𝑨𝑨𝑨𝑨𝑨𝑨 weighted average cost of capital List of mathematical and economic variables <?page no="15"?> 15 𝒛𝒛 𝒊𝒊 𝒁𝒁 List of mathematical and economic variables willigness to pay (reservation prices) Lagrange-function Greek variables 𝜶𝜶 partial production elasticity of capital 𝜷𝜷 partial production elasticity of labour 𝜸𝜸 industry-specific degree of horizontal product differentiation 𝝏𝝏 partial derivative ∆ absolute difference 𝜺𝜺 𝑰𝑰 income elasticity of demand 𝜺𝜺 𝑿𝑿𝑿𝑿 cross-price elasticity of demand 𝜺𝜺 𝒑𝒑 price elasticity of demand 𝜽𝜽 strenght of direct network effect 𝝀𝝀 Lagrange-mulitiplier 𝝁𝝁 degree of loss aversion 𝝅𝝅 profits 𝝈𝝈 𝟐𝟐 variance of a distribution 𝝎𝝎 degree of risk aversion <?page no="17"?> 17 1 Microeconomics, competition and strategic behaviour Making business decisions is a little like cooking - it somehow works, even if you are not good at it, but usually not very well. As a beginner, you watch others cooking first. Sometimes you leaf through a cookbook, check out the back of a packet soup or use an app. But most of the time you simply use the ingredients that you already have in your kitchen. Cookbooks usually are not scientific textbooks. Cooking instructions describe, depending on the level of difficulty, a step-by-step procedure for heating up a pre-cooked tomato soup, or perhaps even preparing more elaborate dishes (Kolmar 2017 and Barham 2001). If you follow the instructions on the back of the package or in a cookbook, you will end up having a decent meal on the table afterwards - although you rarely understand why. And as pretty much everyone has to arrange food or even cook somehow, even if they cannot do it properly, similarly, managers have to make decisions. Of course, you can get your food delivered, (thanks to Deliveroo, Just Eat or Uber Eats) and managers in firms can hire McKinsey & Company, Bain & Company or BCG for delivering strategy advice and decision support, but in the long run this could easily cost an arm and a leg. In this sense, this book on microeconomics is not a cookbook: microeconomics does not attempt to provide step-by-step instructions for managerial action or decisions. Instead, microeconomics is focused on explaining observable behaviour and decisions of people in economic situations in order to show effects on markets, firms and competition. Nevertheless, the observations naturally provide guidance for future decisions and a classification of competitive situations, especially from a management perspective. Any microeconomic analysis has at least three dimensions: describe and explain decisions made by customers in different situations and circumstances - e.g. when buying a smartphone; derive and explain strategic decisions of managers in firms - e.g. setting production capacities or prices for the next smartphone generation; understand and quantify the interplay of decisions, i.e., the interaction between firms and customers in markets and potential effects on prices, quantities, market structure or profits and the fate of firms - e.g. the decline of Nokia and Research in Motion and at the same time the enormous success of Samsung and the profitability of Apple. Microeconomics analyses decisions of customers and firms, their interactions and how markets work. The need for decisions is always due to scarcity: people have to choose between alternatives because resources, money or time are limited. Microeconomics can provide important support for at least two target groups: on the one hand, firms that want to develop strategies for various competitive situations, on the other hand, governmental institutions such as competition authorities that aim to improve the functioning of markets. In contrast, macroeconomics looks at the whole economy and tries to provide explanations for unemployment, inflation, business cycles and growth. <?page no="18"?> 1 Microeconomics, competition and strategic behaviour 18 Learning Objectives This chapter deals with: topics of microeconomics from a management perspective, connection of theory and empirical research to derive 'maps for markets and competition', the basic relationship between supply and demand and how to identify a market equilibrium, and price elasticity of demand and marginal revenues as measures of effects of price changes on quantity demanded and revenues of a firm. 1.1 Microeconomics between empirical research, theory and experiments Why is it that Burger King and McDonald’s restaurants in many cities around the world are in close proximity to each other? Why is Microsoft selling its products Word, Excel and PowerPoint separately, but also as an Office package? And at which price? How many employees will Osram have to lay off because the EU Commission has banned the production of conventional light bulbs? Why does Deutsche Bahn offer a Bahncard50, why is there an OysterCard offered by Transport for London - and how does it optimally set its price? Can BMW use strategic behaviour to block market entry to its new competitor Tesla? What is the value of the 50/ 50 lifeline in Who Wants to be a Millionaire? Microeconomics tries to answer these types of question. In the following chapters, necessary concepts and frameworks are developed and applied to tackle these challenges. Figure 1.1: Water bottle market. 4,50 2,00 0,50 3,00 5,00 4,00 1,50 2,50 3,50 1,00 1,50 2,50 3,50 4,50 0,50 2,00 3,00 4,00 5,00 1,00 👤 👤 👤 👤 👤 👤 👤 👤 👤 👤 👤 👤 👤 👤 👤 👤 👤 👤 👤 👤 <?page no="19"?> 1.1 Microeconomics between empirical research, theory and experiments 19 Microeconomics analyses markets and the behaviour of buyers and sellers. Some of the basic questions can be illustrated looking at the following example. On a very hot public holiday, with all shops closed, 20 people are in a large park. Half of them own a closed 1 litre water bottle each (no gas, absolutely identical quality and temperature) and is in no way thirsty, the other half is very thirsty but does not own any water bottles. The situation can be described roughly as in ► Figure 1.1: the ten owners of the water bottles (black characters) would in principle be willing to sell their water bottles, the ten potential customers (grey characters) would in principle be willing to pay for the bottles. However, the price expectations of the sellers and the willingness to pay of the thirsty individuals are different. Neither sellers nor buyers know the price expectations and willingness to pay of the other side of the market. What will happen now? How many bottles will be sold? And at which prices? How do potential sellers and buyers find each other? How can a seller achieve the highest price? How can a buyer achieve the lowest price? Think about this for a moment - we will return to a solution later in this ► Chapter 1. Models, maps and microeconomics Reality in its complexity cannot be fully described, therefore, science is based on abstraction. Science tries to identify regularities of reality, explain them and - where appropriate - make them applicable. To make regularities visible and seizable in a complexity-reduced form, microeconomics, as well as all other sciences, makes use of models. Models simplify reality to make important features and structures visible. In addition, models add to perceived reality, in particular decidability. This is obvious for every city map in Google Maps. In this sense, microeconomics provides city maps for understanding competition and markets for future managers. With empirically robust, albeit abstractly modelled frameworks, decidability is added. This is crucial if microeconomics is understood to be explaining decisions made by managers in firms. In the analogy of city maps, however, the following is also true: no model can cover all the aspects of reality at the same time. Google Maps, for example, has different views for car drivers (with traffic and congestion information), for users of public transport (with departure times and transfer connections), and for tourists (with references to places of interest) - because otherwise the view for essential information is obscured. Models always simplify and leave things out to make reality manageable. In addition, connections and elements are highlighted which seem to be essential for decision-making. For example, in city maps, main traffic routes are marked in green or orange, although this does not correspond to reality and are shown at a much larger scale (Meyer 1996). Similarly, microeconomic models highlight different aspects from the perspective of various decision-making settings - for example, in ► Chapter 3 decisions with bounded rationality, in ► Chapter 8 pricing strategies or in ► Chapter 10 strategic competition. Microeconomic analyses are based on regularities in decision-making and typical developments in markets. In order to identify regularities scientifically, we adopt two complementary science-theoretical views, as shown in ► Figure 1.2: deduction and induction. Deduction means that a theory is developed from a set of plausible assumptions by logical inference and reasoning. This can be refuted by falsification - the proof of wrong inference or violation of logical rules. Hypotheses can then be formulated from given basic conditions and considerations (a model), i.e., general theoretical considerations are used to infer a special case <?page no="20"?> 1 Microeconomics, competition and strategic behaviour 20 that is to be expected in reality. For example, Einstein theoretically (i.e., by means of deduction) predicted gravitational waves in 1918, but the proof (i.e., empirical confirmation) in reality was only successful 98 years later (Einstein 1918 and Abbott et al. 2016). Induction takes exactly the opposite approach: starting with empirical research (a systematically, however, partially observed reality) researchers assume a regular pattern and a general model is then developed by abstracting and consolidating. Empirical models can be falsified by observing a contradictory individual case, however, an inductive conclusion can never be generalised (Popper 1934). As an example: in Europe, based on consistent empirical observations, it has been assumed for a very long time, that all swans are white and scientists have also developed explanations as to why it cannot be otherwise at all. In 1697, the Dutch sailor de Vlamingh observed black swans for the first time in Australia. The entire theory of white swans was wrong (Taleb 2007). Figure 1.2: Empirical research, hypotheses and theory as foundations for managerial decision-making. With the interplay of induction and deduction, hypotheses emerge which allow a comparison between models (the theoretical and abstract notion of reality) and facts (the empirical observation and classification of reality) (Blaug 1992 and Münter 1999). Decisions of managers are always hypotheses based on a model that has to prove its validity in reality. Without a model, logical decisions are impossible, since the mere perception of reality does not help, when it comes to decision-making. For example, it does not make sense at all for a manager to spend hours looking at a balance sheet if he is not familiar with the model (double-entry bookkeeping, assets vs. liabilities, accrual accounting and the relevant accounting rules). Looking at reality without a model does not allow any conclusions to be drawn and does not support any decisions. models (theory) facts (empirics) induction deduction decisions (hypotheses) costs modelled as cost function costs from P&L statements of recent years <?page no="21"?> 1.1 Microeconomics between empirical research, theory and experiments 21 Good to know │ Rationality - why do people pick up 10 EUR bills off the street? In economics and business administration, it is a usual assumption that people behave or decide rationally. Rationality means that people think and act in a meaningful way, and that this rational action is directed towards a goal consistent with a person's preferences. A decision or behaviour is rational if an action is comprehensible and based on a goal, independent of whether that goal is achieved (effectiveness) or if resources are used efficiently. Rationality thus, implies freedom of decision and choice. From the variety of philosophical explanations of rationality, an understanding has developed in economics which essentially describes logical action directed towards a goal with minimal (i.e., efficient) use of resources, as rational. It is thus clear that individual rationality (at least empirically) includes a wide variety of behavioural patterns. People do not all have the same goals or preferences, nor do they have the same talents or abilities for logical thinking and acting. However, the assumption of rationality makes human behaviour predictable in economic situations: people will typically pick up a 10 EUR bill off the street. On this basis, hypotheses are developed for both customers and managers about their decision-making behaviour in markets, so that predictions for prices or market structure can be developed. Rationality is often under attack, with evidence of people merely following their habits or imitating other people in their consumption patterns or following arbitrary trends. However, all this can be completely rational, depending on individual goals and individual capacity for logic (Smith 2003). That said, the tight and clean concept of rationality cannot be maintained when empirical insights from psychology, sociology, or institutional economics on people's behaviour and decision-making are taken into account (Arthur 1994, DeMartino et al. 2006, Simon 1993 and Conlisk 1996). People still behave predictably - and not erratically or purely by chance - but only boundedly rational. This further expands the theoretical diversity of behaviour and decision-making patterns likely to be observed in markets. Inconsistencies in decision-making behaviour and thus bounded rationality become all the more apparent the more complex the decision-making situations are, the less clear the objectives of human action are, and the more emotions or individual values play a role. In addition, uncertainty and risk, as well as dynamic expectations directed towards a distant future, often have a restrictive effect on the rationality of individual behaviour or decisions. Furthermore, misunderstanding opportunity costs and sunk costs play a decisive role in cognitive distortions of perception, which lead to a boundedly rational decision-making behaviour (► Chapter 3 on behavioural economics as well as Simon 1955 and Kahneman 2003). Models, data, econometrics and stylized facts From a microeconomic perspective, models are theory-driven ideas and mappings of reality that are empirically informed and justified. Models describe the relationship between two or more economic variables: For example, an effect of advertising expenditure on the sales of a product. Empirical data on decisions, be it of customers or managers, can come from market research and competitive analyses. Numbers, data and facts on individual firms from corporate reporting (annual reports with profit and loss statement, balance sheet, but especially manage- <?page no="22"?> 1 Microeconomics, competition and strategic behaviour 22 ment reports and annual strategy presentations at investor’s days) always reflect decisions in markets. If Villeroy & Boch's market share in China increases, this is a reflection of numerous decisions made by customers, the management of Villeroy & Boch and, indirectly, the repercussions of decisions made by Villeroy & Boch's competitors, whose market shares are declining. In order to systematically analyse the connection between advertising expenditure and the profit of a firm, empirical facts can be taken from the profit and loss account or balance sheet. Statistical analyses and methods allow to identify regular patterns in the data, which are then summarised into a model, as shown in ► Figure 1.3. Figure 1.3: Empirical research and theory. This combination of empiricial data and theoretical models is called econometrics. It involves mathematical data analysis and statistical methods to empirically test, quantify, or calibrate theoretical models based on empirical data. Typically, various forms of regression analysis based on cross-sectional data (individual or many different firms in one or more industries) or time series data (development of certain variables over time) are used and carried out using statistical software or Excel (Kennedy 2008 and Davis and Pecar 2013). Good to know │ Learning from data - will machines take over decision-making in the future? Firms use a variety of statistical and mathematical methods to recognise patterns in data in order to prepare and improve decision-making. Many of these methods and procedures analyse relationships in data, i.e., correlations between variables, the variance of data or conditional probabilities as in Bayesian statistics. With increasing digitalisation and data availability in firms, two mutually interdependent developments are taking place: big data and artificial intelligence are by now used more and more by firms to support decision-making. Big data comprises large and complex data sets from very different sources inside and outside the firm, which are weakly or even unstructured, and which are growing rapidly or sometimes exponentially. This enables far better analyses to prepare decision-making, models (theory) to explain perceived reality facts (empirical data) as perceived reality statistical analysis of time series or cross section data (econometrics) 3 1 2 π 0 0 marketing expenditure 0 π = π (marketing expenditure) marketing expenditure 0 <?page no="23"?> 1.1 Microeconomics between empirical research, theory and experiments 23 but also requires new methods for analysis. Some of the methodological development in statistical and mathematical algorithms is summarised as artificial intelligence. AI-based analytical methods learn independently from big data, develop possible explanations and offer decision support, or even decide without further human intervention. Solutions based on data and algorithms for automated price determination for personal pricing or dynamic pricing are already in use (see ► Chapter 8). At the core of these applications of data-driven artificial intelligence are the processes of pattern recognition ('classification') and prediction, in order to first derive options for decisions and finally decide based on causal explanations from the data (Pearl and Mackenzie 2018 and Spiegelhalter 2019). Pattern recognition by means of supervised, machine or deep learning aims to classify a decision situation. Prediction then uses the presumed causal relationship to derive the future development from the data. In this sense, big data and artificial intelligence as logical conclusions are merely new forms of induction and deduction. Obviously, the roles of human beings as decision-makers in interaction with data and algorithms are changing as well and thus, also the object of investigation in economics (see also Brynjolfsson et al. 2017, Loebbecke and Picot 2015, Mihet and Philippon 2019, Currie et al. 2020 as well as Acemoglu and Restrepo 2018). In economics and social sciences, data and results of empirical studies are often summarised into stylised facts. Stylised facts describe regularly observable (‘typical’) findings of empirical studies that are considered essential, for example, between advertising expenditure and the profit of a firm, which then can become the starting point of a model. The understanding of the concept of stylised facts goes back to Kaldor (1961): 'Any theory must necessarily be based on abstractions; but the type of abstraction chosen cannot be decided in vacuum: it must be appropriate to the characteristic features of the economic process as recorded by experience. Hence the theorist, in choosing a particular theoretical approach, ought to start off with a summary of the facts which he regards as relevant to his problem. Since facts, as recorded by statisticians, are always subject to numerous snags and qualifications, and for that reason are incapable of being accurately summarized, the theorist, in my view, should be free to start off with a 'stylized' view of the facts - i.e., concentrate on broad tendencies ignoring individual detail, and proceed on the 'as if' method, i.e., construct a hypothesis that could account for these 'stylized' facts, without necessarily committing himself on the historical accuracy, or sufficiency, of the facts or tendencies thus summarized.' (Kaldor 1961, pp. 177-178). It is inevitable that observations contradicting the stylised facts always are around and identified, so that it is essential to make a careful distinction between true regularities and exceptions. A theoretical model provides a possible explanation of a regular and empirical observation that is considered typical. This implies, in contrast to random coincidence, a description of a causal relationship, i.e., a functional cause-and-effect relationship to explain an observation. Effects that can be explained or can be determined within the framework of a model are called endogenous variables: in microeconomic models, for example, profits of firms. The variables used for explanation are exogenous variables, for example, the market structure or the intensity of competition in an <?page no="24"?> 1 Microeconomics, competition and strategic behaviour 24 industry. Exogenous variables can also be variables that are controlled by the firm, such as marketing expenditures or R&D strategy. The functional relationship between exogenous and endogenous variables might only provide a partial aspect of an explanation. Additionally, the relationship can be mutual, i.e., the level of profits affecting the marketing budget for the next year. Both endogenous and exogenous variables can also be influenced by variables outside the model, i.e., macroeconomic conditions or regulatory restrictions. ► Figure 1.3 on the right shows that there is apparently an effect of advertising expenditure on profits 𝜋𝜋 of Villeroy & Boch - profits 𝜋𝜋 are mathematically a function 𝜋𝜋 = 𝜋𝜋(𝑀𝑀) of advertising expenditures 𝑀𝑀 . If advertising expenditure is very low, profits are negative. With increasing advertising expenditure, profits initially skyrocket. However, this effect does not last. After a certain point, increasing advertising expenditure reduces profits. A manager at Villeroy & Boch can now use this empirically based model to develop future decisions on advertising expenditure in order to increase profits. Experiments and behavioural economics Decisions can only be observed empirically ex post and indirectly, for example, by analysing purchase decisions made by customers or competitive strategies implemented by firms. The indirect analysis of decisions based on market and competition data has several deficits. These include data collection and quality, comparability of data from different industries, markets, regions and time periods, and the selection and specification of appropriate econometric methods. A major drawback, however, is that the decision-making process itself is not directly observable. Although the consequences of the decisions, are visible in market shares or sales (e.g., numbers of smartphones), the actual decision, along with the decision environment or products considered as alternatives, is unobservable. To tackle this challenge, microeconomists have been conducting more and more experiments since the 1980s. The aim is to observe and analyse decision behaviour directly and in particular to see and understand the interaction of people in strategic decision situations under controlled laboratory conditions. A strong driver for these experiments is to check whether people actually decide and act completely rational, or whether deviations from perfect rationality can be observed. In fact, the insights gained in experiments have made significant contributions to the emergence of behavioural economics, an analysis of economic decisions from the perspective of behavioural sciences (► Chapter 3). In recent decades, a large number of experiments have been conducted to investigate individual decision-making behaviour, its influencing factors and dynamic interaction between decision-makers. In order to ensure the robustness of derived insights, currently standardised, computerand simulation-based laboratory experiments are leading edge for further research. An advantage of economic laboratory experiments is that decision behaviour can be observed directly, under supervised and modifiable conditions (similar to 'randomised controlled trials' used in medical and psychological research). Thus, a simple check of the consistency of decisions and actions, as well as the exclusion of alternative explanations, becomes possible. In addition, explanations can be found that are impossible to be identified within abstract data <?page no="25"?> 1.2 Markets, demand and supply 25 (such as a balance sheet or profit and loss account) or cannot be assigned to individual decisions. However, laboratory experiments have a number of disadvantages and challenges. Participants in an experiment are of course aware of their participation. Thus, laboratory situations evoke certain behaviours and suppress others. There are indications that participants want to behave according to the expectations of the researcher or according to the 'usual results' of an experiment. Laboratory situations differ drastically in monetary and social consequences from decisions in real life. In addition, numerous experiments are carried out, mainly for cost reasons, with students at universities who are not representative of decision-makers in real markets. However, tests with managers, albeit rare, confirm typical results of experiments with students (Kagel and Roth 2016 and Davis and Holt 1993). The challenges of the laboratory experiments can partly be solved by field experiments. Here the experiment is carried out, unknown to the participants, in real environments, for example with real customers using different product offers or prices in otherwise identical supermarket branches. Some field experiments are also carried out by firms online. For example, the insurance company CosmosDirekt shows potential customers different product packages or product arrangements (‘A/ B testing’) on their website depending on the IP address, in order to better understand the customers' decision-making behaviour (Levitt and List 2009, List and Reiley 2007 as well as Gneezy and List 2006). 1.2 Markets, demand and supply Microeconomics is focused on the functioning of markets and market outcomes. Colloquially, in a market, products are traded between buyers and sellers. In order to provide conceptual clarity for the following chapters: markets are institutions (systems, rules, patterns and structures) in which transactions between market participants are sought or carried out. In specific cases, a market can be defined narrowly in terms of location, time or content, for example the German automobile market in 2018, where passenger cars are sold by firms and purchased by retail customers. However, automated trading between high-performance computers using computing capacity also describes a market, as does global trading of CO2 emission rights between firms and states. Markets as institutions Markets differ in their respective systems, patterns and structures in many dimensions, including traded products and services - e.g., the stock market selling shares is different from a supermarket selling yoghurt; role and number of market participants - e.g., medium-sized suppliers with lots of competitors to the automobile industry behave differently than a monopolist like Facebook in social media; location - e.g., the famous fish market in Hamburg is different from the famous fish market in Sydney; <?page no="26"?> 1 Microeconomics, competition and strategic behaviour 26 institutional degree of organisation - e.g., approved participants in a mobile phone licence auction of the German Bundesnetzagentur (Federal Network Agency) behave differently than people at a spontaneous flea market buying memorabilia; level of information and the exchange of information between market participants - e.g., negotiations in the context of a M&A transactions are different from dealing drugs on the street; as well as how price and quantity are determined - e.g., pricing for pay-as-you-go data packages works different than pricing of antiques at an auction. The way in which markets operate has evolved over many centuries within the respective legal systems and against the background of technological developments, which significantly affects the behaviour of market participants. In this sense, the stock market naturally works differently from an M&A transaction, although in both cases shares of firms are traded and transferred. Taken together, institutions are the name given to these systems, patterns and structures of markets. Institutions describe well-known patterns of behaviour and rules that are applied by market participants in transactions. They help to understand and explain observable behaviour in markets. Institutions range from non-binding conventions and generally accepted ‘rules of the game’ between market participants, which have proved useful in the past, to the legal system, which in particular defines the notion and transfer of property rights in markets. The purpose and aim of institutions is on the one hand to reduce uncertainty in markets and on the other hand to structure options for action in such a way that transactions can take place. Institutions thus, play a decisive role for the functioning as well as for the efficiency of markets and at the same time explain different growth paths of economies in international comparison, for example, against the background of very different ownership and decision-making systems under capitalism, communism or absolutist and dictatorial systems (Voigt 2019 and Acemoglu and Robinson 2012). If transactions are not carried out in generally accessible markets, but within the family, on black markets, within firms or by the state, it is often due to missing or malfunctioning institutions and excessively high transaction costs. Institutions of a market and in particular the process of how prices and quantities are determined can take many different forms. These range from spontaneous transactions in which market participants repeatedly agree on how to settle the transaction and negotiate prices, to auctions with variable prices (such as the auction of antiques or advertising space on Google), to regulated markets in which the state or the operator of the market determines both possible transactions and procedures for determining the price and settlement of transactions. The latter is the case, for example, in the trading of securities in various countries. Taking Germany as an example, the operation of the market (the stock exchange) and the participation is subject to approval and registration. Additionally, the types of transactions that can be carried out, the permissible products as well as the mode of pricing and settlement of a transaction are defined by Börsenand Wertpapierhandelsgesetz (Stock Exchange and Securities Trading Act). In addition, each transaction is monitored by the Börsenaufsicht (German Stock Exchange Supervisory Authority) and the Bundesanstalt für Finanzdienstleistungsaufsicht (Federal Financial Supervisory Authority). <?page no="27"?> 1.2 Markets, demand and supply 27 Beyond this, markets change, sometimes drastically, over time due to changes in customer preferences, innovations and technological transformation driven by firms and changing regulatory regimes defined by the state and changed institutions, as well as repercussions from changes in other markets. Product markets, factor markets and transaction costs What is a market? An initial helpful and simplifying approach is to classify market participants as buyers (the demand side) and sellers (the supply side) of a product, service or any other good (e.g., rights, information, data, or derivatives of original products). If the supply side is provided by firms, they are collectively referred to as an industry (e.g., telecommunications industry, pharmaceutical industry, or financial services industry). At a higher level are economic sectors (such as manufacturing, education, or construction), at a lower level are products (e.g., based on the SIC/ ISIC classification such as cars, current accounts, or smartphones). In product markets, physical or digital products are traded between firms (B2B) or between firms and customers (B2C). In factor markets labour and capital are traded as input factors for production and services on the labour market and capital market. Carrying out a transaction in a market is not for free. Typical costs include search and information costs of market participants, commissions, services of intermediaries to initiate the conclusion of a contract or the execution of a transaction, and the use of the market as such. All these costs are known as transaction costs. Since transaction costs are usually linked to the transfer of property rights, a reliable and generally accepted legal system leads to a reduction in transaction costs. Conversely, a poor legal system requires extensive contract negotiations or, if necessary, bribes and increases in transaction costs. If these transaction costs are too high, a transaction either does not take place at all or happens in another institution - e.g., a firm (cf. ► Chapter 4). Conversely, zero transaction costs mean that all market participants have complete information - searching for better offers or the negotiation of prices and conditions would then be unnecessary. Market supply, market demand, prices and quantities A key question in microeconomics is how prices and quantities are determined in a market by the interaction of buyers and sellers. The analysis of markets in the context of supply and demand tries - from different perspectives - to find out how many transactions are going to happen, which quantities will be traded and at what price and how large the revenues (i.e., price multiplied by quantity) in a market are. In 2016, for example, 23.7 million smartphones were sold in Germany at an average price of EUR 407 through the interaction of supply and demand (IDC 2017). This results in total revenues in the market of EUR 9.65 billion. In order to gain some initial explanation for this market outcome, demand side and supply side are now looked at separately. <?page no="28"?> 1 Microeconomics, competition and strategic behaviour 28 Figure 1.4: Willingness-to-pay and demand function for smartphones. ► Figure 1.4 on the left shows the (hypothetical) demand for smartphones in Germany, which was identified, for example, by a customer survey of willingness to pay ("Would you be willing to buy a smartphone for 150 EUR? Would you also buy it for 250 EUR? "). The individual willingness to pay by a customer indicates the maximum price that an individual customer would be prepared to pay for a smartphone of a certain quality or brand. In ► Figure 1.4 on the right, all individuals willingness-to-pay are connected and drawn as a straight line - in a linear approximation for simplicity - as the demand curve of the entire market. However, demand curves can take almost any other functional form, depending on data from market research. For a linear demand function, two important pieces of information can now be determined through the intersections with the price and quantity axis: (1) no customer is willing to pay more than 1000 EUR for a smartphone and (2) 40 million customers would accept a smartphone at a price of 0 (i.e., as a gift). The demand curve has a downward slope: the higher the price, the lower the quantity demanded for a product. From this information, a demand function can be reconstructed, which is assumed to be linear in its simplest case. An (inverse) demand function describes the reciprocal functional relationship between the quantity 𝑞𝑞 𝐷𝐷 (in millions of smartphones) demanded as a function of the price 𝑝𝑝 with otherwise unchanged market conditions, in this example as (1.1) 𝑝𝑝 𝐷𝐷 (𝑞𝑞) = 𝑎𝑎 − 𝑏𝑏𝑞𝑞 = 1000 − 25𝑞𝑞 𝐷𝐷 . Here 𝑎𝑎 denotes the maximum willingness to pay in the market (1000 EUR) and 𝑏𝑏 indicates the slope of the demand function, which equals 𝑏𝑏 = 1000 40 = 25 . Hence, with each price reduction of 25 EUR, the quantity of smartphones demanded would increase by 1 million. Conversely, by dividing the maximum willingness to pay by the slope of the demand function, the maximum demand of the market can be determined as 𝑞𝑞 𝑚𝑚𝑚𝑚𝑚𝑚 = 𝑎𝑎/ 𝑏𝑏 = 1000/ 25 = 40, i.e., a maximum of 40 million units. 1/ 𝑏𝑏 is an indicator of the size of the demand in a market - the smaller 𝑏𝑏 is, the higher is the maximum demand at each price. quantity q price p 0 1 10 individual willingness to pay of different customers 1000 40 quantity q price p 0 1 10 highest willingness to pay in the market is 1000 1000 40 maximum number of customers if price is reduced to 0 demand curve as depiction of willingness to pay of all customers <?page no="29"?> 1.2 Markets, demand and supply 29 Figure 1.5: Changes of the demand function. The customers' willingness to pay, which defines the run and location of the demand curve in a market, is determined not only by price, but also by customers’ income, by prices of other products, the behaviour of other customers, the quality of the products, and the marketing of the firms (► Chapters 2 and 3). So, for example, the demand for smartphones increases with rising income, but especially with low-price mobile or data plans. ► Figure 1.5 shows that a change in the willingness to pay of all customers shifts the demand curve in parallel, a change in the size of the market leads to a rotation of the demand curve. There are two reasons for the downward slope of the demand curve which are based on price changes: the substitution effect and the income effect. The substitution effect describes how customers switch to comparable products (which are called substitutive products) when prices increase, because these other products become relatively cheaper, for example, some customers switch from Samsung smartphones to cheaper alternatives when prices of Samsung smartphones increase. The income effect occurs because buying power decreases when income remains constant and prices increase at the same time. Price increases in mobile or data plans then lead to less data being consumed at the same income. In the same way as on the demand side, it is possible to determine the individual willingness of firms to offer a smartphone, depending on the price that can be achieved, by interviewing firms, or conducting market research. Besides the achievable price, crucial determinants of supply are technology, production capacity and cost structures of a firm, as well as profit expectations associated with a certain price, but also the behaviour of competitors (► Chapters 5, 6, 7 and 10). If production costs of touch screens decrease, most smartphone manufacturers will expand quantity q price p 0 40 1 10 1000 quantity q price p 0 40 1 10 quantity q price p 0 40 1 10 quantity q price p 0 40 1 10 1000 25 50 32 original demand function decrease of willingness to pay (reduction of a) increase of market size (decrease of b) decrease of market size (increase of b) 1000 1000 800 p = p q = a − bq = 1000 − 25q p = p q = a − bq = 800 − 25q p = p q = a − bq = 1000 − 20q p = p q = a − bq = 1000 − 40q <?page no="30"?> 1 Microeconomics, competition and strategic behaviour 30 their production and supply, whereas production stoppage at chip suppliers will lead to a decline in supply of smartphones. ► Figure 1.6 on the left shows the (hypothetical) willingness of various firms to offer smartphones on a market at a certain price. The individual willingness of firms to offer smartphones increases the higher the achievable price. In ► Figure 1.6 on the right, the individual offers of firms are connected and form a supply curve of the entire market. The supply function describes the functional relationship between the quantity 𝑞𝑞 𝑆𝑆 supplied (again in millions of units) as a function of the price 𝑝𝑝 (with all other parameters unchanged). This supply curve could be reconstructed - analogous to the demand function - and is given as (1.2) 𝑝𝑝 𝑆𝑆 (𝑞𝑞) = 360 + 2𝑞𝑞 𝑆𝑆 . Figure 1.6: Individual supply and supply curve of smartphones. Smartphones will be offered starting at a price of EUR 360. With each additional market potential of 1 million units, firms will enter the market, expand production and raise their prices by EUR 2. Market equilibrium and market price In a highly simplified model, supply and demand curves can be inspected together. Obviously in ► Figure 1.7 on the left there is an intersection of supply and demand which, as a market equilibrium, mutually determines quantity and price. A market equilibrium is the only pricequantity combination where the quantity demanded 𝑞𝑞 𝐷𝐷 exactly corresponds to the quantity supplied 𝑞𝑞 𝑆𝑆 . The point of intersection of the two curves determines the equilibrium price, which balances the quantity demanded and the quantity supplied. An essential finding is: any customer whose willingness to pay is larger than or equal to the equilibrium price can buy a smartphone at this price and any firm willing to supply at this equilibrium price or below can sell a smartphone at this price. It is precisely under this condition that transactions are taking place. quantity q price p individual supply of different firms depending on price 0 40 1 360 1000 quantity q price p 0 40 1 demand function depicting individual supply of firms below 360 EUR, no firm will supply a smartphone at a price of 440 EUR, 40 mill. smartphones are supplied 360 440 1000 <?page no="31"?> 1.2 Markets, demand and supply 31 By equating the two functions (1.1) and (1.2) with (1.3) 𝑝𝑝 𝐷𝐷 (𝑞𝑞) = 𝑝𝑝 𝑆𝑆 (𝑞𝑞) as (1.4) 1000 − 25𝑞𝑞 𝐷𝐷 = 360 + 2𝑞𝑞 𝑆𝑆 and solving for 𝑞𝑞 𝐷𝐷 = 𝑞𝑞 𝑆𝑆 we get a quantity of 𝑞𝑞 𝐷𝐷 = 𝑞𝑞 𝑆𝑆 = 23.704 million smartphones being offered and bought in market equilibrium. If this quantity is now reinserted either in the demand function (1.1) or in the supply function (1.2), a market price of 𝑝𝑝 𝐷𝐷 (𝑞𝑞) = 𝑝𝑝 𝑆𝑆 (𝑞𝑞) = 407.41 EUR results - both values of this simple model confirm the quantities and prices actually observed in Germany in 2016 for the entire market. Obviously, the model is highly simplified: it is not possible to see how many smartphones individual firms sell at what price. Moreover, as shown in ► Figure 1.7 on the right, based on the original market research data, price dispersion can be observed in the real market, i.e., not every smartphone of the same quality and brand is sold at the same price. Figure 1.7: Market equilibrium as an intersection of market supply and market demand. A uniform market price and a market equilibrium is usually to be expected if: the products traded are absolutely identical, so that there is no product differentiation or customers having preferences for a particular firm or brand; all market participants have the same complete information, i.e., each market participant knows all other market participants and/ or is aware of their willingness to pay and their price expectations; there is competition between sellers, i.e., there is no agreement on prices or any kind of other strategic behaviour; both sides of the market are populated by a large number of participants, so that none of them can influence or even fix prices or quantities through their behaviour; and coordination towards an equilibrium price is possible, i.e., prices are per se flexible and transaction costs of market participants are negligible. quantity q price p 0 40 360 1000 407,4 23,7 demand curve supply curve quantity q price p 0 40 1 360 1000 407,4 23,7 <?page no="32"?> 1 Microeconomics, competition and strategic behaviour 32 If one or more of these requirements are not sufficiently met, it is possible that different market prices exist at the same time. In the following chapters, these simplifying assumptions are successively removed in order to obtain a more detailed picture of the market mechanism and dynamics, the decisions of customers and, in particular, the strategic choices made by firms. However, a market for which the above requirements are relatively well met is the stock market: A Daimler share is a homogeneous product due to its ISIN DE0007100000, the share prices on the various stock exchanges are almost identical at all times for shares with very high liquidity. In addition, it does not matter to customers which stock exchange or broker they buy the share from. Market participants, i.e., buyers and sellers, can observe the order book of the respective stock exchanges at any time - identical to a supply and a demand function, all currently available bids and asks including respective quantities are listed there (www.onvista.de/ aktien/ orderbuch/ Daimler-Aktie-DE0007100000 or within Deutsche Börse www.boersefrankfurt.de/ aktien/ orderbuch). All market participants compete - each of them wants to realise either the optimum selling or respectively the optimum purchase price. Both sides of the market consist of many thousands of market participants and each market participant is, relative to the size of the market, small and cannot significantly influence the share price. Deutsche Börse (as well as other stock exchange operators) coordinates supply and demand for each share at all times via market makers or designated sponsors towards the equilibrium price. As a consequence, however, share prices might fluctuate strongly: a market equilibrium does not mean stable or rigid prices, but an almost immediate balancing of supply and demand through price flexibility leads to constant price adjustment and fluctuating prices. Market mechanism and dynamics of market equilibrium Whether a market equilibrium is reached depends on, among other things, whether deviations from a market equilibrium disappear over time due to adjustments in quantities and prices. In ► Figure 1.8 on the left, a price 𝑝𝑝 1 happens to be above the equilibrium price 𝑝𝑝 0 . As a result, at this price 𝑝𝑝 1 the supply of the firms exceeds the demand of the customers, so that there is an excess supply, for example, in the form of too many produced goods or a high level of inventory. Firms can eliminate this excess supply by reducing prices from 𝑝𝑝 1 to 𝑝𝑝 0 . This increases the quantity demanded by customers and the market outcome moves towards market equilibrium. <?page no="33"?> 1.2 Markets, demand and supply 33 Figure 1.8: Excess supply and excess demand. Similarly, as shown in ► Figure 1.8 on the right, a price 𝑝𝑝 2 < 𝑝𝑝 0 results in excess demand. The firms can now increase prices, and as a result the quantity demanded decreases and market equilibrium is reached. Without price flexibility and the balancing effect of prices, no market equilibrium can be achieved. Figure 1.9: Shift of supply and demand curves and adjustments of market equilibrium. Similarly, a new market equilibrium is determined if basic conditions in a market change and the supply and/ or demand curve shifts. In ► Figure 1.9 on the left, for each price firms increase the quantity offered, so that the supply curve shifts to the right. The reason for this could, for example, be cost reductions of input suppliers. As a result, increased supply could also be sold, but firms have to reduce prices to sell the higher production volume. In ► Figure 1.9 on the right, demand is increasing, for example due to increasing income. As a result, the demand curve shifts to the right - firms can now sell more and also raise prices. How strong these effects quantity q price p 0 q 0 p 0 D S p 1 excess supply: S > D q D q S quantity q price p 0 q 0 p 2 D S p 0 q S q D excess demand: D > S quantity q price p 0 q 0 p 0 D S S‘ p 1 q 1 quantity q price p 0 q 0 p 0 D S p 1 q 1 D‘ <?page no="34"?> 1 Microeconomics, competition and strategic behaviour 34 are is largely determined by the location and slope of the demand and supply curves and is explained in ► Section 1.3. Adaptation to a (new) market equilibrium described in ► Figures 1.8 and 1.9 naturally requires time and a learning process of the market participants. Depending on a current market situation, the level and quality of information of market participants, and the necessary extent of adaptation, one cannot assume that every market is in equilibrium at all times. However, one will usually observe dynamic processes of adaptation towards an equilibrium. Good to know │ Why do markets work so well, why is the state a bad entrepreneur? To achieve good or even the best possible economic solutions in a society, in principle three different institutions compete: markets, firms and the state. Which institution generates the best solutions efficiently cannot be said in general - it depends mainly on the problem to be solved, the incentives given, and whether the market mechanism works. In many countries, dissatisfaction with solutions that markets provide is leading to calls to push back supposedly 'capitalist markets' in favour of 'more state' (Fuest 2020). The first fundamental thing to recognise is that markets themselves are a social innovation. Markets allow new products or business models to be tried out, where good ideas prevail over bad ideas and thus, provide incentives for firms or consumers to look for best solutions (von Hayek 1968 and 1975). This works so well because firms and customers have clear objective functions. Firms want to make profits in order to be viable. Customers want to consume according to their preferences in order to increase their benefits and satisfaction. This is why firms and entrepreneurs also implement new ideas and innovations: they expect to make a profit. Both sides of the market thus, have incentives to seek better solutions. This decentralised process of search and coordination is the centrepiece of the invisible hand, first described by Adam Smith (1776), which coordinates markets and becomes visible in the impact of the price mechanism. Markets generate information, for example, through prices. When a customer negotiates a price with a seller, both of them condense all available information (production costs or the competitive situation on the seller's side, willingness to pay, and alternative products on the customer's side) into a single number. This number then contains all relevant information in the form of a price. High prices indicate scarcity and possible profits, whilst low prices indicate a lack of attractiveness of the solution and hardly any profit opportunities. In this way we see the invisible hand steering the actions of firms and consumers in markets. High prices signal incentives for firms to innovate or create new market segments, whilst low prices indirectly direct resources (ideas, efforts, capital and labour) towards more efficient uses with higher profitability. The customers' search for good solutions is also visible at supermarket checkouts. The waiting time at all checkouts is about the same, because all customers look for the fastest checkout and thus, optimise their behaviour in competition. As a result, all queues are almost the same length at all times (Frank 2011). Similarly, all customers queue up at the 'best ice-cream gelateria' and in doing so signal high quality or best value for money to other customers. <?page no="35"?> 1.2 Markets, demand and supply 35 In this sense, markets repeatedly create spontaneous order in situations that a central coordinator could hardly overlook or control (Sugden 1989 and Preda 2009). Markets thus, coordinate transactions between firms and customers and create the necessary order or structure in an evolutionary and decentralised way. Nobody invents a market - a market emerges and organises itself because it is advantageous for all market participants. Coordination in the market takes place without central control or supervision: instead, the structures adapt themselves decentrally and dynamically to the requirements of the market participants. Markets that can generate or process a large amount of information then lead to efficient solutions and optimal transactions. Conversely, information asymmetry and lack of information hinder the functioning and efficiency of markets. Low information density, bad or wrong information can lead to speculation and bubble formation in markets, or cause a market to collapse. Unlike firms and customers, the state does not have a clear economic objective. Increasing welfare in the entirety of a society is a far more complex problem. It includes, among other things, objectives of equality, distributive fairness, freedom and safety. This is especially the case since the choice of measures and methods to achieve those objectives are subject to political decision-making. Moreover, the state knows neither all production possibilities (e.g., technologies or resources) nor the preferences of all consumers. There is thus, a substantial information deficit that makes central planning of economic activities by the state impossible. This becomes all the more apparent when customer preferences or technological options change over time. This explains why the state should only act in markets where it is socially necessary and where firms and consumers cannot find solutions or cannot agree on transactions. Due to a lack of sufficient incentives, information deficits and the absence of clear objectives, the state often does not find good solutions - as a result, governmental action in markets is often inefficient and not in the customers' interests. In this sense, the state is often said to be not a good entrepreneur. Firms owned by the state or with significant state participation generally perform worse, are less efficient and less profitable than comparable firms owned by private or institutional shareholders (Dewenter and Malatesta 2001, Boardman and Vining 1989 and ► Section 7.4). In extreme cases (as recent history in communist or socialist states shows) the state reduces economic freedom of choice and bureaucratically determines the education and professions of individuals, defines the purpose of the firm, the quality of products, specifies production quantities and sets prices. Frequently, consequences are a lack of goods and supply bottlenecks, poor quality, long queues for sought-after products, black markets, and finally a lack of innovation and economic and social decline. However, it is also possible that a market equilibrium may not be achieved. Supply and demand curves do not intersect - this can happen when, in principle, the willingness to pay of all customers is lower than the firms' price expectations. This is repeatedly the case in the pharmaceutical industry and with drug prices, so that in this case drugs are not developed and cannot be bought. Non-exclusion of consumption - it may be that a firm cannot exclude those customers who do not pay but are nevertheless using the product: this is the case, for example, with public broadcasting or lighthouses. A firm will not offer these products or services because <?page no="36"?> 1 Microeconomics, competition and strategic behaviour 36 either no price can be charged, or the pricing model would not be efficient. Therefore, no profits can be made. Such products are called public goods and are usually provided by the state. Information deficits - buyers and sellers either do not have complete transparency about products or supply and demand (incomplete information), or market participants are differently well informed (asymmetric information). The conclusion of a transaction is endangered or even prevented by these two information deficits. This can happen with secondhand goods. Akerlof (1970) has described that sellers typically have better information about the product before the contract is concluded, for example, a used car. If buyers cannot clearly determine the quality of a used car before signing the contract, buyers will per se assume poor quality of the used car and will therefore have a low willingness to pay. As a result, 'high quality used cars' are almost unmarketable at high prices and are forced out of the market, so that only 'low quality used cars' are offered at low prices: thus, both sellers and potential buyers of 'high quality used cars' are disadvantaged (► Chapter 3 on risk and uncertainty, ► Chapter 4 on principal-agent theory and ► Chapter 7 on market failure). In all these cases, there is a market failure - the market either does not exist or works only to a very limited extent. Governmental action can occur, either by providing the market itself, by regulating the market, or by acting as a public state-owned firm. Such cases are discussed in the context of economic policy and competition policy (see Fritsch 2018). In the case of public broadcasting in Germany, the state intervenes in the market by obligatory fees; lighthouses are set up on the coast as public goods financed by taxes; in the health care market various economic policy instruments are combined, e.g. the price-fixing of drugs, compulsory membership of health insurance schemes and subsidies of R&D expenditure. For the same reason, dealer ratings on auction platforms such as ebay also prove their worth: the product quality is indirectly identifiable via the rating of the dealers. Case Study │ Water Bottle Market Market equilibria and equilibrium prices can be achieved, inter alia, if (as mentioned above) there is no information asymmetry, no transaction costs are incurred and a coordination process towards the equilibrium price takes place. The following example highlights these points - what happens if these conditions are not fulfilled? The ten potential sellers and the ten potential buyers of water bottles in ► Figure 1.1 are still in the park. A quick market research, carried out individually with the respective market participants, shows a willingness to pay among the thirsty people as shown in ► Figure 1.10 on the left: Customer K1 would be prepared to pay 5 EUR, customer K2 4.50 EUR, and so on in steps of 0.50 EUR down to customer K10 with 0.50 EUR. The owners of the water bottles were also asked at what price 𝑝𝑝 they would sell their water bottle - the values are also shown in ► Figure 1.10 on the top: Seller V1 would sell for 0.50 EUR, and so on in steps of 0.50 EUR to Seller V10 with a price expectation of 5 EUR. How many bottles of water are now traded at what price? The answer to this question depends significantly on the market institutions and the level of information of the market participants. In a market without transaction costs and with full information, a (hypothetical and also free of charge) market coordinator would work out supply and demand curves based on the willingness to pay and the price expectations as <?page no="37"?> 1.2 Markets, demand and supply 37 (1.5) 𝑝𝑝 𝑆𝑆 (𝑞𝑞) = 0 + 0.5𝑞𝑞 𝑆𝑆 and (1.6) 𝑝𝑝 𝐷𝐷 (𝑞𝑞) = 5.5 − 0.5𝑞𝑞 𝐷𝐷 and calculate the equilibrium price at the intersection of both functions as 𝑝𝑝 𝑆𝑆 = 𝑝𝑝 𝐷𝐷 = 2.75 . He then communicates this value to all potential buyers and sellers: K1 to K5 and V1 to V5 will immediately carry out transactions at this price. The reason for this is that their willingness to pay is above the price or their price expectations are below the price. All of them will carry out transactions that at least meet their original expectations. This means that there will be five transactions at a price of EUR 2.75 and total revenues of EUR 13.75. This result can actually occur in reality, but only if all market participants are fully informed and there is a coordination process on the equilibrium price, which can arise in particular through mutual learning over time. This is the case in many markets in which there is a high degree of transparency on the willingness to pay of customers and the price expectations of firms. This could be due to a large number of previous business relations, or due to big data and the availability of prices on the internet, for example. Figure 1.10: Demand and supply of water bottles. However, transactions could happen if the willingness to pay is at least equal to the expected price. ► Figure 1.10 shows two additional (out of a very large number) possible institutions that describe how this market could work and how prices and quantities are determined. In the middle, a situation is described in which buyer K1 randomly meets seller V10 horizontally. The willingness to pay 5.00 EUR corresponds exactly to the expected price of 5.00 EUR, so that a transaction takes place. Subsequently (as shown K2 with V9, K3 with V8, etc.) an additional nine of these horizontal transactions can occur by pure chance, in which the price demanded by each seller is exactly met by the customers’ individual willingness to pay. In total, ten transactions are carried out with total revenues of 27.50 EUR. The average price of each water bottle is 2.75 EUR, exactly the same as the equilibrium price - but now each customer pays a different price. Another possible situation is shown in ► Figure 1.10 right - instead of horizontal transactions, vertical transactions are now taking place. In fact, only five transactions are happening instead of ten. Which price is actually paid in these pairs depends on how and in which sequence information is exchanged. If the seller always states their low price expectation first and the buyers respond to it, a total revenue of 7.50 EUR and an average price of 1.50 EUR is achieved; if customers always state their willingness to pay first, total revenue is 20 EUR with an average price of 4.00 EUR; if negotiations between the two take place, the price can also vary in between (see also ► Table 1.1). <?page no="38"?> 1 Microeconomics, competition and strategic behaviour 38 Market results in the water bottle market model market equilibrium horizontal transactions vertical transactions (1) vertical transactions (2) without TK including TK without TK including TK without TK including TK without TK including TK number of transactions 5 4 10 8 5 4 5 4 average price 2.75 2.75 2.75 2.75 4.00 3.75 1.50 1.75 maximum price 2.75 2.75 5.00 4.50 5.00 4.50 2.50 3.00 minimum price 2.75 2.75 0.50 1.50 3.00 3.00 0.50 1.00 revenues 13.75 11.00 27.50 22.00 20.00 15.00 7.50 7.00 Table 1.1: Market outcomes in the water bottle market. Finally, in all three cases, transaction costs may be incurred: for example, by way of information gathering, information exchange or price negotiations, which in fact corresponds to a participation fee for the market. ► Table 1.1 shows the resulting effects for all market mechanisms, assuming that each market side has to pay an amount of 0.50 EUR per potential transaction - thus, of course, the price expectation of each seller increases by 0.50 EUR, while at the same time the willingness to pay of each potential buyer decreases by 0.50 EUR. As a result, the number of transactions and the revenues in the market decrease, and low and high prices tend to be eliminated from the market. Which of the possible market outcomes actually will occur in reality depends on the exchange of information and the competitive dynamics in the market. Different market mechanisms, information asymmetry or incomplete information, the sequence of transactions, and the level of transaction costs can lead to very different market outcomes - different prices, different quantities and different revenues. The two cases of vertical and horizontal transactions at different prices can only arise or persist if there is strong and persistent information asymmetry between market participants and a lack of competition between sellers. Once more this clearly shows that an equilibrium price can only emerge with competition and learning market participants over time. 1.3 Price elasticity and marginal revenues A crucial management task - just as making profits - is to increase sales (or revenues). This applies in particular to business models with high fixed costs and low variable costs. For example, for Deutsche Bahn, for a museum, but also for a mobile network operator, the cost structure is unchangeable in the medium term and independent of the actual number of customers. Decisions are then substantially focused on increasing sales and revenues. Especially at the end <?page no="39"?> 1.3 Price elasticity and marginal revenues 39 of the year, when planned sales targets have not yet been reached, price reductions are often implemented to increase sales. Whether the goal of increasing sales through price reductions really can be achieved, depends on the price elasticity of demand: the price elasticity of demand describes how much a price change affects quantity demanded. Firms often try to increase revenues through price reductions. This idea is based on the downward slope of the demand curve: if prices are reduced, the quantity demanded by customers typically increases. Firms pursue different strategies, apparently with very different levels of success, as the following cases demonstrate. Praktiker: for many years, the German DIY chain, which at that time was active throughout Europe, advertised the marketing slogan "everything 20% off - except pet food" with corresponding recurring price reductions. After a loss of EUR 555 million in 2011, Praktiker filed for insolvency at the Saarbrücken local court in 2013, the business was completely discontinued in 2014 and as a result around 20,000 people lost their jobs. Edeka: for many decades, German based food retailer Edeka has been distributing regional “special offer flyers” via mailings or as a supplement in free newspapers. These announce differentiated price reductions for individual products on a weekly basis - for example 4% off for glass cleaners, 7% off for dishwashing detergents, or 3% off for coffee. Edeka had achieved a profit before taxes of EUR 355 million in 2014 and is one of the most successful retailers worldwide. Price elasticity of demand An analytical framework for checking how strong the effect of a price reduction is on the quantity demanded by customers is the price elasticity of demand. The basic idea of a price change is shown in ► Figure 1.11 on the left: depending on the slope of the demand curve, a price reduction will lead to a (disproportionately low or high) increase in demand and a price increase will lead to a (disproportionately low or high) decrease in the quantity demanded. For a demand function (1.7) 𝑝𝑝(𝑞𝑞) = 𝑎𝑎 − 𝑏𝑏𝑞𝑞 = 100 − 0.02𝑞𝑞 Figure 1.11: Changing prices and impact on quantity demanded. q p 0 A B 4000 3990 20,0 20,2 C 19,8 4010 +1 % -1 % -0,25 % +0,25% q p 0 price increase (1 % up) price reduction (1 % down) A B C increase of quantity (x % up) reduction of quantity (y % less) <?page no="40"?> 1 Microeconomics, competition and strategic behaviour 40 these effects are calculated starting at an initial price 𝑝𝑝 0 = 20 : triggered by a 1% price change the quantity changes by 0.25% each. The price elasticity of demand 𝜀𝜀 𝑝𝑝 describes the percentage change in the quantity demanded as a result of a 1% price change. It can be expressed as the finite difference between the quantity demanded ∆𝑞𝑞 and prices ∆𝑝𝑝 in relation to the original level of quantity 𝑞𝑞 and price 𝑝𝑝 (1.8) 𝜀𝜀 𝑝𝑝 = 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑚𝑚𝑝𝑝𝑝𝑝 𝑝𝑝ℎ𝑚𝑚𝑝𝑝𝑝𝑝𝑝𝑝 𝑖𝑖𝑝𝑝 𝑞𝑞𝑞𝑞𝑚𝑚𝑝𝑝𝑝𝑝𝑖𝑖𝑝𝑝𝑞𝑞 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑚𝑚𝑝𝑝𝑝𝑝 𝑝𝑝ℎ𝑚𝑚𝑝𝑝𝑝𝑝𝑝𝑝 𝑖𝑖𝑝𝑝 𝑝𝑝𝑝𝑝𝑖𝑖𝑝𝑝𝑝𝑝 = ∆𝑞𝑞/ 𝑞𝑞 ∆𝑝𝑝/ 𝑝𝑝 = 𝑝𝑝𝑞𝑞 ∆𝑞𝑞 ∆𝑝𝑝 or, for ∆→ 0, as a differential as (1.9) 𝜀𝜀 𝑝𝑝 = 𝑝𝑝𝑞𝑞 𝜕𝜕𝑞𝑞 𝜕𝜕𝑝𝑝 . Elasticities in general measure a marginal percentage change (the effect) of a change in one variable in the denominator on another variable in the numerator - here, a change in price and its effect on the quantity demanded. The price elasticity in the previous example is (1.10) 𝜀𝜀 𝑝𝑝(20→20.2) = −10/ 4000 0.2/ 20 = 20 4000 −10 0.2 = −0.25 and (1.11) 𝜀𝜀 𝑝𝑝(20→19.8) = 10/ 4000 −0.2/ 20 = 20 4000 10 −0.2 = −0.25, i.e., values are both negative, but equal. The reason for the negative sign is the downward slope of the demand curve - in this case, price changes always result in quantity effects that go in the opposite direction. The fact that the effects are equally large is due to the fact that the starting point of the price change is the same. This can be clearly seen in ► Figure 1.12 - here price elasticities are calculated for the demand function (1.7) at different positions of the demand function (1.7). The price elasticity of demand approaches zero for very small prices: a percentage price change has almost no effect on the quantity demanded. Conversely, the price elasticity of demand approaches −∞ , if prices are close to the maximum willingness to pay of customers. Figure 1.12: Price elasticity alongside a linear demand curve. p inelastic elastic q 0 5000 100 Δp ΔQ ε p → 0 ε p → - ∞ ε p = - 1 50 2500 ε p = - 1,5 2000 60 20 4000 ε p = - 0,25 <?page no="41"?> 1.3 Price elasticity and marginal revenues 41 Apparently, price elasticity can take two different forms. Inelastic demand - if the price elasticity of demand is between 0 and −1, a price change leads to a less than proportional effect on the quantity demanded: the percentage change in price is larger than the percentage change in quantity. Elastic demand - if the price elasticity of demand is between −1 and minus infinity, a price change leads to a less than proportionate effect on the quantity demanded: the percentage change in price is smaller than the percentage change in quantity. A firm can use information on the level of price elasticity of demand to estimate the extent to which demand changes if a firm increases or decreases prices. To do this, equation (1.9) can be rearranged to (1.12) 𝜕𝜕𝑞𝑞 𝑞𝑞 = 𝜀𝜀 𝑝𝑝 ⋅ 𝜕𝜕𝑝𝑝 𝑝𝑝 , i.e., the percentage change in quantity 𝜕𝜕𝑞𝑞/ 𝑞𝑞 results from multiplying the percentage change in price 𝜕𝜕𝑝𝑝/ 𝑝𝑝 with price elasticity of demand 𝜀𝜀 𝑝𝑝 (assumed to be constant here). If the price elasticity is 𝜀𝜀 𝑝𝑝 = −4, then a planned price reduction of 𝜕𝜕𝑝𝑝/ 𝑝𝑝 = −5% results in (1.13) 𝜕𝜕𝑞𝑞 𝑞𝑞 = 𝜀𝜀 𝑝𝑝 ⋅ 𝜕𝜕𝑝𝑝 𝑝𝑝 = −4 ⋅ −5% = 20% , i.e., quantity demanded will increase by 20%. Accordingly, a price increase of 3.5% would result in a 14% drop in quantity due to 𝜕𝜕𝑞𝑞 ⁄ 𝑞𝑞 = −4 ⋅ 3.5% = −14% . In general, price elasticity of demand can be measured given a demand function (1.14) 𝑝𝑝 = 𝑎𝑎 − 𝑏𝑏𝑞𝑞 using a total differential (i.e., the differentiation with respect to all variables) as (1.15) 𝑑𝑑𝑝𝑝 = 𝑑𝑑𝑎𝑎 − 𝑏𝑏𝑑𝑑𝑞𝑞 − 𝑞𝑞𝑑𝑑𝑏𝑏 . If the maximum willingness to pay and the size of the market remain unchanged, so that 𝑑𝑑𝑎𝑎 = 0 and 𝑑𝑑𝑏𝑏 = 0, equation (1.15) simplifies into (1.16) 𝑑𝑑𝑝𝑝 = −𝑏𝑏𝑑𝑑𝑞𝑞 . The absolute effect of a price change on the quantity demanded is then (1.17) 𝜕𝜕𝑞𝑞 𝜕𝜕𝑝𝑝 = − 1 𝑏𝑏 , so that, based on a given price-quantity combination 𝑝𝑝/ 𝑄𝑄 , price elasticity is (1.18) 𝜀𝜀 𝑝𝑝 = 𝑝𝑝 𝑄𝑄 𝜕𝜕𝑞𝑞 𝜕𝜕𝑝𝑝 = − 1 𝑏𝑏 𝑝𝑝 𝑄𝑄 . So, given a linear demand function, the price elasticity is determined by the inverse slope 1/ 𝑏𝑏 and the actual price-quantity combination 𝑝𝑝/ 𝑄𝑄 . Determining factors and empirical values of price elasticity Empirically, the price elasticity of demand is determined mainly by four variables - these should be analysed carefully before implementing a new pricing strategy. <?page no="42"?> 1 Microeconomics, competition and strategic behaviour 42 Urgency of demand - if a product (factual or perceived by the customer) is urgently needed, then a price increase changes demand only slightly or not at all. The higher the urgency, the lower the price elasticity of demand. This is particularly true for essential products such as water or medicines, but also for products on which people are dependent or even addicted: drugs, cigarettes and alcohol. If, for example, a customer is dependent on Wi-Fi or broadband access, she will not change demand significantly, if the price is reduced. On the other hand, if prices rise, consumption will only decrease by a small amount or not at all. Options for substitution and degree of product differentiation - if customers can easily and quickly switch to an alternative product (offered by another firm or from another product category) in the event of a price increase, then price changes will have a major impact on the quantity demanded. The more extensive the substitution possibilities, the higher the price elasticity of demand. For a coffee producer the range to determine prices is relatively small; customers either switch to other brands or manufacturers or drink tea instead, although the effect may be delayed. If, however, there is a high degree of product differentiation, for example because customers have a high level of brand loyalty, firms can use low price elasticity to raise prices: Apple can, even with a small market share, enforce very high prices for smartphones without losing customers to cheaper providers. Share of wallet - if only very small amounts of income are spent on a specific product, price changes have very little effect on quantity demanded. For example, a price increase for a smartphone app is price-inelastic and will hardly change the quantity demanded. On the other hand, for products or services on which a high proportion of income is spent (rent for a flat, taxes, clothing, or food) price rises will lead to stronger and lasting adjustments in demand and, if necessary, lifestyle. With permanently rising rents, for example, people leave booming cities and look for cheaper flats in other areas; with rising income tax, people move to regions with lower regional tax. Competitive intensity and marketing expenditures - the higher the competitive intensity of a market (e.g., because of a larger number of competing firms), the higher the price elasticity of demand. Although firms can highlight special features of their own product, particularly through marketing expenditures, to reduce price elasticity, this effect is all the smaller, the higher the intensity of competition is. These four determinants also drive the numbers in ► Table 1.2. Of course, the empirical studies are only comparable to a limited extent, since data comes from different countries and different periods of time. For example, business flights are more urgent than private holiday flights; visiting a paediatrician in the US has - unlike medication - no alternative; it seems people cannot live without mobile phones or broadband connections; addiction to cigarettes is higher than to marijuana or beer; and in the short term fuels can hardly be replaced, in the medium term they will be more so. <?page no="43"?> 1.3 Price elasticity and marginal revenues 43 Empirical values of price eleasticity of demand personal computer (long run) -2,74 high price elasticity numerous substitutes and low urgency soft drinks -2,59 personal computer (short run) -2,17 tooth paste -2,00 air travel (leisure and holiday) -1,90 beer -1,20 breakfast cereals -1,14 Marihuana -1,00 cinema -0,90 low price elasticity few substitutes and high urgency air travel (business) -0,80 medicine and drugs -0,68 groceries -0,63 broadband access -0,43 mobile access -0,41 petrol (long run) -0,31 cigarettes -0,30 petrol (short run) -0,09 Table 1.2: Empirical price elasticities (data sources: Bijmolt et al. 2005, Goldman and Grossman 1978, Hoch et al 1995, Havranek et al. 2012, Huang and Lin 2000, Farelly and Bray 1998 as well as Deaton 1990). In a meta study of 1851 product markets, Bijmolt et al. (2005) have calculated an average value of the price elasticity of 𝜀𝜀 𝑝𝑝 = −2.62, with 50% of the effects lying between −3 and −1, and 81% of all values between 0 and −4 . As a simple rule of thumb this implies that a 1% price reduction increases demand by 2.62% on average, in 50% of the cases the effect will be between 1% and 3% increase in demand. From a management perspective, this means that values of this magnitude should also typically be found in business cases (the financial planning of a new product, business model, or project) or strategic mid-term planning. If this is not the case, but planning consists e.g. of a 12% increase in quantity in line with a simultaneous price increase of 4% (corresponding to a positive price elasticity 𝜀𝜀 𝑝𝑝 = 12% 4% = +3.0 ), there must either be a striking other explanation (a drastic shift of the demand curve to the top right or a market exit of competitors), or this planning may not unfold as intended. <?page no="44"?> 1 Microeconomics, competition and strategic behaviour 44 Price elasticity of demand is determined not only by the current price, but in particular by the slope of the demand curve. The steeper the demand curve, the weaker quantity demanded reacts to price changes. ► Figure 1.13 illustrates this with two extreme demand curves. For an almost horizontal demand curve (left), the price elasticity converges towards minus infinity: a small price reduction would give rise to a strong increase in quantity; conversely, a price increase would be associated with a strong decrease in demand due to the low urgency of the product. For an almost vertical demand curve (right), price elasticity tends towards zero: a price reduction has almost no effect on quantity; accordingly, a price increase can be applied very effectively without a significant decrease in quantity demanded. The smaller the price elasticity of demand, the greater the willingness of customers to pay and the larger the range for firms to set prices. This also applies in relation to the competitive situation. If there are many competitors, the demand curve is very flat, so that in extreme cases only one price is possible in the market and a pricing strategy is unnecessary. If, on the other hand, the number of competitors is small, the demand curve is steep - firms can now define pricing strategies very effectively (see ► Chapter 7 and ► Chapter 8 for further details). Figure 1.13: Effects of a price reduction with high and low price elasticity. The price elasticity of demand is of course not constant over time. In many markets, price elasticity increases over time: during a product life cycle, customers become familiar with alternative products, new firms offer substitutes and often intensity of competition increases. So, as a matter of fact or just by perception of customers, the product appears after some time less innovative or fancy, so that willingness to pay decreases and price elasticity increases. Firms, therefore, use marketing in particular to emphasize the uniqueness of the product and to increase the urgency for the customer, i.e., to reduce the price elasticity of demand over and over in order to ultimately be able to impose higher prices. Empirical studies show, for example, that the more up-to-date a product is, the lower the price elasticity and the higher the brandor model-specific marketing expenditures. Shortly after q p 0 high price elasticity q p 0 low price elasticity <?page no="45"?> 1.3 Price elasticity and marginal revenues 45 market introduction, early adopters with low price elasticity and high willingness to pay buy at relatively high prices. Furthermore, whenever brand loyalty of existing customers is high, it can be observed that although price elasticity is lower, price changes (e.g., for coffee) lead to significantly stronger reactions of quantity demanded. Numerous studies have shown that price elasticity, despite severe strategic initiatives by firms, halves within four years after the introduction of a new product. However, price elasticity can rise again in later stages due to increasing product differentiation and stronger customer loyalty (Bijmolt et al. 2005, Berry et al. 2004, Krishnamurthi and Raj 1991, Simon 1979, Vakratsas and Ambler 1999 as well as Tellis 1988). Revenues and marginal revenues The key importance of price elasticity of demand for decision-making becomes clear once revenues are included in the analysis. Revenues 𝑅𝑅 of a firm are generally obtained by multiplying price 𝑝𝑝 and quantity 𝑞𝑞 as (1.19) 𝑅𝑅 = 𝑝𝑝𝑞𝑞 . For values given in the demand function (1.7) in ► Figure 1.11 on the right, revenues are (1.20) 𝑅𝑅 𝑝𝑝=20.0 = 20.0 ⋅ 4,000 = 80,000 (1.21) 𝑅𝑅 𝑝𝑝=19.8 = 19.8 ⋅ 4,010 = 79,398 and (1.22) 𝑅𝑅 𝑝𝑝=20.2 = 20.2 ⋅ 3,990 = 80,598 . As shown above, price changes of 1% have symmetrical effects on quantities, but not on revenues - obviously, in this example, a price reduction increases revenues by 0.7475% , while a price increase reduces revenues by −0.7525% . Once more, the reason can be found looking at the price elasticity of demand. Inelastic demand - if price elasticity of demand is between 0 and −1, a price change leads to a disproportionately small effect on the quantity demanded. In this case, revenues increase when prices are increased and a firm is able to increase revenues by increasing prices. Elastic demand - if the price elasticity of demand is between −1 and minus infinity, a price change leads to a more than proportionate effect on quantity demanded. In this case, revenues increase when prices are reduced and a firm is able to increase revenues by reducing prices. Analytically, this effect can be demonstrated by means of the revenue function (1.23) 𝑅𝑅 = 𝑝𝑝(𝑞𝑞)𝑞𝑞 . In order to generally determine a change in revenues as a function of prices and quantities, one can look at marginal revenue 𝑴𝑴𝑴𝑴 - marginal revenue is the additional revenue generated by increasing the quantity 𝑞𝑞 sold minimally. This value can be positive or negative because price and quantity are interdependent alongside the demand curve. Marginal revenue for the demand function (1.14) is obtained by first differentiating the revenue function with respect to quantity as (1.24) 𝜕𝜕𝜕𝜕 𝜕𝜕𝑞𝑞 = 𝑀𝑀𝑅𝑅 = 𝑝𝑝(𝑞𝑞) + 𝜕𝜕𝑝𝑝 𝜕𝜕𝑞𝑞 𝑞𝑞 . <?page no="46"?> 1 Microeconomics, competition and strategic behaviour 46 Extending this equation, we obtain (1.25) 𝜕𝜕𝜕𝜕 𝜕𝜕𝑞𝑞 = 𝑝𝑝 + 𝑝𝑝𝑝𝑝 𝑞𝑞𝑞𝑞 𝜕𝜕𝑝𝑝 𝜕𝜕𝑞𝑞 𝑞𝑞 = 𝑝𝑝 + 𝑝𝑝𝑞𝑞 1 𝜀𝜀𝑝𝑝 𝑞𝑞 = 𝑝𝑝 �1 + 1 𝜀𝜀𝑝𝑝 � , where (1.26) 𝑝𝑝 �1 + 1 𝜀𝜀𝑝𝑝 � > 0 for 𝜀𝜀 𝑝𝑝 < −1 and (1.27) 𝑝𝑝 �1 + 1 𝜀𝜀𝑝𝑝 � < 0 for 𝜀𝜀 𝑝𝑝 > −1 . Marginal revenue is positive if 𝜀𝜀 𝑝𝑝 < −1, so revenues of a firm increase in the event of a price increase exactly if the demand function is inelastic. If, on the other hand, 𝜀𝜀 𝑝𝑝 > −1, then revenues decrease in case of a price increase due to elastic demand. Furthermore, marginal revenue is equal to price if, as can be seen from (1.25), price elasticity of demand is heading towards -∞. Figure 1.14: Revenues, demand curve and marginal revenues. inelastic elastic p, R in EUR 0 5000 100 31 R = p q = 100 q - 0,02q 2 30 57 56 q D inelastic elastic p, MR. R in EUR 0 5000 100 q R MR> 0 MR<0 D <?page no="47"?> 1.3 Price elasticity and marginal revenues 47 ► Figure 1.14 shows the demand curve and revenue curve for a demand function (1.7). Because of the quadratic term in this revenue function (1.28) 𝑅𝑅 = 𝑝𝑝(𝑞𝑞)𝑞𝑞 = (𝑎𝑎 − 𝑏𝑏𝑞𝑞)𝑞𝑞 = 𝑎𝑎𝑞𝑞 − 𝑏𝑏𝑞𝑞 2 = 100𝑞𝑞 − 0.02𝑞𝑞 2 , each revenue curve of a linear demand curve is an inverted parabola. In the elastic area of the demand curve, revenues rise when quantities are increased: if prices are reduced, for example, from EUR 57 to EUR 56, quantities and revenues rise - a firm achieves higher revenues as a result of the price reduction. In the inelastic area of the demand curve, revenues decline when quantities increase: if prices are decreased, for example, from EUR 31 to EUR 30, quantities also increase, but the price effect dominates the quantity effect on revenues - revenues decline. Marginal revenues for a revenue function (1.28) are calculated as (1.29) 𝜕𝜕𝜕𝜕 𝜕𝜕𝑞𝑞 = 𝑎𝑎 − 2𝑏𝑏 = 100 − 0.04𝑞𝑞 , this means that a marginal revenue function 𝑀𝑀𝑅𝑅 is linear with exactly half the slope of the demand function and corresponds to the slope of the revenue function for each quantity. ► Figure 1.14 bottom shows marginal revenues, revenues and the corresponding demand function. Revenues are maximised if a price-quantity combination is chosen where marginal revenue is 𝑀𝑀𝑅𝑅 = 0 and price elasticity of demand is 𝜀𝜀 𝑝𝑝 = −1 : no change in prices or quantities can then increase revenues. This leads to a simple rule of thumb, especially for sales departments, for increasing sales: a price reduction leads to an increase in revenue if price elasticity 𝜺𝜺 𝒑𝒑 < −𝟏𝟏 , and vice versa. Whether this is actually a good strategy depends on the competitive environment and cost situation of the firm (► Chapters 7, 8 and 10). Marginal revenue can be broken down into two effects, as shown in ► Figure 1.15. Before a price reduction the revenue on the left is 𝑅𝑅 1 = 𝑝𝑝 1 𝑞𝑞 1 and equals area 𝐴𝐴 + 𝐵𝐵 . With a price reduction from 𝑝𝑝 1 to 𝑝𝑝 2 the quantity increases from 𝑞𝑞 1 to 𝑞𝑞 2 and the new revenue is 𝑅𝑅 2 = 𝑝𝑝 2 𝑞𝑞 2 and equals area 𝐵𝐵 + 𝐶𝐶 . The revenue decreases by −𝐴𝐴 due to the lower price (price effect) but increases by +𝐶𝐶 due to increasing quantity (quantity effect), so that the total effect as marginal revenues are 𝑀𝑀𝑅𝑅 12 = −𝐴𝐴 + 𝐶𝐶 > 0 is positive, because the price change takes place in the elastic part of the demand curve. Figure 1.15: Marginal revenues as gains and losses of revenues due to price changes. p, R, in EUR 0 a/ b a A B C p 1 p 2 q 1 q 2 q p, R, in EUR 0 a/ b a A B C p 3 p 4 q 3 q 4 q <?page no="48"?> 1 Microeconomics, competition and strategic behaviour 48 ► Figure 1.15 on the right shows the case where marginal revenues 𝑀𝑀𝑅𝑅 34 = −𝐴𝐴 + 𝐶𝐶 < 0 are negative: the revenue loss −𝐴𝐴 from a price reduction from 𝑝𝑝 3 to 𝑝𝑝 4 dominates the effect of a revenue increase +𝐶𝐶 due to the increase in quantity from 𝑞𝑞 3 to 𝑞𝑞 4 , since the price reduction takes place in the inelastic range of the demand function. New revenues 𝑅𝑅 4 = 𝑝𝑝 4 𝑞𝑞 4 < 𝑝𝑝 3 𝑞𝑞 3 = 𝑅𝑅 3 are lower than previous revenues 𝑅𝑅 3 . 1.4 Summary and key learnings Why is microeconomics relevant for managers? Microeconomics attempts to explain the causes of observable behaviour and decisions of human beings in economic situations in order to specify effects on markets, firms and competition. Microeconomic models provide guidance for future decisions and a classification of competitive situations, especially from a management perspective. Microeconomics analyses the decisions of customers and firms, their interaction and how markets work using empirical methods (econometrics, laboratory and field experiments) and theoretical models. In addition to providing a simplified mapping of reality, models are necessary to decide in otherwise undecidable situations. In this sense, microeconomics provides future managers with city maps for competition and markets: using empirically robust, albeit abstractly modelled frameworks, we add decidability to a specific market or competitive situation. At the heart of analysis are markets and the interplay between supply and demand. Markets are institutions (systems, rules, patterns and structures) in which transactions between market participants happen. A market equilibrium, which is determined by an equilibrium price, is the only price-quantity combination where quantity demanded exactly equals quantity supplied. Whether and how quickly this actually occurs depends on the exchange of information, the competitive dynamics in the market and the learning processes of market participants. Different market mechanisms, information asymmetry, the sequence of transactions, and the level of transaction costs can lead to very different market outcomes - i.e., different prices, different quantities and different revenues. The demand side of a market can be described as price-quantity combinations along a demand function, which is empirically identical to the willingness to pay of potential and actual customers. The price elasticity of demand describes the effect of a price change on the quantity demanded. Price elasticity and marginal revenues are key management tools in the analysis of demand and to develop pricing and sales strategies: if marginal revenue is positive and/ or the price elasticity is less than −1, a firm can increase revenues by reducing prices; if marginal revenue is negative, the firm should increase prices to increase revenues. Recommendations for further reading For a broad introduction to how markets and institutions work, see Voigt, S., Institutional Economics, Cambridge 2019. Excellent textbooks that go well beyond the content presented in this book are Belleflamme, P. and Peitz, M., Industrial organisation: markets and strategies, London 2015, with a strong analytical approach, and Mas-Colell, A. Winston, M.D. and Green, J.R., Microeconomic Theory, New York 1995, for those who want to go deeper into mathematics. <?page no="49"?> 1.4 Summary and key learnings 49 Questions for review [1] Provide a precise definition for microeconomics. Why is microeconomics a foundation for all business and management courses? [2] What are the advantages and disadvantages of economic models? What role do stylised facts play? How are microeconomic models developed? Explain why solutions developed in markets are often superior to government solutions. [3] What is the typical shape of supply and demand curves? Which factors determine the course of supply and demand curves? [4] What is an equilibrium market price? Refer to products whose prices are in equilibrium or not in equilibrium. Explain why it could be that markets for used cars do not work well. [5] Find market results for market equilibrium, horizontal and vertical transactions for the water bottle market from the case study with eight market participants each, willingness to pay in 1 EUR steps from 1 EUR to 8 EUR and potential offers in 1 EUR steps from 3 EUR to 10 EUR. Explain your results. [6] Describe the concept of price elasticity of demand. What does it depend on, are there differences between the short and long run? How can you empirically determine the price elasticity of demand of a product or service? [7] Explain the concept of marginal revenue and refer to the price and quantity effect. Which rule of thumb can be derived from the analysis of marginal revenue from a management perspective? [8] Given a price elasticity of -1.3, an ice cream seller increases prices by 4% - what happens qualitatively and quantitatively to her revenues? Also determine the price and quantity effect of marginal revenues. [9] Determine the demand function, the price elasticity of demand for bread rolls of a bakery on the main shopping street in your town as well as revenues and marginal revenues. [10] You are the director of the Guggenheim Museum in Bilbao. Currently the entrance fees are 20 EUR, you have 400,000 visitors per year, revenues are 8 million EUR per year. You know the demand function from market research as p = 100 - 0.0002 q (where q is the number of visitors). The Guggenheim Foundation in New York demands higher sales revenues from you and presses you to make a quick decision - should you increase or decrease prices? Answer the question using the price elasticity of demand and a figure. What would be the optimal price to maximise revenues? Literature Abbott, B.P. et al., Observation of gravitational waves from a binary black hole merger, Physical Review Letters, 2016, 116, 6, 61-102. Acemoglu, D. and Restrepo, P., The race between man and machine: implications of technology for growth, factor shares, and employment, American Economic Review, 2018, 108, 6, 1488-1542. Acemoglu, D. and Robinson, J., Why nations fail: the origins of power, prosperity and poverty, New York, 2012. Akerlof, G., The market for ‘lemons’: quality uncertainty and the market mechanism, Quarterly Journal of Economics, 1970, 84, 3, 488-500. <?page no="50"?> 1 Microeconomics, competition and strategic behaviour 50 Arthur, W.B., Inductive reasoning and bounded rationality, American Economic Review, 1994, 84, 2, 406-411. Barham, P., The science of cooking, New York 2001. Berry, S., Levinsohn, J. and Pakes, A., Differentiated products demand systems from a combination of micro and macro data: the new car market, Journal of Political Economy, 2004, 112, 1, 68-105. Bijmolt, T.H., Heerde, H.J.V. and Pieters, R.G., New empirical generalizations on the determinants of price elasticity, Journal of Marketing Research, 2005, 42, 2, 141-156. Blaug, M., The methodology of economics: or, how economists explain, Boston 1992. Boardman, A.E. and Vining, A.R., Ownership and performance in competitive environments: a comparison of the performance of private, mixed, and state-owned enterprises, Journal of Law and Economics, 1989, 32, 1, 1-33. Brynjolfsson, E., Rock, D. and Syverson, C., Artificial intelligence and the modern productivity paradox: a clash of expectations and statistics, WP 24001, National Bureau of Economic Research, Cambridge 2017. Conlisk, J., Why bounded rationality? . Journal of Economic Literature, 1996, 34, 2, 669-700. Currie, J., Kleven, H. and Zwiers, E., Technology and big data are changing economics: mining text to track methods, WP 26715, National Bureau of Economic Research, Cambridge 2020. Davis, D.D. and Holt, C.A., Experimental economics, Princeton 1993. Davis, G. and Pecar, B., Business statistics using Excel, 2. Auflage, Oxford 2013. De Martino, B., Kumaran, D., Seymour, B. and Dolan, R.J., Frames, biases, and rational decision-making in the human brain, Science, 2006, 313, 5787, 684-687. Deaton, A., Price elasticities from survey data, Journal of Econometrics, 1990, 44, 281-309. Dewenter, K.L. and Malatesta, P.H., State-owned and privately owned firms: an empirical analysis of profitability, leverage, and labor intensity, American Economic Review, 2001, 91, 1, 320-334. Einstein, A., Über Gravitationswellen, Königlich-Preußische Akademie der Wissenschaften, Sitzungsberichte, 1918, Mitteilung vom 31. Januar 1918, 154-167. Frank, R.H., The economic naturalist: why economics explains almost everything, London 2008. Fritsch, M., Marktversagen und Wirtschaftspolitik - mikroökonomische Grundlagen staatlichen Handelns, München 2018. Fuest, C., Neodirigismus, FAZ, 7. Februar 2020, 16. Gneezy, U. and List, J.A., Putting behavioral economics to work: testing for gift exchange in labor markets using field experiments, Econometrica, 2006, 74, 5, 1365-1384. Goldman, F. and Grossman, M., The demand for pediatric care: an hedonic approach, Journal of Political Economy, 1978, 86, 2, 259-280. Havranek, T., Irsova, Z. and Janda, K., Demand for gasoline is more price-inelastic than commonly thought, Energy Economics, 2012, 34, 1, 201-207. Hoch, S.J., Kim, B.-D., Montgomery, A.L. and Rossi, P.E., Determinants of store-level price elasticity, Journal of Marketing Research, 1995, 32, 1, 17-29. Huang, K.S. and Lin, B.H., Estimation of food demand and nutrient elasticities from household survey data, US Department of Agriculture, 2000, Economic Research Service. Kagel, J.H. and Roth, A.E. (eds.), The Handbook of Experimental Economics, Princeton 2016. Kaldor, N., Capital accumulation and economic growth, in: Lutz, F.A. and Hague, D.C. (eds.), The theory of capital, London 1961, 177-222. Kennedy, P., A guide to econometrics, 6. Auflage, Boston 2008. Kolmar, M., Principles of microeconomics - an integrative approach, Berlin 2017. Krishnamurthi, L. and Raj, S.P., An empirical analysis of the relationship between brand loyalty and consumer price elasticity, Marketing Science, 1991, 10, 2, 172-183. <?page no="51"?> 1.4 Summary and key learnings 51 Levitt, S.D. and List, J.A., Field experiments in economics: the past, the present, and the future, European Economic Review, 2009, 53, 1, 1-18. List, J.A and Reiley, D., Field experiments in economics, in: Durlauf, S.N. and Blume, L.E. (eds.), The New Palgrave Dictionary of Economics, 2. Auflage, New York 2007. Loebbecke, C. and Picot, A., Reflections on societal and business model transformation arising from digitization and big data analytics: a research agenda, Journal of Strategic Information Systems, 2015, 24, 3, 149-157. Meyer, M., Operations Research - Systemforschung: eine Einführung in die praktische Bedeutung, München 1996. Mihet, R. and Philippon, T., Disruptive innovation in business and finance in the digital world, International Finance Review, 2019, 20, 29-43. Münter, M.T., Empirical facts, economic theory and everyday experience in industrial organization - three sides of the same story? , in: Mueller, D.C., Haid, A. and Weigand, J. (eds.), Competition, Efficiency, and Welfare, Boston 1999, 279-297. Pearl, J. and Mackenzie, D., The book of why - the new science of cause and effect, New York 2018. Popper, K., Logik der Forschung - zur Erkenntnistheorie der modernen Naturwissenschaft, Wien 1934. Preda, A., Information, knowledge and economic life - an introduction to the sociology of markets, Oxford 2009. Prince, J.T., Repeat purchase amid rapid quality improvement: structural estimation of demand for personal computers, Journal of Economics & Management Strategy, 2008, 17, 1, 1-33. Simon, H., Dynamics of price elasticity and brand life cycles: an empirical study, Journal of Marketing Research, 1979, 16, 4, 439-452. Simon, H.A., Decision-making: rational, nonrational, and irrational, Educational Administration Quarterly, 1993, 29, 3, 392-411. Smith, A., An inquiry into the nature and causes of the wealth of nations, London 1776. Smith, V.L., Constructivist and ecological rationality in economics, American Economic Review, 2003, 93, 3, 465-508. Spiegelhalter, D., The art of statistics - learning from data, London 2019. Sugden, R., Spontaneous order, Journal of Economic Perspectives, 1989, 3, 4, 85-97. Taleb, N.N., The black swan, New York 2007. Tellis, G.J., The price elasticity of selective demand: a meta-analysis of econometric models of sales, Journal of Marketing Research, 1988, 25, 331-341. Vakratsas, D. and Ambler, T., How advertising works: what do we really know? , Journal of Marketing, 1999, 63, 1, 26-43. Voigt, S., Institutional economics, Cambridge 2019. Von Hayek, F.A., Der Wettbewerb als Entdeckungsverfahren, Institut für Weltwirtschaft, Kiel 1968. Von Hayek, F.A., Die Anmaßung von Wissen, ORDO: Jahrbuch für die Ordnung von Wirtschaft und Gesellschaft, 1975, 26, 12-21. <?page no="53"?> 53 2 Customer behaviour, products and network effects When a customer looks for a travel guide for China on a platform such as Amazon, after clicking on the search result the customer is also shown a line of products labelled “customers who bought this item also bought” a phrase book for Chinese, a picture dictionary, city guides for Beijing and Shanghai and a book on rules on appropriate behaviour in China. However, almost every visitor to the Amazon website receives different product offerings - the selection of suggested products and the price level is personalised for each potential customer. In order to accurately suggest the right products, Amazon analyses numerous parameters such as previous customer behaviour (i.e., both purchasing and non-purchasing decisions), internet usage in general, browsing and search behaviour via cookies. Amazon then draws conclusions about the customer’s preferences and tastes based on social status, age and other demographic details from IP addresses, location and movement data. In addition, Amazon will take into account the customer’s income based on selected payment methods and credit checks. From a management perspective, numerous decisions can be derived from such information: Product range and portfolio - to be able to offer appropriate products to individual customers or customer groups in sales and marketing. Prices of individual products and pricing structure - in order to be able to determine the right price for each customer or market segment or the pricing structure across several products or to be able to estimate the repercussions of price changes of competitors’ products on one’s own demand. Market segments and product differentiation - in order to be able to either better address existing market segments based on vertical or horizontal product differentiation or to establish or delimit these segments through differentiation and pricing models (in marketing usually referred to as fencing). The predictive analytics algorithm used by Amazon is based, among other things, on the microeconomic rationale of maximising utility based on customer preferences (Seal 2016, Nichols 2013 and aws.Amazon.com/ en/ aml). In a similar way to Amazon, firms such as Netflix, Facebook and Google also aim to predict future customer behaviour in order to suggest the right products to each individual customer for future purchases. In particular, this becomes possible if customers disclose their respective preferences on social networks, in rating portals or through their browsing behaviour either directly or indirectly. In this ► Chapter 2 human decisions are considered and analysed as essentially rational. Regardless of this, customer behaviour and purchase decisions are of course often affected by routines, previous decisions, adaptation of other people’s behaviour or decisions and psychological influences. These aspects of boundedly rational behaviour are focused on in ► Chapter 3. <?page no="54"?> 2 Customer behaviour, products and network effects 54 Learning Objectives This chapter deals with: the nature of customer behaviour and demand decisions based on preferences and budget constraints; the definition of a market, possibilities of market delineation and product categories; characteristics of demand behaviour in direct and indirect network effects; and their repercussions on two-sided and multisided platforms. 2.1 Customer behaviour and decisions on demand The behaviour of customers and the decision to buy a product can be analysed from a number of perspectives, as shown in ► Figure 2.1. Of course, customer behaviour does not only include the consumption of a product, but extends to the expectation of use, the gathering of information, the execution and experience of the purchase, the actual use, ownership and possession, as well as memories, e.g. when travelling (Deaton 1992 and Kahneman and Tversky 2003). Moreover, a customer may not ask for a product or service per se, but for a mixture of product characteristics (Lancaster 1971). Obviously, customer behaviour has a temporal dimension on the one hand, and it may be strongly influenced or overlaid by emotions on the other. Behaviour of customers can essentially be explained by the desire to achieve utility (satisfaction) from the consumption or ownership of certain products. People do things that increase their utility and avoid things that reduce their utility, but must take into account the limitations of income and time. Utility can result from actual consumption or functional use of the product, but also from the appreciation of other people because of possession or ownership, from a feeling of belonging to a group or a certain lifestyle or attitude towards life connected with it, as well as from signalling certain values or attitudes (Solomon 2014 and e.g. Sinus Milieus of the Sinus Institute). Figure 2.1: Customer behaviour and influences. <?page no="55"?> 2.1 Customer behaviour and decisions on demand 55 It is precisely these interpersonal and socio-cultural influences that are increasingly shaping demand behaviour: in many developed economies, most basic necessities (food, housing, etc.) are sufficiently fulfilled in good quality for most people, so that observable demand behaviour is based more on the realisation of desires and one’s own self or the search for experiences than on the actual satisfaction of needs. Preferences Generally, customer behaviour is characterised by people’s will to survive and vital needs, but decisions are also influenced or determined to a very large extent by preferences. From a microeconomic perspective, preferences describe the likings (or dislikes) towards alternative products based on a ranking, which then enables a best possible choice out of various alternatives. The need to select and to choose between alternatives arises from scarcity: limited income and given prices, limited time for consumption, and possibly limited compatibility of products. Preferences are partly intra-personal - people are different, they have different preferences - but they are also endogenous: preferences are shaped and changed by marketing to build brand loyalty, by interpersonal comparisons in social media and social groups, through habits, by culture and markets or even by income and wealth situation (Bagwell 2007 and Bowles 1998). The phenomenon "keeping up with the Joneses" describes that the comparison, the imitation and the desire for social affiliation and at the same time differentiation between customers in a social group or neighbourhood, is a major determinant of consumption patterns. Alfred Sloan, former CEO of the General Motors Corporation, derived a strategy for product differentiation in the automobile industry from this in the 1930s, which is still working today (at least in Germany): an accurate estimate of the income and social status of the resident population can be made by analysing the type and value of cars parked in a residential area (Gali 1994 and Ghemawat 2002). Preferences are not only shaped by external influences, but also intrapersonal by one’s own behaviour. From a neuroscientific perspective, it can be seen that although individual preferences are random in their basic disposition, they are formed, shaped and consolidated over time, especially through one’s own decisions and consumption - people prefer what they own or what they have decided to do. Especially in the case of important decisions (for a subject and location of study, for a job, for a partner), choices between objectively equivalent alternatives can lead to changes in preferences over time. As a result, the alternatives already chosen are assessed more positively, the rejected alternatives are devalued. This phenomenon is often interpreted as a consequence of the reduction of cognitive dissonance after a decision, which is triggered by the discrepancy between initial preferences and decision results. However, neuroscientific analyses using MRI observations of the brain also show that preferences are already updated during the decision-making process (Voigt et al. 2019). Experiments also show that individual preferences are generally not stable over time - neither in the shortnor the long-term (see ► Chapter 3). <?page no="56"?> 2 Customer behaviour, products and network effects 56 Rational decisions, utility and marginal utility An unambiguous and rational decision based on given preferences with limited income and given prices is possible if the following requirements are met. Complete preferences - in order to be able to make a decision, a customer must be able to create a definite ranking of all products and their possible combinations. Furthermore, this ranking must be transitive, i.e., in the case of smartphones, if Apple is preferred over Samsung and Samsung over Huawei, then Apple must also be preferred over Huawei. Maximising utility - to maximise utility, people must make decisions according to their preferences and given budget constraints (income and prices). They consider all relevant information, utility, or costs within the framework of a completely rational cost-benefit analysis and thus, at least try to increase or maximise their utility (satisfaction) through their chosen consumption behaviour. Opportunity costs - apart from the explicit and obvious costs, there are often invisible or implicit costs (lost utility) associated with the fact that a decision in favour of a product means that the best possible alternative product cannot be consumed. These costs arise because a specific sum of money has only one use and is therefore not available for other expenditures. The reasons for this are essentially limited time and income. Opportunity costs are not real costs to pay, but they are implicit in rational decisions and must always be taken into account. If a student decides to go to a restaurant and pays 30 EUR for dinner, the opportunity cost of going to a restaurant is made up of these costs and the lost benefit of the best alternative (tutoring one hour with earnings of 20 EUR) - and is 30 EUR plus 20 EUR equals 50 EUR. Sunk costs - many decisions are made against the background of expenses (costs) that occurred in the past and cannot be reversed or even recovered by a new decision. Sunk costs are real costs, but due to their specific use (e.g., last year's semester contribution) they are irrelevant for future decisions and must not be taken into account for a rational decision (see further ► Chapters 3 and 6). Case Study │ Deciding to study business or economics The four conditions for completely rational decisions can be illustrated clearly by the decision to study business or economics: Complete preferences - before applying at a university, a student first has to put all study programmes into a clear ranking order and then put all universities and business schools of the preferred study programme in a clear ranking order as well; Maximising utility - a student then chooses, within a given budget, a study programme and a place of study that maximises his or her future income, career opportunities, well-being and reputation; Opportunity costs - a student has weighed up all the opportunities that are not feasible in parallel (e.g., studying to become a dentist) when choosing a business administration programme and has taken these opportunity costs (in particular the possible loss of higher income) into account when making a decision; <?page no="57"?> 2.1 Customer behaviour and decisions on demand 57 Sunk Costs - if, contrary to expectations, a student in the sixth semester finds out shortly before submitting his or her Bachelor's thesis that the choice of study does not correspond to his or her preferences, he or she will immediately change the study programme, because previous investments in learning and tuition fees are sunk costs. It is obvious that many decisions, not only the choice of study programme, are made based on routines or without completely weighing up all alternatives - effects of bounded rationality on decision-making are considered in the following ► Chapter 3. Figure 2.2: Utility, income and prices. As shown in ► Figure 2.2, the relationship between purchasing decisions, income, prices and utility can be used to derive forecasts of purchasing decisions. For this purpose, it is necessary to make utility tangible: utility is a theoretical construct and cannot be directly observed or measured. However, it is possible to determine utility indirectly: customers can be asked about the change in utility whenever they consume a certain product (e.g., an additional piece of chocolate) whether the last piece tasted better/ just as good/ worse than the piece of chocolate they had eaten immediately before. For most people it is observed in experiments that, at least after a larger amount of chocolate, the satisfaction does not increase and the anticipated joy for the next piece decreases. This change in the utility if consumption of a product is marginally expanded is called marginal utility: it describes the additional utility from the last unit of a product consumed (Luce 2014). With a positive marginal utility, total utility increases; with a negative marginal utility, overall utility decreases. The top section of ► Figure 2.3 shows the marginal utility as a function of the two separate products - a microeconomics textbook and beer. For many people, repeated reading in microeconomics textbooks leads (hopefully) to increasing satisfaction due to increasing knowledge. However, as you reach a higher level of knowledge, marginal utility is decreasing: the degree preferences budget line given implicitly and explicitly model for explanation revealed behavior and preferences prediction of consumer decisions purchase decisions budget constraints indifference curves <?page no="58"?> 2 Customer behaviour, products and network effects 58 of increase in knowledge in the left example decreases, so that the utility still increases, but with a decreasing growth rate. In contrast, increasing consumption of beer above a certain quantity often leads to negative marginal utility. Typically, people stop consuming a product before the marginal utility becomes negative. Figure 2.3: Utility and marginal utility. Marginal utility describes the change in utility as a function of the quantity consumed - mathematically this is the first derivative of a utility function 𝑢𝑢(𝑞𝑞) on quantity 𝑞𝑞 , so that in general (2.1) 𝜕𝜕𝑞𝑞 𝜕𝜕𝑞𝑞 > 0 and 𝜕𝜕2𝑞𝑞 𝜕𝜕𝑞𝑞2 < 0 shows, that marginal utility is positive but decreases as consumption of a product increases - equivalent to utility that increases, yet at a decreasing rate. Since every person has different preferences, the individual utility functions, which generally differ - e.g., for two products leasing a flat W and partying P - is given by (2.2) 𝑢𝑢(𝑊𝑊, 𝑃𝑃) and can take various mathematical forms, e.g., 𝑢𝑢(𝑊𝑊, 𝑃𝑃) = 𝑊𝑊 + 𝑃𝑃 or 𝑢𝑢(𝑊𝑊, 𝑃𝑃) = 𝑊𝑊 0.7 𝑃𝑃 0.3 , that satisfy condition (2.1). utility microeconomics beer utility 0 0 marginal utility microeconomics marginal utility beer 0 0 <?page no="59"?> 2.1 Customer behaviour and decisions on demand 59 Marginal utility and utility from leasing a flat and partying In order to illustrate the connection between marginal utility and utility across several products, the decisions of one student are analysed. The student has a monthly disposable income (budget) of 800 𝐸𝐸𝐸𝐸𝑅𝑅 , of which 200 𝐸𝐸𝐸𝐸𝑅𝑅 are used for essential purchases at the grocery. The student spends the remaining 600 𝐸𝐸𝐸𝐸𝑅𝑅 entirely on the two products leasing a flat 𝑊𝑊 and partying 𝑃𝑃 . ► Figure 2.4 shows the products leasing a flat (depending on the number of square metres) and partying (going out in the evening per month). The student in question currently lives in a small flat of 25 square metres and goes out in the evening 16 times a month on average. Although the utility would increase with a larger flat as well as with more frequently going out, it would only increase less than proportionately. To see this, compare two situations (1) in case of a flat with 25 sqm, an increase of 10 sqm to 35 sqm would be significant, (2) in case the flat is currently 60 sqm, an increase to 70 sqm would definitely increase utility, but to a smaller extent. Figure 2.4: Utility and marginal utility for leasing a flat and partying. The first derivative of the respective utility functions (determined by observation or market research and econometrically estimated) then indicates the marginal utility of each product, as shown in ► Figure 2.4. u(sqm) increase of utility decrease of marginal utility 0 10 20 30 40 50 60 70 80 sqm example of a specific utility function MU(sqm) MU(parties) u(parties) increase of utility decrease of marginal utility 0 2 4 6 8 10 12 14 16 18 20 parties 0 10 20 30 40 50 60 70 80 sqm 0 2 4 6 8 10 12 14 16 18 20 parties u = u sqm = sqm 0,5 u = u party = party 0,7 <?page no="60"?> 2 Customer behaviour, products and network effects 60 Indifference curves and marginal rate of substitution The utility that the student gains from both products together can be determined by comparing possible product combinations called bundles. ► Figure 2.5 on the left shows any combination of flat sizes and partying per month. Some points can be easily compared and ranked: 𝐶𝐶′ is clearly better than 𝐴𝐴′ - with the same number of parties, the apartment is larger. In the same way, 𝐵𝐵′′ is clearly worse than 𝐴𝐴 - with the same number of square metres, you can party less often. However, some points - especially from the perspective of individual preferences - cannot be clearly ranked. It is more difficult to compare 𝐴𝐴′ with 𝐵𝐵′′ : in one case the flat is bigger but you can party less often; in the other case the flat is smaller but you can party more often. However, if you find points that have the same level of utility from an individual perspective, you are indifferent between the alternatives - you cannot make a decision because of the equal value of utility. Though, these points can be connected by means of an indifference curve. Along this indifference curve - shown in ► Figure 2.5 on the right - utility from combinations of the two products is equal. Indifference curves can either be determined by market research (people are asked about alternative combinations of products) or reconstructed based on actual decisions made and observed. Figure 2.5: Identification of an indifference curve. An indifference curve depicts the preferences of a customer and shows points of equal utility levels from the consumption of several products, analogous to an isobar in a weather forecast that describes points of equal air pressure. In this example, the student would be willing to move into a smaller flat with only 20 square metres, if she can now party 20 times a month instead of 16 times - and thus, trade five square metres into four times partying. ► Figure 2.6 top right shows that the trade-off ratio changes along the indifference curve: the more of one product is available, the more is traded against another product. The exchange along an indifference curve can be measured more precisely as the marginal rate of substitution, i.e., the flat size in sqm partying per month 20 25 40 12 20 16 10 B‘ A‘ C‘ A A‘‘ B‘‘ partying per month B C‘‘ 20 25 40 12 20 16 10 flat size in sqm B‘ A‘ C‘ A B‘‘ B A‘‘ C‘‘ <?page no="61"?> 2.1 Customer behaviour and decisions on demand 61 relative willingness to substitute one product for another. In order to identify this marginal rate of substitution, the total differential for a utility function 𝑢𝑢(𝑊𝑊, 𝑃𝑃) can be determined as (2.3) 𝑑𝑑𝑢𝑢 = 𝜕𝜕𝑞𝑞 𝜕𝜕𝜕𝜕 𝑑𝑑𝑊𝑊 + 𝜕𝜕𝑞𝑞 𝜕𝜕𝜕𝜕 𝑑𝑑𝑃𝑃 . Here 𝜕𝜕𝑢𝑢/ 𝜕𝜕𝑊𝑊 describes the marginal utility of leasing a flat and 𝑑𝑑𝑊𝑊 denotes the change in space due to a consumer decision, so that 𝜕𝜕𝑢𝑢/ 𝜕𝜕𝑊𝑊 𝑑𝑑𝑊𝑊 describes the change in utility. Since the utility is constant along an indifference curve, the absolute change described by the total differential is 𝑑𝑑𝑢𝑢 = 0, so (2.3) can be converted to (2.4) 𝑑𝑑𝜕𝜕 𝑑𝑑𝜕𝜕 = − 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 . The marginal rate of substitution as an exchange ratio 𝑑𝑑𝑊𝑊 ⁄ 𝑑𝑑𝑃𝑃 of the two products corresponds to the marginal utility ratio and measures the changes in the quantities of the two products at one point of an indifference curve: the willingness to give up something of a product depends on how large the relative marginal utilities of the products under consideration are. Figure 2.6: Utility functions and indifference curves. <?page no="62"?> 2 Customer behaviour, products and network effects 62 Indifference curves are convex towards the origin if a customer would not completely give up either of the two products. This in turn implies that customers prefer variety of consumption options. If the indifference curves were linear (and hence intersecting both axes), a customer would be willing to give up one of the two products in favour of the other: both products would then be perfect substitutes. ► Figure 2.6 bottom right shows that of course there is not just one but an infinite number of indifference curves for a customer. The further to the top right an indifference curve lies, the higher the level of utility. The indifference curve through the points 𝐶𝐶 , 𝐶𝐶′ and 𝐶𝐶′′ indicates a higher level of utility - this is immediately clear when comparing 𝐶𝐶′ and 𝐴𝐴 : the flat is the same size at 25 square metres, but the student can go for partying 20 times instead of 16 times - the question is now why she does not do so. Budget constraints and consumption possibilities This is due to limited disposable income and prices for leasing a flat and partying. The student's consumption possibilities are subject to a budget constraint, which results from her monthly income (budget) of 𝐼𝐼 = 800, the essential grocery shopping 𝑆𝑆 = 200 as well as the prices for renting at 𝑝𝑝 𝜕𝜕 = 8 per sqm and partying at 𝑝𝑝 𝜕𝜕 = 25 . Consumption possibilities then result from a budget constraint in the form of (2.3) 𝐼𝐼 = 𝑆𝑆 + 𝑝𝑝 𝜕𝜕 𝑊𝑊 + 𝑝𝑝 𝜕𝜕 𝑃𝑃 as (2.4) 800 = 200 + 8𝑊𝑊 + 25𝑃𝑃 or (2.5) 600 = 8𝑊𝑊 + 25𝑃𝑃 . The budget constraint in (2.5) shows that, after deducting expenditures 𝑆𝑆 of the grocery shopping, there are 𝐸𝐸𝐸𝐸𝑅𝑅 600 left, which can be spent on renting a flat and/ or partying according to individual preferences. For example, one can solve equation (2.5) for 𝑃𝑃 and obtain (2.6) 𝑃𝑃 = 600 25 − 8 25 𝑊𝑊 , i.e., showing that the maximum number of parties 𝑃𝑃 depends on the size of the flat 𝑊𝑊 . If the flat would not exist, 𝑊𝑊 = 0, then the student can obviously party given the available budget 𝑃𝑃 = 600/ 25 = 24 times per month. Conversely, equation (2.5) can be changed to (2.7) 𝑊𝑊 = 600 8 − 25 8 𝑃𝑃 , so that for 𝑃𝑃 = 0 a maximum flat size of 𝑊𝑊 = 600/ 8 = 75 sqm results. In addition, the slope of the budget line is determined by the relative prices of the two products, in this case (2.8) 𝑝𝑝 𝜕𝜕 / 𝑝𝑝 𝜕𝜕 = 8 25 . <?page no="63"?> 2.1 Customer behaviour and decisions on demand 63 Figure 2.7: Budget line utility maximization. ► Figure 2.7 in the top left-hand corner now shows equation (2.5) with the information from (2.6) and (2.7) as a budget line. The budget line describes the consumption possibilities with given income and prices - all points below and on it are possible, the area above is impossible. The location of the budget line is determined by income. With lower income it shifts parallel downwards, with rising income it shifts parallel upwards, because the slope of the budget line remains unchanged with constant prices. If the price ratio changes, as shown in ► Figure 2.7 bottom left, due to a price reduction for partying, the budget line turns at the point of the product that has remained constant. In ► Figure 2.7 top right, the budget line is combined with the indifference curves the student obviously chooses point 𝐴𝐴 because, given prices and disposable income, she can reach a highest possible indifference curve and the budget line just touches this indifference curve. ► Figure 2.7 at the bottom right-hand corner shows the effect of a price reduction for partying: the student stays in her flat and will go out more often in the future - by choosing 𝐶𝐶′ she increases her utility level compared to 𝐴𝐴 . Using ► Figure 2.7 top right, four observations can now be made, and key conclusions drawn. 600 / 25 = 24 = available budget divided by price of partying flat size in sqm partying per month 24 75 flat size in sqm partying per month 24 75 lower income: 400 30 partying becomes cheaper: 20 16 50 C‘ A 25 20 16 flat size in sqm 24 75 30 partying per month B‘ A‘ A A‘‘ B‘‘ partying per month 20 25 40 12 20 16 10 flat size in sqm 24 75 C‘ Slope = - 8 / 25 = relation of prices 600 / 8 = 75 = available budget divided by price per sqm <?page no="64"?> 2 Customer behaviour, products and network effects 64 Utility maximisation - at point 𝐴𝐴 and an actual combination of living on 25 sqm and partying 16 times per month, the student maximises her utility according to her preferences, given prices and disposable income Clearly the best possible product combination - point 𝐴𝐴 is obviously preferred to point 𝐵𝐵′′ , because with the same size of flat there is less partying. But this also means that 𝐴𝐴 is preferred to 𝐵𝐵′ : 𝐵𝐵′′ and 𝐵𝐵′ have the same level of utility, so that 𝐴𝐴 is also better than 𝐵𝐵′ from the student's perspective. Price ratios correspond to the marginal utility ratios - the slope of the indifference curve at point 𝐴𝐴 corresponds to the slope of the budget line. The slope of the budget line is determined by the relative price ratio of leasing a flat and partying, the slope of the indifference curve is determined by the ratio of both products’ marginal utility. Opportunity costs - if the student wants to party more often, she has to give up flat size: the opportunity costs per party at 25 EUR for a price of 8 EUR per square metre are 3.125 square metres - so she has to give up 3.125 square metres of living space for each time she wants to party more often. Since at point 𝐴𝐴 the price ratios correspond to the marginal utility ratios, i.e., (2.8) 𝑝𝑝𝜕𝜕 𝑝𝑝𝜕𝜕 = 𝜕𝜕𝑞𝑞 𝜕𝜕𝜕𝜕 ⁄ 𝜕𝜕𝑞𝑞 𝜕𝜕𝜕𝜕 ⁄ , rearranging, we get (2.9) 𝜕𝜕𝑞𝑞 𝜕𝜕𝜕𝜕 ⁄ 𝑝𝑝𝜕𝜕 = 𝜕𝜕𝑞𝑞 𝜕𝜕𝜕𝜕 ⁄ 𝑝𝑝𝜕𝜕 . In general terms, with rational purchasing decisions the ratio of marginal utility to price as in (2.9) is the same for all products - of course, this applies not only for two products, but in general to all products that a customer consumes. In other words: if the prices of several products are given, with the last available Euro a customer always buys the product which currently has the highest marginal utility per Euro. If a student goes out on a Friday evening, she will split the budget and spend her money so that the marginal utility per Euro is the same for all products (e.g., Aperol Spritz in her favourite pub and burgers at the St. Johanner Markt). This provides a further explanation why point 𝐵𝐵′ is not chosen. At 𝐵𝐵′ the slope of the indifference curve is obviously steeper than the budget line, so that because of (2.10) 𝜕𝜕𝑞𝑞 𝜕𝜕𝜕𝜕 ⁄ 𝑝𝑝𝜕𝜕 < 𝜕𝜕𝑞𝑞 𝜕𝜕𝜕𝜕 ⁄ 𝑝𝑝𝜕𝜕 , apparently the marginal utility of 20 times partying is not large enough to compensate living in a 10 sqm flat at a given price. The underlying concept of the marginal rate of substitution may seem theoretical at first glance, but can be observed empirically in many situations, especially in international trade or at stock exchanges when trading expected returns on different securities. But even children in the schoolyard behave accordingly when, for example, they swap football pictures between each other and trade relatively expensive pictures of Ronaldo for a large number of relatively cheap pictures of less well-known players. From a management perspective, empirical consumption patterns - such as those published by the Statistisches Bundesamt (Federal Statistical Office) or market researchers firms such as GfK or Nielsen - can be used to draw indirect conclusions about customer preferences at given <?page no="65"?> 2.1 Customer behaviour and decisions on demand 65 prices. In particular, decisions for one's own product portfolio or pricing strategies can be derived if price changes are observed for products which make up a significant proportion of a customer's budget or income, such as housing rent or food prices, or if products are subject to taxes. Willingness to pay, utility and consumer surplus The concept of indifference curves and budget lines has numerous applications in market research and estimating relative price elasticities, but it also allows (at least in cases where the demand for a product is not strongly influenced by income) a explanation for demand functions and makes utility approximately measurable. ► Figure 2.8 top left shows that for a given set of indifference curves of a customer, the quantity of both products demanded is adjusted if price relations change. If the price of product 2 falls from 𝑝𝑝 2 = 2 over 1 to 0.5 with the price 𝑝𝑝 1 of product 1 unchanged, the budget line turns outwards and becomes flatter - product 2 becomes cheaper relative to product 1. Demand thus, increases along the price-consumption curve, which is created via tangential points of indifference curves and budget lines, from 5 over 15 to 25 units. If these price/ quantity coordinates are now transferred to the bottom left, a demand curve can be constructed for product 2 along 𝐴𝐴′ , 𝐵𝐵′ and 𝐶𝐶′ from the sequence of the utility maximising tangential points 𝐴𝐴 , 𝐵𝐵 and 𝐶𝐶 - with decreasing prices, demand is increasing. If income or prices of the other product change, the position of the demand curve in ► Figure 2.8 on the lower left changes. Figure 2.8: Individual demand schedule and utility maximisation. If a demand curve can be derived from utility maximisation, then utility is also indirectly reflected in the demand function: in the right part of ► Figure 2.8, a simplified linear and continuous demand function (2.11) 𝑝𝑝 = 𝑎𝑎 − 𝑏𝑏𝑞𝑞 0 0 q 1 q 2 p 2 q 2 A B C A‘ B‘ C‘ 6 54 5 15 25 2 1 0,5 p 2 =2 p 2 =1 p 2 =0,5 5 15 25 q p ... 321 0 q* p* a z i 1 2 3 … individual demand curve price-consumption-curve consumer surplus <?page no="66"?> 2 Customer behaviour, products and network effects 66 is depicted. At a price 𝑝𝑝 ∗ the customer seems willing to buy a quantity 𝑞𝑞 ∗ . If the price was higher the customer would buy less according to the demand curve, in line with his individual utility maximisation based on his willingness to pay 𝑧𝑧 𝑖𝑖 . The individual utility 𝑢𝑢 𝑖𝑖 = 𝑝𝑝 ∗ − 𝑧𝑧 𝑖𝑖 arises from the difference between willingness to pay and price for each unit demanded. In general, the (net-)utility of a customer from the consumption of a certain quantity 𝑞𝑞 ∗ of a product at a price 𝑝𝑝 ∗ results from the sum of these individual utilities measured by the triangular area between the demand function and the price as (2.12) 𝐶𝐶𝑆𝑆 = 𝑚𝑚−𝑝𝑝∗ 2 ⋅ 𝑞𝑞 ∗ and is referred to as consumer surplus (► Chapters 7 and 8). Good to know │ What is the willingness to pay for Instagram, what is the consumer surplus of Facebook? Numerous social media platforms offer membership and use free of charge - implicitly customers pay with shared data, which in turn is passed on and sold by the platforms to third-parties. But what would be the willingness to pay for a Facebook account? In an empirical study, this was indirectly analysed: European and US American students were asked how much money they would be willing to pay in order to prevent social media services from being blocked for the next month. The average implicit monthly willingness to pay for Facebook calculated in this way was 97 EUR, for WhatsApp 536 EUR, for Instagram 6.79 EUR and for LinkedIn 1.52 EUR - the median value of an annual disallowance of search engines would have to be compensated with 17,530 USD, a renunciation of free email services for one year with 8,414 USD and a renunciation of YouTube with 1,173 USD (Brynjolfsson et al. 2019). While this result might still be valid for the US, relative willingness to pay between platforms may change over time and in geographic areas. UK students are not fans of Facebook, German students move over from Facebook to LinkedIn and Instagram. In another study, the necessary average compensation for not using Facebook for one month was 203 USD, the median value was 100 USD, but 20% of the customers had values of more than 500 USD. If one constructs a demand function from these values, the result is a monthly consumer surplus of approx. 31 billion USD for approx. 200 million Facebook members in the US alone (Allcott et al. 2020). 2.2 Market delineation and product categories Market is probably one of the most used economic terms, in firms and in the media - market is usually where the customer is or is suspected to be. But apart from this general definition, markets are difficult to grasp and describe. On the one hand, they change because customers change their behaviour or competitors actively influence the market. On the other hand, it is easy to discuss at length whether coffee and tea belong to the 'same market' or not - or whether instant coffee and coffee pods are in ‘different markets’. In the following section, possible definitions of markets and related categories of products are developed in order to establish a better understanding. <?page no="67"?> 2.2 Market delineation and product categories 67 Changes in demand by individual customers or of entire market segments can result, for example, from changes in preferences (increased preference for healthy food); changes in relative prices (price reductions for telecommunications and data usage); changes in income (net salary increases based on an income tax reduction); the availability of new products based on innovations (smartphones and apps); or due to deregulation of markets (long-distance buses in Germany). Often changes in demand in one market arise from changes in other market segments. In addition, the interplay of changed preferences or willingness to pay and product innovations can lead to the emergence of new market segments. At the same time, firms can also actively address a new market segment or try to establish it. Size of the market The demand curve, and with that the size of a market, can be determined using information concerning the number of customers and their individual willingness to pay. The willingness to pay itself can be determined by four (combinable) approaches of market research: direct customer survey, direct/ indirect observation of actual decisions and behaviour of customers, econometric estimation of demand with price variations, and conclusions by analogy from purchase decisions of other products. ► Figure 2.9 on the left shows an example of results of market research to determine the combination of number of customers and willingness to pay for a tablet computer. Using empirical data on the left, an empirical demand function - which can per se take any linear or nonlinear functional form - can directly be estimated with statistical software such as Excel or SPSS via a regression analysis as (2.13) 𝑝𝑝 = 𝑝𝑝(𝑞𝑞) = 𝑎𝑎 − 𝑏𝑏𝑞𝑞 = 1000 − 0.0025𝑞𝑞 which in this case for simplicity reasons is assumed to be a linear function. Figure 2.9: Empirical estimate of demand curve and demand function. A demand function describes the interdependence of quantity demanded and the price of a product, where 𝑎𝑎 indicates the maximum willingness to pay in this market - in this case it appears that exactly one customer is willing to pay up to 1000 EUR - and 1/ 𝑏𝑏 being an indicator of the horizontal size of the market: the smaller 𝒃𝒃 , the larger the market. If each customer <?page no="68"?> 2 Customer behaviour, products and network effects 68 buys a maximum of one tablet computer, then for the maximum number of customers at a price of 𝑝𝑝 = 0, after rearranging (2.13), we get (2.14) 𝑞𝑞(𝑝𝑝) = 𝑚𝑚−𝑝𝑝 𝑏𝑏 = 1000−0 0.0025 = 400,000 i.e., a maximum of 400,000 tablet computers will be demanded. The market volume measured by revenues 𝑅𝑅 depends on the size of the market determined by parameters 𝑎𝑎 and 𝑏𝑏 . For a market price 𝑝𝑝 ∗ and the quantity 𝑞𝑞 ∗ determined by the demand function, the market volume is given by 𝑅𝑅 = 𝑝𝑝 ∗ 𝑞𝑞 ∗ , as shown in ► Figure 2.10. Figure 2.10: Market size and market volume. Market size, market segments and market shares Often a market can be divided into different segments, which either exist anyway due to demographic, socio-economic, psychographic, or behavioural characteristics of the customers, or are actively created by firms' product differentiation strategies (Dickson and Ginter 1987). The relationship between market segments and the size of a market can be illustrated for three firms offering high-priced tablet computers. ► Figure 2.11 shows the hypothetical results of market research in a specific region - assuming that due to perfect product differentiation, every potential customer is a loyal customer of his brand and is not willing to switch to another brand. Based on the demand functions 𝐷𝐷 for the individual firms (2.15) 𝐷𝐷 𝐴𝐴 : 𝑝𝑝(𝐴𝐴𝑝𝑝𝑝𝑝𝐴𝐴𝐴𝐴) = 𝑝𝑝(𝑞𝑞) = 𝑎𝑎 𝐴𝐴 − 𝑏𝑏 𝐴𝐴 𝑞𝑞 𝐴𝐴 = 1000 − 0.0025𝑞𝑞 𝐴𝐴 , (2.16) 𝐷𝐷 𝑆𝑆 : 𝑝𝑝(𝑆𝑆𝑎𝑎𝑆𝑆𝑆𝑆𝑢𝑢𝑆𝑆𝑆𝑆) = 𝑝𝑝(𝑞𝑞) = 𝑎𝑎 𝑆𝑆 − 𝑏𝑏 𝑆𝑆 𝑞𝑞 𝑆𝑆 = 800 − 0.001𝑞𝑞 𝑆𝑆 and (2.17) 𝐷𝐷 𝐻𝐻 : 𝑝𝑝(𝐻𝐻𝑢𝑢𝑎𝑎𝐻𝐻𝐴𝐴𝐻𝐻) = 𝑝𝑝(𝑞𝑞) = 𝑎𝑎 𝐻𝐻 − 𝑏𝑏 𝐻𝐻 𝑞𝑞 𝐻𝐻 = 800 − 0.004𝑞𝑞 𝐻𝐻 it can be seen that although Apple's potential customers with 𝑎𝑎 𝐴𝐴 = 1000 have a higher willingness to pay than Samsung's and Huawei's customers, potentially most customers buy a Samsung tablet, since with 𝑏𝑏 𝑆𝑆 = 0.001 the market segment is larger than Apple's market segments with 𝑏𝑏 𝐴𝐴 = 0.0025 and Huawei's with 𝑏𝑏 𝐻𝐻 = 0.004 . Conversely, the price elasticity of demand 𝜀𝜀 at any price level is lowest for Apple and highest for Samsung. <?page no="69"?> 2.2 Market delineation and product categories 69 Figure 2.11: Willingness to pay for different brands. Figure 2.12: Horizontal addition of market segments to determine total market demand. The size of the total market can be calculated by adding the three demand curves horizontally, i.e., for each price the maximum demand of the three segments is added up. ► Figure 2.12 shows the demand functions (2.15) to (2.17) accordingly - any aggregated demand curve is flatter than individual demand curves. At a price of 600 EUR, 160,000 tablet computers would be sold by Apple, 200,000 by Samsung and 50,000 by Huawei, so that the total market would sell 𝑄𝑄 = 𝑞𝑞 𝐴𝐴 + 𝑞𝑞 𝑆𝑆 + 𝑞𝑞 𝐻𝐻 = 410,000 tablets. The market shares 𝑆𝑆 𝑖𝑖 = 𝑞𝑞 𝑖𝑖 ⁄ 𝑄𝑄 of the number of tablets are calculated as 𝑆𝑆 𝐴𝐴 = 39.0% , 𝑆𝑆 𝑆𝑆 = 48.8% and 𝑆𝑆 𝐻𝐻 = 12.2% . If the price were only EUR 250, then firms would have a total market of 𝑄𝑄 = 𝑞𝑞 𝐴𝐴 + 𝑞𝑞 𝑆𝑆 + 𝑞𝑞 𝐻𝐻 = 300,000 + 550,000 + 137,500 = 987,500 tablets, so that market shares would change to 𝑆𝑆 𝐴𝐴 = 30.4% (𝑑𝑑𝑑𝑑𝐻𝐻𝑆𝑆 𝑏𝑏𝑏𝑏 − 8.6%) , 𝑆𝑆 𝑆𝑆 = 55.7% (𝑢𝑢𝑝𝑝 𝑏𝑏𝑏𝑏 + 6.9%) and 𝑆𝑆 𝐻𝐻 = 13.9% (𝑢𝑢𝑝𝑝 𝑏𝑏𝑏𝑏 + 1.7%) , because the different market segments are of different size at different prices - Apple would grow in absolute terms but would lose most market share if prices go down, due to the lowest price elasticity of demand. At a price of EUR 800 or more, Apple would q ,Q in thsnd. p D H p=1000 p=800 D A D S p=600 1.400 410 p=250 988 D H + D A + D S <?page no="70"?> 2 Customer behaviour, products and network effects 70 benefit greatly: its market share would increase to 𝑆𝑆 𝐴𝐴 = 100% , as shown in ► Figure 2.12, because none of the other firms has customers with such a high willingness to pay. In this case, the market shares 𝑆𝑆 𝑖𝑖 = 𝑅𝑅 𝑖𝑖 ⁄ 𝑅𝑅 of the revenues are identical to the market shares in units, as all firms achieve the same price of 600 EUR or 250 EUR per unit. Market shares in terms of units differ from those in revenues if prices of products are different and firms offer a different product portfolio, as is the case in the automobile industry, for example (see also ► Chapter 10 on strategic competition and the effects on market shares and profits). Although the location of the demand curve is partly predetermined for firms of an industry based on income, population and prices of other products, firms can influence the location of the demand function. ► Figure 2.13 illustrates the two polar cases: on the left, firms can increase the willingness to pay of all potential customers to the same extent, so that the demand curve shifts upwards in parallel, and on the right, the market grows by expanding to new customer segments while maintaining the willingness to pay. Figure 2.13: Impact and demand and market size by firms’ strategies. There are two main strategies that firms can employ to influence location and slope of demand curves: marketing as a direct or indirect influence on preferences, advertising in existing target groups, or addressing new customer segments; and product quality via innovative features, better technology or additional services (Chatmi and Elasri 2017, Sridhar et al. 2014 as well as Dorfman and Steiner 1954). Both strategies can lead to higher values of 𝑎𝑎 (higher willingness to pay) or lower values of 𝑏𝑏 (more customers). However, empirical studies show that marketing often has a stronger effect on 𝑏𝑏 , for example, by addressing new customer segments, whereas, an improvement in technology, quality and features of the products has a stronger effect on the willingness to pay 𝑎𝑎 of all customers (Levin and Reiss 1989, Johnson and Myatt 2006 and Ashley et al. 1980). From a management perspective, it is essential to identify these separate effects in a firmor industry-specific way in order to be able to decide on optimal R&D and marketing investments and their relation to each other to influence demand. <?page no="71"?> 2.2 Market delineation and product categories 71 Product life cycles Obviously, firms change their portfolio over time through product innovation (see ► Chapter 4 for more details). A product life cycle describes the development of a product's revenues based on introduction, growth, maturity, saturation and decline stages (Levitt 1965, Cox 1967 and Simon 1979), as shown in ► Figure 2.14. The product life cycle can be identified at a firm level (e.g., Apple's iPod Mini) as well as at an industry level (sales of MP3 players by various firms) and can lead to industry life cycles when aggregated (see ► Chapter 4). Along the product life cycle, the quantity sold 𝑞𝑞 initially increases at a more than proportionate rate, then slows down, before finally declining. Typically, firms lower prices 𝑝𝑝 along an S-shaped curve over time, so that revenues 𝑅𝑅 initially rise faster than quantity 𝑞𝑞 , but then decrease faster during a saturation and decline stage. This pattern is caused by mutual changes of the determinants of demand: the maximum willingness to pay 𝑎𝑎 increases during the introduction and growth stage and then decreases; the size of the market 1/ 𝑏𝑏 increases until the end of the saturation stage and then decreases; the price elasticity of demand 𝜀𝜀 initially decreases during the introduction stage (equivalent to an increase in the willingness of the product’s early adopters to pay) and then rises gradually. Figure 2.14: Product life cycle. The basic assumption for such a view is the observation that firms regularly renew their products - i.e., successively launch a sequence of product generations into the market in order to either gain market share from competitors, to address new customers with new product features, or leverage new technological options. Empirically, the length of product life cycles varies both between firms and with regard to the respective product generation of a firm. Often, as with Apple's smartphones, it can be observed that individual generations of a product are offered overlapping and the generational changes <?page no="72"?> 2 Customer behaviour, products and network effects 72 between firms are not synchronised (Wiecek-Janka et al. 2017). Key factors influencing the pattern and duration of product life cycles are a combination of firm-specific, demand-driven and technological determinants: the interplay between the speed of product development and innovation, a firm's product strategy, behaviour of competitors and customers is decisive (Klepper 1996 as well as Anderson and Zeithaml 1984). The variety of patterns observed, across different industries and also within industries, reflects the different industry and firm-specific factors, which are also subject to evolutionary change and trends over time. For example, product life cycles can be accelerated by technological innovations. Conversely, in the absence of product innovations and stable customer preferences, product life cycles can be weak or even non-existent (Gruber 1995, Clark et al. 1987 and Clark and Fujimoto 1992). Good to know │ Product life cycles and product generations - why are people buying records again? These patterns of product life cycles are clearly visible in the music industry: ► Figure 2.15 shows revenues of music in Germany differentiated by music formats and channels over time (Bundesverband Musikindustrie 2020 and Lehman-Wilzig and Cohen-Avigdor 2004). Figure 2.15: Development of revenues in the German music industry 1984 to 2019 in mill. EUR (Source: Bundesverband Musikindustrie 2020). Obviously the importance of different formats shifts over time: in the mid-1980s the LP dominated but was displaced by the CD developed by Philips, Polygram and Sony. The reason being that despite significantly higher prices, the CD had a longer playing time, better durability, easier handling and partly better sound quality. At the same time, the cassette tape was able to maintain its position as a music media for older target audiences and car drivers until the 1990s. The reasons are indirect network effects and switching <?page no="73"?> 2.2 Market delineation and product categories 73 costs: car radios installed in cars with cassette players and easy-to-use cassette recorders in private households would have required immediate device replacement when switching to CDs and cassettes already purchased would no longer have been usable. Following the introduction of MP3 and other digital data-reduced formats in the mid- 1990s and the emergence of illegal online exchange platforms such as Napster around the year 2000, the CD has lost a great deal of sales. With the establishment of broadband networks, the increasing use of smartphones since 2007 and cheaper mobile phone plans, streaming has become the music format with the highest revenues in Germany since 2018. But the LP is alive and kicking - since 2011 the absolute revenues and relative market share have been rising continuously. Customers can be found in very different market segments: on the one hand, older men with high-quality stereo systems, on the other hand, young customers in the hip-hop and DJ environment. Figure 2.16: Product life cycles of CD-single, music video, CD and downloads and revenue patterns in Germany 1980 to 2020 in mill. EUR (Source: Bundesverband Musikindustrie 2020). ► Figure 2.16 shows the respective sales trends between 1980 and 2000 for the four formats CD, music video (DVD, VHS and Blu-Ray), CD single and downloads. Behind all of the almost prototypical product life cycle trends are not only the competitive effects of substitutes but also changes in complementary playback devices (away from stand-alone solutions such as CD players or the iPod, towards software-based integration in laptops or smartphones) and a shift away from proprietary files towards subscriptions. Especially the relatively short product life cycle of downloads (e.g., iTunes or Amazon from 2012 to 2017) shows that people initially considered ownership of music to be a common variant - similar to the behaviour with LPs and CDs - but now prefer streaming services such as Spotify, Tidal, or Deezer due to the virtual use across different devices. As a result, Amazon, Apple and Google have also shifted their business models from download to streaming (Datta et al. 2018 and Kretschmer and Peukert 2020). <?page no="74"?> 2 Customer behaviour, products and network effects 74 Product categories and their empirical identification Before developing strategies to influence demand, it is necessary to analyse the factors influencing a demand function, which also allows to delineate markets and market segments. The questions shown in ► Figure 2.17, among others, are central to this process and are answered below. Figure 2.17: Options for classification of products and structure of demand. Changes in the level of income, i.e., both the income of individual customers as well as the average per capita income of entire nations, influence the quantity demanded. The effect can - in analogy to the price elasticity of demand - be assessed by means of the income elasticity of demand (2.18) 𝜀𝜀 𝐼𝐼 = ∆𝑞𝑞 𝑞𝑞 � ∆𝐼𝐼 𝐼𝐼 � ≅ 𝐼𝐼 𝑞𝑞 𝜕𝜕𝑞𝑞 𝜕𝜕𝐼𝐼 for ∆ → 0 . Income elasticity describes a percentage increase in the quantity 𝑞𝑞 demanded for a 1% increase in income 𝐼𝐼 - this value can be positive or negative. 𝜀𝜀 𝐼𝐼 > 0 : normal products - with increasing income the quantity of these products in demand increases: If 𝜀𝜀 𝐼𝐼 ∈ [0; 1] , then products are necessary products whose demand grows with increasing income, but less than proportionally - this applies, for example, to men's clothing and basic food. If 𝜀𝜀 𝐼𝐼 > 1, then products are luxury products whose demand rises at a more than proportionate rate with rising income - typically these products are only demanded above a certain income level, such as luxury watches from Panerai, or overnight stays in luxury hideaways such as Schloss Elmau. <?page no="75"?> 2.2 Market delineation and product categories 75 𝜀𝜀 𝐼𝐼 < 0 : inferior products - with increasing income the demand for these products decreases: the reason is that with increasing income the purchase of products of higher quality or better reputation becomes possible. A typical example are rooms in shared flats. People do not rent larger rooms in shared flats as their income rises but move to their own flat in terms of quantity and quality. Empirical values of income eleasticity of demand product/ service elasticity category automobile 2.08 luxury products furniture 1.44 dining in restaurants 1.40 healthcare 1.18 take away food 1.12 women’s apparel 1.07 water 1.02 men’s apparel 0.75 (necessary and) normal products cigarettes and tobacco 0.64 petrol and fuel 0.48 meat and fish 0.45 electricity 0.20 cash -0.23 inferior products pork -0.29 public transport (cities) -0.36 public transport (rural areas) -0.77 Table 2.1: Empirical income elasticities (data sources: Blanciforti 1982, Houthakker and Taylor 1970, Taylor and Halvorsen 1977, Ward and King 1973 as well as Wold and Jureen 1953, see also Frank and Cartwright 2013). ► Table 2.1 shows income elasticities from a selection of empirical studies. With significantly rising income, the demand for cars (both the value and the number of cars) increases disproportionately, the consumption of tap water increases disproportionately with rising income when changing from shower to bathtub and for watering one's own garden. Furthermore, the consumption of pork and journeys on public transport are reduced with rising income in favour of higher quality meat and journeys in one's own car. All these effects will of course take place <?page no="76"?> 2 Customer behaviour, products and network effects 76 successively over time and only if there are sufficient and lasting changes in disposable income and the degree to which households are equipped with consumer electronics or household appliances will change accordingly (Statistisches Bundesamt 2017). The quantity demanded is of course directly influenced by the price of the product (see ► Chapter 1 on the price elasticity of demand), but also indirectly by prices of other products. This indirect effect can be explained, measured and analysed by the cross-price elasticity of demand (2.19) 𝜀𝜀 𝑚𝑚𝑞𝑞 = ∆𝑞𝑞𝑥𝑥 𝑞𝑞𝑥𝑥 � ∆𝑝𝑝𝑦𝑦 𝑝𝑝𝑦𝑦 � ≅ 𝑝𝑝𝑦𝑦 𝑞𝑞𝑥𝑥 𝜕𝜕𝑞𝑞𝑥𝑥 𝜕𝜕𝑝𝑝𝑦𝑦 ( for ∆ → 0) . The cross-price elasticity (2.19) describes the percentage change in the quantity demanded of a product 𝑥𝑥 following a 1% change in the price 𝑝𝑝 𝑞𝑞 of another product 𝑏𝑏 - this value can be positive or negative. 𝜀𝜀 𝑚𝑚𝑞𝑞 > 0 : substitute products - with a price increase of product 𝑥𝑥 the demand for product 𝑏𝑏 increases. From the customer's point of view, these products are obviously relevant alternatives, so that if there is a relative change in price, demand increases for the now relatively cheaper product: a price increase for beef indirectly increases the demand for pork and poultry; if public transport fares are significantly increased, customers switch to substitute products such as bicycles or cars. 𝜀𝜀 𝑚𝑚𝑞𝑞 < 0 : complementary products - with a price increase of product 𝑥𝑥 the demand for the other product 𝑏𝑏 decreases. From the customer's point of view, the products seem to have a relationship of use, so that if the price of one of the products increases, demand for both products decreases: a price increase or a tax increase on fuel reduces the demand for cars with high consumption in the medium term; if the price of a game console increases, the demand for complementary games decreases. ► Table 2.2 shows the cross-price elasticities from a number of empirical studies showing typical results - if price increases are significant, customers switch from their existing product to a substitute product that can be used or consumed in a similar way. For example, the comparison of margarine and butter, as well as the asymmetry of effects, shows that of course basic preferences play an essential role: people switch from margarine to butter to a greater extent than vice versa when prices rise. On the other hand, the complementary character of consumer electronics and food with 𝜀𝜀 𝑚𝑚𝑞𝑞 = −0.72 is an indication of a lifestyle: the larger the flat screen at home, the more often people eat in front of the TV. All these effects also occur over time and are influenced by general conditions such as personal preferences, marketing and of course the cultural and socio-economic environment. <?page no="77"?> 2.2 Market delineation and product categories 77 Empirical values of cross-price elasticity of demand product product with price change cross-price elasticity category fish meat 1.61 substitutive products butter margarine 0.81 margarine butter 0.67 beef pork 0.28 electricity natural gas 0.20 vegetables fruit 0.05 home entertainment food -0.72 complementary products DVD players DVDs -0.77 breakfast cereals fish -0.87 Table 2.2: Empirical cross price elasticities (Data source: Deaton 1987, Deaton 1990, Taylor and Halvorsen 1977 as well as Wold und Jureen 1953, see also Frank and Cartwright 2013). From a management perspective, the concepts of income and cross-price elasticities, in addition to the price elasticity of demand, are always important if changes in the consumption patterns of customers are observed. On the one hand, these might signal changed preferences; on the other hand, these changes may be triggered by income levels, or changes in the relative prices of competing products, or products used in context. For firms such as Amazon, negative cross-price elasticities and income elasticity are particularly relevant: if a customer is looking for a product, Amazon will in particular propose products with complementary characteristics in addition to this product; the absolute price level of the proposed products depends on the customer-specific income elasticity. Horizontal and vertical product differentiation Products within a specific product class - bicycles, yoghurt, or business schools as well as microeconomics textbooks - are rarely homogeneous products with the possibility of perfect substitution. Firms try to differentiate products or services, i.e., to create differences within their own product portfolio or vis-à-vis competitors and then position them via marketing. The objectives of product differentiation are as follows. Increasing the willingness to pay - if preferences of customers for a product, e.g., cars, are heterogeneous, then the willingness to pay of individual customers or of a market segment can be increased through greater differentiation between products, thus achieving a better match with existing preferences. As a result, all other conditions being equal, higher prices can be achieved. <?page no="78"?> 2 Customer behaviour, products and network effects 78 Reduction of competitive intensity - if customers regard competitors' products as close substitutes and consider switching, or actually switch to competing products at will, the intensity of competition is high. Product differentiation can then reduce the intensity of competition (the repercussions of competitors' strategic actions) and increase profits through brand loyalty based on marketing, switching barriers due to technological incompatibility, or the positioning of the firm (see also ► Chapter 10). Product differentiation can take on two different dimensions which have a significant impact on the market structure and competitive intensity of an industry (Shaked and Sutton 1987, Gabszwicz and Thisse 1986 and further on ► Chapter 10). Horizontal product differentiation - product characteristics and/ or quality may differ slightly but are functionally almost identical. However, customers do not agree on the subjective ranking of preferences (socks in black versus red, yellow or green, Coke and Pepsi, beer types, car brands, etc.). Customers seem to have different tastes or otherwise different preferences. As a consequence, some customers are willing to pay more for Coke than for Pepsi - for other customers it is the other way around. Firms can establish and strengthen horizontal product differentiation primarily through marketing - e.g., through branding and brand awareness. Technological differences play a minor or subordinate role. Vertical product differentiation - product characteristics and/ or quality differ functionally; and all customers agree on the objective ranking of preferences (DSL with 20 Mbit vs. 50 Mbit vs. 100 Mbit, ...), but have a different willingness to pay due to different usage behaviour or requirements. As a consequence, some customers are willing to buy DSL with 100 Mbit, others not. Firms can establish and expand vertical product differentiation primarily through qualitative and technological features, but the effect can also be strengthened through marketing. Horizontal differentiation alone usually does not allow price discrimination if customer groups are similarly large: the prices of ice cream at the local gelataria, downloads of songs, cinema visits or beer types are very close or even identical both across firms and within the portfolio of a firm. In case of vertical differentiation (e.g., fast vs. slow broadband), price discrimination is typical; however, it is not objective differences in performance that explain the price differences, but the differences in willingness to pay (see further ► Chapters 7 and 10). Typically, mixed strategies are used to mutually reinforce the effects of horizontal and vertical product differentiation (Degryse 1996 as well as Ferreira and Thisse 1996). Banks have long tried to differentiate horizontally and vertically between evidently standardised and homogeneous products and services, such as payment transactions, current accounts or cash and debit cards; and food firms are investing heavily in the product differentiation of mineral water. However, only product differentiation perceived by a large group of customers leads to a reduction of competitive intensity and a relevant increase in willingness to pay - fans of a brand or technology nerds often distort the picture here. Blind testing of products shows that product differentiation disappears when branding, colouring, or the look and feel of the products are no longer recognisable, so that in many industries - especially consumer products and food - a large proportion of supposedly perceived product differentiation is based solely on strategic marketing investments (Nenycz-Thiel and Romaniuk 2014 as well as Yamada et al. 2014). <?page no="79"?> 2.2 Market delineation and product categories 79 Strategically and from a management perspective, for product differentiation to be effective, customer preferences must either be discernably different and addressable or at least influenced by marketing. The greater the perceived differentiation, the stronger the impact on willingness to pay, the lower the intensity of competition through substitute products, and finally the larger the possibilities of market segmentation. Product differentiation is an essential strategy to significantly reduce the intensity of competition - but typically significant investments in marketing and technology are necessary. Case Study │ Horizontal and vertical product differentiation in the automobile industry In the automobile industry - as sketched in ► Figure 2.18 for the respective model platforms in 2017 - vertical and horizontal product differentiation has long been well established through extensive and long-term investment in technology and marketing and has a significantly positive effect on firms’ profits (Goldberg 1995, Guajardo et al. 2015 and Berry et al. 2004). Typical vertical product differentiation in the automobile industry is based on engine and driving performance, durability and reliability, size, interior equipment such as multimedia, safety equipment as well as service and warranty. Horizontal product differentiation is achieved through branding, design, look and feel, status and relative positioning of a brand to differentiate or approximate other brands. Both dimensions will be linked and reinforced by product proliferation (i.e., the occupation of almost any market niche by coupe, convertible, SUV, sports car, van, or station wagon models) and by the possibility for customers to personalise their car - from the usual extras to specialist suppliers and tuners (Fujimoto 2014). Figure 2.18: Horizontal and vertical product differentiation (product categories according to EU Commission classification). <?page no="80"?> 2 Customer behaviour, products and network effects 80 A fundamental shift in preferences in recent decades is due to an increasing desire of customers to individualise or personalise products. Firms can offer products in modules or variants (via mass customisation and mass personalisation) to increase the willingness to pay for individualised products and to address niches in market segments. Starbucks has established high margins for essentially a homogenous product, coffee, through different cup sizes and strong vertical and horizontal differentiation. Since 2010, Starbucks customers have also been able to personalise their cup of coffee with How-Ever-You-Want-It-Frappuccino, including the (often misspelled) customer name on the cup. With this strategy, Starbucks achieved a gross margin of around 40% in 2016, growing at a CAGR of around 15%. Firms like MyMuesli in Germany can drive the price of cereals from EUR 0.39 for 500 grams to well over EUR 10 by personalisation based on extensive configuration options (Abraham et al. 2017, Vesanen 2007, Starbucks 2016 and FAZ 2016). Uncertainty about product characteristics Particularly in the case of vertical and horizontal product differentiation, customers are often not fully informed about all product features, or the diversity of product information is greater than the ability of customers to evaluate all information correctly. Products can then also be distinguished and classified by the degree of uncertainty before and during consumption (Nelson 1970 and Zeithaml 1988). This uncertainty can be related to the product itself, the transaction to be carried out or other market participants, and increases from search goods to experience goods to credence goods, as explained below. Search goods - here a reduction or elimination of uncertainty is possible through market analysis before the purchase and consumption. Of course, gathering of information is associated with costs and especially time. Examples are clothing, bicycles, or laptops. All products can be researched before purchase and essential product features can be tested. Experience goods - obtaining information before the purchase does not work or is too expensive relative to the purchase price. Here customers can only get to know the characteristics or quality of the product by trying it out after purchase. Examples are the Wi-Fi speed in an Airbnb flat, a visit to a new restaurant or the fuel consumption of a used car. Credence goods - here the product quality cannot be checked either before purchase or during consumption. This applies to airbags in cars, condoms and organic eggs from happy chickens as well as medical diagnoses, preventive medicines, or sports programmes and also to products that can change their quality over time (e.g., financial advice for investment products or school education). Uncertainty of product characteristics is partly due to different preferences and consumption habits of customers, but uncertainty is also caused by unclear and poorly described product characteristics. Firms can try to reduce uncertainty by appropriate marketing measures such as informative or suggestive advertising. However, firms can also intentionally create uncertainty through marketing measures in order to increase the willingness to pay or strategically conceal product quality. This applies to taxi rides for people from out of town: taxi drivers - before Uber started - often use an information advantage over customers for detours and thus, achieve higher sales (Balafoutas et al. <?page no="81"?> 2.2 Market delineation and product categories 81 2013). In the tourism industry, on the other hand, incompletely specified products are used to offer trips where the customer only finds out the actual hotel at the holiday destination: the aim here is to exploit fluctuations in demand and/ or capacity and establish new market segments for customers with a low willingness to pay (Gönsch 2020). The uncertainty of product characteristics can be increased because manufacturing and working conditions cannot be checked, especially with labour-intensive or imported products from developing countries. For all product categories, market failure can occur due to asymmetric information. Uncertainty and information asymmetry can be significantly reduced in particular through test and inspection reports (Stiftung Warentest in Germany, consumer centres or real evaluations by other customers in rating portals), and also through certificates, quality seals, and product guarantees or the reputation of the seller (► Chapter 7). Market delineation From a management perspective, a major topic is to define the relevant market for a firm and its products: on the one hand, to include all relevant customer groups and market segments for a market analysis; and on the other hand, to consider all competitors and their strategies, business models and products when doing a competitive analysis. ► Figure 2.19 shows that a delineation of a market or market segments has to be done along several dimensions. Technological dimension - the focus is on physical, technical or functional product similarities as well as on production and manufacturing processes and business models: from a technological perspective, sports cars, small cars and SUVs, but also tractors, trucks and buses are in the same market. Demand-side dimension - all products that are regarded as alternative solutions from the customer's point of view are grouped together in one market. This can be measured by a significantly positive cross-price elasticity, as customers move to substitute products when prices change: cars, bicycles, local public transport, long-distance buses, or car-sharing agencies are one market here. Substitution gaps then determine the boundaries of a market. Supply-side dimension - in many industries market definition is based on competitive behaviour of firms over time: Mercedes, BMW, or Lexus are competitors in one market, while Dacia, Kia, or Lada compete in another segment. Market definition is based on the strategic perception and behaviour of firms. All delineations along these three dimensions must also take into account spatial or temporal aspects. Spatial distance plays a major role in real estate - a home in Melbourne is a different product than a house in Hamburg - or in products with high transport costs but loses importance in the case of digital or virtual products. Typically, however, it is still not possible to make a clear distinction: multi-product firms cannot be clearly assigned, customer groups do not match in their cross-price elasticities or willingness to switch to other brands. In addition, product innovations, e.g., the convergence of products such as email, city maps, navigation systems, hotel catalogues, telephony, photography and music into a smartphone, and in particular digitisation (does Google offer cheese because cheese is placed in Google Shopping search results? ) shift and change previous market boundaries again and again, or establish new markets and ecosystems (Schmidt et al. 2016, Fiegenbaum and Thomas 1995, Gambardella and <?page no="82"?> 2 Customer behaviour, products and network effects 82 Torrisi 1998, Malhotra and Gupta 2001, also ► Chapter 4). Consequently, each and every market definition is situation-specific and firm-specific. Figure 2.19: Market delineation along various dimensions. From the point of view of competition policy, the definition of a relevant market is also very important: here it must be assessed whether, for example, a merger in one market leads to an increase in market power (see ► Chapter 7). Competition policy now uses the SSNIP test (small but significant and non-transitory increase in price) based on cross-price elasticities. This test checks whether customers react to an assumed (or simulated or carried out in a defined test market) small, significant and permanent price increase (often assumed to be 5% or 10%) for a product by switching to available substitutes. All products that customers switch to are then belonging to the relevant market: all the above dimensions are thus, empirically covered. For example, a market analysis in the shaving industry would have to determine whether different electric or wet shavers, hair removal creams or waxes, electric epilators, or laser hair removal devices belong to one and the same market, i.e., whether customers would switch to other products or services if the price were to be permanently increased by 5% to 10%. The test is applied to available products until a number of products is identified for which a price increase does not lead to a substitution effect - these products and the firms offering them are then outside the relevant market (Motta 2004, Monteiro and Foss 2017 as well as Filistrucchi et al. 2014). If, for example, in the case of a 10% permanent price increase, customers do not switch from one product to another, the firm under consideration has market power or a monopoly position at least in this market segment. The problem with the SSNIP test is that it only works if positive prices are possible: in the case of free services such as internet search or social <?page no="83"?> 2.3 Network effects and multisided markets 83 media membership, it is not possible to define the market in this way (see ► Section 2.1 for an indirect identification of the willingness to pay for social media services). 2.3 Network effects and multisided markets In many markets and industries, network effects play a key role in products and business models. The utility for customers arises in part from the use or consumption of a product, but mainly because other customers also use the same product and it is possible to network with them. No customer would like to be the only member of Facebook - but because ‘everyone’ is on Facebook, a self-reinforcing effect leads to the fact that with an increasing number of members and a growing market share of Facebook, everyone wants to be on Facebook (Rohlfs 1974). This effect of increasing utility based on network effects can lead to a displacement of competitors: for example, relatively quickly after entering the market, Facebook ousted the former market leaders in Germany Wer-kennt-Wen, SchülerVZ, StudiVZ and Stayfriends from the market or into niches. Similar developments happened in a large number of countries. Network effects can occur in two forms. Direct network effects arise if the utility of a customer increases with the number of members of a network. The benefit arises from the possibility of direct communication with other members of the network, e.g., in communication markets (telecommunications network, WhatsApp, Skype, etc.), but also in social networks such as Facebook, Instagram, or Snapchat - if a customer were the only member of one of these social media platforms, no utility would arise. Markets in which direct network effects dominate are called network markets. Indirect network effects occur if a growing number of users of a product or service promotes the creation and offer of complementary products - and thus, indirectly the utility of being a member of the network increases through the number of other customers. Customers do not communicate directly with each other here, but they use the same complementary products: the more customers buy a certain media format (e.g. BluRay vs. HD-DVD for DVDs, Beta vs. Video 2000 vs. VHS for video cassettes), the more different videos for purchase and rental are offered. The same is true for game consoles (e.g. Xbox vs. Playstation vs. Wii) for the number and variety of games offered; or in the case of operating systems and software for PCs; or apps for smartphones. Markets which are characterised by indirect network effects are referred to as system markets. In many industries and business models, direct and indirect network effects are linked: a mobile phone customer chooses a mobile network operator due to direct network effects of a family-and-friends pricing model and free calls within a network (so-called on-net tariffs). Indirect network effects lead to the choice of Android or iOS as operating system to either use certain apps from Google Play or Apple’s AppStore or to establish compatibility with existing devices or data formats. <?page no="84"?> 2 Customer behaviour, products and network effects 84 Decisions on joining a network Decisions taken by customers and firms in network or system markets differ significantly from those taken in normal product markets, as the utility to a customer now consists of two components: a stand-alone utility resulting from the use of a particular technology or product, and the actual network utility, which increases the utility depending on the number of members. ► Figure 2.20 on the left shows a situation where a single network (which is not in direct competition with other networks) already exists but currently has only a few members - examples are the fax machine shortly after its launch (initially under the product names Bildtelegraph or Hellschreiber) in 1930 or the Sony Minidisc in 1992. Figure 2.20: Decisions to join a single network. Before deciding to join a network, a potential new customer will, in addition to the costs of joining (e.g., the cost of purchasing the fax machine and learning how to operate it), in particular form expectations about the future size of the network and evaluate the stand-alone utility of the network. Obviously, the stand-alone utility of a fax machine does not exist, because the utility only comes from other customers with whom it is possible to communicate via the fax machine. On the other hand, a minidisc player also has a stand-alone utility because no other users of the technology are required for private music recordings or back-up copies. In the early stages of a market with a small number of network members, only customers with a high stand-alone utility will buy the product and thus, access to the network. Whereas in later stages, after a critical mass of customers has been reached and the technology has penetrated or spread, the network effect will dominate, as shown in ► Figure 2.20 on the right (Economides 1996). Customers' expectations regarding the long run number of members in a network determine whether a network can establish itself. Expectations of low membership - if the vast majority of potential customers expects that a network will have few members in the long run, they will not join. As a consequence, only customers with a high stand-alone utility will buy the product. Usually, the critical mass is not reached, and the product disappears from the market. This happened to the minidisc: in 2013, Sony has finally stopped producing and distributing minidisc players. <?page no="85"?> 2.3 Network effects and multisided markets 85 Expectations of high membership - if a large number of potential customers expect that a network will have a large number of members in the long run, many will join based on this (then self-fulfilling) expectation. Thus, (independent of the stand-alone utility) the number of members actually increases and many other customers see their expectations of increasing membership confirmed - the critical mass is reached and the network establishes itself in the market. Once the critical mass is reached, the number of members of the network often grows exponentially. This has been the case with CDs and CD players - many customers have recognised the superiority of the CD over LPs (longer playing time, better durability, lower weight and smaller format) and have (correctly) expected that the CD including CD player will prevail over LPs and turntables and establish itself as a system market. As a consequence, only large networks exist in the long run. If the critical mass is not reached, the network collapses; if the critical mass is exceeded (often called tipping point), the network stabilises due to self-reinforcing growth with large numbers of members or users. Network effects do not only play a role for decisions of customers: firms also have to decide whether to join networks, e.g., concerning IT operating systems depending on the variety of software, banks for ATM networks, or hotels for booking systems. In markets with network effects a chicken-egg problem arises: the critical mass describes the necessary number of members to make the network attractive for further customers - if the network is small, nobody joins the network due to its current size. For example, the fax machine has not reached critical mass for a long time because potential customers have not developed a willingness to pay because of the small number of connected machines. Reaching critical mass for fax machines took until the 1960's: success was initiated when Xerox added fax functionality to printers in 1964 - the proliferation of fax machines in multifunction printers grew rapidly, critical mass was reached and subsequently the chicken-egg problem was solved. Decisions on joining competing networks If several platforms or network operators compete, customers' decisions are more complex. ► Figure 2.21 shows a competitive situation in which two platform operators (e.g., social media platforms A and B, for example in Germany currently the career networks Xing and LinkedIn) with different numbers of customers are already established on the market. A number of potential new customers have different preferences: some have a preference for platform A, others for platform B. In addition, another platform C, which may also be of interest to customers, is considering entering the market. <?page no="86"?> 2 Customer behaviour, products and network effects 86 Figure 2.21: Decisions of customers and firms with competing platforms. In an initial situation of low membership and similar market shares of both platforms A and B, all customers will decide whether to join one of the networks according to their original preferences, based on their respective stand-alone utility. In case of Xing and LinkedIn, these standalone ratings will of course be shaped by the surrounding bubble of colleagues or friends in addition to the functionality of the platforms. If preferences of potential new customers are equally distributed, then as a consequence the market shares of both platforms will fluctuate randomly, so that for two platforms the market shares will oscillate around 50%, as shown in ► Figure 2.22 on the left. Figure 2.22: Competition between incompatible networks. However, if one platform (by chance due to the order of accession of new members) achieves a relevant lead in market share, network effects due to relative market shares now occur in addition to the stand-alone valuations of customers. Once the network effect dominates the stand-alone effect, many potential new members will no longer decide according to their own <?page no="87"?> 2.3 Network effects and multisided markets 87 preferences but in favour of the larger network (► Figure 2.22 right). As a result, the market share of the market leader will move towards 100% (winner-takes-it-all markets), while competitors will lose market share. Such competitive processes explain the displacement of Werkennt-Wen by Facebook; the overwhelming market share of Microsoft DOS and subsequently Windows over all other operating systems in the 1990s and 2000s; and that VHS was able to prevail over the competing video formats Beta and Video 2000 (Arthur 1989, David 1985 as well as Liebowitz and Margolis 1995). Once a critical mass of customers has chosen a network, a lock-in effect may occur. The customers are then strongly bound to their decision due to investments in devices, learning, or habituation effects and also influence the decisions of other customers. If network effects do not clearly dominate the stand-alone effect or customers have very heterogeneous preferences, competing networks can also coexist permanently - this is (at least in the early 2020s) the case with game consoles or operating systems for smartphones. Network effects often lead to a de facto standard of the technology being used, a tendency towards a natural monopoly (► Chapter 7). This market dominating role of a single firm is often the result of an accidental or small event in the past which, via path dependencies, determines the future development of a technology and market leadership (Arthur 1989). A path dependency restricts possible future strategies based on past decisions and investments, or makes a change time-consuming or costly. Future decisions are then not independent of the status quo and past decisions. In addition, significant strategic barriers to entry arise (► Chapter 4), so that the dominant technology prevents or blocks alternative technological developments through a lock-in effect. Customers cannot switch to competing IT platforms or media formats even in the long run and are bound to a firm or technology. A permanent coexistence of several networks is only possible if the networks are fully or partially compatible with each other. Customers then do not have to make a permanent or unambiguous decision and may even be customers of several platforms at the same time (so-called multihoming). Strategies in markets with network effects From a management perspective, competition in network and system markets requires a number of special strategic decisions which are mutually dependent on one another (Farrell and Klemperer 2007, Katz and Shapiro 1986 and 1994, Shy 2011 as well as Koski and Kretschmer 2004). Building the customer base and achieving critical mass: in network and systems markets, rapid growth in the number of customers must be achieved through suitable strategies to establish critical mass or even a lock-in effect. In many cases, this is achieved by freemium models, in which a free variant of the product is offered (Adobe Acrobat Reader, Skype, or Spotify); by permanently free offers (Facebook, Google Search, or Android); or by precisely addressing early adopters in niche markets, which are then decisive for the decisions of late adopters. Expectation management: as customer expectations are essential for the establishment of a network or a large market share, firms try to influence customer expectations regarding the current or future size of a network through marketing investments: Parship and Tinder as the market leader in dating apps and platforms in Germany directly signal the greatest benefit to customers and thus, attract more customers. <?page no="88"?> 2 Customer behaviour, products and network effects 88 Complementary products: disclosure of interfaces supports the creation and distribution of complementary products and services, which in turn increase the utility of a platform and attract new members. Due to its use of the Android operating system, Google has been able to increase its lead in the variety of apps since 2015 from 1.3 million to 2.8 million, compared to 1.2 million to 2.2 million for Apple in 2017. During the same period Android's market share increased from 76% to 82%; iOS' market share decreased from 18% to 14% (data worldwide; source idc.com and statista.com). Switching costs: if customers have invested in devices of a certain technology, or have spent years maintaining their Facebook profile, then technical, monetary, or social switching costs prevent them from switching to a more powerful network or better technology. If firms with a large market share can increase the switching costs towards a competing solution, the market share is stabilised. Switching costs are caused by incompatibility. Apple, for example, increases its customers' switching costs every year with new connector systems, chargers, or necessary software upgrades; banks deliberately make it more difficult for customers to change their banking account; and mobile network operators have prevented or delayed mobile number portability for many years. The more pronounced the lock-in effect and the lower the compatibility, the higher the switching costs. Standard wars versus co-ordinated standardisation: firms have to decide whether they will try (possibly with high investments) to establish their own technology or product as a standard in the market (as in the case of VHS vs. Beta vs. Video 2000) or whether they will enter a joint standard with other firms (as in the case of CDs and CD players by Sony and Philips based on the Redbook standard), foregoing market dominance and high market shares, yet increasing the probability to survive. Degree of compatibility: if other firms have already gained relevant market shares, partial or total compatibility of the technology used can provide access to their customer base. Conversely, a leading firm can limit or eliminate compatibility. In Germany, for example, the savings banks are trying to stabilise their market shares by charging high fees (i.e., limited compatibility) at ATMs; mobile network operators have stabilised their market shares over many years by charging termination and roaming fees. Both strategies have meanwhile been restricted by competition policy. Competition in the market or for the market: as network effects can lead to market dominance in terms of technology or market shares, firms have to decide whether to compete "in the market" or "for the market". This decision is significantly influenced by switching costs, compatibility and the market share already achieved by other networks. The relevance and interdependencies of these challenges become obvious when analysing the market entry of PayDirekt (the online payment system of German banks and savings institutions) in 2016. An online payment system is based on direct and indirect network effects: the more customers use the payment system, the more merchants are interested in implementing the payment system as well, so that self-reinforcing network effects take place. PayDirekt competes with PayPal, Amazon Payments and Sofortüberweisung, among others, all of which have already reached a critical mass. Both the connected shops and the customers have extensive switching costs and in particular PayPal and Amazon Payments prevent compatibility with their networks. As a consequence, PayDirekt's market success did not materialise, and the firm was taken over by its competitor Giropay in 2020 (Nestlér 2016 and Atzler 2017). <?page no="89"?> 2.3 Network effects and multisided markets 89 Platforms as twoor multisided markets Why do many firms try to place advertising on Facebook? Why does Amazon offer other firms the opportunity to sell products on Amazon Marketplace? The explanation is that indirect network effects can also arise because the utility of one group of market participants is increased by the existence and size of another group of market participants. Firms which operate platforms bring together several user or customer groups in such a way that communication or transactions are only possible due to the existence of the platform, which does not act as an intermediary but instead provides the marketplace in the shape of a multisided market. Amazon Marketplace can be considered as a two-sided platform based on indirect network effects: the more B2C customers visit or use the Amazon platform, the more third-party shops are opened, and therefore the more attractive the platform becomes for both market sides; the more participants are on one market side, the more participants are attracted on the other market side and vice versa. The mutual interaction of indirect network effects across several market sides or customer groups often leads to platform business models that are referred to as twoor multisided markets (Evans et al. 2006, Rysman 2009 as well as Rochet and Tirole 2003). The logic of a twoor more generally multisided market is shown in ► Figure 2.23. There are two (or more) complementary and interacting market sides A and B or user groups - members of Facebook and advertising firms, or game developers. Both market sides benefit from the platform through mutual indirect network effects with growth and size of the other group - the more members Facebook has, the more attractive the platform is for game developers and advertisers. Within one or both sides of the market there are strong direct or indirect network effects to achieve a critical mass of customers or providers - in case of Facebook strong positive direct network effects of members among themselves. There is a free or subsidised market side A, which attracts numerous members based on significant direct (and sometimes indirect) network effects - albeit with an absolutely low willingness to pay and a relatively high price elasticity of demand, for members Facebook is permanently free of charge. A second paid or high-priced market side B, where services are provided or developed due to the large number of members on market side A. Typically the market side with relatively weak network effects subsidises the market side with stronger network effects with correspondingly lower price elasticity of demand. For example, game developers pay royalties to Facebook to sell games to members, and advertising space is also booked. A platform or platform firm that enables and controls interactions or transactions between and mutual access to the market sides based on indirect and direct network effects (Evans and Schmalensee 2016 as well as Bolt and Tieman 2008). <?page no="90"?> 2 Customer behaviour, products and network effects 90 Figure 2.23: Two-sided market, network effects and prices. ► Figure 2.24 illustrates both the Facebook and Google platforms. One market side of Facebook consists of members who become or have become members of Facebook as a result of direct network effects. The larger the number of members, the larger the utility of each individual member. However, as the number of members increases, third party-firms have incentives to offer products to Facebook members on the Facebook platform (be it games, apps, advertising), or by using member data. These products are complementary to the actual platform and represent indirect network effects for all Facebook members. Figure 2.24: Business models of Facebook and Google as multisided markets. <?page no="91"?> 2.3 Network effects and multisided markets 91 Similarly, Google's business model can be illustrated as a multisided market in ► Figure 2.24 on the right. Customers use search functions, maps and route planners, the video platform YouTube, the smartphone operating system Android and many other functions free of charge. As the number of users of the platform increases, self-reinforcing indirect network effects arise. The Google search algorithm is improved by a growing number of search queries and attracts more customers, so that more firms are now interested in optimising their website for search queries on Google. This in turn leads to better search results for all customers. In addition, there are further indirect network effects as firms are encouraged to develop more apps or services for Google users. The more varied and high-quality the services offered, the more new members are attracted, so that a self-reinforcing effect is created. As a consequence, an ecosystem based on indirect network effects is created, for which a clear market definition is difficult (Filistrucchi et al. 2014). Platform business models platform market side 1 market side 2 free / lowpriced side paid / highpriced side example game consoles retail customer game developer console games xBox, Playstation, Wii smartphone operating system user app developer operating system apps Android credit cards retail customer merchant card usage terminals and payments Visa, Mastercard data and file transfer users firms PDF reader PDF writer Adobe dating portals users advertising firms women men Parship, ElitePartner, match.com etc. social-mediaplatforms members advertising firms members ads and data Facebook, Instagram etc. job portal job seekers employers search ads and data Monster, Stepstone etc. newspapers readers advertisers newspapers ads and campaigns FAZ, Süddeutsche Zeitung etc. B2Bmarket places firms platform provider membership fulfilment and payments Amazon Marketplaces, Alibaba, Expedia, booking etc. travel and holiday portals travelers hotels reviews bookings TripAdvisor etc. Table 2.3: Platform business model based on multisided markets and indirect network effects. <?page no="92"?> 2 Customer behaviour, products and network effects 92 Following the same business model logic, other twoor multisided platforms such as Open- Table, ebay, LinkedIn, Flixbus, Delivery Hero, booking.com, Parship, Uber, Airbnb, or TripAdvisor operate. However, these business models are in no way limited to digital platforms as shown in ► Table 2.3: also, credit card systems such as Visa or Master, newspapers such as the FAZ or The Guardian, traditional shopping centres such as the Main-Taunus Zentrum or game consoles such as the Xbox or PlayStation apply the same strategic rationale. Competition and strategies in multisided markets The number of competing platforms and the competitive situation depends on whether customers are members of only one of the platforms (singlehoming) or of several platforms simultaneously (multihoming), and if communication is possible across platform boundaries and at least partial compatibility exists. If multihoming is possible and additional costs of using several platforms are low, competition between platforms might be possible even in the longrun - this is the case with booking platforms or job portals as well as with the use of transport service providers such as Uber, Lyft, or MyTaxi. If the investment or switching costs are high or one of the platforms has already reached a critical mass, singlehoming is typical. This is the case with social media platforms or game consoles. In the long run, market dominance of one of the platforms is not unlikely. With increasing product differentiation between the platforms, the intensity of competition decreases. In Germany, the market for dating platforms is therefore still heavily fragmented into individual market segments, and people can also use free apps such as Tinder or even Facebook. In contrast, competition in Germany for career platforms has now been reduced to LinkedIn and Xing: product differentiation is low, but costs of multihoming increase for each customer as the number of contacts increases, especially due to rising switching costs. The price structure between and on both sides of the market is of crucial importance: a platform firm must choose prices for both sides of the market in such a way that, in addition to generating profits, there are in particular incentives on both sides of the market for new members to join in large numbers. Empirically, depending on the business model (► Table 2.3), pricing models are observed which use asymmetrical pricing for two market sides 𝐴𝐴 and 𝐵𝐵 (Rochet and Tirole 2003, Krüger 2009 and Armstrong 2006). The demand 𝑞𝑞 𝐴𝐴 of market side 𝐴𝐴 and the demand 𝑞𝑞 𝐵𝐵 of market side 𝐵𝐵 can be explained by (2.20) 𝑞𝑞 𝐴𝐴 = 𝐷𝐷 𝐴𝐴 (𝑝𝑝 𝐴𝐴 , 𝑞𝑞 𝐵𝐵 ) and (2.21) 𝑞𝑞 𝐵𝐵 = 𝐷𝐷 𝐵𝐵 (𝑝𝑝 𝐵𝐵 , 𝑞𝑞 𝐴𝐴 ) . Demand on one side of the market is obviously dependent on the price of using the platform (in the form of a membership fee) and the demand or number of members on the other side of the market. The direct price elasticity of both market sides 𝐻𝐻 can then, with the number of members of the other market side remaining unchanged, be determined by differentiating the demand function with respect to a price change (2.22) 𝜀𝜀 𝑖𝑖 = 𝑝𝑝𝑖𝑖 𝑞𝑞𝑖𝑖 𝜕𝜕𝐷𝐷𝑖𝑖 𝜕𝜕𝑝𝑝𝑖𝑖 for 𝐻𝐻 = 𝐴𝐴, 𝐵𝐵 . <?page no="93"?> 2.3 Network effects and multisided markets 93 The mutual strength of the indirect network effect can be estimated as (2.23) 𝜃𝜃 𝐴𝐴 = 𝑞𝑞𝐵𝐵 𝑞𝑞𝐴𝐴 𝜕𝜕𝐷𝐷𝐴𝐴 𝜕𝜕𝑞𝑞𝐵𝐵 and 𝜃𝜃 𝐵𝐵 = 𝑞𝑞𝐴𝐴 𝑞𝑞𝐵𝐵 𝜕𝜕𝐷𝐷𝐵𝐵 𝜕𝜕𝑞𝑞𝐴𝐴 as the elasticity of the network effect: the larger e.g., 𝜃𝜃 𝐵𝐵 , the stronger the growth in demand 𝑞𝑞 𝐵𝐵 from market participants on market side 𝐵𝐵 due to increasing membership 𝑞𝑞 𝐴𝐴 on market side 𝐴𝐴 . The price elasticity of both market sides 𝐻𝐻 in relation to the number of members of the other market side can be determined by the total differential of the demand function as (2.24) Ε 𝑖𝑖 = 𝑝𝑝𝑖𝑖 𝑞𝑞𝑖𝑖 𝑑𝑑𝑞𝑞𝑖𝑖 𝑑𝑑𝑝𝑝𝑖𝑖 for 𝐻𝐻 = 𝐴𝐴, 𝐵𝐵 . If we now differentiate both demand functions (2.20) and (2.21) according to the prices of both sides of the market (2.25) 𝑑𝑑𝑞𝑞𝐴𝐴 𝑑𝑑𝑝𝑝𝐴𝐴 = 𝜕𝜕𝐷𝐷𝐴𝐴 𝜕𝜕𝑝𝑝𝐴𝐴 + 𝜕𝜕𝐷𝐷𝐴𝐴 𝜕𝜕𝑞𝑞𝐵𝐵 𝑑𝑑𝑞𝑞𝐵𝐵 𝑑𝑑𝑝𝑝𝐴𝐴 and 𝑑𝑑𝑞𝑞𝐵𝐵 𝑑𝑑𝑝𝑝𝐵𝐵 = 𝜕𝜕𝐷𝐷𝐵𝐵 𝜕𝜕𝑝𝑝𝐵𝐵 + 𝜕𝜕𝐷𝐷𝐵𝐵 𝜕𝜕𝑞𝑞𝐴𝐴 𝑑𝑑𝑞𝑞𝐴𝐴 𝑑𝑑𝑝𝑝𝐵𝐵 and plug in (2.22) and (2.23), then price elasticities of both sides of a market depending on both price elasticities and the strength of the network effects are (2.26) Ε 𝐴𝐴 = 𝜀𝜀𝐴𝐴 1−𝜃𝜃𝐴𝐴𝜃𝜃𝐵𝐵 and Ε 𝐵𝐵 = 𝜀𝜀𝐵𝐵 1−𝜃𝜃𝐴𝐴𝜃𝜃𝐵𝐵 . So, an absolute price elasticity 𝛦𝛦 𝐴𝐴 of market side 𝐴𝐴 is determined by the direct price elasticity 𝜀𝜀 𝐴𝐴 depending on the mutual elasticity of network effects 𝜃𝜃 𝐴𝐴 𝜃𝜃 𝐵𝐵 and the number of participants on the other market side. As explained in ► Chapter 1, the smaller the price elasticity, the larger the willingness to pay of customers and the larger the pricing range of firms, all other things being equal. From (2.26) it can be seen that a platform firm sets asymmetric prices vis-à-vis the market sides depending on the relative price elasticities. In extreme cases, prices are set at zero on one market side or even members are payed for participation in a platform; the other market side then subsidises via correspondingly high prices and solely creates revenues (Evans and Schmalensee 2007 and 2016). From a management perspective, the following decisions are crucial when building a platform business model. Pricing strategy and price structure across market sides: numerous clubs, bars and dating platforms (at least those trying to initiate heterosexual partnerships) offer women free or discounted entry or membership, thus, increasing the willingness to pay of men. Pricing models should not only take into account the willingness to pay of one market side, but in particular ensure that the other market side has sufficient incentives to join or participate in the platform through the elasticity of the network effect. For example, Voigt and Hinz (2015) show that revenues of an online dating platform are maximised if 36.2% of the platform's members are women, revenues are then 17.2% higher as if the platform is split 50/ 50 between men and women. Number, interaction and relative size of market sides: ebay has two market sides with buyers and sellers, LinkedIn has three market sides with professionals, employer firms, and recruiting firms. In addition, a platform firm has to determine which transactions (e.g,. communication and products) are possible via the platform, especially to prevent members of <?page no="94"?> 2 Customer behaviour, products and network effects 94 the two market sides from bypassing the platform in the future. The relative size of the market sides must ensure the incentive structure for membership of the platform. 2.4 Summary and key learnings How do customers behave? This is not only a complex question, yet the answer depends on whether we assume customers to decide fully rational or not. Rational customer behaviour is characterised by decisions that, based on preferences, increase utility and are limited by budget constraints. Utility-maximising behaviour can be traced back to the demand of a customer or, in an aggregated way, of a market segment or an entire market. Preferences and demand are of course neither stable over time nor exogenous: firms can influence the location of a demand curve, in particular through product quality and marketing, and change the willingness to pay of customers and the size of the market. In addition, demand changes as a result of changes in tastes, income and relative prices, so that different product categories and their interrelationship can be identified. From a management perspective, the forms of product differentiation and personalisation are particularly crucial, since the intensity of competition decreases with increasing product differentiation perceived by customers. Firms and managers can then develop two strategies: either learn about the preferences of the customers and position their products appropriately, or change the customers’ tastes and preferences so that the firms’ products appear more attractive. In numerous new business models, utility does not depend on the consumption of a product, but via network effects on the number of other members on the same platform. Direct communication on social media platforms is more attractive, the more users are on the platform. In many network effects-driven industries, few firms dominate after reaching a critical mass. In addition, firms can realise business models based on indirect network effects by combining several market sides on platforms as an ecosystem. With that, a range of platform-specific competitive strategies become important, such as quickly creating a customer base, decisions on compatibility and complementary products, or establishing switching costs. Recommendations for further reading If you want to delve deeper into utility theory, Perloff, J.M., Microeconomics - theory and applications with calculus, Harlow 2018, is a good choice. A comprehensive account of customer behaviour from a business perspective with a focus on marketing applications is Solomon, M.G, Consumer behavior: buying, having, and being, New York 2016. A very readable introduction on platforms and multisided markets can be found in Evans, D.S. and Schmalensee, R., Matchmakers - the new economics of multisided platforms, Boston 2016. <?page no="95"?> 2.4 Summary and key learnings 95 Questions for review [1] Describe practical areas of the analysis of customer behaviour from a microeconomic perspective. [2] Concisely define utility, what is marginal utility - how can both be measured? What is a budget constraint, what does it depend on? [3] Briefly describe the difference between normal and inferior products and give two examples each. What are substitutive and complementary products, how can their relationship be measured? [4] Explain different degrees of uncertainty in product characteristics and give three examples each. [5] What does the income elasticity of demand describe, and how can it inform management decisions? [6] Explain the concept of a product life cycle. Provide three examples to explain how products have been displaced by other products. [7] How can you define a market, what possibilities do firms have to influence the demand function? [8] Name the goals, forms and effects of different forms of product differentiation. [9] What are typical characteristics of network and system markets? [10]Define direct and indirect network effects and give two examples of each. [11]What does the formation of a network depend on? What is the role of expectations here? Describe typical strategies of firms in network markets. [12]Explain the basic idea of platforms as multisided markets. What is the competitive advantage based on? Name and explain three examples of digital platforms with reference to single-/ multihoming and asymmetry of the pricing structure. Literature Abraham, M., Mitchelmore, S., Collins, S., Maness, J., Kistulinec, M., Khodabandeh, S., Hoenig, D. and Visser, J., Profiting from personalization, BCG Perspectives 2017. Allcott, H., Braghieri, L., Eichmeyer, S. and Gentzkow, M, The welfare effects of social media, American Economic Review, 2020, 110, 3, 629-676. Anderson, C.R. and Zeithaml, C.P., Stage of the product life cycle, business strategy, and business performance, Academy of Management Journal, 1984, 27, 1, 5-24. Armstrong, M., Competition in two-sided markets, Rand Journal of Economics, 2006, 37, 3, 668-691. Arthur, W.B., Competing technologies, increasing returns, and lock-in by historical events, Economic Journal, 1989, 99, 394, 116-131. Ashley, R., Granger, C. W. and Schmalensee, R.H., Advertising and aggregate consumption: an analysis of causality, Econometrica, 1980, 48, 5, 1149-1167. Atzler, E., Paydirekt droht erneut verkorkstes Jahr, Handelsblatt, 31. Mai 2017, 8. Bagwell, K., The economic analysis of advertising, in: Armstrong, M. and Porter, R. (eds.), Handbook of Industrial Organization, Volume 3, 2007, 1701-1844. <?page no="96"?> 2 Customer behaviour, products and network effects 96 Balafoutas, L., Beck, A., Kerschbamer, R. and Sutter, M., What drives taxi drivers? A field experiment on fraud in a market for credence goods, Review of Economic Studies, 2013, 80, 3, 876-891. Berry, S.T., Levinsohn, J. and Pakes, A., Differentiated products demand systems from a combination of micro and macro data: the new car market, Journal of Political Economy, 2004, 112, 1, 68-105. Bolt, T. and Tieman, A.F., Heavily skewed pricing in two-sided markets, International Journal of Industrial Organization, 2008, 26, 5, 1250-1255. Bowles, S. Endogenous preferences: the cultural consequences of markets and other economic institutions, Journal of Economic Literature, 1998, 36, 1, 75-111. Brynjolfsson, E., Collis, A. and Eggers, F., Using massive online choice experiments to measure changes in well-being, Proceedings of the National Academy of Sciences, 2019, 116, 15, 7250-7255. Bundesverband Musikindustrie (eds.), Musikindustrie in Zahlen, Berlin 2019. Chatmi, A. and Elasri, K., Competition among vortex firms: marketing, R&D or pricing strategy, Journal of High Technology Management Research, 2017, 28, 1, 29-46. Clark, K.B. and Fujimoto, T., Automobilentwicklung mit System: Strategie, Organisation und Management in Europa, Japan und USA, Frankfurt-New York 1992. Clark, K.B., Chew, W.B. and Fujimoto, T., Product development in the world auto industry, Brookings Papers on Economic Activity, 1987, 3, 729-771. Cox, W.E., Product life cycles as marketing models, Journal of Business, 1967, 40, 4, 375-384. Datta, H., Knox, G. and Bronnenberg, B.J., Changing their tune: how consumers’ adoption of online streaming affects music consumption and discovery, Marketing Science, 2018, 37, 1, 5-21. David, P.A., Clio and the Economics of QWERTY, American Economic Review, 1985, 75, 2, 332-337. Deaton, A., Estimating own-and cross-price elasticities from survey data, Journal of Econometrics, 1987, 36, 7-30. Deaton, A., Price elasticities from survey data, Journal of Econometrics, 1990, 44, 281-309. Deaton, A., Understanding consumption, Oxford 1992. Degryse, H., On the interaction between vertical and horizontal product differentiation: an application to banking, Journal of Industrial Economics, 1996, 44, 2, 169-186. Dickson, P.R. and Ginter, J.L., Market segmentation, product differentiation, and marketing strategy, Journal of Marketing, 1987, 51, 3, 1-10. Dorfman, R., and Steiner, P.O., Optimal advertising and optimal quality, American Economic Review, 1954, 44, 5, 826-836. Economides, N., The economics of networks, International Journal of Industrial Organization, 1996, 14, 6, 673-699. Evans, D.S and Schmalensee, R., The industrial organization of markets with two-sided platforms, Competition Policy International, 2007, 3, 1. Evans, D.S. and Schmalensee, R., Matchmakers - the new economics of multisided platforms, Boston 2016. Evans, D.S., Hagiu, A. and Schmalensee, R., Invisible engines: how software platforms drive innovation and transform industries, Cambridge 2006. Farrell, J. and Klemperer, P., Coordination and lock-in: competition with switching costs and network effects, in: Armstrong, M. and Porter, R. (eds.), Handbook of Industrial organization, Volume 3, 2007, 1967-2072. Ferreira, R.D.S. and Thisse, J.F., Horizontal and vertical differentiation: the Launhardt model, International Journal of Industrial Organization, 1996, 14, 4, 485-506. Fiegenbaum, A. and Thomas, H., Strategic groups as reference groups: theory, modeling and empirical examination of industry and competitive strategy, Strategic Management Journal, 1995, 16, 461-476. <?page no="97"?> 2.4 Summary and key learnings 97 Filistrucchi, L., Geradin, D., van Damme, E. and Affeldt, P., Market definition in two-sided markets: theory and practice, Journal of Competition Law and Economics, 2014, 10, 2, 293-339. Frank, R.H. and Cartwright, E., Microeconomics and behavior, Maidenhead 2013. Fujimoto, T., The long tail of the auto industry life cycle, Journal of Product Innovation Management, 2014, 31, 1, 8-16. Gabszwicz, J.J. and Thisse, J.F., On the nature of competition with differentiated products, Economic Journal, 1986, 96, 381, 160-172. Gali, J., Keeping up with the Joneses: consumption externalities, portfolio choice, and asset prices, Journal of Money, Credit and Banking, 1994, 26, 1, 1-8. Gambardella, A. and Torrisi, S., Does technological convergence imply convergence in markets? Evidence from the electronics industry, Research Policy, 1998, 27, 5, 445-463. Ghemawat, P., Competition and business strategy in historical perspective, Business History Review, 2002, 76, 1, 37-74. Goldberg, P.K., Product differentiation and oligopoly in international markets: the case of the US automobile industry, Econometrica, 1995, 63, 4, 891-951. Gönsch, J., How much to tell your customer? - A survey of three perspectives on selling strategies with incompletely specified products, European Journal of Operational Research, 2020, 280, 3, 793-817. Gruber, H., Market structure, learning and product innovation: evidence for the EPROM market, International Journal of the Economics of Business, 1995, 2, 87-101. Guajardo, J.A., Cohen, M.A. and Netessine, S., Service competition and product quality in the US automobile industry, Management Science, 2015, 62, 7, 1860-1877. Houthakker, H.S. and Taylor, L., Consumer demand in the United States: analyses and projections, Cambridge 1970. Johnson, J.P. and Myatt, D.P., On the simple economics of advertising, marketing, and product design, American Economic Review, 2006, 96, 3, 756-784. Kahneman, D. and Tversky, A., Experienced utility and objective happiness: a moment-based approach, Psychology of Economic Decisions, 2003, 1, 187-208. Katz, M.L. and Shapiro, C., Systems competition and network effects, Journal of Economic Perspectives, 1994, 8, 2, 93-115. Katz, M.L. and Shapiro, C., Technology adoption in the presence of network externalities, Journal of Political Economy, 1986, 94, 4, 822-841. Klepper, S., Entry, exit, growth, and innovation over the product life cycle, American Economic Review, 1996, 86, 562- 583. Koski, H. and Kretschmer, T., Survey on competing in network industries: firm strategies, market outcomes, and policy implications, Journal of Industry, Competition and Trade, 2004, 4, 1, 5-31. Kretschmer, T. and Peukert, C., Video killed the radio star? Online music videos and recorded music sales, Information Systems Research, 2020. Krüger, M., Elasticity rules in two-sided markets, Review of Network Economics, 2009, 8, 271-278. Lancaster, K., Consumer demand: a new approach, New York 1971. Lehman-Wilzig, S. and Cohen-Avigdor, N., The natural life cycle of new media evolution: inter-media struggle for survival in the internet age, New Media & Society, 2004, 6, 6, 707-730. Levin, R.C. and Reiss, P.C., Cost-reducing and demand-creating R&D with spillovers, NBER Working Paper No. 2876, 1989. Levitt, T., Exploit the product life cycle, Harvard Business Review, 1965, 43, 81-94. Liebowitz, S.J. and Margolis, S.E., Path dependence, lock-in, and history, Journal of Law, Economics and Organization 1995, 11, 1 205-226. <?page no="98"?> 2 Customer behaviour, products and network effects 98 Luce, R.D., Utility of gains and losses: measurement-theoretical and experimental approaches, Boston 2014. Malhotra, A. and Gupta, A.K., An investigation of firms’ strategic responses to industry convergence, Academy of Management Proceedings, 2001, 1, 1-6. Monteiro, G.F.A. and Foss, N., Resources and market definition: rethinking the “hypothetical monopolist” from a resource-based perspective, Managerial and Decision Economics, 2017, 38, 1, 1-8. Motta, M., Competition policy - theory and practice, Cambridge 2004. Nelson, P., Information and consumer behaviour, Journal of Political Economy, 1970, 78, 2, 311-329. Nenycz-Thiel, M. and Romaniuk, J., The real difference between consumers' perceptions of private labels and national brands, Journal of Consumer Behaviour, 2014, 13, 4, 262-269. Nestler, F., Deutsche Banken versagen mit Paydirekt, FAZ, 13. August 2016, 14. Nichols, W., Advertising Analytics 2.0, Harvard Business Review, 2013, 91, 3, 60-68. o.V., Die Cola, die deinen Namen trägt, FAZ, 16. Januar 2016, 17. Rochet, J.C. and Tirole, J., Platform competition in two-sided markets, Journal of the European Economic Association, 2003, 1, 4, 990-1029. Rohlfs, J., A theory of interdependent demand for a communications service, Bell Journal of Economics and Management Science, 1974, 5, 1, 16-37. Rysman M., The economics of two-sided markets, Journal of Economic Perspectives, 2009, 23, 3, 125-143. Schmidt, J., Makadok, R. and Keil, T., Customer-specific synergies and market convergence, Strategic Management Journal, 2016, 37, 5, 870-895. Shaked, A. and Sutton, J., Product differentiation and industrial structure, Journal of Industrial Economics, 1987, 36, 2, 131-146. Shy, O., A short survey of network economics, Review of Industrial Organization, 2011, 38, 2, 119-149. Siegel, E., Predictive analytics: the power to predict who will click, buy, lie, or die, in: Siegel, E. (ed.), Predictive analytics, Hoboken 2013, 103-110. Simon, H., Dynamics of price elasticity and brand life cycles: an empirical study, Journal of Marketing Research, 1979, 16, 4, 439-452. Solomon, M.R. Consumer behavior: buying, having, and being, Upper Saddle River 2014. Sridhar, S., Narayanan, S. and Srinivasan, R., Dynamic relationships among R&D, advertising, inventory and firm performance, Journal of the Academy of Marketing Science, 2014, 42, 3, 277-290. Starbucks, Starbucks Fiscal Annual Report, Seattle 2017. Statistisches Bundesamt, Statistisches Jahrbuch 2017 - Kapitel 6: Einkommen, Konsum, Lebensbedingungen, Wiesbaden 2017. Taylor, L. and Halvorsen, R., Energy substitution in US manufacturing, Review of Economics and Statistics, 1977, 59, 4, 381-388. van Alstyne, M.W., Parker, G.G. and Choudary, S.P., Pipelines, platforms and the new rules of strategy, Harvard Business Review, April 2016, 2-9. Vesanen, J., What is personalization? A conceptual framework, European Journal of Marketing, 2007, 41, 5/ 6, 409-418. Voigt, K., Murawski, C., Speer, S. and Bode, S., Hard decisions shape the neural coding of preferences, Journal of Neuroscience, 2019, 39, 4, 718-726. Voigt, S. and Hinz, O., Network effects in two-sided markets: why a 50/ 50 user split is not necessarily revenue optimal, Business Research, 2015, 8, 1, 139-170. <?page no="99"?> 2.4 Summary and key learnings 99 Ward, F. and King, G.A., Interfiber competition with emphasis on cotton, US Department of Agriculture, Economic Research Service, Washington 1973. Wiecek-Janka, E., Papierz, M., Kornecka M. and Nitka, M., Apple products: a discussion of the product life cycle, International Conference on Management Science and Management Innovation (MSMI 2017). Wold, H. and Jureen, L., Demand analysis, New York 1953. Yamada, A., Fukuda, H., Samejima, K., Kiyokawa, S., Ueda, K., Noba, S. and Wanikawa, A., The effect of an analytical appreciation of colas on consumer beverage choice, Food Quality and Preference, 2014, 34, 1-4. Zeithaml, V.A., Consumer perceptions of price, quality, and value: a means-end model and synthesis of evidence, Journal of Marketing, 1988, 52, 3, 2-22. <?page no="101"?> 101 3 Decisions under risk and behavioural economics In many firms, important decisions are prepared based on a business case (a strategic financial plan), where revenues and costs are projected and possible profits are determined. In this context, senior management executives often request a base case, a best case and a worst case in order to better assess potential effects of uncertain outcomes of a decision. Similarly, customers buy a new car or a flat screen TV and additionally purchase a warranty or even a warranty extension. People also take out insurances for funeral costs, wedding cancellation costs, or for their mobile phone. This behaviour is based on the perception of risk: people are aware that future developments are not fully predictable and want to protect themselves against possible consequences before making a decision. Decisions under risk and uncertainty always circle around the question of how much risk a person is willing to take and whether this risk is worth taking. In this sense, people form expectations about likely future developments to derive their decisions against the background of their individual risk attitude and risk aversion. People who are supposedly unwilling to take risks often stick to existing solutions or strategies - especially in firms: routines ("we've always done it this way") are often applied. This so-called status quo effect describes that people value strategies, coffee mugs, or habits higher if they own them or have been using them for a longer time. Such behaviour is often described as irrational and analysed in the context of behavioural economics. Behavioural economics studies economic decisions and action based on insights from psychology and behavioural sciences. It originates from the work of Simon (1955) on boundedly rational decisions in firms, and Tversky and Kahneman (1971 ff.) conducting experiments on risk attitude, preferences, and decisions. People seem to systematically, regularly and predictably, deviate from maximising behaviour, although not always and not in every situation. For example, numerous empirical studies and experiments on framing describe that preferences are not absolutely stable, but can be influenced by a changed arrangement of products. Emotions, bias of perception and loss aversion help to understand how decisions are made - both as a boundedly rational customer when making purchasing decisions and as a boundedly rational manager when making strategic decisions (DellaVigna 2009 and Powell et al. 2011). A situation that illustrates the typical interplay of both risk and behavioural economics is that of a football goalkeeper in advance of a penalty kick. A goalkeeper usually knows about the probability distribution of the kicks and the strengths of the player, at the same time spectators and teammates do not expect the goalkeeper to catch the ball - at least in the German Bundesliga this only happened in 18.9% of cases between 1963 and 2004 (Dohmen 2008). The player, in turn, is generally expected to score. First, the goalkeeper will form an expected value and make a decision based on given probabilities. The player likewise weighs all conceivable strategies against the possible probabilities of the goalkeeper's behaviour. Both the goalkeeper and the player take a decision under uncertainty. However, if the goalkeeper thinks back to the last penalty kicks to derive an optimal decision, he is subject to a gambler's fallacy. Probabilities are based on the law of large numbers and, due to the independence of successive events in the absence of serial correlation, they do not immediately equalise or balance each <?page no="102"?> 3 Decisions under risk and behavioural economics 102 other out in small samples. Even in casinos, an odd number does not "finally" come because a series of even numbers has come before, since random events have no memory (Tversky and Kahneman 1971). In addition, the situation is heavily superimposed by emotions and expectations: players shoot in the middle at about 28% - but goalkeepers remain standing there in only about 10% of cases. A goalkeeper could, perhaps, increase his chances by simply standing still. However, if the player scores, the goalkeeper is booed because of an obvious inactivity. The goalkeeper is thus, subject to an emotional action bias, which gives him an incentive to jump to one of the corners and is significantly influenced by the current result and situation of the game (Bar-Eli et al. 2007, Dohmen 2008 as well as Misirlisoy and Haggard 2014). Managers are subject to a similar urge to act driven by quarterly, monthly, or weekly reports - typically, as a manager, you cannot tell the board that 'nothing' was done during the last quarter (Cyert and March 1963, Gavetti and Rivkin 2007 and Brunsson 1982). What goalkeepers and players should do from a strategic and game theoretical perspective is continued in ► Chapter 9. Learning Objectives This chapter deals with: decisions under the influence of risk and uncertainty, the role of risk attitude and risk aversion and the perception of risk; deviations from optimising or maximising behaviour due to bounded rationality of customers and managers; and psychological effects and behavioural regularities such as loss aversion, endowment effects, or framing. 3.1 Decisions under risk and uncertainty Almost all human decisions are influenced by risk and uncertainty. The reason for this is the limited possibility to fully comprehend any present situation in all its states, to anticipate all possible actions of all market participants and to perfectly predict the future in all conceivable forms. General economic conditions - typically, neither managers in firms nor customers are able to capture all current conditions and environmental states, as in a PEST analysis, or their dynamics in the context of decision-making. This information is either not available, or the information gathering and processing takes too long or is simply too expensive. Competitive behaviour - many decisions of firms must take into account the current behaviour of the relevant competitors, but also anticipate their future decisions which is only possible to some extent (see also ► Chapter 9 on game theory and ► Chapter 10 on strategic competition). Repercussions of own decisions - own decisions can have intended and expected consequences, but may have also unintended and unexpected effects, or cause (sometimes surprising) reactions from others. <?page no="103"?> 3.1 Decisions under risk and uncertainty 103 Coincidences and chance - numerous developments do indeed appear to be coincidental in decision-making situations because a dependence on events, strategies, or actions (in the present and also in retrospect) either does not exist, or cannot be determined and thus a causal explanation is impossible. However, managers and customers still have to make decisions. As a consequence, firms make assumptions about possible future developments and these forecasts or plans are backed with probabilities in scenarios. In the future, of course, only one out of these possible developments will then become a concrete reality. Toner et al. (2015) have listed the core risks and influencing factors for strategic decisions from a management perspective along the four dimensions of macroeconomic environment, technology, competitive environment and customers: a firm must keep an eye on at least 80 mutually influencing and uncertain developments and take them into account when making decisions. From an economics perspective, this means incomplete information must be taken into account whenever making decisions. In summary, the overall macroeconomic conditions, competitive behaviour of firms, repercussions of own decisions, as well as coincidences and pure chance can be described as possible events which occur with probabilities. The type and degree of incomplete information is classified following Knight (1921): Risk - describes a decision situation for which all possible future events are known and for which objective probabilities exist. Decisions can then be derived based on expected values. Uncertainty - describes a decision situation in which either not all possible events are known and/ or no objective probabilities exist. Decisions under risk with objective probabilities happen, for example, in games of chance such as roulette or rolling dice (there is a 1 in 6 probability of a player throwing a one with a dice) and where empirical data from the past or repeatable situations allow the estimation of probabilities or a probability distribution. Decisions under uncertainty are typical when events occur only once, or if only subjective probabilities exist (e.g., based on estimates or assumptions by a manager). In reality, it is sometimes difficult to judge whether a situation is risky or uncertain. On the one hand, people's ability to deal with risk and probabilities seems to be generally limited and the perception of probabilities is distorted by psychological effects. On the other hand objective probabilities only apply if the law of large numbers and structural stability of the probability distribution apply (Tversky and Kahneman 1971, 1973 and 1974, Gigerenzer 1990, March and Shapira 1987 as well as Taleb et al. 2009). If experiments are used to investigate whether people prefer to make decisions based on subjective probabilities and uncertainty, or based on objective probabilities and risk, then risk (with a known probability distribution) is regularly preferred over uncertainty, even if perceived probabilities are identical (Ellsberg 1961 and Camerer and Weber 1992). Expected value and variance An initial way to make decisions under risk is to determine the expected value of all possible events, taking into account the distribution of probabilities, and to assess risk based on variance. <?page no="104"?> 3 Decisions under risk and behavioural economics 104 Events - describe completely all conceivable future situations - in the case of six-sided dice, for example, the numbers 1 to 6. Probability - describes the relative frequency that a certain event occurs or will happen. Probabilities can be given as distributions or discrete values - in case of rolling dice the probabilities are 1/ 6 for each event and add up to 1. Expected value - corresponds to the value of the payoffs ( 𝑥𝑥 1 to 𝑥𝑥 𝑝𝑝 ) of all 𝑆𝑆 possible events, weighted with probabilities 𝑝𝑝𝑟𝑟 𝑖𝑖 (probability): 𝐸𝐸𝐸𝐸(𝑥𝑥) = 𝑝𝑝𝑟𝑟 1 ⋅ 𝑥𝑥 1 + 𝑝𝑝𝑟𝑟 2 ⋅ 𝑥𝑥 2 + ⋯ + 𝑝𝑝𝑟𝑟 𝑝𝑝 ⋅ 𝑥𝑥 𝑝𝑝 - in case of a perfect six-sided dice 𝐸𝐸𝐸𝐸(𝑥𝑥) = 1 ⁄ 6 ⋅ 1 + 1 ⁄ 6 ⋅ 2 + 1 ⁄ 6 ⋅ 3 + 1 ⁄ 6 ⋅ 4 + 1 ⁄ 6 ⋅ 5 + 1 ⁄ 6 ⋅ 6 = 3.5 . Variance - as a typical measure of risk, variance measures the squared and weighted deviations between the expected value and the events that actually occur: 𝜎𝜎 2 = 𝑝𝑝𝑟𝑟 1 [𝑥𝑥 1 − 𝐸𝐸𝐸𝐸(𝑥𝑥)] 2 + 𝑝𝑝𝑟𝑟 2 [𝑥𝑥 2 − 𝐸𝐸𝐸𝐸(𝑥𝑥)] 2 + ⋯ + 𝑝𝑝𝑟𝑟 𝑝𝑝 [𝑥𝑥 𝑝𝑝 − 𝐸𝐸𝐸𝐸(𝑥𝑥)] 2 - in case of the six-sided dice 𝜎𝜎 2 = 16 [1 − 3.5] 2 + 1/ 6[2 − 3.5] 2 +. . . +1/ 6[6 − 3.5] 2 = 2.92 . The lower the variance, the smaller are deviations from the expected value, the lower the risk. High values of variance, on the other hand, indicate that events are widely dispersed from the expected value and that there is a high risk. Strategic decisions involving risk - investments to build a new airport and its opening date, defining a global marketing strategy for a consumer goods manufacturer, or focusing on electric mobility for an automobile manufacturer - are structurally similar to game situations based on chance (coin tosses, lotteries, dice, or roulette). ► Figure 3.1 shows three games of chance: depending on heads and tails, the specified payoffs will happen with 50% probability each. Figure 3.1: Gambling with heads and tails. If students are offered a one-off participation in each of the three games, an overwhelming majority of around 85% takes part in game 1; with game 2, the willingness to participate declines sharply; and only around 5% of students want to play game 3. Looking at the expected value and variance of the three games in Table 3.1, a first explanation for these decisions be- <?page no="105"?> 3.1 Decisions under risk and uncertainty 105 comes evident. Although expected values rise from game 1 to game 2 to game 3, perceived risk expressed by the variance increases significantly and apparently dominates decisions of potential players - the risk of losing a relatively large amount of EUR 2,500 is a deterrence to participating in game 3. Coin tossing game 1 game 2 game 3 heads 199 300 5000 tails -1 -100 -2500 expected value 99 100 1250 variance 10,000 40,000 14,062,500 expected utility (initial wealth of 3,000 EUR) 24.913 24.907 24.211 Table 3.1: Tossing coins, expected value and variance. Utility function and risk aversion However, the decision on one-time participation in these games is (in the vast majority of cases) not made once players have explicitly calculated expected values and variances. Rather, the decision is based on weighing up the utility of participating in the game and the resulting possible change in wealth. Figure 3.2: (Expected) utility and (expected value of) wealth. <?page no="106"?> 3 Decisions under risk and behavioural economics 106 ► Figure 3.2 on the left outlines a situation based on a von Neumann-Morgenstern utility function in which two people - one with previous wealth of EUR 100, the other with previous wealth of EUR 100,000 - each receive a gift of EUR 100. The person with low wealth will certainly be more pleased about the doubling of assets and an increase of 100%, compared to the wealthy person whose wealth will increase by just 0.1%. The reasoning behind this is that wealth or income (in almost all societies) provides a utility, but marginal utility decreases with increasing wealth (von Neumann and Morgenstern 1944) - the same applies to decisions by managers on profits in firms. ► Figure 3.2 on the right shows that the marginal utility of an increase in wealth is larger the smaller wealth is - the utility function 𝑢𝑢 of wealth 𝑊𝑊 is then typically concave and can generally be described as (3.1) 𝑢𝑢 = 𝑢𝑢(𝐻𝐻𝐴𝐴𝑎𝑎𝐴𝐴𝑡𝑡ℎ) = 𝑊𝑊 𝜔𝜔 in which 𝜔𝜔 < 1 describes the curvature of a person's utility function. Using the utility function (3.1), the expected utility of participation in the games of chance shown in ► Figure 3.1 can now be checked. Suppose a female student currently has a total wealth 𝑊𝑊 of 3,000 EUR with an arbitrarily assumed value 𝜔𝜔 = 0.4, so that her current utility level is (3.2) 𝑢𝑢 = 𝑊𝑊 𝜔𝜔 = 3000 0.4 = 24.595 . The expected utility EU in terms of gambling can generally be described as (3.3) 𝐸𝐸𝐸𝐸 = 𝑝𝑝𝑟𝑟 1 (𝑊𝑊 + 𝑥𝑥 1 − 𝑝𝑝) 𝜔𝜔 + 𝑝𝑝𝑟𝑟 2 (𝑊𝑊 + 𝑥𝑥 2 − 𝑝𝑝) 𝜔𝜔 + ⋯ + 𝑝𝑝𝑟𝑟 𝑝𝑝 (𝑊𝑊 + 𝑥𝑥 𝑝𝑝 − 𝑝𝑝) 𝜔𝜔 where 𝑝𝑝 is the price of participation in the game of chance and 𝑥𝑥 1 to 𝑥𝑥 𝑝𝑝 are possible risky gains or losses. If the student participates free of charge, i.e., 𝑝𝑝 = 0, in one of the three games shown in ► Figure 3.1, her assets will increase by 𝑥𝑥 1 or decrease by 𝑥𝑥 2 with 50% probability in each case, so that expected utilities are (3.4) 𝐸𝐸𝐸𝐸(𝑆𝑆𝑎𝑎𝑆𝑆𝐴𝐴 1) = 0.5(3,000 + 199) 0.4 + 0.5(3,000 − 1) 0.4 = 24.913 (3.5) 𝐸𝐸𝐸𝐸(𝑆𝑆𝑎𝑎𝑆𝑆𝐴𝐴 2) = 0.5(3,000 + 300) 0.4 + 0.5(3,000 − 100) 0.4 = 24.907 and (3.6) 𝐸𝐸𝐸𝐸(𝑆𝑆𝑎𝑎𝑆𝑆𝐴𝐴 3) = 0.5(3,000 + 5,000) 0.4 + 0.5(3,000 − 2,500) 0.4 = 24.211 . Two observations can be made: the valuation based on the ranking of expected utilities deviates from the ranking of expected values of the games (see also ► Table 3.1) and by comparing her current utility level (3.2) with the possible expected utilities from (3.4) to (3.6), the student would want to participate in games 1 or 2, but not in game 3. If the student is free to choose which game to play, she would participate in game 1 - it has the highest expected utility given her initial wealth. Game 3 would definitely be rejected because the expected utility is less than the utility from her current wealth. The reason for this decision is that the student seems to shy away from the risk of losing 2,500 EUR. This behaviour is generally described as risk aversion with a preference for secure versus insecure wealth situations of equal expected values: when deciding between several alternative strategies with identical expected values, the strategy with the lowest risk is always chosen; <?page no="107"?> 3.1 Decisions under risk and uncertainty 107 when deciding between several alternative strategies with identical risks, the strategy with the highest expected value is always chosen; when deciding between several alternative strategies with different expected values, the strategy with the highest expected utility is always chosen; the utility function of people with risk aversion is concave, the marginal utility increases disproportionately with increasing wealth and 𝜔𝜔 < 1 describes the degree of risk aversion - the smaller 𝜔𝜔 , the larger the risk aversion. Empirical measurement of risk attitude An experimental procedure for measuring attitude towards risk or risk aversion is to offer potential players participation in two lotteries in ten consecutive rounds, the attractiveness of which (measured by the expected values) shifts successively from lottery A to lottery B (Holt and Laury 2002). Empirical assessment of risk preference based on Holt-Laury lotteries lottery A lottery B difference between expected values of both lotteries round probability payoff probability payoff expected value probability payoff probability payoff expected value pr x pr y EV(A) pr x pr y EV(B) Δ EV = EV(A) - EV(B) 1 0,10 2,00 0,90 1,60 1,64 0,10 3,85 0,90 0,10 0,48 1,17 2 0,20 2,00 0,80 1,60 1,68 0,20 3,85 0,80 0,10 0,85 0,83 3 0,30 2,00 0,70 1,60 1,72 0,30 3,85 0,70 0,10 1,23 0,50 4 0,40 2,00 0,60 1,60 1,76 0,40 3,85 0,60 0,10 1,60 0,16 5 0,50 2,00 0,50 1,60 1,80 0,50 3,85 0,50 0,10 1,98 -0,18 6 0,60 2,00 0,40 1,60 1,84 0,60 3,85 0,40 0,10 2,35 -0,51 7 0,70 2,00 0,30 1,60 1,88 0,70 3,85 0,30 0,10 2,73 -0,85 8 0,80 2,00 0,20 1,60 1,92 0,80 3,85 0,20 0,10 3,10 -1,18 9 0,90 2,00 0,10 1,60 1,96 0,90 3,85 0,10 0,10 3,48 -1,52 10 1,00 2,00 0,00 1,60 2,00 1,00 3,85 0,00 0,10 3,85 -1,85 Table 3.2: Holt-Laury lottery, numbers are partially rounded (cf. Holt and Laury 2002, p.1645). In ► Table 3.2 lotteries A and B are described - lottery A has a permanently lower variance than lottery B and until the fourth round of the game expected value of A is higher than expected value of B. A risk neutral participant should choose the lottery with the highest expected value without taking variance into account, so that up to round 4 lottery A is chosen and from round <?page no="108"?> 3 Decisions under risk and behavioural economics 108 5 lottery B is chosen. If a participant switches to lottery B earlier, he is obviously willing to take risks (also called risk-loving). On the other hand, the later a participant switches to lottery B, the higher his degree of risk aversion. In fact - as shown in ► Table 3.3 - no clear picture emerges from numerous experiments and studies. On average, about two thirds of the participants in such Holt-Laury lotteries show risk-averse behaviour, about 20% are risk-neutral decision-makers and about 15% show riskloving behaviour, but the results of the studies vary greatly. Inconsistent or even irrational behaviour can also be observed. Empirical risk preferences Study risk seeking risk neutral risk averse irrational other Participants Holt and Laury 2002 7.1 % 23.0 % 58.3 % 13.2 % students Anderson and Mellor 2009 4.7 % 20.8 % 74.5 % students Mascalet et al. 2009 2.7 % 10.2 % 87.1 % students, entrepreneurs and employees Pennings and Smidts 2000 61.0 % 4.0 % 35.0 % firm owners Bellemare and Shearer 2006 17.6 % 13.7 % 37.3 % 31.4 % employees Günther and Detzner 2012 18.7 % 17.0 % 29.4 % 35.0 % managers Günther and Detzner 2012 18.0 % 27.0 % 26.0 % 29.0 % employees of controlling department Table 3.3: Empirical attitudes to risk in Holt-Laury lotteries (cf. also Vanini 2016). Risk aversion and risk premium People who make risk-averse decisions reject fair games. Fair games have an expected value of zero: the chances of winning and losing are equal. In this case, a currently available secure wealth can be compared with an uncertain expected value of the same amount. Such a fair game is illustrated in ► Figure 3.3 on the left. A potential player with a utility function 𝑢𝑢 = 𝑢𝑢(𝑊𝑊) = 𝑊𝑊 0.5 is offered to bet 50 EUR in a game of chance via a coin toss that ends with a 50% probability at 10 EUR or at 90 EUR for him/ her. The chance of winning and the risk of loss are (3.7) 𝐸𝐸𝐸𝐸(𝑥𝑥) = 0.5 ⋅ (+40) + 0.5 ⋅ (−40) = 0 the same size, so that the expected value of wealth is (3.8) 𝐸𝐸𝐸𝐸(𝑊𝑊) = 0.5 ⋅ (50 + 40) + 0.5 ⋅ (50 − 40) = 50, <?page no="109"?> 3.1 Decisions under risk and uncertainty 109 which is exactly the same as initial wealth, i.e., represents a fair game. There are three possible decisions: rejecting the fair game - thus, valuing the secure 50 EUR higher than the risky expectation of 50 EUR - the player is risk-averse and does not wager his 50 EUR in the game; accepting the fair game - rating the secure 50 EUR lower than the risky expected value of 50 EUR - this player is risk-seeking, wagers 50 EUR in the game and has either 10 EUR or 90 EUR once the coin is tossed; indifferent to participating in a fair game - the secure 50 EUR is obviously valued just as high as the risky expected value of 50 EUR - so this player is risk-neutral. Figure 3.3: Faire game and risk premium. ► Figure 3.3 on the left shows that a risk-averse player with the utility function 𝑢𝑢 = 𝑢𝑢(𝑊𝑊) = 𝑊𝑊 0.5 rejects this fair game. The two possible events, loss of 40 EUR and gain of 40 EUR, result in points A and B on the utility function. Their respective utility level is 𝑢𝑢(𝑊𝑊 = 10) = 10 0.5 = 3.16 and 𝑢𝑢(𝑊𝑊 = 90) = 90 0.5 = 9.49, but each with a probability of 50%. The expected utility of a fair game can then be read off from the connecting line AB of uncertain events at the level of the expected value 𝐸𝐸𝐸𝐸(𝑊𝑊) = 50 at point 𝐶𝐶 as (3.9) 𝐸𝐸𝐸𝐸(𝑊𝑊) = 0.5 ⋅ 10 0.5 + 0.5 ⋅ 90 0.5 = 6.32 . However, this expected utility is less than the utility from the current wealth of EUR 50, (3.10) 𝑢𝑢(𝑊𝑊) = 50 0.5 = 7.07, so that the player refuses to play the game because of 𝐸𝐸𝐸𝐸(𝑊𝑊) = 6.32 < 7.07 = 𝑢𝑢(𝑊𝑊) . In ► Figure 3.3 on the left this can also be seen from the fact that point C is below point B, because the connecting line AB of the uncertain events is below the concave utility function of secure wealth. <?page no="110"?> 3 Decisions under risk and behavioural economics 110 A risk-averse person is generally willing to pay for the exclusion or reduction of risks. This amount is known as the risk premium: a risk premium is the difference between a secure and an insecure wealth situation at the same utility level. In ► Figure 3.3 on the right, the risk premium is drawn as a horizontal line from point C to point E on the utility function. At point E the utility is apparently the same as at point C, but the utility here is secure - unlike at point C. The secure utility at E is based on the certainty equivalent wealth 𝑊𝑊 𝑆𝑆 , which provides the same utility as a participation in a fair game. The risk premium results from (3.11) 𝑢𝑢(𝑊𝑊 𝑆𝑆 ) = 𝑊𝑊 𝑆𝑆 0.5 = 6.32 and the certainty equivalent wealth (3.12) 𝑊𝑊 𝑆𝑆 = 6.32 2 = 39.90 as (3.13) 𝑅𝑅𝑃𝑃 = 𝑊𝑊 𝜕𝜕 − 𝑊𝑊 𝑆𝑆 = 50 − 39.90 = 10.10, i.e., the player would be willing to pay up to 10.10 EUR in order not to participate in the risky game. This consideration is not only correct in this abstract case: many people take out insurance policies to limit or exclude risks - the insurance fee corresponds to a risk premium. People also buy public transport tickets to limit the damage if they are caught without a ticket. The same applies to pay parking fees - the risk of receiving a fine is weighed against the price of parking. Figure 3.4: Degree of risk aversion and size of risk premium. The amount of the risk premium depends on the degree of risk aversion. The higher the level of risk aversion, the larger the amount a risk-averse decision-maker would want to spend to limit or eliminate a risk - conversely, of course, a lower level of risk aversion also tempts people to park illegally and use public transport illegally. ► Figure 3.4 shows a situation that is 0 A B C E EV(W) EU 0 A‘ B‘ C‘ E‘ EV(W) W A W B W R W A W B W S ‘ W R W S RP‘ RP EU weak risk aversion strong risk aversion <?page no="111"?> 3.1 Decisions under risk and uncertainty 111 fully described by the risky situation 𝑊𝑊 𝐴𝐴 and 𝑊𝑊 𝐵𝐵 , and the expected value of wealth 𝑊𝑊 𝜕𝜕 - the two figures only differ in the curvature of the utility function due to the different degrees of risk aversion of two decision makers. The risk aversion on the left is relatively low, so the risk premium is lower than in the case of relatively high-risk aversion on the right, since the degree of risk aversion is reflected in the certainty equivalent wealth, so that (3.14) 𝑅𝑅𝑃𝑃 = 𝑊𝑊 𝜕𝜕 − 𝑊𝑊 𝑆𝑆 < 𝑊𝑊 𝜕𝜕 − 𝑊𝑊 𝑆𝑆 ′ = 𝑅𝑅𝑃𝑃‘ . Case Study │ Who Wants to Be a Millionaire The game show Who Wants to Be a Millionaire, first on television in 1998 in the UK, has been running very successfully in numerous countries for decades - a candidate gets to answer 15 questions, can make use of several lifelines (called jokers in Germany) and win up to 1 million EUR in Germany (for the rules, see https: / / en.wikipedia.org/ wiki/ Who_ Wants_to_Be_a_Millionaire%3F). Although the game is partly based on knowledge, usually situations arise in which the candidate is totally lost - and therefore at risk - and they have to decide whether to end the game and keep the previous profit amount (depending on the question value), or to choose one of the remaining answers at random (if no lifelines are left). If the candidate is right, his question value rises and he can continue playing, if he is wrong his gain drops to a lower level and he is eliminated from the game. Numerous studies have taken the game as an opportunity to empirically examine the risk aversion and decision-making behaviour of candidates: men and women do not differ significantly in their risk attitude, younger candidates are more risk averse than older candidates; and the relative risk aversion remains constant during the game as the amount of gains increases (Hartley et al. 2014, Daghofer 2007 as well as Franzen and Pointner 2009). In order to follow Who wants to be a Millionaire with a microeconomic eye, the situation can be structured as follows: A participant in Who Wants to Be a Millionaire has a utility function 𝑢𝑢(𝑊𝑊) = 𝑊𝑊 0.6 . She is at a question value of 64,000 EUR and can drop back to 16,000 EUR if the answer is wrong. With two remaining answering alternatives she has no idea about the 125,000 EUR question and no more lifelines - what will she do? What is the risk premium? What does it mean in this case? If we first transfer the current profit level of EUR 64,000 and the two possible random events EUR 16,000 or EUR 125,000 to a curved utility function, the situation can be analysed using ► Figure 3.5. The utility for the candidate at the profit level of 64,000 EUR is then 𝑢𝑢(𝑊𝑊) = 64,000 0.6 = 765.08 . Although the expected value of continuing to play 𝐸𝐸𝐸𝐸(𝑊𝑊) = 1 ⁄ 2 ⋅ 16,000 + 1 ⁄ 2 ⋅ 125,000 = 70,500 is larger than the current profit level, the expected utility decreases to 𝐸𝐸𝐸𝐸(𝑊𝑊 𝜕𝜕 ) = 12 ∙ 16,000 0.6 + 1/ 2 ∙ 125,000 0.6 = 738.14, if they randomly choose one of the two remaining answers. Given the risk aversion, the candidate would not want to gamble, but choose to end the game with EUR 64,000. <?page no="112"?> 3 Decisions under risk and behavioural economics 112 Figure 3.5: Decision-making in Who-wants-to-be-a-millionaire. If the host asks the candidate and urges her to continue playing, this contradicts the candidate's risk preference - in the end she would be willing to pay a risk premium to be allowed to stop. This is shown in ► Figure 3.5 as the distance between points E and F and above 𝐸𝐸𝐸𝐸(𝑊𝑊 𝜕𝜕 ) = 738.14 with (3.15) 𝑊𝑊 𝑆𝑆 = 𝐸𝐸𝐸𝐸(𝑊𝑊 𝜕𝜕 ) 1 0.6 = 60,288.32 with (3.16) 𝑅𝑅𝑃𝑃 = 64,000- 60,288.32 = 3,711.68 . Television stations could, with a small change of rules, get kickbacks of the prize money from the candidates in the form of the risk premium. Risk premium, capital markets and firm value A risk premium might also play a central role in an opposite scenario: as a necessary payment to a decision-maker so that he will take or run a risk. This consideration is, for example, key when analysing portfolio decisions in financial and capital markets, in which a decision has to be made between a supposedly risk-free investment (overnight money or fixed-term money with a low but stable return) and riskier forms of an investment (shares with a higher but fluctuating return). The risk premium in this case expresses the necessarily expected excess return of the risky investment over the risk-free investment, so that a private or institutional investor is prepared to bear the risk. This consideration is fundamental to the capital asset pricing model to determine the value of risky investments or the prices of risky stocks and indirectly the cost of capital and the value of a firm (more on this in ► Chapter 4). Entrepreneurial activity means taking risks and investing equity capital and is carried out in particular if a positive risk premium can be achieved compared to other alternatives (Markowitz 1952, Sharpe 1964). Bonus payments to managers linked to the share price of a firm are incentives based on a risk premium. <?page no="113"?> 3.1 Decisions under risk and uncertainty 113 Risk preference and managerial behaviour Almost all people are risk averse in many decision-making situations. This does not mean that dangerous sports are not practised, that people do not play lottery, or that managers do not take risks: risk aversion merely describes the basic attitude of weighing up risks against the background of a diminishing marginal utility of assets or wealth. People differ interpersonally in their risk aversion, and risk aversion does not appear to be stable over time. Numerous studies show that risk aversion increases with the absolute and relative level of income; and with demographic characteristics such as individual age or the size of the family. In particular, assessment of risk is influenced by experience (success and failure) in certain risky situations; the decision-making framework; and, of course, the absolute level of risk and maximum possible loss (March and Shapira 1987, Pratt 1964, Holt and Laury 2002 as well as Kahneman and Tversky 1979). In addition, behaviour in risky situations is influenced by the possibility to manage risk or to insure against risks. From a management perspective, a central question is what degree of risk aversion decisionmakers have; how risky and uncertain decisions are influenced; and how an organisational setting of a firm, and in particular remuneration systems, influence the behaviour of managers in risky situations. The aim is always to make risk transparent by designing both the organisation and the decision-making processes and then to make decisions. In many organisations and firms, risk management and risk culture, as well as the separation of risk from uncertainty, are underdeveloped. For example, variable incentives in remuneration systems (performance-related bonus payments as a form of risk premium) are seen as an explanation for excessive risk taking by managers and as an indirect driver of the financial and public debt crisis from 2007 onwards. Empirically, this relation has not yet been finally clarified. Some studies show a positive correlation between managers' risk appetite and firm success, since lower risk aversion leads to riskier investments with higher returns. On the other hand, the variance of firm success increases. The willingness to take greater risks depends not only on positive incentives in the event of success, but also in particular on negative sanctions (repayment of salaries or withholding of bonuses) and liability for managerial action in the event of failure. In particular, the perception and assessment of potential risks and their symmetrical or asymmetrical repercussions on one's own income have a major influence on the risk attitude and behaviour of managers (Ross 2004, Holmström 1999, Carpenter 2000 and Coles et al. 2006). Risk management and reduction of risk and uncertainty Firms conduct strategic risk management in order to optimise decisions under risk and uncertainty. The focus is on identifying and analysing potential risks and assessing (via expected values, variances and simulations) whether the risk appears acceptable or should be reduced by appropriate measures. In addition, there are numerous regulatory requirements to manage or reduce risks. For example, a risk management system for capital market-related firms in Germany is prescribed by the German Act on Corporate Control and Transparency (Gesetz zur Kontrolle und Transparenz im Unternehmensbereich); and banks and financial service providers must also comply with the minimum requirements for risk management of the German <?page no="114"?> 3 Decisions under risk and behavioural economics 114 Financial Supervisory Authority (Bundesanstalt für Finanzdienstleistungsaufsicht) in line with the respective competitive strategy. All methods and measures to reduce risks try to achieve a better assessment of the probabilities or the exploitation of the law of large numbers. In case of firms, this can be done either internally within the firm, by making use of the market via insurance or the capital market, or through signalling as follows. Information: obtaining additional and supplementary information enables a better assessment of probabilities and possible events and, if necessary, the transformation of uncertainty into risk or even certainty. This is typically done by monitoring market prices for risks (e.g., risk premiums for securities), by collaboration with management consultants or doing market research. The maximum price for additional information is again depending on the risk premium. Insurance: insurance excludes risks or reduces possible financial losses. The availability of insurance is based on the law of large numbers. Although some risks will materialise, by pooling all the risks taken on independently of each other, an insurance firm can exist based on collected risk premiums and can carry risks and cover claims. Diversification: proverbial "not putting all your eggs in one basket", diversification means to spread risk in order not to lose everything. Diversification is typical in all markets with technological uncertainties or high R&D intensity. In financial investments, diversification is generally achieved by allocation and spreading of overall risk across a portfolio of several projects. Diversification is based on negatively correlated events, i.e., investments are made in projects with (expected) opposite event characteristics in order to reduce the absolute risk. If no such projects exist within the firm, a firm can alternatively invest indirectly via the capital market in another firm or directly through an acquisition of another firm in order to spread and thus reduce the risk (see also ► Chapter 6). Signalling describes a way of credibly communicating information to another market participant or competitor that cannot be observed or determined in any other way (Spence 1973 and 2002). Signalling reduces asymmetric information, or the degree of incomplete information. This information can be used to assure customers or suppliers that a collaboration or a business relationship does not involve any (or at least better assessable) risks. Typical forms are detailed annual reports on risk-situation, in particular the preparation of credit ratings or certificates of compliance with environmental or quality standards. Examples of indirect signals include skyscrapers of banks in the centre of the most expensive cities around the world - banks produce intangible services and products, but expensive real estate signals financial solidity. To make sure a receiver classifies this information as credible, the creation or transmission of the signal must be costly or complex; or it must be difficult or risky for third parties to forge or imitate the signal. Signalling is the main economic explanation for expensive engagement rings or expensive tailor-made suits: both help to reduce the uncertainty of the other side of the market about the financial situation of a potential spouse or a potential new business partner. Signalling thus, also has an immediate meaning for students, because hiring a new employee means a lot of uncertainty for the employer. By means of an elaborately acquired degree, which is associated with opportunity costs and corresponding skills, the future employer receives risk-reducing information for the selection of applicants - thus, the risk premium is reduced and in return the salary and probability of employment increases. <?page no="115"?> 3.1 Decisions under risk and uncertainty 115 Weighing of risk and black swans Even when objective probabilities exist, people sometimes do not decide in favour of a maximum expected utility as described above. The reason for this is the limited ability to process all relevant information, understand risks and correctly perceive probabilities: low probabilities are perceived as subjectively larger, high probabilities as smaller. For example, while many people consider car and air travel to be risky, the subjective fear of a plane crash is typically high, contrary to the actual probabilities. Figure 3.6: Actual and perceived probabilities. ► Figure 3.6 shows a weighting function of actual and perceived probabilities, which results from experiments in the assessment or evaluation of risk situations. In very simplified terms, three areas of risk perception can be identified: very low probabilities are overestimated (area A); moderate probabilities are clearly underestimated (area B); and for high probabilities (area C) perception of risk is growing rapidly (Kahneman and Tversky 1979 as well as Tversky and Kahneman 1992). The weighting function runs relatively steep near the extreme values of 0% and 100%, so that people here react very strongly to a change in probabilities. For example, people are prepared to pay significantly higher amounts to reduce a risk from 1% to 0% than from 50% to 40%, although the risk reduction is larger in the second case. The overestimation of low probabilities can explain, for example, both participation in lotteries with low expected values, and also too much protection against small risks in the form of insurance. The perception of small differences in probabilities changes along the S-shaped, non-linear course of the weighting function. This is one possible explanation for the so-called Allais paradox: in experiments it is regularly observed that people evaluate and make decisions differently if the probability increases from 10% to 11% than if it increases from 99% to 100% (also called a certainty effect, Allais 1953). <?page no="116"?> 3 Decisions under risk and behavioural economics 116 From a managerial point of view, it is central to appreciate the perception of risk by managers in addition to the actual risk aversion. In the context of risk or uncertainty, if one offers managers the choice of three different scenarios for an investment case or a business model (base case, best case and worst case), they will tend to choose the middle option due to an aversion against extreme cases (so called extremeness aversion). Thus, there is an unsubstantiated expectation that this base case will also occur. In the same way, it is observed that managers are surprised by events that lie towards tails of probability distributions. This realisation of supposedly unlikely events is now also proverbially referred to as black swans, which were long thought not to exist (Simonson and Tversky 1992, Garbuio et al. 2014 and Taleb 2007). 3.2 Bounded rationality and behavioural economics In ► Chapter 2, decisions made by people, whether as customers or managers, were described as being completely rational based on preferences without any emotion or bias of perception that maximise utility. This view seems to be based on people like Mr Spock (perfect rationality without any emotion) from Star Trek, Sheldon Cooper (stable preferences) from Big Bang Theory and Sherlock Holmes (perfect cognitive abilities) in the novels of Arthur Conan Doyle - extremely clever guys who are constantly maximising. Mr Spock, Sheldon Cooper and Sherlock Holmes are the prototypes of the (often criticised) homo oeconomicus, but Spock is a Vulcan, Cooper suffers from Asperger's syndrome, and Holmes is a drug addict - all three are fiction. In reality, however, many decisions are not made completely rationally and deviate significantly and in systematic patterns from maximising behaviour. These decisions are analysed in the context of behavioural economics and explanations are based in particular on psychological and behavioural experiments and observations. A central explanation for boundedly rational decisions is based on two different modes of functioning of the human brain - fast and slow thinking (Kahneman 2003 and 2011) explained below. Thinking fast (system 1): decisions are based on intuition, instinct and routines. It summarises fast, simple, automated decisions and behaviours that ensure and facilitate survival. The explanations for this are essentially evolution (it has proven effective to jump to the side when hearing squealing tyres on the road without thinking about it for a long time), emotions and routines (adaptive repetition of decisions made in the past). Fast thinking is easy for people: it costs little brain capacity (in the limbic system), uses little energy and is not strenuous. Thinking slow (system 2): decisions are based on thinking, reflection and the application of logic. It summarises decisions and behaviour that require mental concentration, detailed processing of information, or the application of complex decision-making rules. Slow thinking is difficult for people because it requires a lot of brain capacity (in the frontal lobe) and energy. The human brain actively avoids slow thinking, it "automatically" switches to fast thinking to save energy. The human brain essentially decides on its own whether a task is solved by fast or slow thinking: many people cannot prevent their brain from automatically solving a task like 4 + 2 by fast thinking, whereas a task like calculating 231 × 86 - 19,865 requires slow thinking and many people immediately feel effort (or even pain) encountering this task. <?page no="117"?> 3.2 Bounded rationality and behavioural economics 117 Figure 3.7: Observed decisions based on perfect and bounded rationality vs. thinking fast and slow. The tendency to think fast restricts people in the process of information perception, information processing and in actual decision-making. As a result, many decisions are made boundedly rational (Simon 1955 and 1957, March and Simon 1958 as well as March 1991) due to: incomplete understanding of the situation - many decision-making situations are complex, and the possible strategies and their interactions with conceivable objectives cannot be fully described; incomplete information - many decision-making situations are a mixture of uncertainty, risk and a complete lack of information, cognitive limitations - people have limited intellectual abilities and are limited or biased in perception, learning, memory and planning and time constraints in decision-making - many strategic decisions, especially in firms, cannot be fully thought through due to limited time or deadlines. People (decision-makers as customers or managers) certainly recognise the limitations of their rationality - against this background, only a few alternatives are examined and the decisionmaking process is based on heuristics. In addition, instead of maximising utility or profit, people search for good-enough solutions (so called satisficing), in which a certain level of entitlement or satisfaction, a level of planned profit, or a preservation of the status quo is achieved. This behaviour can be observed in many situations in life: with management decisions, when looking for a new job, on someone’s daily way to university or when choosing a partner for life. Fundamentally new decisions are only made if a clear deviation from the level of aspiration is detected; otherwise routines dominate, which continue and keep previous decisions or improve them incrementally based on local search and thus, consolidate or secure the status quo (Lindblom 1959 as well as Levinthal and March 2007). A good-enough solution can be quite consistent with maximising behaviour. Satisficing can also occur when one sticks with the current choice, having taken account of all opportunity costs of an alternative, search costs for complete rationality irrationality bounded rationality observed decisions ‚slow’ thinking ‚fast‘ thinking <?page no="118"?> 3 Decisions under risk and behavioural economics 118 better alternatives and, in particular, time constraints - i.e., maximisation with numerous constraints. Such a good-enough solution provides an explanation for some observed behaviour of taxi drivers in New York. Taxi drivers in New York must decide two days in advance whether to rent a car - then they pay a rental fee for 12 hours of use and can keep all revenues. On the respective day, the rental fee becomes an irrelevant sunk cost and therefore, the only important decision is how many hours the taxi is operated. There are two possible environmental conditions: good days (rain, snow, underground breakdowns) with many passengers and bad days (sun, scheduled underground operation) with rather few passengers. Maximising behaviour of taxi drivers and a positive elasticity of achievable income requires that - as shown in ► Figure 3.8 on the left - on good days the full 12 hours are worked, but on bad days the car is returned to the taxi firm early because no additional revenues are possible. Figure 3.8: Schematic behaviour of taxi drivers in New York. An empirical study by Camerer et al. (1997) shows that a behaviour such as that shown in ► Figure 3.8 centre can be observed instead. The explanation is that a mentally set target income (as a reference point) per day is aimed at in order to cover the costs of renting the car. Once this is reached or exceeded, work is stopped on that day - if it is not reached, work continues. In a recent study Farber (2015) showed that not all taxi drivers act in this way - especially older and more experienced drivers show stronger tendency towards maximising behaviour and optimise their income not only on a daily basis, but over a month or a year. Heuristics and rules of thumb Many boundedly rational decisions can be attributed to the application of heuristics (Gigerenzer and Selten 2002 as well as Kahneman 2003). A heuristic consists in the application of mostly simple rules for decision-making where knowledge is limited, information is incomplete, or if there are time constraints. The emphasis of heuristics is on efficiency (speed, low cost and complexity reduction), not on finding a best possible decision, and often only a limited number of alternatives or strategies are considered. Decisions based on heuristics typically deworking hours on good days working hours on bad days 12 h good days bad days income working hours on good days working hours on bad days 12 h working hours working hours target income = satisficing prediction of behavior real behavior <?page no="119"?> 3.2 Bounded rationality and behavioural economics 119 viate from an optimal decision (and maximising behaviour) - the distance from a theoretically best possible decision then determines the quality of the heuristic. Heuristics cover a range from systematic and logical but simplifying decision rules to trial and error - the simple iterative toying with potential solutions and aborting the search for a best solution once a good-enough solution is found. Simplifying decision rules are based on rules of thumb: there is a systematic or arbitrary simplification of the decision problem and the application of a decision rule that was functional in the past. For example, many people replace the calculation of 19×21 with 20×20 and are only 0.25% wrong in the result; other individuals (mostly men) do not read operating instructions and thus, cannot achieve an optimal solution with a washing machine, but can achieve a good-enough solution by "putting everything in together at 40 degrees and always on maximum spin speed". Bias of perception and decisions Experiments and research in the line of Tversky, Kahneman and Thaler focus on the understanding of behaviour-related heuristics. These heuristics have evolved through evolutionary processes and people are subject to psychologically induced cognitive biases. These biases affect judgements and decisions made in complex situations - typically, this results in goodenough solutions, which at least partially contradict maximising behaviour. Behavioural heuristics in this sense provide explanations for systematic deviations from optimal decisions. From a microeconomic perspective, biases of perception are relevant during perception and processing of information, and of course in the subsequent decision-making. In particular there are the following biases of perception, although they cannot be clearly distinguished and are mutually influencing in many real life situations (DellaVigna 2009, Camerer et al. 2011, Kahneman 2011 and Thaler 2015). Confirmation bias (or selective perception) - information is selected and interpreted in such a way that one's own beliefs and assumptions are confirmed, explained and reinforced. In return, inappropriate information is hidden or suppressed (cognitive dissonance), so that firms adhere to previous decisions or strategies and routines and path dependencies emerge. This is often observed in market research and competitive analysis: market research is often used to confirm existing assumptions and, if necessary, to convince employees and managers of one’s own assumption. Even purely stochastic events are attributed significance, so that managers - depending on their point of view - interpret confirming or refuting information into random patterns. The quality of a decision, on the other hand, could be improved by consciously searching for refuting information (Nickerson 1998). Availability bias - the assessment of a fact or information is dominated by easily available data, clear recollection of certain facts (what is easily recalled must obviously be important) or analogy inferences, even if these are not related to the question or decision. This creates a tendency to systematically overvalue singular events or recent experiences. When asked whether the population of Manchester or Leicester is larger, many German students draw on the completely irrelevant but easily available information that there are two well-known football clubs in Manchester - and happen to answer correctly. British students, on the other hand, have a great deal of information about both cities and are less likely to answer this question correctly (Tversky and Kahneman 1973). <?page no="120"?> 3 Decisions under risk and behavioural economics 120 Figure 3.9: Availability bias and the size of towns (own data collected in various lectures in Saarbrücken, Munich, Hatfield and London between 2014 and 2017). Representativeness - the classification of information is dominated by 'obvious facts': one considers these facts wrongly as statistically representative. Representative heuristics are based on the fact that people violate the Bayesian rule of probabilities due to an attentionrelated overweighting of certain characteristics of an observed event or a wrong classification of the population of possible events (Kahneman and Tversky 1972 as well as Tversky and Kahneman 1974). Case Study │ Tilman the lorry driver Consider this scenario. A person is described as follows: "Tilman is an ascetic man, 42 years old, wears rimless glasses, loves to listen to Mozart and plays chess regularly". Alongside the description comes a picture of a white man who obviously corresponds to stereotypical beauty ideals. When interviewees are asked "Is Tilman a) a lorry driver or b) a professor of cosmetic surgery? ”, the vast majority will answer b). However, the probability that he is a professor of cosmetic surgery is only about 0.8% - in 99.2% of cases he is a lorry driver. With about 1 million lorry drivers between 25 and 65 years of age, and about 40 professors of cosmetic surgery between 40 and 60 years, there is a population of about 25,000 lorry drivers and about 2 professors of plastic surgery aged 42 in Germany. Even if one assumes that the combination of rimless glasses, Mozart and chess is 100 times over-represented among professors of cosmetic surgery, for example, i.e., among professors of plastic surgery in 50% of cases, and among lorry drivers in only 0.5% of cases, the result is 0.5% * 25,000 = 125 lorry drivers who fit the description, but only 50% * 2 = 1 professor of plastic surgery. These and many more heuristics in the perception and processing of information significantly shape and bias decisions of people (including customers and managers) as follows. Overconfidence - describes a systematic overestimation of one's own abilities and knowledge and in particular the precision or correctness of one's decisions. Numerous experiments show that people believe that they can drive a car better than average; make better investment decisions than others; increase their chances of winning the lottery by <?page no="121"?> 3.2 Bounded rationality and behavioural economics 121 selecting their own numbers; or even predict random coin tosses. In less 'random' situations, this tendency towards overoptimism and the illusion of control is even more pronounced and is significantly reinforced by a confirmation bias. Empirical studies show that managers tend to attribute success to their own abilities, while failures are attributed to bad luck and unexpected strategies of competitors, or 'injustice of the market' (Svenson 1981, Russo and Shoemaker 1992, Klayman et al. 1999, Garbuio et al 2014 as well as Malmendier and Tate 2005). Managers' overconfidence and overoptimism increases in particular if a firm’s profits or share price development in previous years was above average (Malmendier and Tate 2015, Chen et al. 2015 and Ben-David et al. 2007). Two business models that are based on overconfidence and overoptimism and increase the willingness to pay of customers are gyms and actively managed mutual funds. On the one hand, people overestimate their discipline (the ability to stick to decisions made over time) to go for regular workouts and pay too much, while on the other hand, fund managers implicitly claim to be able to predict random developments in efficient capital markets. Empirically, however, it has been proven again and again that share price developments correspond to a (yet complicated) random walk - i.e., are essentially purely driven by chance (DellaVigna and Malmendier 2006 as well as Pütz and Ruenzi 2011). Herd behaviour - describes convergent social behaviour or decision-making, in which the judgements or behaviour of individuals in a group are aligned without centralized coordination: everyone does what everyone else is doing, even if own preferences or private information suggest a different decision (Banjeree 1992). Herd behaviour arises and is encouraged when there is uncertainty about one's own decisions, when the decisions of others are readily observable and adaptable, and when utility arises or is expected from conforming or assimilating behaviour. Herd behaviour includes the adaptation of trends and fashions in purchasing or consumption behaviour, the behaviour of investors on the capital market, but also the confirmation and adoption of opinions or statements on social media platforms (Leibenstein 1950, Asch 1956, Shiller 1990, Ngai et al. 2015, Bayer et al. 2020 and Pavlovic 2020). The influence of social media is strengthened by network effects. The stronger the influence of social media (whether for consumption decisions influenced by friends or influencers on Instagram, Facebook, or Pinterest, or for professional decisions in career networks such as LinkedIn or Xing), the stronger the effect on perception and distortion of information and decision-making behaviour. Group think - decisions are decisively influenced by whether they are made alone and independently or in groups. Often, especially in groups of well-informed individuals, there is approval for decisions that turn out to be wrong. Two causes seem to be dominant: the presumption concerning the competence of the others and the lack of energy to lead every discussion. Both causes are reinforced by confirmation bias and lead to wrong agreement on common decisions ("go-with-the-flow"). Often herd behaviour grows with the number of managers or the size of the board and the division of responsibilities. One possible way out of this, as applied by airlines and management consultancies, is the so-called "obligation-to-dissent" principle, i.e., to veto against decisions even in large groups or to hierarchical superiors in a justified way (Esser 1998). Sunk cost fallacy - rational decisions require that sunk costs are not taken into account for future decisions (see ► Chapter 2). However, in many decision situations people do not stick to this: Thaler (1980) showed in an experiment that the number of pizza slices eaten <?page no="122"?> 3 Decisions under risk and behavioural economics 122 in an all-you-can-eat restaurant depends on whether and how much people have paid. If you give money back to a group of customers immediately after entering the restaurant and paying for it at the entrance, they will eat until they are full. Members of a control group who do not get their money back, on the other hand, continue to eat until they feel that at least the value of the payment has been achieved - on average about 40% more. Just and Wansik (2011) confirm this: with a 50% price reduction, customers eat 27.9% less in an all-you-can-eat restaurant - in absolute contradiction to not take sunk costs into account. Similarly, shoe cabinets are piled up with shoes that are either out of fashion or do not fit - the fact that these shoes did cost a lot of money in the past prevents them from being given away. However, sunk costs are costs incurred in the past which cannot be recovered, i.e., these costs are irrelevant for future decisions - but people irrationally stick to past decisions. For example, people will watch bad movies in the cinema to the end (but not at home because they didn't pay) or finish reading bad books they have bought (but not borrowed ones or gifts because they didn't pay) - so that the mental justification of a purchase price paid in the past is the reason for not considering sunk costs (see also ► Chapter 6 on managerial decisions involving sunk costs). Loss aversion and asymmetry of risk In addition to the biases of perception and decision-making, experiments show that people assess risks and uncertainty asymmetrically - potential losses regularly weigh more heavily than potential gains - and use reference points for valuation which take a current status quo or an initial situation into account. The basic idea can be understood looking at the following decision situations (Tversky and Kahneman 1981): Situation 1 - Players can choose between the following options: (A) a safe win of 240 EUR. (B) a win of 1,000 EUR with a 25% probability or a win of EUR 0 with a probability of 75 %. In situation 1, 84% of the participants chose option (A) - the sure win of EUR 240 against an uncertain expectation value of 𝐸𝐸𝐸𝐸(𝑊𝑊) = 0.25 ⋅ 1000 + 0.75 ⋅ 0 = 250 which indicates the participants' risk aversion. Situation 2 - Players have the choice between the following options: (C) a safe loss of −750 EUR (D) a loss of −1,000 EUR with a 75% probability or a loss of EUR 0 with a probability of 25% . In situation 2, 87% of the participants have chosen option (D) - there seems to be a preference to prefer a risky expectation value of 𝐸𝐸𝐸𝐸(𝑊𝑊) = 0.75 ⋅ (−1000) + 0.25 ⋅ 0 = −750 to a safe event of the same amount. This result is not consistent with risk aversion but describes a loss aversion: in case of possible losses, the risk propensity increases, and an uncertain event is preferred to a safe event with the same expected value - apparently the hope of having no loss at all with 25% probability dominates. <?page no="123"?> 3.2 Bounded rationality and behavioural economics 123 Figure 3.10: Loss aversion vs. risk aversion and a value function. ► Figure 3.10 shows the basic idea of prospect theory using a value function (in analogy to the utility function). Given the basic observation of loss aversion, it describes how people value profits and losses for given initial endowments (reference points). Through numerous experiments, Kahneman and Tversky have shown the course of the value function 𝑣𝑣 as (3.17) 𝑣𝑣(𝑥𝑥) = � 𝑥𝑥 𝜔𝜔 𝑓𝑓𝑑𝑑𝑟𝑟 𝑥𝑥 > 0 − 𝜇𝜇(−𝑥𝑥 𝜔𝜔 ) 𝑓𝑓𝑑𝑑𝑟𝑟 𝑥𝑥 < 0 with values of 𝜔𝜔 = 0.88 and 𝜇𝜇 = 2.25 - of course, similar to a utility function, the parameters and thus, the preferences of people also differ here. In general, starting from a reference point, a possible gain is judged positively, but a loss of the same amount is not judged equally - along the dotted line - but about 2.25 times more negatively. So, if a customer comes into a kiosk to cash in her lottery prize from last weekend and receives 40 EUR, she is initially happy - but if she then finds a parking ticket worth 40 EUR on her car immediately afterwards, the emotion and thus, the assessment of the overall situation is negative - although the absolute amount of the assets has remained the same. To compensate for the negative utility of the 40 EUR fine, a lottery win of around 90 EUR would have been necessary according to (3.17). The valuation of gains and losses is done relative to a reference point that describes the current situation of a customer or a manager. Loss aversion indirectly explains the status quo bias: people prefer the preservation of assets, of reputation or keeping a current activity and position in a firm to a risky change away from this reference point, because even with the same probability of positive or negative changes, possible losses from a changed situation are valued stronger than possible gains. This is especially true for customers: due to a default effect a product or equipment option is preferred which is already offered by the firm - in fact, people <?page no="124"?> 3 Decisions under risk and behavioural economics 124 find it difficult to opt out of special equipment and extras after test driving a car. The default effect has a significant influence on the selection of a product and systematically increases the willingness to pay of customers (Brown and Krishna 2004 and Johnson et al. 2012). Government institutions in particular try to "nudge" customers towards a decision that is preferred from an economic policy perspective by setting default options (predefined standard specifications) with corresponding positioning or description of product alternatives. This is done through symbolic references concerning food (traffic light systems) to support a healthy lifestyle, as well as the positioning of healthy and unhealthy food in a canteen or through the automated presetting (so called opt-out solutions) for organ donation or firm pension schemes, which must be actively deselected (Sunstein and Thaler 2008, Bruttel et al. 2014 as well as Dams et al. 2015). Framing and the architecture of choice Prospect theory has numerous implications for business models and marketing strategies. Many are based on the fact that framing (the description or arrangement as well as the context of a decision problem) can influence the decision. Human decisions are for example siginificantly influenced by the presentation and arrangement (and the wording) of a decision situation. Tversky and Kahneman (1981) have described the following situation in two alternative settings. Imagine the US preparing for the outbreak of an unusual Asian disease, which is expected to kill 600 people. Two alternative programs have been developed to combat the disease. Assume that the consequences of the two programs are well known: Setting 1 If program A is implemented, 200 people will be saved. If program B is implemented, there is a one-third chance that 600 people will be saved and a two-thirds chance that no one will be saved. Setting 2 If program C is implemented, 400 people will die. If program D is implemented, there is a one-third chance that no one will die and a twothirds chance that 600 people will die. Obviously, programs A and C as well as B and D are identical - however, in setting 1 a majority of 72% chose program A, while in setting 2 only 22% chose the equivalent program C. The reason for this is positive vs. negative framing. People decide in favour of a positively presented safe option and against an uncertainly presented negative option. Framing also plays a central role in purchase and demand decisions. In ► Figure 3.11 on the left, two student flats A and B are positioned in an experiment, where 50% of students prefer A, 50% prefer B (Simonson and Tversky 1992). A is closer to the campus, but has a higher rent, while the opposite is true for B. If you now add in a third flat C to the right in ► Figure 3.11, nothing should change: flat B dominates flat C in terms of cost and proximity to campus, so that 50% of students should still choose A and 50% should choose B according to their preferences - C is an irrelevant alternative. <?page no="125"?> 3.2 Bounded rationality and behavioural economics 125 Figure 3.11: Framing of student flats. In fact, nobody takes C - but the existence of flat C influences the decision between A and B: due to B and C being easy to compare, B becomes more attractive and demand is now distributed A: 30% and B: 70%. This effect is called a decoy effect (Ariely and Wallsten 1995 as well as Ariely 2008): an irrelevant decision option that is asymmetrically dominated influences preferences and demand behaviour. This effect is all the stronger the weaker the original preferences are, the greater the uncertainty about the decision situation and the better the comparability of the irrelevant option. Figure 3.12: Framing with bikes. <?page no="126"?> 3 Decisions under risk and behavioural economics 126 A decoy effect can be observed in various alternatives in numerous business models: for example, a new market for electric bicycles is currently emerging. Potential customers only have a vague idea of the relevant features (battery capacity, charging time, weight, range, etc.) and correspondingly have weak preferences. In markets of this kind when differences are not clearly discernible, customers have a tendency to buy the cheaper of two bicycles offered - as assumed in ► Figure 3.12, A: 80% and B: 20%. If a third, better-equipped but also considerably more expensive bicycle is added, a maximum of 20% of customers (only those who had previously chosen bicycle B) would consider switching to bicycle C if their preferences were complete and consistent (a proportion of those with a high willingness to pay and an interest in quality). However, in experiments many customers are now switching from bike A to bikes B and C (this means that preferences are not stable and robust) but it also leads to an increase in revenues for the bicycle dealer of about 40%. Similarly, framing can be found in the mug sizes of Starbucks, memory options of Apple devices, or the range of wines offered at a vineyard: firms always attempt to distort preferences, increase willingness to pay, and influence decisions by adding a most expensive, potentially irrelevant option. Media industries and magazines also use framing. In 2006, the Economist offered the product options and prices shown in ► Figure 3.13 - at first glance it appears that the digital option is included free of charge with the print and digital package. In an experiment, Ariely (2008) first examined how large the demand for the three options is and confirmed that nobody chooses the print version alone. Figure 3.13: Pricing structure and offerings of Economist. However, if this supposedly irrelevant print version is removed, then a simple comparison between the combined print and digital offering on the one hand and digital stand-alone offering three versions print and digital print digital two versions product variant price share of demand share of demand price <?page no="127"?> 3.2 Bounded rationality and behavioural economics 127 is no longer possible. As a result, demand shares shift drastically. By adding a supposedly irrelevant option, the share of expensive subscriptions can be increased from 32% to 84%, which is equivalent to a 42% increase in revenue when switching from two to three options. Endowment effect, mugs, elevator and IKEA Numerous business models - gyms, streaming services, software packages, or newspapers - offer so-called 30-day free trial subscriptions. The economic justification for this is not that customers need to be convinced of the service or the product quality, but that free access triggers an endowment effect (based on ownership or possession) that permanently increases the willingness to pay. An endowment effect has been identified and verified in numerous experiments, amongst these especially with coffee mugs. It describes that both valuation and appreciation of objects is influenced by whether a person currently owns the object or not, so that the willingness to pay is affected (Knetsch 1989, Kahneman et al. 1990 and 1991). In experiments with a group of students, mugs were randomly distributed to half of the participants, while the other half of the group received no mugs. If the (involuntary) owners of the mugs are asked after a while at what price they would sell it, the average value over various experiments is around 6 USD; if the mugless participants are asked about their willingness to pay for one of the mugs, the average value is just under 3 USD. If purchases and sales are possible between the two groups, only about 15% to 20% of the mugs are traded across all experiments according to differences in individual willingness to pay and sell; but if the willingness to pay and sell is random and equally distributed, it would be expected that 50% of the mugs change hands through transactions (see also ► Chapter 1 with the example of a Water Bottle Park: there, 50% of the water bottles were traded in market equilibrium). As soon as a participant owns almost any object or product, appreciation increases - such experiments have subsequently been conducted with pieces of chocolate, admission tickets, lottery tickets, pens, Lego figures, paper cubes, or unknown contacts on Facebook. The central reason for this is again the loss aversion, which is generated by a random allocation of an object to an individual. The appreciation itself can be increased by the following factors, among others. Time - as time that a participant spends with the arbitrarily allocated product increases, the appreciation rises disproportionately (Strahilewitz and Loewenstein 1998). Shared Experience - if shared experiences are made with the allocated product (e.g., an elevator trip with the mug), the appreciation also increases (Kahneman et al. 1990). IKEA effect - if a product is at least partially assembled by the customer himself, appreciation increases. In anticipation of this increased appreciation, customers are willing to pay more in advance, so that de facto higher prices can be charged for self-assembled furniture than for furniture assembled by the firm (Norton et al. 2012 and Mochon et al. 2012). The extent to which preferences can be distorted by the endowment effect is shown by another experiment with three separate (but representative) groups of students. The participants in group 1 - as a benchmark and reference group - are free to choose between a mug or chocolate of the same value: 59% of the participants choose the mug, so that a prediction for the preferences of groups 2 and 3 is now available. In group 2 all participants will first receive a <?page no="128"?> 3 Decisions under risk and behavioural economics 128 mug, in group 3 all will first receive chocolate (which must not be eaten). After 90 minutes, participants in groups 2 and 3 will each be offered to exchange their mugs for chocolate or chocolate for mugs. Now 89% in group 2 keep the mug, 90% in group 3 keep the chocolate - the original distribution of preferences is not stable, but the 'assigned' ownership of coffee mugs or chocolate distorts the preferences so much that there is no switching to the preferred product (Knetsch 1989 and List 2004). It is precisely these shared experiences, even over a short period of time, that explain to a large extent why car dealers allow test driving, why expensive loudspeakers can be tried out at home for a weekend or why boutiques offer to put clothes aside - the certainty that a piece of cloth is waiting for the buyer increases the willingness to buy and pay. In the same way, due to the endowment effect, people sometimes value their own assets - securities, real estate, used cars - much higher than potential buyers and on the one hand hold on to them for too long, and on the other hand excessive price expectations can arise (Barberis and Thaler 2003 as well as Shleifer 2000). Conversely, the endowment effect itself is reduced by frequent market transactions (each market participant owns a product only briefly), by the experience of buyers and sellers and by strategic incentives for efficient trading (List 2003 and Tontrup 2017). ► Figure 3.14 shows the effect of a free trial month with streaming service providers such as Amazon Prime, Netflix, Maxdome, or Watchever - all offered a free trial month in 2017, after which a basic subscription is available at 7.99 EUR. Assuming a willingness to pay, e.g. 3.50 EUR per month, the free trial month will shift a customer's reference point. An endowment effect is created by the free trial month, which increases the willingness to pay. To compensate for the fear of losing access to the service, customers are now prepared to pay around 8 EUR per month - according to formula (3.17). A streaming provider waives 3.50 EUR at once and then, instead of 12 ⋅ 3.50 𝐸𝐸𝐸𝐸𝑅𝑅 = 42 𝐸𝐸𝐸𝐸𝑅𝑅 , generates revenues of 11 ⋅ 7.99 = 87.89 𝐸𝐸𝐸𝐸𝑅𝑅 per year - more than a doubling of revenues already in the first year. Figure 3.14: Endowment effect and loss aversion with streaming service providers. <?page no="129"?> 3.2 Bounded rationality and behavioural economics 129 The endowment effect has numerous implications for business models and strategy: people hold on too strongly and for too long to (coincidentally developed or established) strategies and organisations and complicate change processes. The IKEA effect in particular encourages managers to stick to a self-developed competitive strategy: firms hold on to their product portfolio and divisions for too long and the effects of overoptimism and confirmation bias reinforce each other. Firms tend to focus too much on existing rather than on new business and invest too much in exploiting existing technologies rather than exploring new technologies (Garbuio et al. 2014, Levinthal and March 2003 as well as Teece 2007). Fairness and altruism Maximising a person's utility implicitly involves selfishness. The extent to which selfishness and indirectly the maximisation of a person's utility is pronounced can be tested experimentally by means of the ultimatum game (Güth et al. 1982, Binmore et al. 1985, Güth and Tietz 1990, Kahneman et al. 1986 as well as Thaler 1988). In its simplest form, the game about dividing 10 EUR between two individuals is based on the following rules: Two players - one proposer and one responder, who decide and announce their decision simultaneously; Amount of money - the proposer receives a monetary amount 𝑊𝑊 , e.g., 10 EUR, known to both players, to be split between them; Decision of the proposer - the proposer secretely has to decide what share 𝑣𝑣 𝑉𝑉 of the money he keeps for himself and what share (1 − 𝑣𝑣 𝑉𝑉 ) he is willing to give to the respondent; Responder's decision - the responder must secretely decide what proportion 𝑣𝑣 𝐴𝐴 of the amount of money he expects or demands as a minimum for himself; No communication or negotiation - the game is a one-off game; information cannot be exchanged and no agreements can be made; Distribution of the money - both players receive the respective amounts 𝑣𝑣 𝑉𝑉 ⋅ 𝑊𝑊 and (1 − 𝑣𝑣 𝑉𝑉 ) ⋅ 𝑊𝑊 if and only if the proposer's offer is at least equal to the responder's demand or expectation, i.e., (1 − 𝑣𝑣 𝑉𝑉 ) ≥ 𝑣𝑣 𝐴𝐴 - in this case the money is distributed according to the proposal, otherwise no player receives money. <?page no="130"?> 3 Decisions under risk and behavioural economics 130 Figure 3.15: Ultimatum game. With perfect rationality, maximisation of utility and selfishness, the following result should be observed: a responder expects the minimum possible sum, 0.01 EUR, as this will increase his utility compared to 0 EUR and ensure that he will benefit from any offer from the proposer that is above this amount; the proposer should anticipate this expectation and the behaviour of the responder and offer exactly 0.01 EUR, keeping the rest of the amount - at 10 EUR thus 9.99 EUR for himself; and the distribution of money is (always) according to the proposal (Rubinstein 1982). This result is not confirmed in laboratory experiments. On the one hand, responders typically expect a share that is clearly above a minimum sum; on the other hand, proposers expect this and also offer clearly larger shares; responders regularly expect shares of 20% to 50%; the proposers typically offer shares between 30% and 50%. In such game situations there is no strict maximisation of utility: players expect fairness, but also offer fairness. A propensity towards fairness depends, among other things, on the anonymity or mutual knowledge of the players, role allocation and general conditions (boss, employees and birthday cake), potential repetition of the game and the possibility of building up reputation over time, as well as the amount of money (10 EUR vs. 1 million EUR). However, the allocation of a sum of money, in addition to fairness, can also be based on altruism. In this case, decision-makers include the utility of others in their own utility function in a dynamic calculation. From a management perspective, fairness and altruism have implications for negotiating situations (salary or collective bargaining, mergers, or long-term contracts such as outsourcing) and for building up reputation in organisations or vis-à-vis competitors (see also ► Chapter 9 on game theory and ► Chapter 10 on strategic competition). <?page no="131"?> 3.3 Summary and key learnings 131 3.3 Summary and key learnings Do managers or customers typically make perfect decisions? No - people are not perfect, and decision-making is a complicated task. Decisions and human behaviour are influenced by risk, uncertainty and bounded rationality. Decisions based on incomplete information - risk or uncertainty - are affected by risk aversion depending on possible conditions and situations: people prefer secure over uncertain outcomes for the same expected value, reject fair games, and maximise their expected utility. The degree of risk aversion varies across groups of people and in view of the decision-making situation. Furthermore, perception of probabilities is biased according to the weighting function of Kahneman and Tversky. Risk management by means of insurance, diversification, information gathering, or signalling can help to improve risk assessment or reduce risk. Numerous empirically observable decisions deviate from maximising behaviour in systematic and predictable patterns. This is due to bounded rationality and fast thinking, which becomes visible whenever people apply heuristics. Customers and managers alike are looking for goodenough solutions against the background of limited information or limited cognitive abilities. People regularly make decisions under the influence of cognitive biases and show strong loss aversion. At the same time, decisions are shaped and influenced by endowment effects (arbitrary or random initial configurations) or by framing (arrangement of options). Observations of decisions under risk as well as under bounded rationality show that utility maximisation based on perfect rationality is the exception rather than the rule. On the contrary, human decisions in economic situations can be understood a lot better if all three perspectives - rationality, bounded rationality, as well as risk and uncertainty - on a decision or competitive situation are integrated. From a management perspective, there are many implications: on the one hand, there are possibilities to develop business models and pricing strategies based on boundedly rational behaviour of customers in order to increase profits. On the other hand, an assessment of the risk attitude and rationality of one's own employees as well as of competitors should always be made. Recommendations for further reading Risk and uncertainty are excellently presented in Gigerenzer, G., Risk savvy: how to make good decisions, Munich 2015. For deep insights into the origins, numerous experiments and psychological foundations of behavioural economics, see Kahneman, D., Thinking, fast and slow, Princeton 2011. Rather anecdotal but very entertaining is Ariely, D., Predictably irrational: the hidden forces that shape our decisions, New York 2008. <?page no="132"?> 3 Decisions under risk and behavioural economics 132 Questions for review [1] Describe applications of the analysis of decisions under risk and bounded rationality from a microeconomic perspective as well as their limits, advantages and disadvantages. [2] Determine a 'risk-optimum ticket price' based on a risk premium for a potential fare dodger with a risk aversion 𝜔𝜔 = 0.8 and a wealth of EUR 200, a penalty fee of EUR 150 and a probability of being caught of 3%. How does the ticket price change depending on wealth, risk aversion and probability of being caught? [3] Career-Katja and Poker-Pete currently earn EUR 75,000 fixed. Both receive an offer to move to the sales department, but there salary will include a variable component of EUR 40,000, which can only be achieved with a 40 % probability, i.e.: Probability income 0.60 EUR 60,000 0.40 EUR 100,000 What is the expected value of the income of the new job? Katja is strongly risk averse, her utility function is given as 𝐸𝐸 = 𝑊𝑊 0.5 - will she accept the new job? Pete is less risk averse, his utility function is given as 𝐸𝐸 = 𝑊𝑊 0.8 - will he take the new job? [4] A candidate in Who Wants to Be a Millionaire has a utility function 𝐸𝐸(𝑊𝑊) = 𝑊𝑊 0.8 . He stands at 125,000 EUR, can drop back to 16,000 EUR and still has the 50: 50 joker. With four remaining answer alternatives, he has no idea about the EUR 500,000 question - what will he do? What is the risk premium here, what does it say in this case? Should the candidate take the 50: 50 joker? What value does the joker have for him? Draw both cases and describe the differences. [5] Define uncertainty and risk. Describe the phenomenon of risk aversion. How can risk aversion be measured? What is meant by a so-called risk premium? [6] What are the main differences between traditional microeconomics and behavioural economics? Explain the basic considerations of bounded rationality. [7] What is a heuristic, when is it used? Describe three typical perception biases. [8] What are typical outcomes of the ultimatum game? What role does fairness play in the one-off and repeated ultimatum game? [9] What is framing? Explain three typical consumer products where manufacturers or retailers use this effect to increase profits. What does the endowment effect describe? Explain three typical consumer products where manufacturers or retailers use this effect. [10]Explain the basic statement of the prospect theory by Kahneman and Tversky and give two examples. <?page no="133"?> 3.3 Summary and key learnings 133 Literature Anderson, L.R. and Mellor, J., Are risk preferences stable? Comparing an experimental measure with a validated survey-based measure, Journal of Risk and Uncertainty, 2009, 39, 2, 137-160. Ariely, D. and Wallsten, T.S., Seeking subjective dominance in multidimensional space: an explanation of the asymmetric dominance effect, Organizational Behavior and Human Decision Processes, 1995, 63, 3, 223-232. Ariely, D., Predictably irrational: the hidden forces that shape our decisions, New York 2008. Armstrong, M. and Huck, S., Behavioral economics as applied to firms: a primer, CESifo Working Paper Series, 2010, No. 2937. Asch, S.E., Studies of independence and conformity: a minority of one against a unanimous majority, Psychological Monographs: General and Applied, 1956, 70, 9, 1-70. Barberis, N. and Thaler, R.H., A survey of behavioral finance, in: Constantinides, G.M., Harris, M. and Stulz, R.M. (eds.), Handbook of the Economics of Finance, Volume 1, 2003, 1053-1128. Bar-Eli, M., Azar, O.H., Ritov, I., Keidar-Levin, Y. and Schein, G., Action bias among elite soccer goalkeepers: the case of penalty kicks, Journal of Economic Psychology, 2007, 28, 5, 606- 621. Bayer, J.B., Triệu, P. and Ellison, N.B., Social media elements, ecologies, and effects, Annual Review of Psychology, 2020, 71, 471-497. Bellemare, C. and Shearer, B.S., Sorting, incentives and risk preferences: evidence from a field experiment, IZA Discussion Papers, No. 2227, July 2006, IZA Bonn. Ben-David, I., Graham, J.R. and Harvey, C.R., Managerial overconfidence and corporate policies, Working Paper 13711, National Bureau of Economic Research 2007. Binmore, K., Shaked, A. and Sutton, J., Testing non-cooperative bargaining theory: a preliminary study, American Economic Review, 1985, 75, 1178-1180. Brown, C.L. and Krishna, A., The skeptical shopper: a metacognitive account for the effects of default options on choice, Journal of Consumer Research, 2004, 31, 3, 529-539. Brunsson, N., The irrationality of action and action rationality: decisions, ideologies and organizational actions, Journal of Management Studies, 1982, 19, 1, 29-44. Bruttel, L.V., Stolley, F., Güth, W., Kliemt, H., Bosworth, S., Bartke, S., Schnellenbach, J., Weimann, J., Haupt, M. and Funk, L., Nudging als politisches Instrument - gute Absicht oder staatlicher Übergriff? , Wirtschaftsdienst, 2014, 94, 11, 767-791. Camerer, C.F. and Weber, M., Recent developments in modeling preferences: uncertainty and ambiguity, Journal of Risk and Uncertainty, 1992, 5, 4, 325-370. Camerer, C.F., Babcock, L., Loewenstein, G. and Thaler, R.H., Labor supply of New York City cabdrivers: one day at a time, Quarterly Journal of Economics, 1997, 112, 2, 407-441. Camerer, C.F., Loewenstein, G. and Rabin, M., Advances in behavioral economics, Princeton 2011. Carpenter, J.N., Does option compensation increase managerial risk appetite? , Journal of Finance, 2000, 55, 5, 2311-2331. Chen, G., Crossland, C. and Luo, S., Making the same mistake all over again: CEO overconfidence and corporate resistance to corrective feedback, Strategic Management Journal, 2015, 36, 10, 1513-1535. Coles, J.L., Daniel, N.D. and Naveen, L., Managerial incentives and risk-taking, Journal of Financial Economics, 2006, 79, 2, 431-468. Cyert, R.M. and March, J.G., A behavioral theory of the firm, Englewood Cliffs 1963. Daghofer, F., Financial risk-taking on “Who Wants to Be a Millionaire”: a comparison between Austria, Germany, and Slovenia, International Journal of Psychology, 2007, 42, 5, 317-330. <?page no="134"?> 3 Decisions under risk and behavioural economics 134 Dams, J., Ettel, A., Greive, A. and Zschäpitz, H., Merkel will die Deutschen durch Nudging erziehen, Die Welt, 12. März 2015. DellaVigna, S. and Malmendier, U., Paying not to go to the gym, American Economic Review, 2006, 96, 3, 694-719. DellaVigna, S., Psychology and economics: evidence from the field, Journal of Economic Literature, 2009, 47, 2, 315-372. Dohmen, T.J., Do professionals choke under pressure? , Journal of Economic Behavior and Organization, 2008, 65, 3, 636-653. Ellsberg, D., Risk, ambiguity, and the Savage axioms, Quarterly Journal of Economics, 1961, 75, 4, 643-669. Esser, J.K., Alive and well after 25 years: a review of groupthink research, Organizational Behavior and Human Decision Processes, 1998, 73, 2, 116-141. Farber, H.S., Why you can’t find a taxi in the rain and other labor supply lessons from cab drivers, Quarterly Journal of Economics, 2015, 130, 4, 1975-2026. Franzen, A. and Pointner, S., Wer wird Millionär? Eine empirische Analyse der Erfolgsdeterminanten in der gleichnamigen Quizshow, Zeitschrift für Soziologie, 2009, 38, 3, 239-256. Garbuio, M., Lovallo, D. and Ketenciouglu, E., Behavioral economics and strategic decision-making, The Oxford Handbook of Managerial Economics, Boston-London 2014. Gavetti, G. and Rivkin, J.W., On the origin of strategy: action and cognition over time, Organization Science, 2007, 18, 3, 420-439. Gigerenzer, G. and Selten, R., Bounded rationality, Cambridge 2002. Gigerenzer, G., The empire of chance: how probability changed science and everyday life, Boston 1990. Günther, T. and Detzner, M., Sind Manager und Controller risikoscheu? Ergebnisse einer empirischen Untersuchung, Controlling, 2012, 24, 4/ 5, 247-254. Güth, W. and Tietz, R., Ultimatum bargaining behavior: a survey and comparison of experimental results, Journal of Economic Psychology, 1990, 11, 3, 417-449. Güth, W., Schmittberger, R. and Schwarze, B., An experimental analysis of ultimatum bargaining, Journal of Economic Behavior and Organization, 1982, 3, 4, 367-388. Hartley, R., Lanot, G. and Walker, I., Who really wants to be a millionaire? Estimates of risk aversion from game show data, Journal of Applied Econometrics, 2014, 29, 6, 861-879. Holmström, B., Managerial incentive problems: a dynamic perspective, Review of Economic Studies, 1999, 66, 1, 169-182. Holt, C.A. and Laury, S.K., Risk aversion and incentive effects, American Economic Review, 2002, 92, 5, 1644-1655. Huck, S., Zhou, J. and Duke, C., Consumer behavioral biases in competition - a survey, Office of Fair Trade, 2011. Johnson, E.J., Shu, S.B., Dellaert, B.G., Fox, C., Goldstein, D.G., Häubl, G., Larrick, R.P., Payne, J.W., Peters, E., Schkade, D. and Wansink, B., Beyond nudges: tools of a choice architecture, Marketing Letters, 2012, 23, 2, 487-504. Just, D.R. and Wansink, B., The flat-rate pricing paradox: conflicting effects of “all-you-can-eat” buffet pricing, Review of Economics and Statistics, 2011, 93, 1, 193-200. Kahneman, D. and Tversky, A., Prospect theory: an analysis of decision under risk, Econometrica, 1979, 47, 2, 263-292. Kahneman, D. and Tversky, A., Subjective probability: a judgment of representativeness, Cognitive Psychology, 1972, 3, 3, 430-454. Kahneman, D., Knetsch, J.L. and Thaler, R.H., Anomalies: the endowment effect, loss aversion, and status quo bias, Journal of Economic Perspectives, 1991, 5, 1, 193-206. Kahneman, D., Knetsch, J.L. and Thaler, R.H., Experimental tests of the endowment effect and the Coase theorem, Journal of Political Economy, 1990, 98, 6, 1325-1348. <?page no="135"?> 3.3 Summary and key learnings 135 Kahneman, D., Knetsch, J.L. and Thaler, R.H., Fairness and the assumptions of economics, Journal of Business, 1986, 59, 285-300. Kahneman, D., Maps of bounded rationality: psychology for behavioral economics, American Economic Review, 2003, 93, 5, 1449-1475. Kahneman, D., Thinking, fast and slow, Princeton 2011. Klayman, J., Soll, J.B., Gonzalez-Vallejo, C. and Barlas, S., Overconfidence: it depends on how, what, and whom you ask, Organizational Behavior and Human Decision Processes, 1999, 79, 3, 216-247. Knetsch, J., The endowment effect and evidence of nonreversible indifference curves, American Economic Review, 1989, 79, 1277-1284. Knight, F.H., Risk, uncertainty and profit, New York 1921. Leibenstein, H., Bandwagon, snob, and Veblen effects in the theory of consumers' demand, Quarterly Journal of Economics, 1950, 64, 2, 183-207. Levinthal, D.A. and March, J.G., The myopia of learning, Strategic Management Journal, 1993, 14, 95-112. Lindblom, C.E., The science of muddling through, Public Administration Review, 1959, 4, 79-88. List, J.A., Does market experience eliminate market anomalies? , Quarterly Journal of Economics, 2003, 118, 1, 41-71. List, J.A., Neoclassical theory versus prospect theory: evidence from the marketplace, Econometrica, 2004, 72, 2, 615-625. Malmendier, U. and Tate, G., Behavioral CEOs: the role of managerial overconfidence, Journal of Economic Perspectives, 2015, 29, 4, 37-60. Malmendier, U. and Tate, G., CEO overconfidence and corporate investment, Journal of Finance, 2005, 60, 6, 2661- 2700. March, J.G. and Shapira, Z., Managerial perspectives on risk and risk taking, Management Science, 1987, 33, 11, 1404- 1418. March, J.G. and Simon, H.A., Organizations, New York 1958. March, J.G., A primer on decision-making: how decisions happen, New York 1994. Markowitz, H.M., Portfolio selection, Journal of Finance, 1952, 7, 77-91. Mascalet, D., Colombier, N., Denant-Boemont, L. and Loheac, Y., Group and individual risk preferences: a lotterychoice experiment with self-employed and salaried workers. Journal of Economic Behavior and Organization, 2009, 70, 470-484. Misirlisoy, E. and Haggard, P., Asymmetric predictability and cognitive competition in football penalty shootouts, Current Biology, 2014, 24, 16, 1918-1922. Mochon, D., Norton, M.I. and Ariely, D., Bolstering and restoring feelings of competence via the IKEA effect, International Journal of Research in Marketing, 2012, 29, 4, 363-369. Ngai, E.W.T., Tao, S.S.C. and Moon, K.K.L., Social media research: theories, constructs and conceptual frameworks, International Journal of Information Management, 2015, 35, 1, 33-44. Nickerson, R.S., Confirmation bias: a ubiquitous phenomenon in many guises, Review of General Psychology, 1998, 2, 2, 175-220. Norton, M.I., Mochon, D. and Ariely, D., The IKEA effect: when labor leads to love, Journal of Consumer Psychology, 2012, 22, 3, 453-460. Pavlovic, N., Herd behaviour in consumer buying decisions in the age of social media, Nürnberg 2020. Pennings, J.M.E. and Smidts, A., Assessing the construct validity of risk attitude, Management Science, 2000, 46, 10, 1337-1348. <?page no="136"?> 3 Decisions under risk and behavioural economics 136 Powell, T.C., Lovallo, D. and Fox, C.R., Behavioral strategy, Strategic Management Journal, 2011, 32, 13, 1369-1386. Pratt, J.W., Risk aversion in the small and in the large, Econometrica, 1964, 32, 122-136. Pütz, A. and Ruenzi, S., Overconfidence among professional investors: evidence from mutual fund managers, Journal of Business Finance and Accounting, 2011, 38, 5-6, 684-712. Ross, S.A., Compensation, incentives, and the duality of risk aversion and riskiness, Journal of Finance, 2004, 59, 1, 207-225. Rubinstein, A., Perfect equilibrium in a bargaining model, Econometrica, 1982, 50, 97-109. Russo, J.E. and Schoemaker, P.J., Managing overconfidence, Sloan Management Review, 1992, 33, 2, 7-17. Sharpe, W.F., Capital asset prices: a theory of market equilibrium under conditions of risk, Journal of Finance, 1964, 19, 3, 425-442. Shiller, R.J., Speculative prices and popular models, Journal of Economic Perspectives, 1990, 4, 2, 55-65. Shleifer, A., Inefficient markets: an introduction to behavioural finance, Oxford 2000. Simon, H.A., A behavioral model of rational choice, Quarterly Journal of Economics, 1955, 69, 1, 99-118. Simon, H.A., Models of man, social and rational: mathematical essays on rational human behavior in a social setting, New York 1957. Simonson, I. and Tversky, A., Choice in context: tradeoff contrast and extremeness aversion, Journal of Marketing Research, 1992, 29, 3, 281-295. Spence, M., Job market signaling, Quarterly Journal of Economics, 1973, 87, 3, 355-374. Spence, M., Signaling in retrospect and the informational structure of markets, American Economic Review, 2002, 92, 3, 434-459. Strahilevitz, M.A. and Loewenstein, G., The effect of ownership history on the valuation of objects, Journal of Consumer Research, 1998, 25, 3, 276-289. Sunstein, C. and Thaler, R.H., Nudge - the politics of libertarian paternalism, New Haven 2008. Svenson, O., Are we all less risky and more skillful than our fellow driver? , Acta Psychologica, 1981, 47, 2, 143-148. Taleb, N.N., Goldstein, D.G. and Spitznagel, M.W., The six mistakes executives make in risk management, Harvard Business Review, 2009, 87, 10, 78-81. Taleb, N.N., The black swan, New York 2007. Teece, D.J., Explicating dynamic capabilities: the nature and microfoundations of (sustainable) enterprise performance, Strategic Management Journal, 2007, 28, 13, 1319-1350. Thaler, R.H., Anomalies: the ultimatum game, Journal of Economic Perspectives, 1988, 2, 4, 195-206. Thaler, R.H., Misbehaving: the making of behavioral economics, New York 2015. Thaler, R.H., Toward a positive theory of consumer choice, Journal of Economic Behavior and Organization, 1980, 1, 1, 39-60. Toner, M., Ojha, N., de Paepe P. and de Melo, M.S., A strategy for thriving in uncertainty, Bain Brief, Bain&Company, 12. August 2015. Tontrup, S., Does the endowment effect prevail when traders act strategically? , New York University Law and Economics Working Papers 450, 2017. Tversky, A. and Kahneman, D., Advances in prospect theory: cumulative representation of uncertainty, Journal of Risk and Uncertainty, 1992, 5, 297-323. Tversky, A. and Kahneman, D., Availability: a heuristic for judging frequency and probability, Cognitive Psychology, 1973, 5, 2, 207-232. <?page no="137"?> 3.3 Summary and key learnings 137 Tversky, A. and Kahneman, D., Belief in the law of small numbers, Psychological Bulletin, 1971, 76, 2, 105-110. Tversky, A. and Kahneman, D., Judgment under uncertainty: heuristics and biases, Science, 1974, 185, 4157, 1124-1131. Tversky, A. and Kahneman, D., The framing of decisions and the psychology of choice, Science, 1981, 4481, 453-458. Vanini, U., Risikoneigung und Unternehmenssteuerung - Ergebnisse empirischer Studien, Working Paper, CARF Luzern 2016. von Neumann, J. and Morgenstern, O., Theory of games and economic behavior, Princeton 1944. <?page no="139"?> 139 4 Firms, competition and innovation Since 29 th January 1886, the Kaiserliche Patentamt in Germany (Imperial Patent Office during the time of Wilhelm I., former King of Prussia and later the first German Emperor) has identified Carl Benz as the inventor of the "vehicle with gas engine operation" under the number 37435. Around the same time, Albert Hammel and Knut Johansen in Denmark, Magnus Volk in Great Britain, Siegfried Marcus in Austria and Léon Serpollet in France were also designing motordriven vehicles - an accumulation of similar innovations gave rise to the worldwide automobile industry. In 1926 Carl Benz's firm merged with that of his competitor Gottlieb Daimler, close to bankruptcy and under pressure from Deutsche Bank. Together with the constructor Wilhelm Maybach, Daimler produced a steel wheeled car since 1889 and sold it under the Mercedes brand name from 1901. Léon Serpollet moved from his own workshop to that of the bicycle manufacturer Armand Peugeot in 1889 and both started to collaborate and produce cars. After more than 100 years, Daimler-Benz (in between DaimlerChrysler, then Daimler and since 2022 Mercedes-Benz) and Peugeot are still among the leading car manufacturers. In 2016 Daimler made a profit of 8.8 billion EUR and employs 282,488 people worldwide. Nothing else is known about the careers of Hammel, Johansen, Volk and Marcus - history mostly reports winners (Münter 1999). The growth and prosperity of an economy, nations, cultures and societies is accompanied and shaped by the emergence of firms, and by the development and decline of various industries. The evolution of an industry is only a reflection of the interplay between customers and firms, but above all is competition between firms for the favour of customers, and continuous innovation of products, processes, or business models. Competition is a complex dynamic process. The dynamics of the process are driven by strategic interaction of firms over time, which on the one hand are constantly seizing new opportunities and on the other hand are constantly confronted with competitors that possess some kind of competitive advantage. Threatened by the risk of losses and becoming deprived of their existence, firms try to make profit to ensure their survival. Competition between firms is constantly creating new market structures, which reflect the changing competitive positioning of firms over time. Long-term competition is characterised by innovation, some are a result of research and development within one industry, but some also arise in other industries and are then adapted or are brought forward by start-ups. As such, technologies for autonomous driving, electric engines, digital connection of vehicles, or business models such as car sharing or ride hailing, are currently being developed by firms such as Google, Tesla or in scientific research institutions. As a result the existence of incumbent firms, such as Daimler in the automotive industry, is constantly threatened by possible disruptive innovations. <?page no="140"?> 4 Firms, competition and innovation 140 Learning Objectives This chapter deals with the existence and objectives of a firm in order to develop a basic understanding of how and in what dimensions strategic decisions are made by managers; competitive analysis using structure-conduct-performance and five-forces framework for understanding the relationship between market structure, strategy and profit; market-based view and resource-based view to explain competitive advantages and profit, and the long-term patterns and interaction of innovation, market structure and competition in various technological regimes. 4.1 Firms, objectives and strategies The creation of a new firm and its entry into a market is based either on the combination of new knowledge and the pursuit of profit, often creating new markets or industries, as in the case of Daimler-Benz; or on the identification of profit opportunities in existing markets or industries, as in the case of Google's entry in 1998 into the search engine industry that had already been in existence for six years. The available knowledge base of a society and the institutions (see ► Chapter 1) at any given time determine the rise of markets, firms and industries. The main reason for the emergence of new knowledge, which is ultimately embodied in new products, new markets and new firms, is the quest of individuals for knowledge and their dissatisfaction with existing explanations and solutions - whether it be in R&D departments of firms, in garages in California, or at colleges and universities. New knowledge is implemented in new firms, if entrepreneurship is possible under the given institutional environment, and entrepreneurial action allows the appropriation of profits (Schumpeter 1911, Kirzner 1973 and Audretsch 1995). Good to know │ Starting a business - are entrepreneurs born or made? Starting a business (i.e., foundation of a firm) is driven by individuals - single founders or teams. In essence, setting up a business is always just one career alternative to others, especially compared with a permanent position. Determinants of self-employment are both legal, economic and other external conditions, and intrapersonal factors, which lie in the specifics of the founder's personality. In addition, there are factors including the willingness to take risks, bear uncertainty, seize opportunities and make decisions, which are immanently linked to the role of a founder and entrepreneur (Kirzner 1973, Knight 1921, Alvarez and Parker 2009 and Münter 2020). These individual factors are also influenced by macroeconomic conditions: numerous studies have shown that, in addition to economic drivers, new businesses and start-ups resulting from unemployment play a particularly important role (Block and Wagner 2010). Start-ups and entrepreneurship are emergent phenomena - one is neither born to be a founder nor does one learn to 'start a business' (Rauch and Frese 2012, Simoes et al. 2016 <?page no="141"?> 4.1 Firms, objectives and strategies 141 as well as Acs and Audretsch 2006). Rather, constellations that can lead to the founding of a firm repeatedly arise in permanent employment relationships, at the university, in discussions with friends and family and also based on market opportunities or technological innovations (see also the entrepreneurs in the concrete planning of a start-up ('nascent entrepreneurs') in Arenius and Minniti (2005). Whether this actually happens depends on several of the founder’s personality dimensions as well as socio-demographic factors such as age, gender, marital status, or origin. In the simplest model for explaining entrepreneurship, the occupational choice theory, the focus is on an individual's decision between self-employment and a permanent position in an existing business. The decision to set up a new business then depends on the subjective assessment of the mutual advantages and disadvantages of both options, in particular income, effort, risks and degrees of freedom of action (Lucas 1978, Kanbur 1979, Parker 2007 and Nicolaou and Shane 2010). The probability of becoming an entrepreneur or self-employed increases with benefits from self-employment (the difference between income and necessary input and effort) compared to the wage or salary of a dependent employment. The benefit or profit from entrepreneurial activity usually further increases with the (expected) entrepreneurial abilities of the founder - the more pronounced these are, the more probable it is that the firm will be founded under otherwise equal conditions. In addition to these monetary aspects, a start-up and its success is influenced by the individual skills of the founders. Entrepreneurial skills originate from a combination of experience, personality traits and qualifications. Qualification describes learnable formal or informal knowledge, and skills that can be condensed into cognitive abilities. Qualification typically increases with the duration and quality of education and training. In addition, the breadth, balance, and connectedness of skills (skill balance) often leads to an increase in the relevant qualification of founders (Lazear 2004 and Hayward et al. 2006). Theory of the firm, transaction costs and boundaries of a firm Why do firms exist? To answer this question, Coase (1937) looked at 'the market' and 'the firm' as alternatives for organising transactions. Coase recognised that the market does not work for free - the coordination in a market and the use of the price mechanism generates transaction costs (see also ► Chapter 1). Although transaction costs also occur within a firm, they are typically to some extent lower than those in the market. The existence of firms, therefore, reduces the transaction costs that arise in the market. It is due to these cost advantages - which ultimately determine the size of the firm - that a firm is better able to perform transactions when compared to alternatively performing transactions via markets. In addition to the original approach of transaction costs, meanwhile the resource-based view and the principal-agent theory have become the focus of interest in the theory of the firm. The theory of the firms brings together economic, organisational, sociological and behavioural models of the formation, strategy and behaviour of a firm in order to answer the following questions. Existence - why do firms exist at all, why do people build a firm and start a venture, what is a firm, what are alternatives to a firm, how different are firms? <?page no="142"?> 4 Firms, competition and innovation 142 Boundaries - where do the boundaries of a firm run in vertical (e.g. concerning the value chain) and horizontal (e.g. concerning the size of the firm) dimensions, and how are they determined? Organisation - how are firms organised, why are there (and how many levels of) hierarchies and how are decisions made and implemented, what is the effect of separation of ownership and management in corporations, i.e., firms owned by shareholders but run by managers? Strategy - how do firms determine their strategies, do they differ within an industry, and if so, why? How does a firm interact with the environment and competitors? Related questions are whether, how, and why start-ups, SMEs or international corporations with employed (so-called hired) managers differ; how these differences are reflected in financing, strategy or business success; and whether small start-ups have systematic advantages in innovation over large firms. Firms can be identified in two dimensions described below. Firms as organisations - a firm is a group of owners, managers and employees (i.e., labour) combined with machinery, financial assets, infrastructure, rights and above all knowledge (i.e., capital) organised in such a way that they can produce products or services (i.e., production). From an institutional perspective, firms are organisations in the form of long-term contracts with labour and capital (i.e., managers and employees on the one hand and shareholders and creditors on the other). Firms as alternatives to the market - a firm integrates transactions (tasks, functions, activities, services, etc.) that could alternatively be carried out in the market or via suppliers and inter-firm alliances, as coordination within the firm can be more effective and/ or less costly. The existence and size of a firm can then be regarded as an alternative to realize transactions via the market. The long-term existence of a firm in competition is only possible if the firm can produce products or services, which cannot be produced in higher quality or more efficient in alternative constellations, for example by a public or state-owned authority, by a private household, by any gathering of people in the underground or by a sports club. Both advantages as well as boundaries of a firm result from a combination of: efficiency advantages resulting from size or scope of the firm, in particular cost advantages based on economies of scale or scope, division of labour and specialisation or learning curve effects (see ► Chapter 6) - in all these cases average costs (e.g., unit costs) of a firm decrease with increasing size of a firm; transaction cost advantages due to lower intrafirm coordination or monitoring costs compared to alternative uses of the market - activities are conducted inside a firm instead of being purchased from the market; and firm-specific capabilities based on resources represented by knowledge, employees, technology, patents and strategies, and so on. Transaction costs include costs incurred in the procurement, coordination and control of information and resources in the context of market relations, production, or services. If transaction costs of a regularly required resource or activity are lower within the firm than in the market, a transaction is internalised, i.e. performed inside a firm and a job is created, (e.g., the activity of <?page no="143"?> 4.1 Firms, objectives and strategies 143 an employee in the marketing department); otherwise a transaction is handled via the market (e.g., purchased from a supplier such as a marketing agency). ► Figure 4.1 compares transactions in the market (between individuals) and within firms. The higher the strategic importance and regularity of a transaction, the stronger the incentives to bundle or encapsulate it within a firm. Within the firm transactions then take place based on organisation and hierarchy, and replace search costs and negotiations in the market. Firms grow the easier and more stable transactions can be integrated. Figure 4.1: Transactions in the market vs. transactions within firms. On the one hand, this means that the size of a firm increases with the relative advantage of internal transaction costs of the organisation and hierarchy compared to the transaction costs in the market. The higher the efficiency of a firm's own organisational structure and processes, the faster the firm can grow: if internal transaction costs were zero, one firm alone would cover the entire market. In fact, however, transaction costs are not zero, and they increase with increasing firm size. Accordingly, the maximum size and growth dynamics of the firms themselves are limited. Higher transaction costs within a firm, therefore, limit the growth of a firm, but as a consequence, support the emergence of other firms in the same industry or even a supply industry. Conversely, falling transaction costs in the market - e.g. through the use of new technologies such as blockchain or the networking of market participants on multisided platforms - lead to greater use of decentralised market transactions and firms become smaller. Numerous new business models coordinate their necessary resources 'in the market' instead of 'inside the firm' (e.g. via freelancers on Fiverr instead of permanent employees), so that boundaries of an individual firm are more narrowly defined and the firms are smaller (Chen and Kamal 2016 and Roy and Sarkar 2016). Transaction costs include all costs associated with the initiation, conclusion, and implementation of transactions and contracts. These are costs for information gathering and contract negotiations before the conclusion of the contract and for coordination, monitoring and control of the service. It also includes the possible risks arising from the contract after the conclusion of the contract (Coase 1937, Williamson 1975, Foss 2003, Hart and Holmstrom 2010 as well as Voigt 2019). This becomes clear in make or buy and outsourcing decisions - from strategically <?page no="144"?> 4 Firms, competition and innovation 144 irrelevant areas such as canteen operations, facility management, or payroll accounting to strategically important areas such as payment and securities processing at banks, operation of IT platforms, or software. The extent of the transaction costs then also determines which activities are carried out within the firm itself. From a management perspective, in addition to the level of transaction costs, the quality of the services available on the market and their strategic importance must be assessed, and specifically, against the background of dependency on service providers in connection with the firm's own core competencies (so called hold-up problem). For example, outsourcing the operation of a canteen may cause short-term discontent among employees, but outsourcing the IT of a telecommunications provider can be a decisive factor for survival and profit. Objectives of firms Firms are sometimes free to define their goals: "market leader", "highest customer satisfaction", "growth in Asia" or even "best corporate culture". From an economic perspective, these goals are - if at all - a means to an end. Firms can only exist in the long-term if strategies are aligned towards the goal of making a profit. From a microeconomic perspective the focus is: survival of a firm and its profitability, depending on the competitive environment (market-based view), based on the resources and capabilities of a firm (resource-based view), and its capability for innovation and organizational development. The surviability of a firm is ensured in particular if it generates sustainable economic profits and if the equity investors receive a return on their invested capital in line with their expectations. From a static perspective - and without considering accounting principles and regulation, taxes, or balance sheet effects - profits 𝝅𝝅 of a firm at any given time are equal to the difference between revenues 𝑅𝑅 minus total costs 𝑇𝑇𝐶𝐶 as (4.1) 𝜋𝜋 = 𝑅𝑅 − 𝑇𝑇𝐶𝐶 . Revenues 𝑅𝑅 correspond to the price 𝑝𝑝 multiplied by the quantity 𝑞𝑞 . Total costs 𝑇𝑇𝐶𝐶 comprise all costs of a firm for labour and capital (see also ► Chapter 6). In particular, from an economic perspective, total costs also include the costs of equity, which reflect the return expected by the owners (in case of capital market-oriented or listed firms in the form of dividends): accounting profit (before opportunity costs), therefore, differs from economic profit. In the long-term and from a dynamic perspective, the central objective from an owner's perspective is to maintain and ensure the survival of the firm in order to generate profits on a recurring basis and to maximise the value of the firm. A firm’s value 𝑨𝑨 is the sum of all discounted future profits 𝜋𝜋 𝑝𝑝 as (4.2) 𝐸𝐸 = ∑ 𝜋𝜋𝑡𝑡 (1+𝑝𝑝𝑡𝑡)𝑡𝑡 𝑝𝑝=∞ 𝑝𝑝=0 = 𝜋𝜋 0 + 𝜋𝜋1 (1+𝑝𝑝1) + 𝜋𝜋2 (1+𝑝𝑝2)2 +. . + 𝜋𝜋𝑛𝑛 (1+𝑝𝑝𝑛𝑛)𝑛𝑛 +. . + 𝜋𝜋∞ (1+𝑝𝑝∞)∞ . With increasing time horizon, the discounted contribution of profits to the current firm value decreases. As a simple rule of thumb, it can be deduced that if profit expectations, 𝜋𝜋 0 = 𝜋𝜋 1 = … = 𝜋𝜋 ∞ , remain constant in the future and the discount factor remains the same, 𝑟𝑟 0 = 𝑟𝑟 1 = ⋯ = 𝑟𝑟 ∞ , the firm value (4.2) based on an infinite series boils down to <?page no="145"?> 4.1 Firms, objectives and strategies 145 (4.3) 𝐸𝐸 = 𝜋𝜋 0 + 𝜋𝜋1 (1+𝑝𝑝1) + 𝜋𝜋2 (1+𝑝𝑝2)2 +. . + 𝜋𝜋𝑛𝑛 (1+𝑝𝑝𝑛𝑛)𝑛𝑛 +. . + 𝜋𝜋∞ (1+𝑝𝑝∞)∞ = 𝜋𝜋𝑝𝑝 . The discount factor 𝑟𝑟 𝑝𝑝 of future profits (4.4) 𝑟𝑟 𝑝𝑝 = 𝐸𝐸𝐸𝐸 𝐸𝐸 ⋅ 𝑟𝑟 𝑆𝑆𝐻𝐻 + 𝐹𝐹𝐸𝐸 𝐸𝐸 ⋅ 𝑟𝑟 𝐷𝐷 = 𝑊𝑊𝐴𝐴𝐶𝐶𝐶𝐶 𝑝𝑝 at each point in time 𝑡𝑡 is calculated from the average cost of debt 𝑟𝑟 𝐷𝐷 (e.g, due to the interest rates of loans and bonds) and the cost of equity 𝑟𝑟 𝑆𝑆𝐻𝐻 weighted with the respective shares of equity capital 𝐸𝐸𝐸𝐸 and debt capital 𝐹𝐹𝐸𝐸 in the total capital 𝐸𝐸 = 𝐸𝐸𝐸𝐸 + 𝐹𝐹𝐸𝐸 of a firm. The equity capital costs 𝑟𝑟 𝑆𝑆𝐻𝐻 result as opportunity cost from the profit expectations of the equity capital providers (typically these are shareholders) given alternative possible investments (e.g, in other firms). The discounting factor corresponds to the WACC (weighted average cost of capital), which can be determined in a firm-specific way. From this perspective, a firm achieves a positive economic profit if the return on invested capital (ROIC) exceeds the weighted average cost of capital (WACC). The weighted costs of equity and debt capital in Germany in 2015 were around 9.6 % (with industry-specific differences and depending on the financing structure of the firm, see KPMG 2016). Thus, a firm, assuming constant profits of EUR 100 million (representing free cash flow), then, accordingly, (4.2) and (4.4) with (4.5) 𝐸𝐸 = 100 (1 + 0.096) 0 + 100 (1 + 0.096) 1 +. . + 100 (1 + 0.096) 20 +. . + 100 (1 + 0.096) ∞ = 100 1 + 100 1.096 +. . + 100 6.255 +. . + 100 ∞ = 100 + 91.241+. . +15.988+. . +0 = 1,041.67 has an indicative firm value of approximately EUR 1.042 billion - or more simply following from (4.3) that 𝐸𝐸 = 100/ 0.096 = 1,041.667 . If the current firm value (measured, for example, by market capitalisation on the stock exchange) is higher, then there are obviously rising profit expectations on the stock market, and vice versa. Rising profit expectations or falling capital costs increase the value of a firm and, in the long-term, boost the share price of a firm listed on the stock market. From a long-term perspective, all management decisions are aligned towards increasing the value of the firm. Inside firms, this might lead to decision conflicts over the weighing of short-term and long-term goals, especially because employed managers pursue short-term goals as part of their bonus arrangements, however, owners are solely interested in the longterm value of the firm. Profit maximization and the ability to survive For analytical simplification, it is often assumed that firms maximise short-term profits and long-term firm value - thus mathematical instruments can be used very simply to derive the best possible strategy (also in the literal sense of the word). Yet, a firm must have complete information about all possible strategies and perfect foresight, to be able to choose a strategy rationally and be able to implement it precisely. Accordingly, the behaviour of firms often is interpreted "as if" they maximise profits. In the short term it does not matter whether all firms really maximise profit ex ante, as in the long-term only those patterns of behaviour and firms <?page no="146"?> 4 Firms, competition and innovation 146 will survive which truly develop optimal strategies and adapt to changing competitive and environmental conditions, and finally show profit maximising behaviour. There is no clear empirical evidence for this and firms can, at least temporarily, prioritise other objectives such as market share, corporate social responsibility, empire building via mergers, technology leadership and a few more. Against the hypothesis of profit maximisation itself - whether in the absolute or weakened "as if" version - numerous empirical observations and theoretical considerations are mentioned. On the one hand, for the survival of a firm, profits are necessary, but not profit maximisation. However, profits are not sufficient for the survival of a firm: if strategic or financial investors observe that a firm is not fully exploiting its profit potential, a hostile takeover (by a private equity firm or a competitor) may occur. In this case, owners and management of a firm are replaced in order to increase the profit and value of the firm - at least an attempt to maximise profits can therefore, increase the chances of a firm remaining independent. Good to know │ Profits and environmental sustainability - isn't that mutually exclusive? Firms - especially those with a predominant or sole aim of profit maximisation - are constantly challenged by society and politics to review and explain their objectives, particularly in the case of planned job cuts or against the background of climate change. For stock corporations in Germany, the matter seems simple at first glance. In the German Stock Corporation Act (Deutsches Aktiengesetz), and repeatedly decided upon in case law, management board and supervisory board of firms are obliged to represent only the interests of the firm and thus, to ensure the existence and further development of the firm (see also Hutzschenreuter 2019). In reality, however, firms are always at the intersection of at least two interest groups: shareholders and stakeholders. From a shareholder perspective (the collective viewpoint of owners and senior management), profit and survival objectives of an organisation or firm dominate. However, from a stakeholder perspective (other interest groups such as employees, customers, suppliers, and also trade unions, public authorities, cities and communities, NGOs, or climate activists) objectives are manifold. Often conflicts between stakeholders and shareholders occur, as typically costs increase and profits decrease as a result of implementing stakeholder interests. Depending on the intensity of conflicting objectives and possible consequences, firms’ decisions must then address the interests and objectives of stakeholders and shareholders. Within the firm, stakeholder participation is partly formalised: in Germany, for example, the Works Constitution Act (Betriebsverfassungsgesetz) and the Codetermination Act (Mitbestimmungsgesetz) require the involvement of employees and information as well as participation rights, depending on the number of employees. New concepts attempt to identify an integrative shared value which emphasises the links between the interests of external stakeholders and shareholders within a firm (Hillman and Keim 2001 as well as Porter and Kramer 2011). <?page no="147"?> 4.1 Firms, objectives and strategies 147 A purely altruistic corporate strategy aligned only towards interests of stakeholders leads to a consumption of equity and threatens the existence of the firm. In addition, conflicts between managers and shareholders are inevitable. A regularly argued social responsibility of firms and managers also tends to serve to protect vested interests and maintain power, a reactive sealing off from stronger interests and usually the targeted preference of individual interest groups (see also Wagner 2019). Two mechanisms can be combined as a solution: voluntary ethical commitment by firms and strategic orientation of corporate social responsibility. It is precisely through strategic measures, which are comprehensibly documented, for example through reporting obligations, that firms can build up a reputation. If these measures and strategies then also support objectives of the shareholders - for example, by meeting ESG (environmental, social and corporate governance) or CSR (corporate social responsibility) criteria - then profits and sustainability are not mutually exclusive (Buchanan et al. 2018). Irrespective of this, addressing social goals through strategic CSR measures can of course also increase the attractiveness of a firm for employees or customers. Strategies of firms From a scientific point of view, three - sometimes contradictory, but complementary - perspectives on firms’ objectives and strategies have emerged (Alchian 1950, Alchian and Demsetz 1972, Donaldson 1990, Grant 1996, Nelson and Winter 1982, Rumelt et al. 1991 and Hart 1995) as described below. Industrial organisation perspective and strategic management - a firms’ objective is to maximise profit (short-term) and firm value (long-term). Research in industrial organisation and strategic management centres on the development or validation of optimal decisions in order to provide orientation and guidelines for corporate development and strategy selection - strategies are developed that are perfectly rational and aim to maximise profits. Behavioural science and evolutionary perspective - firms are complex, socio-economic organisations within which different and contradictory objectives exist (e.g., between managers or as silo thinking between departments). The focus of these studies is on behaviour that is actually observed: Managers act boundedly rational and engage in satisficing, decisions are based on path-dependent behaviour. Within the firm there is controversial negotiation about objectives - strategy is based on routines and is influenced by chance. Corporate governance perspective - many firms consist of owners, managers and employees. These separate roles foster conflicts of interest and objectives. In particular, instead of maximising profit or firm value, managers may maximise their status quo and budgets with discretionary room for manoeuvre, without being able to be monitored or sanctioned by owners. The focus of corporate governance research (on rules and principles of corporate governance based on a corporate constitution) is on these potential conflicts of interest resulting from the separation of management and ownership in capital market-oriented firms - strategy here is codetermined by organisational structure, hierarchies and decision-making models. <?page no="148"?> 4 Firms, competition and innovation 148 Organisation of firms, corporate governance and principal-agent problems In many firms, strategic decisions are not made by ‘lonesome’ individuals, but are delegated to committees (board meeting, management board, steering committee, divisional managers' meeting, etc.) - even more, these decisions are often made by employed managers, without participation of owners of the firm. Corporate governance describes internal organisation and regulation of collective actions within a firm. This includes hierarchies, decision-making paths and powers, the flow of information and the legal or factual regulatory framework for the management and monitoring of a firm. Specifically, corporate governance is concerned with how decisions are made and managed in firms on a formal or informal basis; how objectives for the organisation are derived and implemented; how internal and external monitoring is organised; and how shareholders and stakeholders are being informed or involved (Letza et al. 2005). It is precisely the interdependencies of different shareholder and stakeholder interests that influence the decisions and the choice of strategy of firms, especially since these do not only have to meet formal and logical criteria, but are often overlaid by ethical, moral, or emotional perspectives and expectations. For example, from a shareholder perspective, a job reduction motivated by the shut-down of a firm’s location may be a suitable action to reduce costs and increase profits - but from a stakeholder perspective, it reduces the credibility of the firm and may lead to customer churn. To ensure the survival of a firm, corporate governance is therefore used as a key strategic element. Corporate governance is becoming more important with increasing size of a firm; with complexity of stakeholder and shareholder relationships; and with the degree of separation of ownership and management of a firm: an owner managing an auto repair shop with three employees makes decisions in a fundamentally different way to a multinational car manufacturer with 500,000 employees. Organisation of a firm is based on the division or delegation of tasks or objectives. On the one hand this organisation is functional or divisional, on the other hand hierarchical through management and delegation. Ideally, good organisation induces all employees to do exactly the right thing at all times, however, reorganisations that are regularly observed in firms show that this optimal set-up rarely exists or is of lasting value. The causes for these sub-optimal situations from a microeconomic perspective are essentially linked to three challenges: incomplete or asymmetric information, incomplete monitoring and verifiability, and incomplete or wrong incentive structures. These three challenges are discussed in the context of the principal-agent theory. Here a principal is the owner or decision maker, the agent is the contractor or employee - thus structures like those between supervisory board and management board are described as well as situations between team leader and team member. First of all, the organisational structure of firms, in combination with incomplete information, causes significant problems which complicate the implementation of strategies or cause costs. Each larger firm is organised in teams (departments or divisions). This means that a result is achieved jointly within a team, but identification and measurement of the individual contribution of a team member is difficult. Anticipating this, each team member will perform a <?page no="149"?> 4.1 Firms, objectives and strategies 149 non-optimal individual work effort - with the expectation or hope of benefiting from the team effort and result. Firms can counteract this free-rider problem through improved monitoring of individual performance or incentive-compatible combinations of team and individual bonuses (Alchian and Demsetz 1976 and Grant 2016). This problem is not limited to production or service processes: decisions are also made and agreed upon collectively by board members or managing directors, typically in teams. Thus, once more here are no incentives to take risks, to take responsibilities or actually make decisions, and decisions are often delegated to boards and committees (Baysinger and Hoskisson 1990 and Hillman and Dalziel 2003). Moreover, shareholders, managers and employees have different incentive structures and objectives. Shareholders are essentially interested in a high firm value and long-term survival of the firm, bear risks and assume liability. Managers, on the other hand, focus on high income and degrees of freedom to design, create and decide, and are only prepared to take on risks and liability to a limited extent. Thus, the separation of ownership and management of a firm creates conflicts of objectives. Conversely, employees are looking for a long-term secure job, bear no risk and no liability, but are usually satisfied with lower incomes. With complete information, the different incentive structures and goals could be represented by perfect contracts, i.e., a congruence of all goals and incentives could be achieved. However, the contracts between shareholders and managers on the one hand and managers and employees on the other hand cannot be fully specified - neither are the activities clearly and conclusively described, nor are the objectives clearly and fully defined against the background of dynamic and unpredictable environmental conditions. As a consequence, the so-called principal-agent problem (Jensen and Meckling 1976, Fama 1980, Eisenhardt 1989 and Voigt 2019) arises: a contractor (the agent) typically has more or better information than the principal (e.g., concerning the degree of difficulty of a task or the effective achievement of objectives or productivity), and therefore complete control or monitoring of the agent is inefficient or impossible for the principal. This asymmetry of information provides the agent with discretionary room for manoeuvre, which he can use to his advantage and thereby causes two problems in the organisation of a firm - adverse selection and moral hazard - which can only be reduced by appropriate measures. The associated costs are socalled agency costs. Adverse selection can occur when a principal (e.g., an owner of a firm or employer) cannot clearly identify skills, objectives, or motivation of an employee prior to signing a contract. Adverse selection is based on hidden characteristics of the agent, which are not recognisable before the contract is signed, but become visible after the contract is signed. The risk of adverse selection when hiring new employees is usually reduced within firms by two measures: employees explain their individual abilities (signalling) by means of their university degree and previous job references; and firms test true abilities (screening) of applicants using an elaborate selection process and during a probation period. Moral hazard or misconduct can arise after the conclusion of a contract if an agent does not have to bear all the costs or risks associated with the performance of an activity because the principal is not in a position to assess in what way the agent is responsible for the result (e.g., achievement or failure to achieve the goal). Due to information asymmetry, shareholders (the <?page no="150"?> 4 Firms, competition and innovation 150 principals) are usually not in a position to judge whether the decline in a firm's profit is due to the inadequate performance of the managers (the agents) or to external influences such as an economic decline in demand or the strategic behaviour of competitors. Since managers recognise this connection, they will only pass on information that is useful to them and withhold further information (hidden information). In addition, due to the impossibility of complete control by the shareholders, managers will opportunistically take advantage of discretionary room for manoeuvre (hidden action) which is not congruent with the objectives of the shareholders such as off-sites in luxury hotels, employment of family members, settlement of ‘private’ business meals, self-promotion in the media or private use of firm resources. The risk of moral hazard can be reduced if a congruence of objectives can be established based on suitable incentive structures between principal and agent. This can be achieved, for example, by allowing managers to participate in profits and firm value (stock options or shares); supplemented by an improvement in corporate governance (compliance with guidelines for good corporate management, such as those contained in the German Corporate Governance Code, internal and external audits, reporting obligations and an appropriate management board and supervisory board structure); and corresponding sanctions in the event of violations of guidelines and directives. There is also external monitoring: managers who fail to make sufficiently good decisions are valued externally by the share price and by analysts from investment banks. If a firm is then taken over due to poor management performance (market for corporate control), in which the new shareholders want to increase the profitability of the firm, the management is replaced. Evolutionary dynamics and profitability Taken together, it is questionable whether firms really pursue the objective of maximising profits (or other variables such as market share of the firm or the status quo and salary of managers) and whether this is generally possible for them, in particular whether all necessary information is available and is used appropriately to make optimal decisions. In general, one can observe a variety of strategies in industries across firms, which, at least in the short term, co-exist evolutionary without maximising profits (Jovanovic 1982, Malerba and Orsenigo 1996 and Münter 1999). The intensity of competition then determines whether and how quickly sub-optimal strategies are sorted out: “If one thinks within the frame of evolutionary theory, it is nonsense to presume that a firm can calculate an actual 'best' strategy. […] There are certain characteristics of a firm's strategy and of its associated structure, that management can have confidence will enhance the chances that it will develop the capabilities it needs to succeed. […] there is a lot of room in between, where a firm (or its management) simply has to lay its bets knowing that it does not know how they will turn out. Thus diversity of firms is just what one would expect under evolutionary theory. It is virtually inevitable that firms will choose somewhat different strategies […]. Inevitably firms will pursue somewhat different paths. Some will prove profitable, given what other firms are doing and the way markets evolve, others not. Firms that systematically lose money will have to change their strategy and structure and develop new core capabilities, or operate the ones they have more effectively, or drop out of the contest.” (Nelson 1991, S. 69). <?page no="151"?> 4.2 Competitive advantage, market structure and firm-specific capabilities 151 4.2 Competitive advantage, market structure and firm-specific capabilities The lasting success of firms is based on a combination of executing the right strategy and the exploitation of firm or industry-specific competitive advantages. Temporary or permanent competitive advantages translate into relatively higher profits for a firm, because the firm either can charge higher prices than its competitors due to customers' higher willingness to pay, product quality or product differentiation; has lower costs due to economies of scale, economies of scope or more cost-effective access to labour and capital markets; or due to positioning in the competitive environment or firm-specific capabilities, the firm is not vulnerable to attack by competitors. Firms apply strategies to achieve their respective goals based on these competitive advantages. Strategy can generally be described as: the long-term orientation of a firm and guidelines for all decisions and the direction of all activities, taking into account market structure and possible strategies of all competitors, to exploit or realise competitive advantages based on and by shaping firm-specific capabilities in a dynamic competitive environment, with the overall objective of ensuring robust profitability and the firm's ability to survive. Strategy always represents a hypothesis that turns out to be right or wrong in future developments (Porter 1996, Rumelt and Lamb 1997 and Rumelt 2011). How firms determine, develop, and apply their heterogeneous strategies, how competitive advantages are justified, and why profits can be made is explained from two compelementary perspectives - competitive positioning and firm-specific capabilities as follows. Market-based view - the success of a firm is largely determined by an adequate positioning within the market structure, by the strategic behaviour of the firm and the attractiveness of the market. Competitive advantages result from being "in the right market". Resource-based view - the success of a firm is decisively characterised by the firm-specific capabilities, the existing core competencies and their future improvement. Competitive advantages result from having "the right capabilities". Market-based view as an explanation for competitive advantage Competitive advantage can be supported by the competitive environment. From a marketbased view, firms adopt market opportunities that they have identified based on market and competitive analysis. As a consequence, a firm exists as a reflection of market opportunities and positions itself in an attractive market or market segment. The profits themselves are essentially determined by the competitive environment and the attractiveness of the market, e.g., a generally high willingness to pay with low price elasticity of demand, or due to existing trends and market growth, or market entries of new firms do not happen due to extensive entry barriers. Accordingly, firms in the same industry show similar profitability or return on equity, as these are clearly determined by the market and only to a limited extent by the specific capabilities of <?page no="152"?> 4 Firms, competition and innovation 152 the firms. For example, from 2012 to 2016, all private hospital operators in the Greater Munich area benefited from positive market conditions as a result of Chinese and Arab medical tourism. In the same way, no German energy supplier was able to escape the negative market effects of the energy turnaround in the years 2010 to 2016. Conversely, profitability varies across different markets and industries - for example, firms in the pharmaceutical, telecommunications, or financial services industries systematically show higher profits than airlines or call centres. Strategy development from a market-based view focuses on finding the "right market": by analysing the competitive and market position; identifying and analysing opportunities and risks in the market environment; and the existence and development of market entry barriers. As a result, firms invest heavily in entry barriers to cover and secure their own positioning, but less in the skills of their own employees through further training and education. The market-based view explains, for example, the extremely high profitability of all “bratwurst stands” at the famous Nuremberg Christmas market at the same level. The market is attractive due to the high willingness to pay of customers; market entry for competitors is prohibited due to official approval; and the necessary core competencies (sourcing and grilling of brat-wursts and receiving cash) are almost irrelevant. Similarly, a beer tent at the Munich Oktoberfest is more profitable than a beer tent of the same size in Melbourne or Hatfield. Market structure, SCP-framework and five-forces framework Analyses from a market-based view are guided by the structure-conduct-performance framework (Bain 1949, Mason 1939, Porter 1981, Barney 1986 and Geroski 1990), which summarises important empirical findings of industrial organisation research on market structure, success factors of firms and their profits, as shown in ► Figure 4.2. Figure 4.2: Structure-conduct-performance framework. <?page no="153"?> 4.2 Competitive advantage, market structure and firm-specific capabilities 153 From a SCP perspective, exogenous factors such as the political, economic, social and legal competitive environment, the overall demand structure and market size as well as technological opportunities influence competition within an industry. The market structure then determines the interplay of firms’ strategies, which in turn determine the market result and the success of individual firms, e.g., profits of firms, but also their growth. Thus, it is clear that market results influence market structure and strategies, so that the SCP framework describes the interdependencies in competition. Market structure describes, as shown in ► Figure 4.3, the number and size distribution of firms within an industry - in its simplest case from one firm (a monopoly) to a few firms (an oligopoly) to many firms (in perfect competition). However, this has far-reaching strategic implications. Monopoly or dominant firm: this firm is alone in the market or has a dominant market share. Such firms in principle are free in their choice of strategy, since there are no repercussions from (even potential) competitors - the intensity of competition is very low (see ► Chapters 7 and 8). Firms can achieve very high profits, but inefficiency is often observed due to the lack of competition. Figure 4.3: Market structure, product differentiation and intensity of competition. Competition in an oligopoly: firms are mutually in strategic interaction, so that each firm takes the strategies of its competitors into account when choosing its own strategy - the intensity of competition increases, all things being equal, with the number of firms and is determined by the strategic behaviour of the firms (see further ► Chapters 9 and 10). Firms’ profits are typical, but the amount depends largely on the type and intensity of competition. Firms in perfect competition: all firms are small relative to the size of the market; products do not differ perceptibly from the customer's point of view. The competitive situation and the behaviour of competitors constrains and determines their own strategic behaviour - the intensity of competition is very high (see further on in ► Chapter 7). Firms achieve profits only temporarily at best and profitability is close to zero. <?page no="154"?> 4 Firms, competition and innovation 154 The environment of the competitive process and the market structure itself are essential determinants of possible strategies and the behaviour of firms: for example, the intensity of competition increases with the number of firms but decreases with a higher degree of product differentiation. Horizontal concentration and competitive intensity Industries differ in their horizontal concentration of the distribution of market shares. Market shares may be highly concentrated in a few firms (as for example in the German mobile telecommunications industry with four providers Telefónica, Vodafone, 1&1 Drillisch and Deutsche Telekom) or less concentrated in many small firms with minimal market shares (such as restaurants, flat rentals, hairdressers or bakeries). Concentration is determined by the dynamics of the number of firms and the change in market shares over time, i.e., whether many firms enter or leave the market and how different the growth rates are. Empirical evidence shows that the higher the horizontal concentration of an industry, the higher the profits of the firms, and profits of individual firms are positively correlated with market shares. High competitive intensity is the strongest driver for many firms to develop new strategies, implement innovations or reduce costs - but how can competitive intensity of an industry be estimated and measured? In fact, the intensity of competition in many industries can only be described vaguely and cannot be measured directly. This is due to the diversity of possible firm strategies and the specific conditions in individual markets. However, the intensity of competition can be inferred indirectly from the often well observable market structure. Market structure describes the number and size distribution of firms in an industry. In order to make market structure measurable and thus comparable across different industries, concentration indices and concentration rates are used. Horizontal concentration describes how strongly market shares are held by a small number of firms in an industry. Two types of measures are common: concentration rates and concentration indices. Concentration rates describe the sum of the market shares 𝑆𝑆 𝑖𝑖 = 𝑞𝑞 𝑖𝑖 / 𝑄𝑄 of the largest firms, e.g., as 𝑨𝑨𝑪𝑪 concentration rate (4.6) 𝐶𝐶4 = 𝑆𝑆 1 + 𝑆𝑆 2 +𝑆𝑆 3 +𝑆𝑆 4 = ∑ 𝑆𝑆 𝑖𝑖 4𝑖𝑖=1 the combined market shares of the four largest firms in an industry. This value tends towards zero if there is a very large number of firms in an industry, since even the four largest firms also have market shares close to 0% - conversely, the value is 1 when there are only four firms in that industry. The main advantage of this concentration rate is that it is usually easy to calculate and easy to comprehend - however, the distribution of market shares of smaller firms is not taken into account, and the value of 𝐶𝐶4 may be significantly distorted by the largest firm (e.g. Google in search engines in Germany). Therefore, if data from all firms are available, horizontal concentration in an industry is often measured by the so-called Herfindahl index (4.7) 𝐻𝐻 = 𝑆𝑆 12 + 𝑆𝑆 22 +𝑆𝑆 32 + ⋯ + 𝑆𝑆 𝑝𝑝2 = ∑ 𝑆𝑆 𝑖𝑖2 𝑝𝑝𝑖𝑖=1 , which equals the sum of the squared market shares 𝑆𝑆 𝑖𝑖 of all 𝑆𝑆 firms. This can be expressed alternatively with a number of firms 𝑆𝑆 and a variance 𝜎𝜎 𝑠𝑠𝑖𝑖 2 = 1 𝑝𝑝 ∑ (𝑆𝑆 𝑖𝑖 − 𝑆𝑆̅) 2 𝑝𝑝𝑖𝑖=1 of the market shares 𝑆𝑆 𝑖𝑖 as <?page no="155"?> 4.2 Competitive advantage, market structure and firm-specific capabilities 155 (4.8) 𝐻𝐻 = 1 𝑝𝑝 + 𝑆𝑆𝜎𝜎 𝑠𝑠𝑖𝑖 2 . The term 𝑆𝑆̅ = 1/ 𝑆𝑆 indicates the average size of the firms and represents for firms of the same size (i.e., 𝜎𝜎 𝑠𝑠𝑖𝑖 2 = 0 ) a minimum and thus, a lower bound of horizontal concentration depending on the number of firms. The larger the variance 𝜎𝜎 𝑠𝑠𝑖𝑖 2 of market shares of firms in an industry, the higher the horizontal concentration - the Herfindahl index converges towards 1 if one firm has a market share close to 100%, it becomes 0 if a large number of very small firms compete. As the Herfindahl index takes into account all firms in an industry, a much more precise estimate of concentration of the market structure and the underlying intensity of competition is possible. In addition, the Herfindahl index allows some further analysis: if the firms are of the same size (and thus a variance 𝜎𝜎 𝑠𝑠𝑖𝑖 2 = 0 ), the change in equation (4.8) results in the expected number of firms (numbers equivalent) (4.9) 𝑆𝑆� = 1 𝐻𝐻 , which corresponds to the reciprocal of the Herfindahl index. If the current number 𝑆𝑆 of firms in an industry is significantly larger than the expected value 𝑆𝑆� , this is a first indicator of a dominant position of the largest firms and reduced intensity of competition (further ► Chapter 6 on minimum efficient firm sizes). Case study │ Intensity of competition in the German food retailing industry ► Table 4.1 first shows the market shares of the eight largest German food retailers for the years 2009 to 2019. At first glance there appear to be four large firms, of which Edeka was able to slightly increase its market share lead in the years under consideration. It is also apparent that the four smaller competitors have slightly lost market share overall. If one calculates the 𝐶𝐶4 concentration rate and the Herfindahl index for the given marketshares, ► Figure 4.4 shows a significant increase in horizontal concentration. In fact, however, these firms cover only about 82% of the total market in 2009, rising to 85% in 2019 - i.e., obviously smaller or regional firms are not included in the statistics. In order to take the missing firms into account, it is possible to estimate their number and size by assuming that the missing market shares are distributed among firms that are on average half the size of the smallest firm 𝑆𝑆 𝑖𝑖 𝑚𝑚𝑖𝑖𝑝𝑝 currently in the data. This gives a number of missing firms 𝑆𝑆′ for 2009 of (4.10) 𝑆𝑆 2009 ‘ = �1−∑ 𝑠𝑠𝑖𝑖 𝑛𝑛𝑖𝑖=1 � 𝑠𝑠𝑖𝑖 𝑚𝑚𝑖𝑖𝑛𝑛/ 2 = 25.29, so that the number of relevant competitors in 2009 was approximately 𝑆𝑆� = 𝑆𝑆 + 𝑆𝑆 ′ ≅ 33 . If one now continues to develop the value of 𝑆𝑆� for the years 2010 to 2019, it can be seen that the obvious shifts in market share by the large firms have led in particular to a significant reduction in the number of firms in the market (see also ► Figure 4.4, bottom left). <?page no="156"?> 4 Firms, competition and innovation 156 Market structure in German food retailing year 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 market shares (revenues) Edeka-Group 24.4 % 24.6 % 25.3 % 25.6 % 25.5 % 25.2 % 25.3 % 25.3 % 23.5 % 26.2 % 26.8 % Rewe-Group 16.1 % 16.2 % 14.9 % 15.0 % 14.9 % 14.8 % 15.0 % 15.1 % 17.5 % 16.1 % 16.2 % Schwarz-Group 13.7 % 13.9 % 13.8 % 13.8 % 14.4 % 14.8 % 14.7 % 15.0 % 15.9 % 15.7 % 16.0 % Aldi-Group 12.5 % 12.1 % 12.0 % 12.0 % 12.3 % 12.1 % 11.9 % 12.0 % 12.2 % 12.0 % 11.5 % Metro-Group 7.3 % 7.0 % 6.8 % 6.5 % 6.0 % 5.8 % 5.4 % 5.2 % 5.6 % 4.8 % 4.6 % Lekkerland 4.8 % 4.7 % 4.5 % 4.7 % 4.6 % 4.7 % 4.7 % 4.6 % 3.8 % 3.9 % 3.7 % Dm 2.1 % 2.2 % 2.4 % 2.6 % 2.9 % 3.1 % 3.3 % 3.5 % 3.2 % 3.5 % 3.6 % Rossmann 1.4 % 1.6 % 1.7 % 2.3 % 2.5 % 2.6 % 2.7 % 2.8 % 2.6 % 2.9 % 3.0 % sum of market shares 82.3 % 82.3 % 81.4 % 82.5 % 83.1 % 83.1 % 83.0 % 83.5 % 84.3 % 85.1 % 85.4 % C4-concentration rate 0.6670 0.6680 0.6600 0.6640 0.6710 0.6690 0.6690 0.6740 0.6910 0.7000 0.7050 Herfindahl-concentration index 𝐻𝐻 0.1281 0.1286 0.1272 0.1291 0.1303 0.1292 0.1292 0.1305 0.1323 0.1395 0.1426 missing firms 𝑆𝑆’ (estimate) 25.29 22.13 21.88 15.22 13.52 13.00 12.59 11.79 12.08 10.28 9.73 adjusted Herfindahl-index 0.1294 0.1300 0.1287 0.1311 0.1324 0.1314 0.1315 0.1328 0.1343 0.1417 0.1448 hypothetical 𝑆𝑆� = 1/ 𝐻𝐻 (numbers equivalent) 7.73 7.69 7.77 7.63 7.55 7.61 7.60 7.53 7.44 7.06 6.91 number of relevant competitors 𝑆𝑆� (estimate) 33.29 30.13 29.88 23.22 21.52 21.00 20.59 19.79 20.08 18.28 17.73 variance of market shares 0.5431 0.5488 0.5542 0.5505 0.5494 0.5355 0.5388 0.5423 0.5432 0.6122 0.6426 Table 4.1: Market structure of German food retailers 2009 to 2019, market shares based on revenues in Germany. Data source: BVE Jahresbericht 2019/ 2020, own calculations. <?page no="157"?> 4.2 Competitive advantage, market structure and firm-specific capabilities 157 Figure 4.4: Analysis of market structure and intensity of competition of German food retailers. The picture becomes even clearer if the number of relevant firms is viewed in relation to the variance of market shares over time (► Figure 4.4 in the lower right-hand corner): obviously, since 2009, the number of firms has initially decreased significantly due to market exits as well as mergers and acquisitions, and from around 2016 the variance of market shares did increase significantly. This is equivalent to a market share growth of the large firms at the expense of the market shares of the smaller firms. For both 𝐶𝐶4 concentration rates and the Herfindahl index, empirical studies show the following robust stylised facts (Böbel 1984, Sutton 1991 and Münter 1999): the higher the horizontal concentration, the lower the intensity of competition; the higher the horizontal concentration, the higher the average profitability of firms; and horizontal concentration is significantly influenced by the interplay of market size and growth, entry barriers and sunk costs in an industry. Against this background, horizontal concentration is a possible indicator of market power or collective dominance (further on ► Chapter 7 and ► Chapter 10). Every second year, the Monopolkommission (Monopoly Commission) in Germany publishes a report on horizontal concentration in German industries in order to analyse the development of competitive intensity and identify possible patterns of dominant market positions (Monopolkommission 2020). The close correlation between the 𝐶𝐶4 concentration rate and the Herfindahl index in ► Figure 4.4 in the upper right-hand corner is not coincidental - it can theoretically be shown that the Herfindahl index has a lower boundary <?page no="158"?> 4 Firms, competition and innovation 158 (4.11) 𝐻𝐻 𝑚𝑚𝑖𝑖𝑝𝑝 = 𝐶𝐶42 4 and an upper boundary (4.12) 𝐻𝐻 𝑚𝑚𝑚𝑚𝑚𝑚 = � 𝐶𝐶4 4 für 𝐶𝐶4 < 14 𝐶𝐶4 2 für 𝐶𝐶4 ≥ 14 (Sleuwaegen and Dehandschutter 1986 as well as Münter 1999). In addition, based on empirical studies, horizontal concentration can be roughly assigned to certain market structures, as shown in ► Figure 4.5. Accordingly, in the German food retailer industry, in which the four largest firms have market shares of more than 70% in 2019, dominant firms can be assumed. Figure 4.5: Market structure and horizontal concentration. In the US, the Merger Guidelines for the assessment of mergers define a classification of horizontal concentration based on the Herfindahl index: at 𝐻𝐻 < 0.15 a market is considered not concentrated, a value of 0.15 ≤ 𝐻𝐻 ≤ 0.25 indicates a medium concentration, at 𝐻𝐻 > 0.25 a market is highly concentrated. H 0,4 0,6 0,8 0,8 1 0,2 C4 0,6 1 0,4 0 0,2 dominant firms oligopoly perfect competition H min H max <?page no="159"?> 4.2 Competitive advantage, market structure and firm-specific capabilities 159 Barriers to entry Market entry conditions play a key role affecting market structure and horizontal concentration of firms in an industry. Entry barriers describe all conditions which either explain how a new entrant is at a competitive disadvantage or has higher costs compared to incumbents, or is unable to enter the market at all, such that entry barriers reduce the intensity of competition for incumbents. Typically, the costs of market entry - e.g., in the form of a productor firmspecific marketing campaign - are also sunk costs, i.e., they are industry-specific and cannot be recovered once a firm exits the market (see further on in ► Chapter 6). Market entry barriers can take the following dimensions, among others (Geroski et al. 1990). Strategic barriers to entry are based on strategic decisions by incumbent firms which aim to block market entries or make market entries more difficult or more costly for new firms. Incumbents, for example, voluntarily invest in sunk costs (extensive marketing investments that promote the building of reputation and brand loyalty; R&D investments that enable rapid technological progress; use of direct and indirect network effects to establish a multisided market, etc.), which are intended to signal potential competitors the unattractiveness of market entry: firms that enter the market must also be able to finance this level of investment in the long-term - many are thus, deterred from entering the market (cf. further ► Chapters 9 and 10). Structural barriers to entry are determined by exogenous industry-specific technology and production functions (increasing returns to scale, economies of scale or vertical integration), which lead to a minimum size of a firm and correspondingly high capital requirements in an industry. In addition, structural barriers to entry can be transformed into strategic barriers to entry (see ► Chapter 6 for further details). Structural barriers to entry can also arise if firms can transform learning curve effects into lower costs based on many years of experience - so age and experience of incumbents are then an indication of the entry barriers for new firms. Legal barriers to entry are determined by property rights, legislation and regulation which only allow a certain number of firms (taxi services, notaries, etc.) to enter the market, or by patents or licences (registered design, copyrights, etc.) which block other firms from entering the market or make it more costly. Entry barriers allow a separation between incumbent firms and new entrants: an incumbent has a competitive advantage, based on sunk cost investment and industry-specific experience, over a new entrant, which may be reflected in reputation, strong customer loyalty, lower absolute costs, or greater ability to innovate. Profitability and the five-forces framework Porter (1980, 1981 and 1985) condensed the SCP framework into the five-forces framework as a management tool to determine the attractiveness of a market or industry, measured by the profitability of firms and the stability of profits, and to derive strategic decisions, e.g., on market entry or investments. <?page no="160"?> 4 Firms, competition and innovation 160 Figure 4.6: Five-forces-framework and PEST-analysis. Based on empirical studies, five mutually interdependent variables, as shown in ► Figure 4.6, turn out to be decisive for the level of profits in an industry. (1) Intensity of competition and choice of strategic parameters: the higher the intensity of competition and the lower the horizontal concentration, the lower the profitability of an industry. The lower the number of firms and the higher the degree of product differentiation, the lower the intensity of competition. If firms compete via prices, mutual price undercutting is the rule - profit margins fall and profitability decreases. In contrast, profitability is higher if firms compete based on long-term capacity and product differentiation (see ► Chapter 10). Market growth can also play a role: if the market grows strongly, for example through new customers, competition is focussed on winning new customers. As a result, the intensity of competition is low. In shrinking markets, however, the intensity of competition tends to increase: firms also compete aggressively for competitors' existing customer base. (2) Threats from new firms and extent of entry barriers: the lower the barriers to entry, the more entries into the industry, the higher the intensity of competition and the lower the profitability. Entry barriers based on sunk costs, however, can also represent exit barriers (Caves and Porter 1977 and Rosenbaum and Lamort 1992): firms are then forced to stay in the industry and the intensity of competition increases - with a negative effect on profitability. Market entry is all the easier if new firms have access to relevant suppliers and distribution partners - or if a new value chain can be established through new business models. (3) Bargaining power and market power of suppliers: if suppliers can exert a strong influence on prices or quality in negotiations, or if the dependency on individual suppliers is high (as for example in complex international supply chains), the profitability of an industry <?page no="161"?> 4.2 Competitive advantage, market structure and firm-specific capabilities 161 is low. This is particularly the case if there is a small number of suppliers facing a large number of customers, or if there are high switching costs from one supplier to another. (4) Bargaining power and market power of customers: the higher the bargaining power of customers, the more they are bundled or the more sporadic the customer relationships are, the lower the profitability of an industry. This applies, for example, to standardised or weakly differentiated products and if customers are well informed about prices and quality. On the other hand, high customer loyalty or dependence on products and services can significantly reduce customers' negotiating power. (5) Threat of substitutes and availability of alternative solutions: the more customers are able to switch to alternative products and services, and the larger the price elasticity of demand, the less able firms are to impose high prices and the lower the profits within an industry. From a management perspective, the five-forces framework always needs to be integrated into a comprehensive analysis of the competitive environment, for example, using the PEST framework, since in many industries relevant influencing factors (e.g., technological changes such as digitalisation or new legal frameworks for energy system transformation and CO2 taxation) arise outside a narrow competitive environment and have an impact on the profitability of an industry. In addition, these exogenous changes also shift and blur market and industry boundaries (Malhotra and Gupta 2001). Thus, the five-forces framework is mainly descriptive in its application, no logical or quantitative statements can be derived, so that in addition game theoretical models as in ► Chapters 9 and 10 have to be applied. Case Study │ Five-forces analysis in the airline industry Numerous industries are regularly screened using the five-forces-framework in the context of market and competitive analyses - for example, the International Air Transport Association IATA had McKinsey repeatedly analyse the low profitability of airlines compared to other industries (IATA 2013): (1) Very high level of competitive intensity - due to a large number of firms, low product differentiation, increasing international competition, seasonally low-capacity utilisation and tendency towards price wars. (2) Strong and growing threat from new firms - due to deregulation of international markets, low sunk costs from leasing (used) aircraft and low investment needs because of small minimum size of a firm and relatively low economies of scale. (3) Massive bargaining power and market power of suppliers - in particular due to only two suppliers for new long-haul aircraft (Boeing and Airbus), a small number of international hubs (Frankfurt, Singapore, Istanbul, etc.) with high charges, and trade unions with market power, with a high risk of strike action by in-flight personnel and airport ground and security staff. (4) Moderate to high bargaining power and market power of the customers - based on weakly perceived product differentiation and high price transparency, which is increasing due to bundling and allocation of demand by booking and flight portals, while the customers' willingness to switch continues to be high due to weakly effective loyalty programmes. <?page no="162"?> 4 Firms, competition and innovation 162 (5) Moderate to high threat from substitutes and alternative services - in national markets increasingly from high-speed trains (Germany, France, Belgium, China, Japan, etc.), and internationally for business travel from substitutes such as web or video conferencing. All factors taken together consistently explain why airlines typically show low profitability, especially in comparison with firms in other industries, and are often not able to cover their weighted average cost of capital (WACC) or to achieve an economic profit (Bundesverband Deutscher Fluggesellschaften 2013). Resource-based view as an explanation for competitive advantage Competitive advantage can be based on firm-specific capabilities. From the resource-based view, firms basically represent capabilities and core competencies - consequently, a firm exists as a representation of tangible and intangible resources and develops strategies based on these (Penrose 1959, Wernerfelt 1984, Prahalad and Hamel 1990, Rumelt et al. 1991, Teece 2007 and Teece et al. 1997). These core competencies are static capabilities such as qualification or experience of employees, or are rooted in technology, organisation and based on management skills of a firm, efficiency and cost structure as well as high product quality and reputation of a firm. These capabilities give rise to higher willingness to pay and/ or lower average or marginal costs. Profits from this point of view are largely determined by these firm-specific capabilities - firms within an industry or market differ significantly in their profitability if capabilities are heterogeneous. Firms with low profits do not reach the profitability of successful firms, either because they lack access to the necessary resources or because they cannot adapt, imitate, or learn these skills. Strategy development from the perspective of the resource-based view essentially attempts to identify and use the "right capabilities". Competitive advantage is thus, derived from an internal analysis of core competence and strengths/ weaknesses of the organization. In addition, strong investments are made in core competencies, skills are further developed, and existing competencies are protected by imitation barriers (patents, IP foreclosure, employee retention, etc.). As a result, firms invest more in employee skills, for example, in order to expand firm-specific capabilities, but less in marketing to build up entry barriers. Profitable growth of a firm and the long-term viability of an organisation depend in particular on dynamic capabilities. These describe an adaptability of a firm and involve learning as an organisation, integrating new core competencies (e.g., new employees, but also an acquisition of a firm), adapting existing skills to new markets or to changing environmental conditions and generating innovations or new business models. It is precisely these intangible, tacit and hard to describe skills - corporate culture, organisational experience, or the ability of a firm to change or develop strategically - that can play a central role in the success of a firm. Dynamic capabilities are difficult for competitors to identify and therefore, difficult to imitate: so called causal ambiguity of a firm's success can be based on hidden firm-specific dynamic capabilities or on externally unrecognisable connections with suppliers or customers (Reed and DeFilippi 1990, Lippman and Rumelt 1982 and King 2007). Dynamic capabilities are particularly crucial for firm success and profitability whenever market or technological environmental con- <?page no="163"?> 4.2 Competitive advantage, market structure and firm-specific capabilities 163 ditions change (Eisenhardt and Martin 2000, Teece 2007 and Henneke 2014). Accordingly, three generic dynamic capabilities are necessary. Sensing - the continuous perception, observation and analysis of changes in opportunities and risks in the market environment and the exploratory development of new solutions across markets and technologies. Seizing - seizing opportunities by trying out new products, business models and technologies in market segments, or by acquiring firms in these market segments or cooperating with start-ups. Reconfiguring - the transformation of the organisation or business model to address the opportunities recognised, the development of necessary new static capabilities and the abandonment of now irrelevant capabilities or business areas. Case Study │ Success of RyanAir The resource-based view explains, for example, the extremely high profitability of the European airline RyanAir, as outlined in ► Figure 4.7, compared to competitors in the runup to AirBerlin's market exit in October 2017: the airline industry has been characterised for decades by very low profitability and a high number of insolvency-related market exits due to high competitive intensity and low entry barriers. Obviously, RyanAir can use nonimitable capabilities, low-cost structures and high strategic skills to generate robust high profits in an otherwise unattractive market (IATA 2013 and Bundesverband Deutscher Fluggesellschaften 2013). Among other things, RyanAir has, against the background of an analysis of transaction costs, drawn the boundaries of the firm more narrowly than its competitors: RyanAir can handle almost 80% of Lufthansa's passenger traffic with one fifth of its employees and half of its aircraft; RyanAir can handle almost three times the number of passengers with approximately the same number of employees as AirBerlin; the efficiency measured in passengers per aircraft or passengers per RyanAir employee is drastically higher than that of its competitors and highlights the strategic need for action at Lufthansa and explains AirBerlin's exit from the market. As a result, RyanAir made a profit of EUR 17 per passenger in 2014, Lufthansa EUR 4 and AirBerlin EUR -9. <?page no="164"?> 4 Firms, competition and innovation 164 Figure 4.7: Differences in profitability based on firm-specific capabilities (Data from corporate reports, share prices 2010 to 2017 comdirect.de; own calculations). Determinants of profitability Of course, an analytical separation of competitive advantages from market-based view and resource-based view is blurry in reality - every firm tries to operate in the right market with the right skills. Both views together are integrated in the so-called SWOT analysis (internal to a firm strengths and weaknesses, external to a firm opportunities and risks). A main focus of the market-based view is the identification of market opportunities, whereas the resource-based view is looking at strengths and capabilities of a firm to evaluate and design strategies. However, it can be empirically investigated which of these models used to explain competitiveness offers a better explanation for the profits of firms. Usually, it is examined which proportions of profits can be explained by belonging to a particular industry and which proportions of profits can be attributed to firm-specific capabilities. Empirical studies (Cubbin 1988, McGahan and Porter 1999 and 2003, Rumelt 1991, Schmalensee 1985 and Bradley et al. 2014) show that profitability (measured in profits, return on equity or cash flow) varies across industries. ► Figure 4.8 shows the return on invested capital of selected industries for the period 1965 to 2007 as a median and within ranges of the 2nd and 3rd quartiles - apparently a confirmation of the market-based view, because pharmaceutical or telecommunications industries have systematically higher profitability than metal processing or airlines for more than four decades. <?page no="165"?> 4.3 Competition and innovation 165 Figure 4.8: Profitability of selected industries between 1965 and 2007, median and 2. and 3. quartile of return on invested capital (left) and determinants of profits 2007 to 2011 (right) (Data source: McKinsey 2014 and Bradley et al. 2014). In contrast, a McKinsey study from 2014 shows that for 2,288 firms in the period 2007 to 2011 (► Figure 4.8 on the right) on average 60% of a firm's absolute profit can be explained by firm effects (the resource-based view) and 40% by industry effects (the market-based view). The share of firm-specific effects is particularly high for very successful firms (quintile 1), as well as for those with the lowest absolute profits (quintile 5). If firms achieve average profits (as in quintile 2 to 4), firm and industry effects have equal influence. In addition, empirical studies show strong evidence of firm-specific persistence of profits and a high positive correlation with market shares within an industry. Apparently, there are firms that can achieve higher market shares coupled with higher profitability or absolute profit levels over many years (Mueller 1990). Competition within an industry therefore neither leads to an equalisation of the firms' profits nor to a reduction in the level of profits. There are two possible causes for the persistence of profits: on the one hand, structural restraints of competition (e.g., entry barriers) within the industry or market power of individual incumbent firms, on the other hand, higher efficiency and competitive advantage based on the resource-based view. Taken together, empirical studies do not show a clear picture: the market-based view explains industry-specific effects on profitability, but within the individual industries there are strong firm-specific effects as an indication of the resource-based view. 4.3 Competition and innovation In the first quarter of 2007, Nokia - according to legend once a successful manufacturer of rubber boots - had a global market share of about 40% in mobile telecommunications devices (Economist 2011). On 9 January 2007, Apple announced the launch of the first iPhone, which <?page no="166"?> 4 Firms, competition and innovation 166 was available in the US from 29 June 2007 and then successively worldwide. Nokia's worldwide market share decreased to about 5% by 2012. Innovations can significantly influence competition and market structure. In the following section, the interaction between innovation and market structure is examined in order to gain an understanding of the shortand long-term patterns of competition. ► Figure 4.9 shows the market shares of some leading smartphone manufacturers from Q4/ 2009 to Q2/ 2020. Apparently, the two initial market leaders Nokia and Blackberry were completely squeezed out of the market within five years, Apple was never able to achieve market share leadership and Samsung was replaced by Huawei at the end of 2019 after more than nine years as the largest manufacturer of smartphones. Figure 4.9: Market shares of leading smartphone manufacturers worldwide, Q4/ 2009 to Q2/ 2020, Data source: IDC Data, data partially estimated. Competition between firms has two characteristics directly related to innovation, first described by Schumpeter (1911) and Hayek (1968). Hayek described competition as a discovery procedure. Any outcome of a dynamic competitive process is at best incompletely predictable because competition between competing firms always aims at the discovery of novel competitive advantage or strategies that are tried out in competition. This creates 'new knowledge' per se, the competitive process is open-ended and cannot be predicted. For this reason, competition policy interventions, for example, must generally be looked at critically, since no institution can precisely predict the consequences of technology or competitive developments. <?page no="167"?> 4.3 Competition and innovation 167 Schumpeter, on the other hand, has stressed the importance of competition as a process of creative destruction: " The opening up of new markets, foreign or domestic, and the organizational development from the craft shop to such concerns as U.S. Steel illustrate the same process of industrial mutation — if I may use that biological term — that incessantly revolutionizes the economic structure from within, incessantly destroying the old one, incessantly creating a new one. This process of Creative Destruction is the essential fact about capitalism.“ (Schumpeter 1942, p. 83). Competition here is essentially a process of innovation, i.e., previous products, market structures and firms are constantly being transformed or even replaced by new ones. A key role is played by entrepreneurs who develop innovative products or business models and bring them to the market. Innovation in general means an implementation of so-called new combinations, for example in products, processes, technologies, business models or the organisation of a firm. This does not require 'something new' per se, but rather emphasises the complementary and combinatorial character of innovation - e.g., digitisation, which essentially supplements existing business models, services and products or enables new combinations (Brynjolfsson and McAffee 2016). Although the repercussions of strategic competitive behaviour of firms on the structure of the industry are often emphasised, the causality is usually seen in accordance with the SCP framework from the structure of the industry on the behaviour of the firms towards the market outcome. However, if we look at the interrelation between competition and innovation, this is only half the truth. “Competition, at least in the most popular sense of the word, is all about rivalry between firms. It is about taking actions to increase, or at least to defend market share. The popular language of competition is often clothed in metaphors of conflict involving gladiators, knights in white and black armour, and so on. Yet not all of the responses that a firm makes to the actions of its rivals are the same. Some are more fundamental or more revolutionary than others. In this context, it is useful to distinguish tactical from strategic responses to rivals. A tactical response to the action of a rival is a response in kind. It is likely to involve use of the same (or a similar) competitive weapon (i.e., meeting a price cut with a price cut), and does not alter the basis on which the two firms compete. By contrast, a strategic response to the action of a rival is an attempt to change the basis of competition between the two firms (i.e., meeting a price cut with a radical change in the way a product is distributed or marketed). It is about breaking the rules of competition, and it corresponds to what most people understand when they talk about innovation.” (Geroski 1998, p. 15). A part of the competitive behaviour of firms - innovation - is therefore, not primarily directed against competitors, but aims at changing market structures and the rules of the competitive process. Moreover, innovations do not follow from the competitive situation, but innovations are the driving force behind the competitive process and can change market structure. In order to understand long-term competitive processes and the evolution of industries, it is therefore, not reasonable to start from given market structures, but it is necessary to focus on the co-evolution of technology, firms and competition. <?page no="168"?> 4 Firms, competition and innovation 168 Industry life cycles and the long run evolution of market structure The long run evolution of industries is characterised by numerous market entries of new firms and market exits of non-viable firms. Empirical studies show a robust pattern of the number of firms along an industrial life cycle and, beyond this, of entries and exits, as shown in ► Figure 4.10 and in ► Table 4.2 for the global automobile industry (Jovanovic and MacDonald 1994, Klepper and McGraddy 1990, Klepper 1997, Münter 1999 and Münter 2013). Figure 4.10: Industry life cycle, technology cycles and technological regimes. Initiated by a few first firms - in the automobile industry, Benz and Peugeot, among others - new, innovative products are launched on the market. Subsequently, during stage 1, numerous other new firms which take up or modify the product innovation enter the industry and the number of firms increases significantly. However, after some time (in 1922 in the automobile industry) this process is reversed and the number of firms in an industry falls dramatically during stage 2 as a result of numerous exits. Eventually, number of entries and exits more or less balances out and the number of firms stabilises. <?page no="169"?> 4.3 Competition and innovation 169 Case Study │ Competition, population dynamics and survival of firms in the automobile industry Population dynamics in global automobile industry all stages stage 1 stage 2 stage 3 start 1875 1922 1945 duration in years 47 23 > 70 average number of market entries per year 18.11 37.47 14.08 6.09 average number of market exits per year 16.71 24.51 34.78 5.17 average entry rate per year in % 8.46 % 16.08 % 5.17 % 3.37 % average exit rate per year in % 7.80 % 10.52 % 12.76 % 2.86 % number of unique firms 2,499 1,761 928 550 incumbents firms in 2012 started in … 13.33 % 8.72 % 77.95 % firms entering in stage x still existed in 2012 … 1.48 % 8.33 % 34.30 % Table 4.2: Population dynamics of the global automobile industry (Münter 2013). Looking at this development using the example of the automobile industry, it is clear that during stage 1 both absolute and percentage numbers of market entries dominate the number of market exits. During stage 2, this pattern is reversed until stage 3, with about 3% annual market entries and exits, the number of firms remains more or less constant. The reason for the pattern described in the example of the automobile industry, which can also be observed in a very large number of other industries, is not a decline in demand: production and sales in the global automobile industry have been growing continuously since its inception. The change in the number of firms is due to the interplay of innovation, technological change and strategic behaviour of firms. Technology cycles in the long run evolution of industries A first explanation for the long-term development of market structure, the number of firms and changes in the competitive process lies in technology cycles, which describe regular patterns of innovation over time as a sequence of an initial experimental phase, an emergence of the dominant design and finally a stage of incremental change (Utterback and Abernathy 1975, Clark 1985, Utterback and Suarez 1993 a and b, Christensen et al. 1998 as well as Henderson and Clark 1990). Here, as shown in ► Figure 4.10, three stages running in parallel to the industry life cycle can be identified. <?page no="170"?> 4 Firms, competition and innovation 170 Experimental stage: at the beginning of an industry life cycle, firms compete with very different product concepts and configurations. There is a high degree of uncertainty regarding both technological opportunities and true customer expectations and needs. Thus, none of the firms is able to offer a completely convincing product from a customer's perspective, but each firm tries to address identified market niches with a firm-specific product and to gain market share in a trial-and-error process. The firms' business models and strategies are very heterogeneous during this stage. Each new firm brings in product innovations which replace previous solutions and render existing knowledge within an industry obsolete in parts. Formation and emergence of a dominant design: with competition between firms over time, a dominant design emerges. One firm is first in combining existing product innovations and features in such a way that expectations of a large number of customers are well met. In the automobile industry, this was achieved by Ford's Model T (produced from 1908 to 1927) - a car with four wheels, front axle control, combustion engine in the front, passenger cabin, rear luggage compartment, brakes on the wheels, front and rear lighting, steering wheel and a few other features still found in almost every car today. The variety of competing product concepts and business models of the experimental stage is now coming to an end, a lock-in effect occurs and the features of this dominant design shape expectations of customers and are henceforth a de facto standard in an industry, which is imitated by other firms or adapted to specific firm requirements. Stage of incremental change: based on a dominant design, path-dependent further developments, and improvements of the product settles in, which extend and deepen existing knowledge in an industry, so that firmand industry-specific knowledge is built up and lasts. However, the focus is now no longer on product innovations, but on increasing efficiency and scaling up production based on process innovations. The emergence of a dominant design provides an initial explanation for a transition from stage 1 to stage 2 of the industry life cycle and changes the competitive process in four main dimensions as follows. Growth in market shares: market shares of those firms that are able to produce the dominant design are now growing strongly. In the smartphone industry, all firms that have been able to reproduce the core elements of the dominant design with the blueprint of the Apple iPhone - touch screen with virtual keyboard, photo camera, app store, music player, email reception and Wi-Fi capability - have grown strongly: Samsung, Huawei, Oppo, HTC and of course Apple itself (Cecere et al. 2015). Shake-out of competitors: other firms that are unable to produce the dominant design or deliberately continue to adhere to their previous product concepts are losing market share and are perspectively being squeezed out into niches or forced to exit the market during stage 2. For example, in the smartphone industry the former market leader Nokia’s firm-specific capabilities were insufficient, and the Blackberry manufacturer, Research in Motion, continued to stick with a combination of small screen and physical keyboard. Reduction of uncertainty: a dominant design significantly reduces uncertainty in an industry, providing incentives for investment in production capacity, branding, customer base and firm size, and building and enhancing firm-specific capabilities. Rise of barriers to entry: these growth processes are accompanied by a creation and severe rise of structural and strategic barriers to entry, so that both absolute number of new <?page no="171"?> 4.3 Competition and innovation 171 firms and entry rates (see also ► Table 4.2) are significantly lower. At the same time, however, the likelihood of a permanent stay in an industry is now increasing - the reason for this is again reduced uncertainty, since basic conditions of the competitive process can now be better assessed. Innovation and technological regimes Comparing the two major publications of Schumpeter, ‘Theorie der wirtschaftlichen Entwicklung’ (1911) and ‘Capitalism, Socialism and Democracy’ (1942), Schumpeter gave very different answers to the question of competitive advantage in the innovation process depending on the size of the firm. In 1911 Schumpeter saw small, innovative and young firms (start-ups) as main drivers of technological progress; in 1942 he emphasised the role of large firms due to high capital intensity of R&D processes and necessary industry-specific knowledge. Adopting a dynamic view, indeed, both views seem to be justified empirically: on the one hand, the relative roles of firms in the innovation process change over time; on the other hand, firm-specific roles in the R&D process emerge in many industries. This creates a cross-industry ecosystem of customers, firms and suppliers in the R&D area, which determines the industry-specific patterns of innovation and interaction between large and small firms. Nelson and Winter (1982) and Winter (1984) picked up on the questions and ascribed central importance to determinants of the emergence of new knowledge for the joint explanation of competition and innovation processes. The crucial questions, which go back to Schumpeter, are: whether new knowledge - product innovations, new business models or technologies - is more likely to be generated in R&D departments of large incumbent firms, or by start-ups or spin-offs; and whether new knowledge remains and consolidates persistently inside new firms and startups, or whether new knowledge is absorbed quickly and directly by large incumbent firms. Figure 4.11: Empirical regularities of entrepreneurial and routinized regime. <?page no="172"?> 4 Firms, competition and innovation 172 As shown in Figure 4.11, the patterns of the emergence of new knowledge and its repercussions on firms and firms‘ strategies found in empirical studies can be assigned to two different competitive environments for innovations called technological regimes: an entrepreneurial and a routinized regime, first proposed by Nelson and Winter (1982). The entrepreneurial regime describes conditions that promote the emergence and entry of new innovative firms into an industry. As a consequence, new knowledge is mainly created by innovative start-ups. Under the routinized regime, innovations are strongly path-dependent and essentially based on existing knowledge - this explains competitive advantage of incumbent firms based on firm-specific skills and experience (Audretsch 1991, Breschi et al. 2000 and Malerba 2007). Technological regimes can be identified on different levels such as products, processes, or business models, and define a set of rules or potential strategies based on technological know-how and novelty or differentness. The regime in which an industry finds itself depends essentially on whether creating new knowledge and grappling of emerging technological possibilities requires extensive accumulated knowledge and industry-specific experience, or whether this can also be done outside existing firms and even more without many years of industry-specific experience (Dosi, 1988, Marsili, 2002, Peneder, 2010 and Dosi and Nelson, 2010). Technological opportunities of an industry are determined by basic scientific research, basic innovations and technology development in other industries. The more extensive and the better identifiable technological opportunities are, the greater the incentives for firms to invest in R&D and vice versa. Cumulativeness of knowledge describes how strongly technological or processual knowledge builds on and mutually requires each other. The higher the cumulativeness and the more path-dependent the development, the higher the R&D investment and incentives. If knowledge cannot be accumulated or adopted in a large variety of applications, products, or services - i.e., has a high rate of obsolescence - little is invested in new knowledge. Appropriability conditions of an industry determine to what extent firms can obtain or collect benefits or advantages (especially profits) from new knowledge. The stronger and more sustainable the ability of firms to reap the benefits of innovation, the greater the incentives for enterprises to invest in R&D. This is influenced, among other things, by property rights for knowledge (patents, etc.), the intensity of competition between firms and the number of technological paths pursued or realized across firms in an industry. Different characteristics of these dimensions then allow a separation of technological regimes. An industry is in an entrepreneurial regime, if for the appropriation of profits from new knowledge and for the use of existing technological opportunities weakly accumulated knowledge is sufficient or no industry-specific experience is necessary. On the other hand, if a high level of accumulated firm-specific knowledge and corresponding industry-specific experience is necessary, an industry is in a routinized regime. Many industries are characterised by a transition from an entrepreneurial to a routinized regime, as shown in ► Figure 4.12, which coincides with the emergence of a dominant design and a peak in the number of firms. During an entrepreneurial regime, small and young firms follow very different technological paths. Many firms - indicated by dotted arrows - have a limited ability to survive and exit the industry after some time due to lack of profits. Some other <?page no="173"?> 4.3 Competition and innovation 173 firms - indicated by solid arrows - are able to build up industry-specific knowledge based on dynamic capabilities, which ensures their survival. At the same time, the variety of technological paths pursued within an industry is significantly reduced. This transition also means an emergence of barriers to entry in the shape of industry-specific knowledge and firm-specific capabilities. Figure 4.12: Survival of firms, reduction of technological trajectories, emergence of a dominant design and barriers to entry. During early stages of an industrial life cycle, young, small, and often science and researchbased start-ups are particularly dominant in the innovation process: they bring more and fundamental innovations to the market. With the transition to the routinized regime, incumbent firms with investment-intensive R&D departments become superior in the innovation process. During later stages of the industrial life cycle, these then incumbents either regularly create new products or processes, or are able to learn, absorb and transfer new knowledge generated in an industry very quickly and effectively into their own organization (Cohen and Levinthal 1990, Zahra et al. 2006 as well as Teece 2007). With the transition from the entrepreneurial regime to the routinized regime, competition between firms is also changing fundamentally along the following two important dimensions. Reduction of uncertainty and risk - for all firms and customers, uncertainty and risk are reduced with the transition to a routinized regime. Essential product features, future-proof technologies, the size of the market and truly relevant competitors become much more visible. The reduction in uncertainty also leads to improved financing conditions via the capital market and through banks, so that, taken together, continuous growth of viable firms is now possible. Onset of strategic behaviour - during the entrepreneurial regime, firms must take into account in particular the possibility of significant technological and/ or market changes and uncertainty. Individual competitors and their strategies play a subordinate role in this <?page no="174"?> 4 Firms, competition and innovation 174 phase, as their viability is not clearly evident, and strategies are subject to very rapid changes. The transition to a routinized regime makes strategic behaviour possible for the first time, since the survivable competitors can now be clearly identified, and their strategies can be analysed. Strategic investments in sunk cost activities (large production plants, marketing and branding, M&A, firm-specific capabilities, etc.) now create 'incumbent firms', which differ from start-ups due to these firm-specific capabilities, assets and investments and these incumbents now leverage competitive advantage. Evolution of knowledge over time Evolution of knowledge, learning about and exploitation of technological opportunities, and the diffusion of innovation takes place along an S-shaped course over time, as shown in ► Figure 4.13 on the left (Foster 1986, Christensen 1992 and Rogers 2010). For each technology, increasing followed by decreasing growth in knowledge or in the performance of a technology is observed, which successively exploit opportunities of a technology. The S-shaped pattern develops, because with every new technology, in an early stage fundamental experiments have to be carried out making only very small advances possible. As soon as uncertainty about opportunities of the new technology decreases and a critical mass of possible products or applications is identified, an industry-wide learning and development process is significantly accelerated by a massive increase in cumulative R&D investment by firms. Finally, a slowdown and reduction of the learning rate sets in, as opportunities of a technology are increasingly exhausted, e.g., due to physical laws. This growth pattern of knowledge can be found at a firmand at an industry-level and has been identified for many products and industries. Examples include load capacity of sailing ships, tensile load of diesel locomotives, rubber abrasion of car tyres, number of seats in passenger planes, storage space on hard disks, processing of cotton, resilience of plastics or weight reduction of racing bikes - but also for human learning of new skills. The rate of innovation (or learning rate) along a technology is influenced, among other things, by the R&D expenditure of firms, the intensity of competition, and the mutual adaptability of technological advances of other firms. This can also be seen in ► Figure 4.13, where an S-curve of industry-specific knowledge emerges alongside the transition from an entrepreneurial into a routinized regime. Figure 4.13: Technology as an S-shaped curve (left) and successive generations of technologies (right). <?page no="175"?> 4.3 Competition and innovation 175 ► Figure 4.13 on the right shows that sequences of technologies can often be observed over time. If technological opportunities are increasingly exhausted, firms invest more heavily in a subsequent generation of this technology and introduce it to the market as soon as its performance has reached at least that of the previous technology generation. This is well illustrated in case of processor technology for PCs: first a processor, e.g., an Intel 80486 in 1989, is introduced to the market in a basic configuration. After the introduction, the possibilities of this technology generation had been successively exhausted by 1994 by means of clock rates (from 16 MHz to up to 100 MHz), cache assembly on the chip and addressability by software. Previous and subsequent generations of Intel computer processors - 8088, 80186, 80286, 80386, 80486, P5, Celeron, i3, i5 and i7 - are going through the same pattern, as well as processor clones of competitors. Sustaining innovation versus disruptive innovation For a sequence of technologies, typically two patterns can be observed: sustaining innovation, which preserves and improves knowledge in a path-dependent way, and knowledge-destroying disruptive innovation (Bower and Christensen 1995 and Christensen and Bower 1996). Both cases are depicted in ► Figure 4.14. In case of knowledge-sustaining innovation and a transition from technology 1 to technology 2, which is characterised by a relatively long transition period and a small technological difference, incumbent firms are generally well able to adapt to the new technology. This is due to pathand experience-dependent knowledge about the previous technology, which makes it possible to understand and evaluate a new technology, reduces uncertainty regarding implementation, and makes the adaptation processes easy to design, especially for incumbent firms. Figure 4.14: Sustaining innovation and disruptive innovation. <?page no="176"?> 4 Firms, competition and innovation 176 A sequence of continuous, knowledge-preserving and knowledge-expanding innovations - over several years or even decades - then takes place within an industry-specific technological paradigm (Dosi 1982, Christensen and Rosenbloom 1995 and Castellacci 2008). In a routinized regime, this is characterised by strategies, business models, processes and products that are essentially understood and applied in a similar way by all firms in an industry. For example, firms in the financial services industry are so similar, that an employee quickly finds his or her way around when moving from BNP Paribas to Deutsche Bank or Banco Santander, because essentially the same things are done in the same way and similar strategic considerations play a role. An industry-specific paradigm summarises - in the sense of Kuhn (1962) - all routines, principles, standard answers, and knowledge used by firms to compete. Knowledge-preserving innovations stabilise an industry-specific paradigm, consolidate positions of incumbent firms and strengthen entry barriers to new firms. Disruptive innovation, on the other hand, destroys industry-specific paradigms, endangers the existence of incumbent firms and change market structures - in a process of creative destruction as originally described by Schumpeter (1942) - so that new firms entering the market regularly displace incumbent firms and replace their market leadership. Disruptive innovation and a transition from technology 2 to technology 3, which - as shown in ► Figure 4.14 on the right - is characterised by a relatively short transition period and a large technological difference; existing firms are often unable to adopt the new technology quickly enough or accurately. The reason for this is that pathand experience-dependent knowledge from the previous paradigm does not support, but rather complicates, the classification and evaluation of a new technology: this new technology differs fundamentally from the previous one in many dimensions and thus, uncertainty arises regarding implementation. Due to path-dependent perceptions, firms are literally blind to disruptive innovations, then hesitate to adapt and are subsequently forced out of the market by new competitors. Disruptive innovations create a new industry-specific paradigm: previous routines, basic principles and existing knowledge become largely obsolete and destroyed, so that incumbent firms lose their resource-based competitive advantages. Mail-order firm Quelle in Germany, for example, did not recognize the importance of e-commerce and the threat of Amazon; encyclopaedia publishers such as Brockhaus or Britannica did not anticipate possibilities of collaborative article creation of Wikipedia; telecommunications providers have lost much of their SMS revenues due to free services like WhatsApp and film manufacturers like Kodak - who actually co-invented digital photography - have completely misjudged possibilities and speed of development of a new technology that has led to new cameras, new customer behaviour and new business models and products. Three mutually reinforcing causes below are regularly identified as factors for the failure of incumbent firms (Klenner 2011, Kelly and Amburgey 1991, Christensen and Bower 1996 as well as Hill and Rothaermel 2003). A general inertia of incumbent firms in change and adaptation processes, i.e., insufficiently developed dynamic capabilities, which are reflected in a lack of flexibility and adaptability to technological changes. Path-dependent investments into a previous technology or market position that make a change of technology unattractive from a financial perspective because the existing business is (still) profitable. In addition, path-dependent investments have a strong sunk cost <?page no="177"?> 4.3 Competition and innovation 177 character, so that depreciations on capital or goodwill of the existing technology endanger the survival of firms. A strategic misjudgement of the rapidly changing market and environmental conditions, as well as uncertainty due to shifting product performance characteristics and emerging competitive advantage for new firms and start-ups. Characteristics of disruptive innovations are similar across many industries and markets. In essence, initially a newly emerging product or business model is not a threat to incumbent firms due to limited functionality - growth is limited and realized only with a few customers in niche markets. However, products based on disruptive innovations are often easier to use and focus on features that were previously unavailable. For example, although the first digital cameras were very heavy and unwieldy, maximum resolution and maximum number of images remained well below analogue cameras - digital cameras have, on the other hand, led to a competitive advantage in the niche market of sports photography (Klenner 2011 and Klenner et al 2013.). However, incumbent firms cannot respond, if the performance parameters of a disruptive competitor's product are continuously improving - an analogue camera cannot store 2,000 images, cost per photo cannot be reduced to zero, it is impossible to create a photo book and it is impossible to upload or share images via social media. As a result, the entire value-chain of analogue photography has been destroyed and replaced by a new digital ecosystem - numerous firms (film manufacturers, camera manufacturers, film developers, photo retailers and distributors) have lost their source of revenue and had been forced out of the market in this process of creative destruction. Classification of innovations Competitive effects of innovation can be described and forecasted based on the transilience map, as shown in ► Figure 4.15 on the left (Abernathy and Clark 1985, Clark 1985 and Henderson and Clark 1990). Here, innovation dynamics are mapped into four possible areas based on their knowledge-preserving and/ or knowledge-destroying effects on customer/ market relations and technology as follows. Path-dependent (or regular) innovations strengthen existing customer relationships based on existing technology. Essentially, incremental changes improve performance characteristics of products or production costs are reduced. Market launches of new product generations in the automobile industry fit into this category. Path-dependent innovations are knowledge-preserving and stabilise the competitive behaviour of firms under the routinized regime and reinforce the role of incumbents - accumulated knowledge of these firms increases. If existing technology is applied to a new market or market segment, niche creation occurs. This is especially true when a firm uses existing technology to address customer segments of a competitor. In this way, Sony has gained large market shares in the 1980s with the Walkman. Similarly, the strategies of car manufacturers to take market share from competitors with SUVs are based entirely on existing technology but address new market segments. If an existing customer relationship is stabilised or expanded, but a completely new technology is used, a revolutionary innovation happens. This category includes the strategies of established banks to transform their branch business model with existing customers to- <?page no="178"?> 4 Firms, competition and innovation 178 wards online banking or of monitor manufacturers in the transition to high-definition formats and technologies. Revolutionary innovations are put forward by incumbents with strong R&D departments, some of which are making their own knowledge base obsolete. At the same time, this reduces barriers to entry, making the technological paradigm and, indirectly, incumbent firms vulnerable. If both the existing market relationship and the existing technology path are destroyed, disruptive innovation (also known as architectural innovation) occurs. This either creates new markets or fundamentally changes existing markets, basic patterns of strategic behaviour and dimensions of competitive advantage, and threatens the existence of incumbent firms. Amazon has implemented a multi-dimensional disruptive innovation according to this definition: the technology dimension in retail has been renewed through e-commerce, the combination of a proliferation and long-tail strategy (expansion into all product categories and occupation of all niches in each category) has destroyed the customer relationships of existing competitors, and by opening the platform to third party suppliers as a multisided market Amazon even delivered a strategic innovation. Figure 4.15: Disruptive innovation and technological regime. <?page no="179"?> 4.4 Summary and key learnings 179 The transilience map is applied in strategy development or investment/ portfolio planning of firms to determine the positioning of new products or to classify innovative activities of competitors. Furthermore, this definition of disruption - a simultaneous destruction of technology and customer relations - allows conclusions about the dynamics of technological regimes, as shown in ► Figure 4.15 on the right. Over time, an industry develops from the entrepreneurial regime to the routinized regime. The main reason is, that in order to seize technological opportunities and to appropriate profits, a low level of accumulated knowledge is initially sufficient. However, with the emergence of the dominant design, more and more industry-specific accumulated knowledge becomes necessary, which comes from a combination of experience and R&D investments. If a disruptive innovation destroys industry-specific knowledge, advantages of incumbent firms are lost, and entry barriers become drastically less important. An industry is then thrown back from a routinized into an entrepreneurial regime and young new firms can now enter the industry in large numbers and try to gain market share. This situation can be observed in the financial services industry since 2010 (Münter and Weisser 2016). Whether fintechs - young technology-driven firms that establish virtual products and digital business models - will really be able to survive in the long run and gain significant market share depends largely on whether and how quickly incumbent banks can learn and adapt to changing environmental conditions based on dynamic capabilities. 4.4 Summary and key learnings Why do firms exist? How would markets look like without organisations and firms? Firms are based on knowledge that is superior to pure or spontaneous market solutions. The ability of firms to survive in competition depends on whether competitive advantage can be transformed into profits that are sufficient in relation to expectations of shareholders. However, the often claimed hypothesis of profit maximisation as such is questionable from a behavioural, strategic and corporate governance perspective. Making profits is essential, maximising profits is the exception rather than the rule. From a managerial view, it is important to understand main competitive advantages and profit drivers of a firm. Competitive advantage of firms can be explained from two complementary perspectives: the market-based view emphasises the strategic positioning of a firm in the right market, the resource-based view emphasises the application and development of appropriate capabilities. Hence, profits of firms are determined by a combination of firmand industryspecific factors. Whether and to what extent competitive advantage can be developed and applied depends to a large extent on the interaction of competition and innovation over time in different technological regimes. An industry life cycle is characterised in early stages by an entrepreneurial regime in which young and small start-ups have advantages in the innovation process. If a dominant design emerges and some firms succeed in augmenting industry-specific knowledge <?page no="180"?> 4 Firms, competition and innovation 180 as an entry barrier, the competitive process switches to a routinized regime. Through knowledge-preserving sustaining innovations a stable technological paradigm can emerge. Through knowledge-destroying disruptive innovations - corresponding to the idea of creative destruction argued by Schumpeter - incumbent firms could be forced out of the market and market structures can be completely changed. Against this background, from a management perspective, it is necessary to assess and understand the technological regime and paradigm in which a firm operates, and to evaluate the risk of knowledge-destroying innovations. Recommendations for further reading Firms, competitive advantage and innovation are a vast subject - for a comprehensive microeconomic insight, see Belleflamme, P. and Peitz, M., Industrial organisation: markets and strategies, London 2015. Perloff, J.M. and Brander, J.A., Managerial economics and strategy, London 2017, use less mathematics and provide a bit more strategic management attitude. If you do not need any formulas at all, Johnson, G., Whittington, R., Scholes, K., Angwin, D. and Regnér, P., Exploring Strategy, Harlow 2017, or Grant, R., Contemporary strategy analysis, Hoboken 2019, are good choices for competitive and corporate strategy following a strategic management approach. Questions for review [1] Describe applications of the analysis of competition, the existence of firms and market structure as well as their limits, advantages and disadvantages. [2] Explain from different perspectives why firms exist. [3] Describe basic considerations of setting up a firm. [4] Explain the differences and similarities between the market-based view and the resourcebased view to explain competitive advantages and the development of strategies. [5] Explain essential statements of the principal-agent theory. [6] Explain differences between new and incumbent firms. [7] Explain the notion of disruptive innovation. Which two conditions must be met for this to happen? Explain examples of three industries in which disruptive innovation has displaced incumbent firms. [8] Explain the importance of entry barriers for the profitability of firm in an industry. Explain three possible forms of entry barriers. [9] How does the number of firms in an industry typically develop in the long-term? What are possible explanations for this pattern? [10] Explain the relationship between market structure, strategy and market outcome. [11] Explain the relationship between strategy, competitive advantage and profits. [12] Describe the influence on competitive strategy in the transition from an entrepreneurial to a routinized regime. <?page no="181"?> 4.4 Summary and key learnings 181 Literature Abernathy, W.J. and Clark, K.B., Innovation: mapping the winds of creative destruction, Research Policy, 1985, 14, 3-22. Acs, Z.J. and Audretsch, D.B., Handbook of entrepreneurship research: an interdisciplinary survey and introduction, Vol. 1, Berlin 2006. Ajzen, I., The theory of planned behavior, Organizational Behavior and Human Decision Processes, 1991, 50, 2, 179- 211. Alchian, A.A. and Demsetz, H., Production, information costs, and economic organization, American Economic Review, 1972, 62, 5, 777-795. Alchian, A.A., Uncertainty, evolution, and economic theory, Journal of Political Economy, 1950, 58, 3, 211-221. Alvarez, S. and Parker, S.C., Emerging firms and the allocation of control rights, Academy of Management Review, 2009, 34, 2, 209-227. Arenius, P. and Minniti, M., Perceptual variables and nascent entrepreneurship, Small business economics, 2005, 24, 3, 233-247. Audretsch, D.B., Innovation and industry evolution, Cambridge-London 1995. Audretsch, D.B., New-firm survival and the technological regime, Review of Economics and Statistics, 1991, 73, 3, 441-450. Bain, J.S., Barriers to new competition: their character and consequences in manufacturing, Boston 1956. Barney, J.B., Types of competition and the theory of strategy: toward an integrative framework, Academy of Management Review, 1986, 11, 4, 791-800. Baysinger, B. and Hoskisson, R.E., The composition of boards of directors and strategic control: effects on corporate strategy, Academy of Management Review, 1990, 15, 1, 72-87. Block, J.H. and Wagner, M., Necessity and opportunity entrepreneurs in Germany: characteristics and earning s differentials, Schmalenbach Business Review, 2010, 62, 2, 154-174. Bloom, N., Brynjolfsson, E., Foster, L., Jarmin, R., Patnaik, M., SAPorta-Eksten, I. and Van Reenen, J., What drives differences in management practices? , American Economic Review, 2019, 109, 5, 1648-1683. Böbel, I., Wettbewerb und Industriestruktur, Berlin-Heidelberg-New York-Tokyo 1984. Bower, J.L. and Christensen, C.M., Disruptive technologies: catching the wave, Harvard Business Review, 1995, 73, 1, 43- 53. Bowman, E.H. and Helfat, C.E., Does corporate strategy matter? , Strategic Management Journal, 2001, 22, 1, 1-23. Bradley, C., Dawson, A. and Smit, S., The strategic yardstick you can’t afford to ignore, McKinsey Quarterly, October 2013, 2, 23-39. Breschi, S., Malerba, F. and Orsenigo, L., Technological regimes and Schumpeterian patterns of innovation, Economic Journal, 2000, 110, 463, 388-410. Brynjolfsson, E. and McAffee, A., The second machine age: work, progress, and prosperity in a time of brilliant technologies, New York-London 2016. Buchanan, B., Cao, C.X. and Chen, C., Corporate social responsibility, firm value, and influential institutional ownership, Journal of Corporate Finance, 2018, 52, 73-95. Bundesverband deutscher Fluggesellschaften, Rentabilität einer Airline spielt sich im Cent-Bereich ab, Pressemeldung 12. Juni 2013. Castellacci, F., Technological paradigms, regimes and trajectories: manufacturing and service industries in a new taxonomy of sectoral patterns of innovation, Research Policy, 2008, 37, 6, 978-994. <?page no="182"?> 4 Firms, competition and innovation 182 Caves, R.E. and Porter, M.E., From entry barriers to mobility barriers: conjectural decisions and contrived deterrence to new competition, Quarterly Journal of Economics, 1977, 91, 241-261. Cecere, G., CorRocher, N. and Battaglia, R.D., Innovation and competition in the smartphone industry: is there a dominant design? , Telecommunications Policy, 2015, 39, 3, 162-175. Chen, W. and Kamal, F., The impact of information and communication technology adoption on multinational firm boundary decisions, Journal of International Business Studies, 2016, 47, 5, 563-576. Christensen, C.M. and Bower, J.L., Customer power, strategic investment, and the failure of leading firms, Strategic Management Journal, 1996, 17, 3, 197-218. Christensen, C.M. and Rosenbloom, R.S., Explaining the attacker's advantage: technological paradigms, organizational dynamics, and the value network, Research Policy, 1995, 24, 2, 233-257. Christensen, C.M., Exploring the limits of the technology S-curve - part I: component technologies, Production and Operations Management, 1992, 1, 4, 334-357. Christensen, C.M., Suarez, F.F. and Utterback, J.M., Strategies for survival in fast changing industries, Management Science, 2001, 44, 12, 207-220. Christensen, C.M., The innovator's dilemma: when new technologies cause great firms to fail, Boston 1997. Clark, K., The interaction of design hierarchies and market concepts in technological evolution, Research Policy, 1985, 14, 5, 235-251. Coase, R.H., The nature of the firm, Economica, 1937, 4, 16, 386-405. Cohen, W.M. and Levinthal, D.A., Absorptive capacity: A new perspective on learning and innovation, Administrative Science Quarterly, 1990, 35, 1, 128-152. Cubbin, J., Is it better to be a weak firm in a strong industry or a strong firm in a weak industry? , London Business School, Centre for Business Strategy, Working Paper No.49, London 1988. Donaldson, L., The ethereal hand: organizational economics and management theory, Academy of Management Review, 1990, 15, 3, 369-381. Dosi, G. and Nelson, R.R., Technological paradigms and technological trajectories, in: Augier, M. and Teece, D.J. (eds.), The Palgrave Encyclopedia of Strategic Management, London 2016, 1-12. Dosi, G., Faillo, M. and Marengo, L., Organizational capabilities, patterns of knowledge accumulation and governance structures in business firms - an introduction, Laboratory of Economics and Management (LEM), Sant'Anna School of Advanced Studies, Pisa, 2011. Dosi, G., Sources, procedures, and microeconomic effects of innovation, Journal of Economic Literature, 1988, 26, 3, 1120-1171. Dosi, G., Technological paradigms and technological trajectories: a suggested interpretation of the determinants and directions of technical change, Research Policy, 1982, 11, 3, 147-162. Economist, iProfit, 10. Februar 2011. Eisenhardt, K.M. and Martin, J.A., Dynamic capabilities: what are they? , Strategic Management Journal, 2000, 21, 10- 11, 1105-1121. Eisenhardt, K.M., Agency theory: an assessment and review, Academy of Management Review, 1989, 14, 1, 57-74. Fama, E.F., Agency problems and the theory of the firm, Journal of Political Economy, 1980, 88, 2, 288-307. Foss, N.J., The strategic management and transaction cost nexus: past debates, central questions, and future research possibilities, Strategic Organization, 2003, 1, 2, 139-169. Foster, R.N., Working the S-curve: assessing technological threats, Research Management, 1986, 29, 4, 17-20. <?page no="183"?> 4.4 Summary and key learnings 183 Geroski, P.A., Gilbert, R.J. and Jacquemin, A., Barriers to new competition and strategic competition, Fundamentals of Pure and Applied Economics: theory of the firm and industrial organization, Chur 1990. Geroski, P.A., Innovation as an engine of competition, in: Mueller, D.C., Haid, A. and Weigand, J. (eds.), Competition, Efficiency and Welfare, Dordrecht 1998, 13-26. Geroski, P.A., Innovation, technological opportunity, and market structure, Oxford Economic Papers, 1990, 42, 3, 586-602. Grant, R.M., Toward a knowledge-based theory of the firm, Strategic Management Journal, 1996, 17, 109-122. Hart, O. and Holmstrom, B., A theory of firm scope, Quarterly Journal of Economics, 2010, 125, 2, 483-513. Hart, O., Corporate governance: some theory and implications, Economic Journal, 1995, 105, 430, 678-689. Hayek, F.A., Der Wettbewerb als Entdeckungsverfahren, Kiel 1968. Hayward, M.L., Shepherd, D.A. and Griffin, D., A hubris theory of entrepreneurship, Management Science, 2006, 52, 2, 160-172. Henderson, R.M. and Clark, K.B., Architectural innovation: the reconfiguration of existing product technologies and the failure of established firms, Administrative Science Quarterly, 1990, 35, 9-30. Henneke, D., Dynamische Fähigkeiten und Unternehmenserfolg, Bern 2014. Hill, C.W. and Rothaermel, F.T., The performance of incumbent firms in the face of radical technological innovation, Academy of Management Review, 2003, 28, 2, 257-274. Hillman, A.J. and Dalziel, T., Boards of directors and firm performance: integrating agency and resource dependence perspectives, Academy of Management Review, 2003, 28, 3, 383-396. Hillman, A.J. and Keim, G.D., Shareholder value, stakeholder management, and social issues: What's the bottom line? , Strategic Management Journal, 2001, 22, 2, 125-139. Hutzschenreuter, T., Das Unternehmen geht vor, FAZ, 18. November 2019, 16. IATA, Profitability and the air transport value chain - an analysis of investor returns within the airline industry and its supply chain, IATA Economics Briefing, 10, 2013. Jensen, M. and Meckling, W., Theory of the firm - managerial behavior, agency costs, and ownership structure, Journal of Financial Economics, 1976, 3, 4, 305-360. Jovanovic, B. and MacDonald, G.M., The life cycle of a competitive industry, Journal of Political Economy, 1994, 102, 322-347. Jovanovic, B., Selection and the evolution of industry, Econometrica, 1982, 50, 3, 649-670. Kanbur, S.M., Risk taking and taxation: an alternative perspective, Journal of Public Economics, 1981, 15, 2, 163-184. Kelly, D. and Amburgey, T.L., Organizational inertia and momentum: a dynamic model of strategic change, Academy of Management Journal, 1991, 34, 3, 591-612. King, A.W., Disentangling inter-firm and intra-firm causal ambiguity: a conceptual model of causal ambiguity and sustainable competitive advantage, Academy of Management Review, 2007, 32, 1, 156-178. Kirzner, I., Competition and entrepreneurship, Chicago 1973. Kirzner, I.M., Competition and entrepreneurship, Chicago 1973. Klenner, P., Die Bestimmung der Disruptionsreife von Märkten, Hamborn 2011. Klenner, P., Hüsig, S. and Dowling, M., Ex-ante evaluation of disruptive susceptibility in established value networks - when are markets ready for disruptive innovations? , Research Policy, 2013, 42, 914-927. Klepper, S. and Graddy, E., The evolution of new industries and the determinants of market structure, Rand Journal of Economics, 1990, 21, 27-44. Klepper, S., Industry life cycles, Industrial and Corporate Change, 1997, 6, 145-182. <?page no="184"?> 4 Firms, competition and innovation 184 Knight, F.H., Risk, uncertainty and profit, New York 1921. KPMG, Cost of capital study 2016 - value measurement, quo vadis? , Frankfurt 2016. Kuhn, T.S., The structure of scientific revolutions, Chicago 1962. Lazear, E.P., Balanced skills and entrepreneurship, American Economic Review, 2004, 94, 2, 208-211. Letza, S., Sun, X. and Kirkbride, J., Shareholding versus stakeholding: A critical review of corporate governance, Corporate Governance, 2005, 12, 3, 242-262. Lippman, S.A. and Rumelt, R.P., Uncertain imitability: an analysis of inter-firm differences in efficiency under competition, Bell Journal of Economics, 1982, 13, 418-438. Lucas, R.E., On the size distribution of business firms, Bell Journal of Economics, 1978, 9, 508-523. Malerba, F. and Orsenigo, L., Technological regimes and firm behavior, Industrial and Corporate Change, 1993, 2, 45-74. Malerba, F. and Orsenigo, L., The dynamics and evolution of industries, Industrial and Corporate Change, 1996, 5, 51-87. Malerba, F., Innovation and the dynamics and evolution of industries: progress and challenges, International Journal of Industrial Organization, 2007, 25, 675-699. Malhotra, A. and Gupta, A.K., An investigation of firms’ strategic responses to industry convergence, Academy of Management Proceedings, 2001, 1, 1-6. Mason, E.S., Price and production policies of large-scale enterprise, American Economic Review, 1939, 29, 61-74. McGahan, A.M. and Porter, M.E., How much does industry matter, really? , Strategic Management Journal, 1999, 18, 3, 15-30. McGahan, A.M. and Porter, M.E., The emergence and sustainability of abnormal profits, Strategic Organization, 2003, 1, 1, 79-108. Monopolkommission, Hauptgutachten XXIII: Wettbewerb 2020, Bonn 2020. Mueller, D.C., The dynamics of company profits, Cambridge 1990. Münter, M.T. and Weisser, N., Digital Leadership in der Finanzdienstleisterindustrie - werden etablierte Banken oder FinTechs die digitale Zukunft gestalten? , IM+io - Magazin für Innovation, Organisation und Management, 2016, 4, 106-111. Münter, M.T., Entry is easy, survival is not - some insights concerning the emergence of incumbent firms in worldwide automobile industry, Working Paper, May 2013. Münter, M.T., Gründerpersönlichkeit und Unternehmensgründung, in: Betzold, R., Emrich, E., Gassmann, F., Heidenreich, S., Jordanov, S., Koch, M., Münter, M.T. and Nguyen, Q. (eds.), Unternehmensgründungen im Umfeld saarländischer Hochschulen - Empirische Ergebnisse und regionalökonomische Effekte, 2021, 111-140. Münter, M.T., Wettbewerb und die Evolution von Industrien, Bayreuth 1999. Nelson, R.R. and Winter, S.G., An evolutionary theory of economic change, Cambridge-London 1982. Nelson, R.R., Why do firms differ, and how does it matter? , Strategic Management Journal, 1991, 12, 2, 61-74. Nicolaou, N. and Shane, S., Entrepreneurship and occupational choice: genetic and environmental influences, Journal of Economic Behavior & Organization, 2010, 76, 3-14. Parker, S.C., Entrepreneurship as occupational choice, in: Minniti, M. (ed.), Entrepreneurship - the engine of growth, 2007, 1, 81-100. Patel, P. and Pavitt, K., Uneven and divergent technological accumulation among advanced countries: evidence and a framework for generalisation, Industrial and Corporate Change, 1994, 3, 759-787. Penrose, E.T., The theory of the growth of firms, Oxford 1959. <?page no="185"?> 4.4 Summary and key learnings 185 Peteraf, M.A., The cornerstones of competitive advantage: a resource-based view, Strategic Management Journal, 1993, 14, 3, 179-191. Porter, M.E. and Kramer, M.R., Creating shared value, Harvard Business Review, 2011, 89, 1-2, 62-77. Porter, M.E., Competitive strategy: techniques for analyzing industries and competitors, New York 1980. Porter, M.E., Competitive advantage: creating and sustaining superior performance, New York 1985. Porter, M.E., The contributions of industrial organization to strategic management, Academy of Management Review, 1981, 6, 4, 609-620. Porter, M.E., What is strategy? , Harvard Business Review, 1996, 74, 6, 71-78. Prahalad, C. and Hamel, G., The core competences of the corporation, Harvard Business Review, 1990, 68, 3, 79-90. Quinn, J.B. and Hilmer, F.G., Strategic outsourcing, Sloan Management Review, 1994, 35, 43-55. Rauch, A. and Frese, M., Born to be an Entrepreneur? Revisiting the personality approach to entrepreneurship, in: Baum, J.R. et al. (eds.), The Psychology of Entrepreneurship, New York 2012, 41-65. Reed, R. and DeFillippi, R.J., Causal ambiguity, barriers to imitation, and sustainable competitive advantage, Academy of Management Review, 1990, 15, 1, 88-102. Rogers, E.M., Diffusion of innovations, New York 2010. Rosenbaum, D.I. and Lamort, F., Entry, barriers, exit, and sunk costs: an analysis, Applied Economics, 1992, 24, 297-304. Roy, R. and Sarkar, M.B., Knowledge, firm boundaries, and innovation: mitigating the incumbent's curse during radical technological change, Strategic Management Journal, 2016, 37, 5, 835-854. Rumelt, R.P. and Lamb, R., Towards a strategic theory of the firm, in: Lamb, R., Resources, firms, and strategies: a reader in the resource-based perspective, Upper Saddler River 1997, 132-145. Rumelt, R.P., How much does industry matter? , Strategic Management Journal, 1991, 12, 3, 167-185. Rumelt, R.P., Schendel, D. and Teece, D.J., Strategic management and economics, Strategic Management Journal, 1991, 12, 2, 5-29. Rumelt, R.P., The perils of bad strategy, McKinsey Quarterly, 2011, 1, 3, 30-39. Schmalensee, R., Do markets differ much? , American Economic Review, 1985, 75, 3, 341-351. Schumpeter, J.A., Capitalism, socialism and democracy, London 1942. Schumpeter, J.A., Theorie der wirtschaftlichen Entwicklung - eine Untersuchung über Unternehmergewinn, Kapital, Kredit, Zins und den Konjunkturzyklus, Berlin 1911. Simoes, N., Crespo, N. and Moreira, S.B., Individual determinants of self-employment entry: what do we really know? , Journal of Economic Surveys, 2016, 30, 4, 783-806. Sleuwaegen, L. and Dehandschutter, W., The critical choice between the concentration ratio and the Herfindahl-index in assessing industry performance, Journal of Industrial Economics, 1986, 35, 193-208. Sutton, J., Sunk costs and market structure: price competition, advertising, and the evolution of concentration, Cambridge-London 1991. Teece, D.J. Explicating dynamic capabilities: the nature and microfoundations of (sustainable) enterprise performance, Strategic Management Journal, 2007, 28, 13, 1319-1350. Teece, D.J., Pisano, G. and Shuen, A., Dynamic capabilities and strategic management, Strategic Management Journal, 1997, 18, 7, 509-533. Utterback, J.M. and Abernathy, W.J., A dynamic model of process and product innovation, Omega: International Journal of Management Science, 1975, 3, 639-655. Utterback, J.M. and Suarez, F.F., Innovation, competition, and industry structure, Research Policy, 1993, 22, 1-21. <?page no="186"?> 4 Firms, competition and innovation 186 Utterback, J.M. and Suarez, F.F., Patterns of industrial evolution, dominant designs, and firms’ survival, Research on Technological Innovation, Management and Policy, 1993, 5, 47-87. Voigt, S., Institutional economics, Cambridge 2019. Wagner, F.W., Wozu ist die Firma gut? , FAZ, 29. Dezember 2019, 22. Wernerfelt, B., A resource-based view of the firm, Strategic Management Journal, 1984, 5, 2, 171-180. Williamson, O.E., Markets and hierarchies - analysis and antitrust implications: a study in the economics of internal organization, New York 1975. Winter, S., Schumpeterian competition in alternative technological regimes, Journal of Economic Behavior and Organization, 1984, 5, 287-320. Zahra, S.A., Sapienza, H.J. and Davidsson, P., Entrepreneurship and dynamic capabilities: a review, model and research agenda, Journal of Management Studies, 2006, 43, 4, 917-955. <?page no="187"?> 187 5 Firm size, technology and decisions on production When strategic or financial investors are considering an investment in a new business model or start-up (whether in real life or formats such as Dragon’s Den in the UK, Shark Tank in the US or Höhle der Löwen in Germany) a crucial question is always whether the firm can be scaled. In other words: are high growth rates possible without having to hire large numbers of additional employees and arranging additional financial resources? Scaling depends on the demand for the product or service and essentially on whether growth dynamics require less than, or more than, a proportional input of resources in order to achieve critical mass or significant market share. If the growth rate of a firm’s output exceeds the growth rate of the number of employees and capital inputs, then there are increasing returns to scale ("scaling the business model"). Assuming all things are otherwise equal in the business environment, profits can be reached easier or faster. However, decisions on growth rate or size of firms and their production capacity are also among the most important strategic decisions in incumbent firms. Whenever a firm opens a new location, makes shortor long-term capacity adjustments due to changes in the demand structure, or merges plants of different locations, size of the firm is adjusted. This translates into a change of main input factors labour and capital, i.e., number and qualification of employees as well as structure and scope of infrastructure, IT hardware and software, machinery and investment. To illustrate the importance and significance of such decisions: in 2014, Volkswagen had around 100 production sites worldwide with more than 500,000 employees. Of these sites, 20 were in China with a total of 89,000 employees, a maximum production capacity of 3.3 million vehicles and total equity and debt invested of around EUR 36 billion. For the period 2014 to 2018, Volkswagen announced an investment of approximately EUR 18.2 billion at Chinese production sites (Manager Magazin 2013). Obviously, such an enormous investment represents both business opportunities and reflects requirements of the Chinese government. In many industries, knowledge concerning possible manufacturing processes in production and organisation of a firm changes over time: through innovations and continuous learning and improvement processes. Technological developments implicate, even with constant production volumes, a change in the interplay between qualification and number of employees, production processes and type and extent of capital employed. Of course, these developments are not exogenously predetermined: empirical studies show that an adaptation of process innovations and the productivity of a firm are significantly influenced by heterogeneous management skills (Bloom and van Reenen 2007). <?page no="188"?> 5 Firm size, technology and decisions on production 188 Case Study │ Productivity development at Audi ► Table 5.1 highlights a couple of these developments and sheds some light on the underlying decisions for the car manufacturer Audi. A comparison of the years 1925 and 2015 shows that the absolute size of the firm (measured by production volume of passenger cars) has increased from 1,200 vehicles to 1.8 million vehicles per annum. This equals an average annual growth rate (so called compound annual growth rate - CAGR) of 8.5%, regardless of increased quality of the vehicles. By contrast, the number of jobs in manufacturing (full time equivalents - FTE) grew at a CAGR of only 3.3%. This translates into an increase in productivity of more than 5% on average each year: from approximately 0.5 vehicles per employee in 1925 to about 42 vehicles per employee in 2015. This was accompanied by a significant increase in jobs outside manufacturing. In 2015 Audi had more than 34,000 jobs in administrative and business areas such as human resources, controlling, marketing, sales, IT and corporate management. Output, employees and productivity at Audi year 1925 2015 CAGR output (cars) approx. 1,200 approx. 1,800,000 ~ 8.5 % FTE in production approx. 2,400 (~ 91 %) approx. 43,000 (~ 56 %) ~ 3.3 % productivity per employee approx. 0.5 Autos p.a. approx. 42 Autos p.a. > 5 % FTE outside production approx. 200 (~ 9 %) approx. 34,000 (~ 44 %) Table 5.1: Technology, organisation and manufacturing (Data source: audi.com/ corporate/ de/ investor-relations.html and Audi archive 2016, own calculations). Drivers for an increase in productivity include innovations that enable new production processes or an improved interaction between human labour and machines (Belitz et al. 2017 on an analysis of productivity and growth drivers in German industries). There are indications that this development will continue during the so-called fourth industrial revolution in the coming decades (Bloem et al. 2014, Bughin 2017 et al., van Tunzelmann 2003 and McGowan et al. 2015). Digitalisation, internet of things (the connectivity and communication of products, services and machines), big data and artificial intelligence, among others, will enable firms to develop and implement new technologies to further improve manufacturing processes and make organisations more efficient. This will go along with changes in firm organisation, decision-making routines, employee qualification profiles and the interaction between human and machine. Frey and Osborne (2017) describe that 47% of all current jobs in the US could be replaced by robots, big data, or intelligent software in the next 20 years. For Germany, Bonin et al. (2015) have determined that 42% of all jobs could be affected by digitalisation. Digitalisation, thus, has a strong influence on the interaction of humans (labour) and machines (capital) in production. The changes here will take both forms: on the one hand, existing human <?page no="189"?> 5.1 Production function and technology 189 activities will be substituted by technology, on the other hand, human work will be complemented by new digital solutions, so that new professional roles and profiles in particular will emerge. Learning Objectives This chapter deals with developing a basic understanding of how decisions concerning the interrelationship between production output and the input of capital and labour are made in a long run versus short run perspective depending on technology; the concept of diminishing marginal product, to understand how a firm can influence its level of production in the short run by changing labour input; and the concept of returns to scale, to show how a firm can derive long-term decisions on the size of business units or location issues or manage growth processes of new digital business models. 5.1 Production function and technology Firms differ in many dimensions: for example, firm size, growth rates, or technology employed are not only heterogeneous across different industries, but also within an industry. Moreover, competitors within an industry often differ significantly in productivity, measured by the amount of output per employee per hour or year. At least three intertwined drivers can be identified: active strategic decisions by management for a certain technology and firm size in order to achieve certain goals based on firm-specific capabilities; market and competition-related external factors such as absolute levels and seasonal or structural fluctuations in demand and number of competitors, which influence or even determine management decisions; and path-dependencies of previous firm-specific decisions, which in particular restrict the choice of technology over time and predetermine specific growth paths of a firm. Production generally describes the transformation process whereby a firm converts input factors (labour and capital) into outputs such as products, services, or business models aimed towards retail customers (B2C) or business customers (B2B). This process is also known as value creation. Production becomes possible by way of an organised combination of employees (i.e., the labour input factor measured by the number of jobs) and machinery, infrastructure, raw materials, supplies and stocks (i.e., the capital input factor measured by equity and debt capital). The level of production depends on the number of employees and machines employed and, in particular, the specific linkage between them. This linkage of capital and labour is called technology - in a broader sense, technology describes the business model, all processes, knowledge and organisation of a firm and its management. <?page no="190"?> 5 Firm size, technology and decisions on production 190 ► Figure 5.1 illustrates the logic of a firm's transformation process. Any production of services or products is made possible from heterogeneous input factors by combining capital and labour - this linkage of labour and capital is called a production function. Of course, the linkage of capital and labour is firm-specific and dynamic: each firm has specific capabilities, experience, firm-specific organisational capital and routines for combining labour and capital, and each firm learns over time (Black and Lynch 2005). Figure 5.1: Transformation process of a firm. Input factors capital and labour and shortvs. long-term decisions Every firm must obtain capital in an appropriate way as part of its corporate finance procedures. This may be in the form of equity from owners, or via capital markets in case of new shareholders, debt via loans from banks or issue of bonds. Capital 𝐸𝐸 refers to and includes all forms of equity and debt (on the asset side of the balance sheet also financed raw materials or supplies) in various qualities, for example in terms of financing structure or maturity. The costs (factor prices) for the use of capital include all economic costs, i.e., opportunity costs and depreciation. In very simplified terms, the cost of capital can be reduced to the respective costs of equity and debt, as described in ► Chapter 4: a firm's cost of capital is then equal to the weighted average cost of capital (WACC) multiplied by the amount of capital employed. Following the same logic, a firm must recruit employees from the labour market: labour 𝐿𝐿 refers to all forms of work (from board members to clerical staff) in various forms. The costs (factor prices) for the use of labour again include all economic costs, e.g., wages, salaries, social insurance contributions or qualification and training. Simplified, labour costs can be reduced to a wage rate 𝐻𝐻 per time unit. Costs of labour then correspond to a wage rate (hourly wage, annual salary, etc.) multiplied by the amount of labour employed. <?page no="191"?> 5.1 Production function and technology 191 By varying the input factors capital and labour, the production level can be adjusted. However, in the short run capital input can only be changed slightly or at very high cost: even the decommissioning of machinery does not reduce the capital input, as the machinery continues to be financed by equity or debt. On the other hand, labour input can be adjusted more easily in the short run, for example, by arranging overtime, short-time work, hiring or reducing temporary workers. So, in order to change the production volume in the short run, this can only be done via labour inputs. A short-term analysis, therefore, checks whether and to what extent output changes if labour input is changed slightly: i.e., how much does output increase if one additional employee is hired? The basic concept of the marginal product of labour is developed in ► Section 5.2. In the long run, however, all factors can be changed: output can be changed by changing input quantity of both input factors or by choosing a new technology. Long-term analysis, therefore, examines whether and how the output changes if all factors (capital, labour and, where appropriate, technology) are changed simultaneously: the basic concept of returns to scale is explained in ► Section 5.3. What this means in the short and long run naturally depends on differences across industries (e.g., consulting vs. banking vs. steel production) and can change over time. In addition, industries differ in their elasticity of substitution, i.e., how easily capital and labour can be substituted for each other and what combinations are generally possible. Only rarely are firms completely free in their possible combinations, as in the case of the decision between a fully automatic car wash vs. pure hand washing of cars. Production function and its empirical identification The functional relationship of a production function in general may be described as (5.1) 𝑞𝑞 = 𝑇𝑇(𝐸𝐸, 𝐿𝐿) , i.e., an output 𝑞𝑞 is determined by firm-specific technology 𝑇𝑇 (for example production/ manufacturing processes, know-how, organisation and management of a firm) as a function of capital input 𝐸𝐸 and labour input 𝐿𝐿 . A production function can - based on empirical data of a firm - take on different functional forms, which can be determined by statistical and econometric methods (e.g. regression analyses). One out of many possible depictions of a production function is the so-called Cobb-Douglas production function (5.2) 𝑞𝑞 = 𝐴𝐴 𝐸𝐸 𝛼𝛼 𝐿𝐿 𝛽𝛽 with 𝐴𝐴, 𝛼𝛼, 𝛽𝛽 > 0 . In a Cobb-Douglas production function, 𝐴𝐴 denotes technological efficiency, i.e., all capabilities of a firm, organisation, or processes - the larger the value of 𝐴𝐴 , the higher the production 𝑞𝑞 for given values of 𝐸𝐸 and 𝐿𝐿 . 𝛼𝛼 and 𝛽𝛽 denote partial production elasticities of labour and capital. These indicate a percentage growth of output given a 1% increase in labour or capital input (while the other input factor is kept constant): the higher the value of 𝛼𝛼 or 𝛽𝛽 , the stronger the effect on output. Any Cobb-Douglas production function is firm specific and changes over time: for example, different values for 𝐴𝐴 , 𝛼𝛼 and 𝛽𝛽 apply to each firm at any given time, yet the structure of the function is similar for firms within one industry. For example, firms with higher values of 𝐴𝐴 have a <?page no="192"?> 5 Firm size, technology and decisions on production 192 technological advantage - i.e., they can produce more than their competitors with the same number of employees and amount of capital. Cobb-Douglas production functions are empirically well grounded, and are therefore, regularly used to analyse or plan production decisions of firms. From a management perspective, it is crucial to understand how technological efficiency and partial production elasticities influence production output when decisions on capital and labour inputs are made. Technological efficiency 𝑨𝑨 of a firm depends largely on the capabilities of the firm (resource-based view), choice of manufacturing processes, as well as organisation and management of the firm. Possibilities to increase 𝐴𝐴 are, for example, the development of new or improvement of existing capabilities, optimisation of organisational structure and streamlining of hierarchies, better leadership/ management or training of employees, switch to just-in-time production, more powerful software in the call centre, or standardisation of processes. Partial production elasticity 𝜶𝜶 of capital depends on the interaction between employees and machines and measures the effect of an expansion of capital input: can an employee also operate several machines (higher capital input) and does output then increase? Will the use of higher quality and more capital-intensive machines, with the same number and qualification of employees, increase production? Partial production elasticity 𝜷𝜷 of labour depends on the interaction between machines and employees and measures the effect of an increase in the number of employees: could the number of employees on a production line be increased to increase production? Will sales increase if we increase the number of sales assistants in a branch? A production function is not a hypothetical construct: the production function of each firm - i.e., an empirical and functional relationship between the output 𝑞𝑞 of a firm and input factors capital 𝐸𝐸 and labour 𝐿𝐿 - can be determined by analysing firm-specific data from internal (controlling, finance, operations) or external (balance sheet, profit and loss statement, annual reports) sources. It describes the microeconomic logic and business model of a firm, abstracting from product categories, divisional designations, machine types or locations - but making it possible for management to make decisions and, in addition, to analyse competitors from publicly accessible balance sheet or P&L data. Econometric methods such as regression analyses are applied to identify a production function, in everyday business life often using Excel, SPSS, or SAP solutions. Alternatively, values e.g. for 𝛽𝛽 can also be determined indirectly via management interviews ("if the production volume doubles with an existing IT platform - what percentage increase of employees do you need then? ”). If values for 𝐴𝐴 , 𝛼𝛼 and 𝛽𝛽 of competitors are known, these can be used as benchmarks for optimisation initiatives in your own firm. Case Study │ Turbine manufacturer with adjustments of capital and labour ► Figure 5.2 shows anonymised data of a medium-sized turbine manufacturer. Apparently, the firm is growing strongly over the period 2004 to 2010, driven by an expansion of capital inputs with a somewhat fluctuating number of employees. <?page no="193"?> 5.1 Production function and technology 193 Empirical analysis of the data of the indexed firm production 𝑞𝑞 , the number of employees 𝐿𝐿 (quarterly average) and the capital input 𝐸𝐸 in million EUR (per quarter) using a regression analysis allows the production function to be determined as (5.3) 𝑞𝑞 = 4.12 𝐸𝐸 0.79 𝐿𝐿 0.50 and thus, identifying values of 𝐴𝐴 = 4.12, 𝛼𝛼 = 0.79 and 𝛽𝛽 = 0.50 . In this example, there are no significant technological changes (e.g., process innovations or improved management methods) in the short period considered, so that 𝐴𝐴 as a technological parameter is constant at a value of 4.12 . 𝛼𝛼 and 𝛽𝛽 are partial production elasticities of labour and capital - if capital input is increased (or decreased) by 1% , then production increases (or decreases) by 0.79% , if labour input is varied by 1% , then production varies by 0.50% . If current capital input is 𝐸𝐸 = 100 and number of employees equals 𝐿𝐿 = 200, then output is (5.4) 𝑞𝑞 = 4.12 ⋅ 100 0.79 ⋅ 200 0.50 = 2215.20 . Figure 5.2: Empirical analysis of firm-specific data to determine a production function. 2000 2200 2400 2600 2800 3000 3200 3400 3600 output q t 80,0 90,0 100,0 110,0 120,0 130,0 140,0 150,0 160,0 capital input K t 150,0 170,0 190,0 210,0 230,0 250,0 number of employees L t q = 245,38 + 20,237 K R² = 0,9377 2000 2200 2400 2600 2800 3000 3200 3400 3600 80 100 120 140 160 output q K q = - 293,5 + 14,23 L R² = 0,3567 2000 2200 2400 2600 2800 3000 3200 3400 3600 170 190 210 230 250 output q L 2000 2200 2400 2600 2800 3000 3200 3400 3600 output q q (estimate) t <?page no="194"?> 5 Firm size, technology and decisions on production 194 As can be seen in ► Figure 5.2, bottom left, this estimate of the production function provides a good approximation of actual production data. Based on this information, the effects of changes in input factors can now be analysed. If, for example, capital input 𝐸𝐸 is increased by 1% , the new output becomes (5.5) 𝑞𝑞 = 4.12 ⋅ 101 0.79 ⋅ 200 0.50 = 100.79 % ⋅ 2215.20 = 2232.67 . Production volume grows by 0.79% due to the partial production elasticity 𝛼𝛼 of capital of 0.79 . If, on the other hand, the firm loses two employees, i.e., labour input reduces by 1% , then production volume decreases by 0.50% due to the partial production elasticity of 𝛽𝛽 = 0.50 : (5.6) 𝑞𝑞 = 4.12 ⋅ 100 0.79 ⋅ 198 0.50 = 99.50 % ⋅ 2215.20 = 2204.12 . Efficiency and cost-minimizing combinations of input factors A production function describes efficient combinations of input factors capital and labour, i.e., a maximum possible output using a given technology. Equally, it means that a firm produces a planned output efficiently if factor inputs and total production costs are minimised. For a given technology and factor prices, the minimum necessary quantities of capital and labour can then be determined. However, most firms in real life do not produce efficiently in this sense. Often existing input factors are not used optimally, for example, skills and number of employees may not match the existing IT infrastructure. Additionally, input factors may not be fully leveraged, such as the qualifications of employees or possibilities of existing IT solutions. Efficiency and inefficiency can then be easily distinguished: if capital or labour can be reduced in a firm without reducing production output, then the allocation of input factors is currently inefficient. Thus, savings of input factors, e.g., through process optimisation, reduction of working capital, or reduction of excessive number of employees, and cost reductions are possible without reducing output. Empirical studies have shown that efficiency of firms is essentially driven by intense competition, effective competition policy, a strong influence of the capital market and good corporate governance. Conversely, closed off markets with trade barriers or extensive governmental intervention in markets and regulation of industries lead to inefficiency (Hay and Liu 1997, Holmes and Schmitz 2010 as well as Andries and Capraru 2014). 5.2 Short run decisions: diminishing marginal product and productivity In the short run, capital input of a firm cannot be changed or can only be changed slightly, so that output can only be influenced by changing labour input. Typically, the output increases with increasing labour input: initially at a more than proportionate rate, but then only less than proportionately, before it may even decrease. The reason for this lies in the technology used: only certain input ratios of labour and capital make sense or are feasible. From a management perspective, two concepts are central to analyse short run production decisions: <?page no="195"?> 5.2 Short run decisions: diminishing marginal product and productivity 195 Productivity 𝑨𝑨𝑨𝑨 (also called average product) describes the output 𝑞𝑞 per employee 𝐿𝐿 . Changes in productivity occur if the rate of change of the output deviates from the rate of change of the employees - for example, if the number of employees increases by 10% from 200 to 220. but output only increases by 5% from 1,000 to 1,050 units, productivity per employee decreases by −4.55% from 𝐴𝐴𝑃𝑃 1 = 𝑞𝑞/ 𝐿𝐿 = 1,000/ 200 = 5 units per employee to 𝐴𝐴𝑃𝑃 2 = 𝑞𝑞/ 𝐿𝐿 = 1050/ 220 = 4.77 . Marginal product MP describes the change of output ∆𝑞𝑞 with a marginal change in the number of employees ∆𝐿𝐿 while keeping capital 𝐸𝐸 of the firm constant, in the example above 𝑀𝑀𝑃𝑃 = ∆𝑞𝑞/ ∆𝐿𝐿 = (1050 − 1000)/ (220 − 200) = 2.5 - one additional employee, thus, increases the output by 2.5 units. ► Figure 5.3 shows the output 𝑞𝑞 in burgers per hour of a restaurant with a given capital endowment, i.e., in particular the equipment in the kitchen and the number of stoves, of a fast food restaurant. Initially, as the number of employees increases, the output rises more than proportionately - mainly due to division of labour - but due to increasingly cramped working conditions and a limited number of cookers, the increase in production with each employee initially reduces until finally the number of employees in the kitchen hampers production: the output decreases in absolute terms. Figure 5.3: Output and number of employees. <?page no="196"?> 5 Firm size, technology and decisions on production 196 Output with variable number of employees number of employees output productivity marginal product 𝐿𝐿 𝑞𝑞 𝐴𝐴𝑃𝑃 = 𝑞𝑞/ 𝐿𝐿 𝑀𝑀𝑃𝑃 = ∆𝑞𝑞/ ∆𝐿𝐿 0 0 8 increasing marginal product 1 8 8 16 2 24 12 30 3 54 18 26 diminishing marginal product 4 80 20 15 5 95 19 13 6 108 18 4 7 112 16 -8 8 104 13 -14 9 90 10 -20 10 70 7 Table 5.2: Productivity and marginal product given a short run change in labour input with fixed capital input. These effects can be analysed quantitatively using ► Table 5.2: in addition to the absolute quantity 𝑞𝑞 of burgers produced with an increasing number of employees 𝐿𝐿 , the productivity 𝑞𝑞/ 𝐿𝐿 and the marginal product ∆𝑞𝑞/ ∆𝐿𝐿 , i.e., the change in the output ∆𝑞𝑞 triggered by a change in the number of employees ∆𝐿𝐿 , are given. Productivity apparently increases up to the fourth employee (5.7) 𝐴𝐴𝑃𝑃 = 𝑞𝑞/ 𝐿𝐿 = 80/ 4 = 20. i.e., each of these four employees produces an average of 20 burgers per hour. Total production increases up to the seventh employee, but this increase is now lower, so that productivity of each employee drops to 𝐴𝐴𝑃𝑃 = 16 . The explanation for this lies in the marginal product: the marginal product is initially positive and increases - each additional employee enables a better <?page no="197"?> 5.2 Short run decisions: diminishing marginal product and productivity 197 division of the work steps, increases the total production more than proportionately and increases the productivity of his colleagues. The additional output with a small (marginal) increase in labour input (5.8) 𝑀𝑀𝑃𝑃 = ∆𝑞𝑞/ ∆𝐿𝐿 = (80 − 54)/ (4 − 3) = 26 illustrates that with a marginal increase in labour input from three to four employees, the output increases by 26 from 54 to 80 burgers per hour. However, the marginal product is not constant: with an increasing number of employees, the marginal product is initially still positive, but declining, and finally becomes negative - additional employees initially increase production less than proportionately, and then output even decreases in absolute terms. The marginal product is influenced among other things by: capacity limits - the more the existing capacity is utilised, the less additional employees can increase production, the lower the marginal product; management skills - the better the management and organisational skills in a firm, the higher the marginal product; technological capabilities - the more flexible the existing technology is, the better additional employees can be deployed, the higher is the marginal product. This observation applies in the short run in the same way to an individual branch of a fast food restaurant as it does to a global corporation - with a diminishing marginal product, an increase in the number of employees leads to a less than proportionate increase in output. ► Figure 5.4 shows typical empirical shapes of marginal product, output and productivity as a function of the number of employees. The marginal product corresponds to the increase in output with an increase in the number of employees - mathematically this is equivalent to the first derivative of the production function 𝑞𝑞 = 𝑇𝑇(𝐸𝐸, 𝐿𝐿) with respect to the number of employees 𝐿𝐿 . Three main areas can be distinguished, in ► Figure 5.4 on the right: 1: marginal product larger than productivity, i.e., 𝑀𝑀𝑃𝑃 > 𝐴𝐴𝑃𝑃 - productivity increases. Additional employees make a higher contribution to production than previous employees - as a result, productivity of all employees increases. 2: marginal product is positive, but smaller than productivity, i.e., 𝑀𝑀𝑃𝑃 < 𝐴𝐴𝑃𝑃 - productivity decreases. Additional employees make a smaller contribution to production than previous employees - as a result, productivity of all employees decreases. 3: marginal product is negative - productivity and output decrease. Additional employees reduce production and the productivity of all employees decreases. In order to increase output and improve productivity, the firm should either lay off employees or choose a different technology and capital endowment. Typically, one empirically observes production processes which show decreasing marginal products (hence often also called "law of diminishing marginal product"). <?page no="198"?> 5 Firm size, technology and decisions on production 198 Figure 5.4: Productivity and marginal product. From a management perspective, identifying the three areas in ► Figure 5.4 by comparing marginal product and productivity ( 𝑀𝑀𝑃𝑃 > 𝐴𝐴𝑃𝑃 in area 1 or 𝑀𝑀𝑃𝑃 < 𝐴𝐴𝑃𝑃 in areas 2 and 3) is relevant for controlling efficiency in production processes and to estimate what effects will arise if the number of employees is increased or reduced. Typical is control in area 2 - there are diminishing marginal products and declining productivity, but by increasing the number of employees 𝐿𝐿 , an attempt can be made to increase or even maximise output for a given technology and capital endowment. Decisions on resource allocation based on marginal products This is especially true when a manager has to assign employees across several factories or branches. ► Table 5.3 shows productivity indicators of two branches of a bank based on the number of transactions processed. If ten employees are to be assigned to both branches in order to maximise total output, the focus must be on marginal products: employees are assigned step by step to the branch with the highest marginal product depending on the number of employees already assigned. This means in this case that the first five employees are assigned to the East branch, because the marginal product from 108 in descending order to 92 is continuously higher compared to the West branch. The sixth employee is then, based on a marginal product of 92, the first to be assigned to the West branch. Employees seven to nine with marginal products of 88, 84 and 80 are assigned to the East branch again, the tenth employee is finally assigned to the West branch - in total, eight employees now work in the East branch, and two in the West branch. The total production of 𝑞𝑞 = 𝑞𝑞 𝑂𝑂 + 𝑞𝑞 𝜕𝜕 = 752 + 172 = 924 cannot be reached or exceeded by any other allocation of employees. It is important to note that the productivity in both branches is different ( 𝐴𝐴𝑃𝑃 𝑂𝑂 = 94.0 in East Branch, 𝐴𝐴𝑃𝑃 𝜕𝜕 = 86.0 in West Branch), but this is not essential for the allocation of employees: it is not important to balance the productivity of both branches, but to equalise marginal products across branches. The marginal product of the last assigned employee in each shop is 𝑀𝑀𝑃𝑃 𝑂𝑂 = 𝑀𝑀𝑃𝑃 𝜕𝜕 = 80 . <?page no="199"?> 5.2 Short run decisions: diminishing marginal product and productivity 199 Bank branches and allocation of employees branch East branch West employees productivity output marginal product employees productivity output marginal product 𝐿𝐿 𝐴𝐴𝑃𝑃 𝑂𝑂 = 𝑞𝑞/ 𝐿𝐿 𝑞𝑞 𝑂𝑂 𝑀𝑀𝑃𝑃 𝑂𝑂 = ∆𝑞𝑞/ ∆𝐿𝐿 𝐿𝐿 𝐴𝐴𝑃𝑃 𝜕𝜕 = 𝑞𝑞/ 𝐿𝐿 𝑞𝑞 𝜕𝜕 𝑀𝑀𝑃𝑃 𝜕𝜕 = ∆𝑞𝑞/ ∆𝐿𝐿 0 - 0 0 - 0 108 (1) 92 (6) 1 108.0 108 1 92.0 92 104 (2) 80 (10) 2 106.0 212 2 86.0 172 100 (3) 60 3 104.0 312 3 77.3 232 96 (4) 40 4 102.0 408 4 68.0 272 92 (5) 20 5 100.0 500 5 58.4 292 88 (7) 0 6 98.0 588 6 48.7 292 84 (8) -20 7 96.0 672 7 38.9 272 80 (9) -40 8 94.0 752 8 29.0 232 58 -60 9 90.0 810 9 19.1 172 50 -80 10 86.0 860 10 9.2 92 Table 5.3: Allocation of employees to branches. For managers, this implies a simple decision-making rule: the marginal product is the key short-term control variable in production. The marginal product must be positive, otherwise production will decline. From a short-term perspective and when allocating employees, a general rule is: always allocate additional employees, resources, or capacity to the factory (or location, branch, team, etc.) where marginal product is highest - i.e., where an additional employee can make the largest additional contribution to production. In laboratory experiments (and in reality) managers often deviate from this rule: in most cases they focus on productivity instead of marginal product, and as a consequence the productivity of factories is seen as the only relevant factor for decision-making and therefore, wrongly decided, this leads to inefficiency with corresponding negative effects on overall production. In typical management consulting projects (e.g., regarding branch sizes of banks with similar products and customer groups, factory sizes of car manufacturers, team sizes of the same activity at different locations, etc.), an increase in production of between 10% and 30% with a <?page no="200"?> 5 Firm size, technology and decisions on production 200 constant number of employees is possible by focusing on marginal products - without investments in new IT infrastructure, without process optimisation and without cost increases. Marginal product and productivity in the Cobb-Douglas production function For each production function, the short-term marginal product and productivity can be determined. For the Cobb-Douglas production function (5.2), productivity is generally obtained by dividing the output by the labour input as (5.9) 𝐴𝐴𝑃𝑃 𝐿𝐿 = 𝑞𝑞𝐿𝐿 = 𝐴𝐴𝐸𝐸𝛼𝛼𝐿𝐿𝛽𝛽 𝐿𝐿 = 𝐴𝐴𝐸𝐸 𝛼𝛼 𝐿𝐿 𝛽𝛽−1 and the marginal product of labour as (5.10) 𝑀𝑀𝑃𝑃 𝐿𝐿 = 𝜕𝜕𝑞𝑞 𝜕𝜕𝐿𝐿 = 𝛽𝛽𝐴𝐴𝐸𝐸 𝛼𝛼 𝐿𝐿 𝛽𝛽−1 , by partially differentiating the production function. Thus, for the production function empirically determined in ► Section 5.1, we get (5.11) 𝐴𝐴𝑃𝑃 𝐿𝐿 = 𝑞𝑞𝐿𝐿 = 4.12 𝐸𝐸 0.79 𝐿𝐿 −0.50 and (5.12) 𝑀𝑀𝑃𝑃 𝐿𝐿 = 𝜕𝜕𝑞𝑞 𝜕𝜕𝐿𝐿 = 2.06 𝐸𝐸 0.79 𝐿𝐿 −0.50 . Obviously neither productivity 𝐴𝐴𝑃𝑃 𝐿𝐿 nor marginal product 𝑀𝑀𝑃𝑃 𝐿𝐿 are constant, but change depending on given capital input if the labour input, i.e., the number of employees, changes. In addition, 𝑀𝑀𝑃𝑃 𝐿𝐿 = 2.06 𝐸𝐸 0.79 𝐿𝐿 −0.50 < 4.12 𝐸𝐸 0.79 𝐿𝐿 −0.50 = 𝐴𝐴𝑃𝑃 𝐿𝐿 immediately shows that there are diminishing marginal products and declining productivity. Case Study │ Refrigerators - short run adjustment of production quantiy A team of management consultants analysed operations and production of a global refrigerator manufacturer. Using internal firm data, they identified the following values of a Cobb-Douglas production function: (5.13) 𝑞𝑞 = 𝐴𝐴𝐸𝐸 𝛼𝛼 𝐿𝐿 𝛽𝛽 with 𝐴𝐴 = 0.10348, 𝛼𝛼 = 0.8, 𝛽𝛽 = 0.6. 𝐸𝐸 = 80,000,000 and 𝐿𝐿 1 = 800 . Currently, 12 million refrigerators are produced per annum - now production is to be expanded to 15 million units in the short run. Management must decide how many employees need to be hired. If we first check via equation (5.13) whether the firm is currently producing efficiently, then we obtain (5.15) 𝑞𝑞 1 = 0.10348 80,000,000 0.8 800 0.6 = 12,000,000. (5.16) 𝑞𝑞1 𝐿𝐿1 = 𝐴𝐴𝐸𝐸𝛼𝛼𝐿𝐿1 𝛽𝛽 𝐿𝐿1 = 0.10348 ⋅80,000,0000.8⋅8000.6 800 = 15,000 and (5.17) 𝜕𝜕𝑞𝑞 𝜕𝜕𝐿𝐿1 = 0.6 ⋅ 0.10348 ⋅ 80,000,000 0.8 ⋅ 800 −0.4 = 9,000 . So, with given capital and workforce, 12 million refrigerators can actually be produced. However, the marginal product (5.17) is smaller than the productivity (5.16), so that the firm is definitely in area 2 of ► Figure 5.4: one additional employee will increase the production by 9,000 refrigerators per year, but the firm's productivity will decrease. <?page no="201"?> 5.3 Long run decisions: technical progress and returns to scale 201 In order to expand production to 15 million refrigerators in the short run, the firm has to increase the number of employees from the current level of 800 employees while keeping capital employed constant. In order to determine the necessary number of employees, the production function (5.13) can be changed to (5.18) 𝑞𝑞 2 = 𝐴𝐴𝐸𝐸 𝛼𝛼 𝐿𝐿 2 𝛽𝛽 = 15,000,000 and solved according to the new number of employees 𝐿𝐿 2 , so that (5.19) 𝐿𝐿 2 = � 𝑞𝑞2 𝐴𝐴 𝐸𝐸𝛼𝛼 𝛽𝛽 = � 𝑞𝑞2 𝐴𝐴 𝐸𝐸𝛼𝛼 � 1 𝛽𝛽 = � 15,000,000 0.10348⋅80,000,0000.8 � 1 0.6 = 1,160 results. Compared to the initial situation and a production volume of 12,000,000 refrigerators, the increase in production volume requires an additional 1,160 − 800 = 360 employees. Firstly, productivity drops from the original 15,000 refrigerators per employee to (5.20) 𝑞𝑞2 𝐿𝐿2 = 𝐴𝐴 𝐸𝐸𝛼𝛼 𝐿𝐿2 𝛽𝛽 𝐿𝐿2 = 0.10348⋅80,000,0000.8⋅1,1600.6 1,160 = 12,931, secondly, the marginal product (5.21) 𝜕𝜕𝑞𝑞2 𝜕𝜕𝐿𝐿 = 𝛽𝛽𝐴𝐴 𝐸𝐸 𝛼𝛼 𝐿𝐿 2 𝛽𝛽−1 = 0.6 ⋅ 0.10348 ⋅ 80,000,000 0.8 ⋅ 1,160 −0.4 = 7,757 decreases from 9,000 to 7,757 due to the increased number of jobs. The results (5.20) and (5.21) again confirm ► Figure 5.4: if productivity is larger than marginal product, then if the number of employees increases, both numbers will fall in the short run - while maintaining the capital input and unchanged technology. 5.3 Long run decisions: technical progress and returns to scale In the long run, input quantities of all input factors can be changed, i.e., not only are decisions made on the optimum use of labour, but capital input is also freely determined within longterm investment planning. This becomes visible, for example, in the formation of a new Tesla battery factory in Sparks, Nevada, or the long-term capacity expansion of Frankfurt Rhein- Main International Airport to establish additional runways or terminals. In the following two key aspects of long-term production functions are developed. Returns to scale - measure the effect of a simultaneous increase in all input factors on production output. Returns to scale as a control parameter for long-term growth processes of firms (e.g., the scalability of a new business model) foster competitive advantage based on the size of a firm depending on changes in capital and labour input. Technical progress - technical progress makes it possible to increase production volume with constant use of input factors. Typical is an improvement in technology, i.e., investment in research and development, which, through digitalisation, artificial intelligence, or automation, is reflected in a change in the interplay of labour and capital and tends to lead to an increase in capital intensity. Long-term production functions can be analysed graphically in the same way as utility functions (see ► Chapter 2): depending on the technology used, production quantities can be realized <?page no="202"?> 5 Firm size, technology and decisions on production 202 with different combinations of labour and capital. An isoquant describes all points of the same output; a set of isoquants then describes the production function (i.e., the quantity of production possibilities with varying factor input), as shown in ► Figure 5.5. The higher an isoquant is located, the larger the output. Figure 5.5: Long run production function and capital intensity. Typically, production can be realized through alternative combinations of capital and labour. The input ratio of capital to labour is referred to as capital intensity - the higher the relative capital input per employee, the larger the capital intensity. In the long run - capital and labour are now totally variable - firms can decide not only on production output but also on capital intensity through choice of technology. Technical progress and capital intensity Industries (and within each industry also firms) differ in their capital intensity, as can be seen in ► Figure 5.6. In addition to existing technologies, this is due to differences in the availability of input factors (e.g., shortage of skilled workers or procurement bottlenecks for machinery) and regional or international differences in factor prices, as well as strategic decisions by firms (e.g., different automation strategies of Volkswagen and Toyota). In the long run, the capital intensity will also increase in almost all industries - in Germany until 2013, on average across industries up to 400,000 EUR, i.e., on average 400,000 EUR per employee or workplace are tied up in equity or borrowed capital. <?page no="203"?> 5.3 Long run decisions: technical progress and returns to scale 203 Figure 5.6: Development of capital intensity from 1991 to 2013 in TEUR in Germany (data source: Institut der deutschen Wirtschaft 2014, own calculations). The reasons for this are initially long-term increases in labour costs (wage rates) as a result of repeated higher collective labour agreements, while capital costs (interest rates) in the capital market remain constant. Subsequently, firms try to substitute the relatively more expensive factor labour at least partially by automation, i.e., higher capital inputs. On the other hand, there is an interaction with technological change taking place in other industries. Due to rising labour costs, firms are actively looking for new production processes, e.g., robotics or artificial intelligence, which will make human labour obsolete and reduce the number of jobs in absolute terms. As a result, however, the productivity of the remaining jobs increases. Technological progress means that either more can be produced with constant input factors, or the same output can be achieved with less factor input. In the long run, technical progress is mostly labour-saving, i.e., it increases capital intensity, while at the same time the marginal product of labour relative to the marginal product of capital increases. Technical progress and innovations increase productivity of employees and can be a major driver for higher wage rates (Aghion and Howitt 2009 for a comprehensive analysis in the context of growth theory and Belitz et al. 2017 for an empirical analysis of drivers in Germany). In food retailing, for example, labour-saving technical progress is constantly taking place: self-service, packaged food, digital price tags, scanner cash registers, or self-checkout have significantly increased employee productivity by increasing capital intensity since the days of the corner shop to modern supermarkets. In addition, in the medium term, efficiency 𝐴𝐴 can be increased by a reorganisation or new organisational structure, training of employees or better management models of a firm with the same capital resources. For each employee (and each number of employees), production, marginal product and productivity then increase. <?page no="204"?> 5 Firm size, technology and decisions on production 204 Figure 5.7: Percentage productivity increase in large EU countries from 1995 to 2016 across various industries (data source: OECD Productivity and ULC by main economic activity (ISIC Rev.4) 2017, own calculations). ► Figure 5.7 shows the productivity development in large EU countries for the period 1995 to 2016 - on average, the productivity of each employee has increased by about 2.2% per year. As a rule of thumb, business plans of incumbent firms can take into account an annual productivity increase of about 2% if all opportunities for technical progress (new software, employee training, automation, etc.) are seized. In addition to technical progress, economic factors related to business cycles - as can be seen from the decline in productivity in the context of the financial and national debt crisis from 2007 and the subsequent recovery in 2010 - obviously play a significant role. In the long run, a decline in productivity growth is observed in many industries: a major reason for this is the increasing exploitation of opportunities in existing technologies. In addition, the complementary innovations in the area of IT and communications technology have so far apparently not been fully reflected in productivity growth in individual industries and firms. This observation is known as the ICT productivity paradox (Brynjolfsson and Hitt 2000, Syverson 2011 and Acemoglu et al. 2014) and apparently continues to apply to big data and artificial intelligence (Bughin et al. 2017 and Tambe 2014). The reasons are to be seen in the longterm and costly adaptation processes of incumbent firms for new technologies and the complementary character of these technological developments. Returns to scale and competitive advantage in production Firms in different industries differ not only in terms of capital intensity, but also in absolute size of the firms and growth dynamics. Some industries are characterised by a few very large firms, others are typically populated by many small firms: in Germany there used to be eleven very large oil refineries in 2015 (cf. Mineralölwirtschaftsverband www.mwv.de/ statistiken/ ) with an average production volume of more <?page no="205"?> 5.3 Long run decisions: technical progress and returns to scale 205 than 8 million tons with an average of just over 800 employees and a capital investment per firm of around EUR 650 million, in contrast. there are around 1,600 manufacturers of musical instruments in Germany in 2014 (cf. Deutsches Musikinformationszentrum www.miz.org) with an average number of employees of less than seven and a capital investment of less than EUR 1 million each. The main factor determining the optimal size of a firm (or, in a dynamic view of the growth rate) is returns of scale, in other words, the correlation between the absolute size of a firm and advantages in efficiency or productivity. Returns to scale are a management concept for estimating and determining the long-term quantitative effect of variations in all input factors on output: they are always relevant when long-term production decisions have to be made regarding technology and input factors to expand or reduce output. Returns to scale measure how much the output changes when all input factors of production (capital and labour) are changed to the same extent. Three types of returns to scale can occur as follows. At constant returns to scale, a higher use of all inputs leads to a proportional growth in output (i.e., output increases by 10% if all inputs are increased by 10%). Growth is then linear with an adjustment of employees and capital. In this case, managers are indifferent between one large or several small factories during growth processes of a firm - both options are equally efficient. If a higher use of all input factors leads to a more than proportionate growth in production volume (i.e., volume increases by more than 10% if all factors are increased by 10%), returns to scale are increasing. In this case, when a firm grows, managers should enlarge the existing factory instead of opening new small factories - a large factory is more efficient. With decreasing returns to scale, a higher use of all input factors leads to a less than proportionate growth in output (i.e., output increases by less than 10% if all factors are increased by 10%). In this case, when a firm is growing, managers should open new small factories instead of expanding the existing one - small factories are more efficient. Figure 5.8: Effect of a change in all input factors on output. <?page no="206"?> 5 Firm size, technology and decisions on production 206 ► Figure 5.8 on the right shows that although increasing returns to scale make a more than proportionate growth possible, the use of resources only decreases less than proportionately when production quantities decline. With decreasing returns to scale in ► Figure 5.8 left, growth requires a more than proportionate increase in resource use. Conversely, a decline in production also releases more than proportionately inputs. The reasons for increasing, constant, or decreasing returns to scale are specific to each industry or firm. The following patterns are often identified in empirical studies or case studies. Increasing returns to scale arise due to: specialisation (division of labour) and learning effects (experience curve) in production; modularisation and platform strategies or high proportion of common parts in production; investment in research and development with a broad product portfolio or high coverage of different stages of the value chain; reduced risks, portfolio strategies for products, parenting advantage for overarching umbrella brands and internalisation of resources (e.g., internal labour market and training, insourcing of value-added stages); positive direct and indirect network effects on the demand side and connection or affiliation to technological or economic ecosystems (multisided platforms in digital business models); and technologically indivisible input factors and thus, better use of technology and opportunities for rationalisation. Decreasing returns to scale are caused by: limited management capacity (leadership span, communication, or project size), organisation (complexity, conglomerate structures, hierarchies, international subsidiaries, etc.) as well as organisational inertia (e.g., in the shape of meeting structures or limited ability to deliver necessary organisational change); more complex auditing (regulatory framework) and inefficient governance structures (coordination and control of internal boards or committees); lack of suitable manufacturing processes and machines to enable large scale or mass production; fully utilised technology and lack of availability of well-trained staff or skills (i.e., lack of additivity). Increasing returns to scale often support rapid and more than proportionate growth: emerging market opportunities or demand growth can be realised because either no or only a few additional resources are required. With decreasing returns to scale, firm growth requires a more than proportionate increase in labour and capital input - which means that seizing every market opportunity directly requires the recruitment of new employees, but declining demand also leads to job losses. <?page no="207"?> 5.3 Long run decisions: technical progress and returns to scale 207 Figure 5.9: Transition from increasing to decreasing returns to scale. Along a typical growth process of a firm, a sequence of increasing and then decreasing returns to scale can be observed depending on size, but also over time. First, an organisation scales and learning curve effects increase production at a more than proportionate rate; then complexity and limited management skills dominate and limit the opportunities for efficient growth as shown in ► Figure 5.9. This explains why small firms typically have higher growth rates (yet with higher variance) compared to larger firms (Münter 1999), but also why firms can achieve exponential growth rates based on new digital business models (Arthur 1996, Nielsen and Lund 2015). The considerations outlined above for firms have effects on industries and on market structure. If increasing returns to scale are not very pronounced, or if the area of decreasing returns to scale is reached relatively quickly, then: there is weak competitive advantage from size, i.e., firms tend to be smaller; for a given demand structure and market size, there tends to be a large number of firms in the market; and relative firm growth is made more difficult by technological conditions and requires a more than proportionate increase in the use of resources. Conversely, if increasing returns to scale are strongly pronounced and the area of decreasing returns to scale is reached late, then: firms have significant competitive advantage from size, i.e., firms tend to be larger; as a result, for a given demand structure and market size, there tends to be a small number of firms in the market; and relative firm growth is supported by technological conditions and requires a less than proportionate increase in the use of input factors. The possibility of achieving increasing returns to scale differs significantly across industries. The more extensive returns to scale there are, the larger the firms in an industry tend to be (i.e., higher absolute production with higher capital and labour input in absolute terms). Examples <?page no="208"?> 5 Firm size, technology and decisions on production 208 of this include telecommunications, energy, or network-related transportation firms. Numerous new business models based on multisided platforms are also characterised by extensive and increasing returns to scale: in order to manage rapidly growing numbers of members or customers on social media platforms, only small increases in capital and employees are necessary (Arthur 1996, van Alstyne et al. 2016 as well as Hirt and Wilmott 2014). Returns to scale in long run production functions The extent and nature of the returns to scale of a firm can be determined empirically by analysing the production function. If, for example, one uses the Cobb-Douglas production function from ► Section 5.1, after some conversions it can be quickly seen that, if one puts 𝑆𝑆 -times inputs into the firm, production does not increase by a factor of 𝑆𝑆 : rather, the real increase depends on the sum of the partial production elasticities 𝜶𝜶 and 𝜷𝜷 , as can be seen from the analysis of (5.22) 𝑞𝑞 0 = 𝑇𝑇(𝐸𝐸, 𝐿𝐿) = 𝐴𝐴𝐸𝐸 𝛼𝛼 𝐿𝐿 𝛽𝛽 and (5.23) 𝑞𝑞 1 = 𝑇𝑇(𝑆𝑆𝐸𝐸, 𝑆𝑆𝐿𝐿) = 𝐴𝐴(𝑆𝑆𝐸𝐸) 𝛼𝛼 (𝑆𝑆𝐿𝐿) 𝛽𝛽 . Solving equation (5.23) as (5.24) 𝑞𝑞 1 = 𝑇𝑇(𝑆𝑆𝐸𝐸, 𝑆𝑆𝐿𝐿) = 𝐴𝐴(𝑆𝑆𝐸𝐸) 𝛼𝛼 (𝑆𝑆𝐿𝐿) 𝛽𝛽 = 𝑆𝑆 𝛼𝛼+𝛽𝛽 𝐴𝐴𝐸𝐸 𝛼𝛼 𝐿𝐿 𝛽𝛽 = 𝑆𝑆 𝛼𝛼+𝛽𝛽 𝑞𝑞 0 , we obtain the new output 𝑞𝑞 1 , which results from a linear increase of the input factors 𝐸𝐸 and 𝐿𝐿 , corresponds to the original output 𝑞𝑞 0 multiplied by the factor 𝑆𝑆 𝛼𝛼+𝛽𝛽 . Thus, in case of marginal changes in input quantities, the rate of change in total output can be determined by adding the partial elasticities of production: if the sum 𝛼𝛼 + 𝛽𝛽 is equal to one, then there are constant returns to scale, if the sum 𝛼𝛼 + 𝛽𝛽 is greater than one, there are increasing returns to scale, if the sum 𝛼𝛼 + 𝛽𝛽 is less than one, the firm has decreasing returns of scale. The sum of 𝛼𝛼 and 𝛽𝛽 is called scale elasticity. Case Study │ Turbine manufacturer with long run change of all input factors In case of marginal changes in input quantities, the rate of change in output can be determined by adding the partial production elasticities. Returning to the turbine manufacturer, the Cobb-Douglas production function is (5.25) 𝑞𝑞 0 = 𝑇𝑇(𝐸𝐸 = 100, 𝐿𝐿 = 200) = 4.12 ⋅ 100 0.79 ⋅ 200 0.50 = 2,215.20 and output 𝑞𝑞 0 equals 2,215.20 . Increasing capital and labour input by 1% provides a new output of (5.31) 𝑞𝑞 1 = 𝑇𝑇(𝐸𝐸 = 101, 𝐿𝐿 = 202) = 4.12 ⋅ 101 0.79 ⋅ 202 0.50 = 2,243.81, i.e., the production volume increases to 101.29 % ⋅ 𝑞𝑞 0 : this increase corresponds to the sum of the partial production elasticities 𝛼𝛼 = 0.79 and 𝛽𝛽 = 0.50 . In the same way, longterm and interdependent effects can be examined, such as an increase in capital input of 1% and a simultaneous reduction in labour input of 1% . The new production volume is (5.32) 𝑞𝑞 2 = 𝑓𝑓(𝐸𝐸 = 101, 𝐿𝐿 = 198) = 4.12 ⋅ 101 0.79 ⋅ 198 0.50 = 2,221.49, an increase of 0.29% compared to the original output 𝑞𝑞 0 . This is again obtained by adding the partial elasticity of production weighted by the change in factor input, in this case as 0.79 ⋅ 1% − 0.50 ⋅ 1% = 0.29% and correspondingly 𝑞𝑞 2 = 100.29% ⋅ 𝑞𝑞 0 . <?page no="209"?> 5.3 Long run decisions: technical progress and returns to scale 209 Case Study │ Refrigerators with long run increase in demand and adjustment of all input factors The refrigerator manufacturer from ► Section 5.2 is now planning - due to positive market research results - to expand production in the long run from 12 million to 36 million refrigerators. The crucial question for management is now whether additional factories should be established for this purpose, or whether production should be increased (and concentrated) in the existing factory. The initial situation is characterised by (5.33) 𝑞𝑞 = 𝐴𝐴𝐸𝐸 1𝛼𝛼 𝐿𝐿 1 𝛽𝛽 with 𝐴𝐴 = 0.10348, 𝛼𝛼 = 0.8, 𝛽𝛽 = 0.6 given a factor input of 𝐸𝐸 1 = 80,000,000 and 𝐿𝐿 1 = 800 . Output is accordingly 𝑞𝑞 1 = 12,000,000 . The goal of the firm is now to triple the output via long-term adjustment by a factor 𝑆𝑆 of the input factors 𝐸𝐸 and 𝐿𝐿 , so that (5.34) 𝑞𝑞 2 = 𝐴𝐴𝐸𝐸 2𝛼𝛼 𝐿𝐿 2 𝛽𝛽 = 36,000,000 with (5.35) 𝑞𝑞 2 = 𝐴𝐴(𝑆𝑆𝐸𝐸 1 ) 𝛼𝛼 (𝑆𝑆𝐿𝐿 1 ) 𝛽𝛽 = 𝑆𝑆 𝛼𝛼+𝛽𝛽 𝐴𝐴𝐸𝐸 1𝛼𝛼 𝐿𝐿 1 𝛽𝛽 ≤≥ 36,000,000 . First, one can check what would happen if input factors are tripled - as a result the output increases to (5.36) 𝑞𝑞 2 (𝑆𝑆 = 3) = 𝑆𝑆 𝛼𝛼+𝛽𝛽 𝐴𝐴𝐸𝐸 1𝛼𝛼 𝐿𝐿 1 𝛽𝛽 = 3 0.8+0.6 ⋅ 0.10348 ⋅ 80,000,000 0.8 ⋅ 800 0.6 = = 3 1.4 ⋅ 0.10348 ⋅ 80,000,000 0.8 ⋅ 800 0.6 = 55,866,589 . A tripling of labour and capital input would lead to a significantly more than proportionate increase in production due to increasing returns to scale with a scale elasticity 𝛼𝛼 + 𝛽𝛽 = 1.4 - so it is clear from a strategic perspective, that the existing factory should be expanded in any case and no additional small factories should be considered. The next question is how much additional capital and how many additional jobs are necessary in order to realize the planned increase in production. From the condition (5.37) 𝑞𝑞 2 = 𝑆𝑆 𝛼𝛼+𝛽𝛽 𝐴𝐴𝐸𝐸 1𝛼𝛼 𝐿𝐿 1 𝛽𝛽 → 𝑆𝑆 𝛼𝛼+𝛽𝛽 𝑞𝑞 1 = 𝑞𝑞 2 it follows that (5.38) 𝑆𝑆 𝛼𝛼+𝛽𝛽 = 𝑞𝑞2 𝑞𝑞1 and (5.39) 𝑆𝑆 = � 𝑞𝑞2 𝑞𝑞1 � 1 𝛼𝛼+𝛽𝛽 must apply. Plugging in values from equation (5.39), one gets (5.40) 𝑞𝑞2 𝑞𝑞1 = 36,000,000 12,000,000 = 3 , hence (5.41) 𝑆𝑆 = � 𝑞𝑞2 𝑞𝑞1 � 1 𝛼𝛼+𝛽𝛽 = (3) 1 0.8+0.6 = 2.19 . This means that for a triplication of the output, the input factors capital and labour only need to be adjusted by the factor 𝑆𝑆 = 2.19 due to prevailing increasing returns to scale. <?page no="210"?> 5 Firm size, technology and decisions on production 210 Of course, in addition to the existing first factory, two identical factories could have been established. This approach obviously leads to an increase in output to 36 million refrigerators, but this is an inefficient way: three "small factories" require a respective capital and labour input of (5.42) 𝐸𝐸 1 = 𝐸𝐸 2 = 𝐸𝐸 3 = 80,000,000 → ∑𝐸𝐸 𝑖𝑖 = 240,000,000 and 𝐿𝐿 1 = 𝐿𝐿 2 = 𝐿𝐿 3 = 800 → ∑𝐿𝐿 𝑖𝑖 = 2,400 . A "large factory", however, triples the production volume by a factor of 2.19, i.e., capital input is 175,200,000 and 1,752 employees are needed, resulting in an inefficiency of (5.43) 𝐸𝐸 ′1 = 2.19 ⋅ 80,000,000 = 175,200,000 𝐿𝐿 ′1 = 2.19 ⋅ 800 = 1,752 and ∆𝐸𝐸 𝐸𝐸 = −27 % and ∆𝐿𝐿 𝐿𝐿 = −27 % . Productivity, capital intensity and returns to scale From a management perspective, the longand short-term concepts of productivity, capital intensity and returns to scale are deeply intertwined, especially in the context of technical progress. Against the backdrop of lower profitability of the Volkswagen Group, especially of the core brand VW, compared to its competitors, management has repeatedly implemented costcutting initiatives and restructuring of the organisation (Süddeutsche Zeitung 2014, Manager Magazin 2014, Handelsblatt 2014 and Frankfurter Allgemeine Zeitung 2017). Process optimisation, plant closures, job cuts, and further automation and digitisation have all been repeatedly discussed in detail. Without looking into details of the cost situation, ► Table 5.4 shows the development of some key performance indicators of the Volkswagen Group. Although the firm has grown in the period under review, the growth rates of capital and employees are significantly higher than those of production - as a consequence, productivity did decrease. Volkswagen group: technological progress, productivity and capital intensity year 2006 2010 2014 CAGR 2006-2010 CAGR 2010-2014 output cars in million 5.66 7.36 10.21 6.80 % 8.50 % employees in thousands 329 389 583 4.30 % 10.60 % capital in billion EURs 136.5 196.7 351 9.60 % 15.60 % capital intensity in thousand EURs 414 505 602 5.10 % 4.50 % productivity cars per employee 17.2 18.9 17.5 2.40 % -1.90 % scale elasticity ~ 0.98 ~ 0.65 Table 5.4: Volkswagen - technical progress and capital intensity. (data: Volkswagen Factbook and annular reports, own calculations). <?page no="211"?> 5.4 Summary and key learnings 211 Figure 5.10: Transition from constant to decreasing returns to scale at Volkswagen 2006 to 2014. A first explanation may be that this significant growth in capital and employees has led to inefficiencies in the group's organisation. Secondly, the increase in staff and capital could be inconsistent with prevailing returns to scale, i.e., Volkswagen either has too many small sites or some of the sites are absolutely too large. If data is depicted in ► Figure 5.10, a possible further explanation for Volkswagen's dilemma becomes apparent: in 2014 the Volkswagen Group may be at the transition from constant to decreasing returns to scale. A significant expansion of capital and labour resulted in a significantly less than proportional increase in production between 2006 and 2014. If the production of 10 million vehicles is extrapolated based on capital intensity and efficiency of the years 2006 and 2010, Volkswagen would have to get by with around 450,000 employees and equity and debt of around EUR 280 billion, if the production function were to become more efficient and returning to constant returns to scale - thus, restructuring requirements could be large. 5.4 Summary and key learnings Why do some firms grow much faster than others? Why are some small firms able to survive, while some large firms struggle with inefficiency? A production function is the key analytical tool and management concept for making strategic decisions on production capacity, and to determine the structure of input factors capital and labour. A production function can easily be <?page no="212"?> 5 Firm size, technology and decisions on production 212 determined from existing firm data (e.g., output, use of equity and debt and number of employees) using econometric methods. From a management perspective, it is essential to distinguish between a shortand a longterm perspective. In the short run, the number of employees, i.e., the input factor labour, can be varied - for example, by ordering overtime or procuring additional staff through a temporary employment agency. In the short run, marginal product of labour and labour productivity indicate how the output changes. Typically, positive but diminishing marginal products are observed, i.e., production increases less than proportionately with additional employees, and productivity decreases accordingly. In the long run, it is of key importance whether, in what way and how easily or quickly capital and labour can be changed or substituted. Typically, in many industries one can observe an increase in capital intensity accompanied by labour-saving technical progress. The reasons are mainly to be found in long run increasing costs of labour (rising wage rates and salaries) in comparison to relatively constant costs of capital (fluctuating but relatively stable interest rates for equity and debt), and an incentive to implement innovations and technical improvements which result in a long-term increase in capital intensity. From a strategic perspective, there are some crucial points. Long-term adjustments in production capacity by adjusting the input of capital and labour can result in increasing, constant, or decreasing returns to scale. The analysis of the type and scope of returns to scale helps to make decisions concerning size and number of factories within a firm but is also an indicator of market structure to be expected, i.e., the number and size distribution of competitors. Moreover, returns to scale are one of the main success levers of new digital business models: during the planning, design of new business models, and often exponential growth processes, the question of scalability (i.e., the strategic realisation of extensive increasing returns to scale) is at the core of management attention. Recommendations for further reading For more on digitalisation from a microeconomic perspective and its impact on labour and capital, see Brynjolfsson, E. and McAfee, A., The second machine age: work, progress, and prosperity in a time of brilliant technologies, New York 2014. For a comprehensive account of production theory, see Mas-Colell, A. Winston, M.D. and Green, J.R., Microeconomic theory, New York 1995. Questions for review [1] Describe applications of the analysis of technology and firm size decisions from a microeconomic perspective as well as their limitations, advantages and disadvantages. [2] How can the production function of a firm be identified? Which parameters can be specified, for example, in a Cobb-Douglas production function? <?page no="213"?> 5.4 Summary and key learnings 213 [3] Describe the input factors capital and labour with reference to the time horizon of managerial decisions. Give two examples each of shortand long-term adjustments of input factors. [4] Describe the relationship between marginal product and productivity when the output is expanded. [5] Determine the following quantities for a Cobb-Douglas production function with the parameters technical efficiency 𝐴𝐴 = 2.5, partial production elasticity of capital 𝛼𝛼 = 0.5 and partial production elasticity of labour 𝛽𝛽 = 0.4 : a. output with capital input of 𝐸𝐸 = 100 (in million EUR) and labour input of 𝐿𝐿 = 200 (in employees per year), b. Marginal product, productivity (at this output level) and returns to scale. c. The firm is now losing 5% of its employees due to the "retirement at 63" regulation and cannot find adequate replacements - what percentage increase of capital input is necessary to keep output constant? What happens to productivity in this firm? [6] In two factories A and B, which produce identical products at different locations, the productivity is 𝐴𝐴𝑃𝑃 𝐴𝐴 = 9 resp. 𝐴𝐴𝑃𝑃 𝐵𝐵 = 10 and the respective marginal products 𝑀𝑀𝑃𝑃 𝐴𝐴 = 11 resp. 𝑀𝑀𝑃𝑃 𝐵𝐵 = 9 - explain with the help of diagrams which decisions a production manager can derive from these values? [7] Explain possible reasons why there are 'many small firms' in some industries but 'few large firms' in other industries and give three examples each (industries). [8] Describe the concept of economies of scale. In what ways does it help in decisions on the number of production sites or factories? Firms often show first increasing, then decreasing returns to scale - what could be the reason for this? [9] Why is there a long-term increase in capital intensity for many industries? [10]You are planning to establish a new business model - how do you explain to potential investors that your business model scales and that increasing returns to scale are possible? Literature Acemoglu, D., Dorn, D., Hanson, G. and Price, B., Return of the Solow paradox? IT, productivity, and employment in US manufacturing, American Economic Review, 2014, 104, 5, 394-399. Aghion, P. and Howitt, P.W., The economics of growth, Boston 2009. Andries, A.M. and Capraru, B., The nexus between competition and efficiency: the European banking industries experience, International Business Review, 2014, 23, 3, 566-579. Arthur, W.B., Increasing returns and the new world of business, Harvard Business Review, July/ August 1996, 31-53. Belitz, H., Eickelpasch, A., Mouel, M.L. and Schiersch, A., Wissensbasiertes Kapital in Deutschland: Analyse zu Produktivitäts- und Wachstumseffekten und Erstellung eines Indikatorsystems, DIW Berlin, 2017. Black, S.E. and Lynch, L.M., Measuring organisational capital in the new economy, in: Corrado, C., Haltiwanger, J. and Sichel, D. (eds.), Measuring capital in a new economy, Chicago 2005, 205-236. Bloem, J., van Doorn, M., Duivestein, S., Excoffier, D., Maas, R. and van Ommeren, E., The fourth industrial revolution - things to tighten the link between IT and OT, Groningen 2014. <?page no="214"?> 5 Firm size, technology and decisions on production 214 Bloom, N. and van Reenen, J., Measuring and explaining management practices across firms and countries, Quarterly Journal of Economics, 2007, 122, 4, 1351-1408. Bonin, H., Gregory, T. and Zierahn, U., Übertragung der Studie von Frey/ Osborne (2013) auf Deutschland, ZEW Kurzexpertise Nr. 57, Mannheim 2015. Brynjolfsson, E. and Hitt, L.M., Beyond computation: information technology, organizational transformation and business performance, Journal of Economic Perspectives, 2000, 14, 4, 23-48. Bughin, J., Hazan, E., Ramaswamy, S., Chui, M., Allas, T., Dahlström, M., Henke, N. and Trench, M., Artificial Intelligence: the next digital frontier, McKinsey Global Institute 2017. Frey, C.B. and Osborne, M.A., The future of employment: how susceptible are jobs to computerization? , Technological Forecasting and Social Change, 2017, 114, 254-280. Hay, D.A. and Liu, G.S., The efficiency of firms: what difference does competition make? , Economic Journal, 1997, 107, 442, 597-617. Hirt, M. and Willmott, P., Strategic principles for competing in the digital age, McKinsey Quarterly, May 2014, 1-13. Holmes, T.J. and Schmitz, J.A., Competition and productivity: a review of evidence, Annual Review of Economics, 2010, 2, 619-642. Läsker, K., Winterkorn verdonnert VW zu radikalem Sparkurs, Süddeutsche Zeitung, 15. Juli 2014. McGowan, M.G., Andrews, D., Criscuolo, C. and Nicoletti, G., The future of productivity, OECD, Paris 2015. Münter, M.T., Wettbewerb und die Evolution von Industrien, Bayreuth 1999. Müssgens, C. and Jung, M., Streit um Sparprogramm wird zur Chefsache, Frankfurter Allgemeine Zeitung, 13. Februar 2017. Nielsen, C. and Lund, M., The concept of business model scalability, Aalborg University, mimeo, 2015. o.V., Alle VW-Marken müssen ihre Kosten senken, Handelsblatt, 24.11.2014. o.V., Volkswagen eröffnet neues Werk in China, Manager Magazin, 24.10.2013. o.V., VW-Betriebsratschef Osterloh will McKinsey aus dem Haus jagen, Manager Magazin, 25.7.2014. Syverson, C., What determines productivity? , Journal of Economic Literature, 2011, 49, 2, 326-365. Tambe, P., Big data investment, skills, and firm value, Management Science, 2014, 60, 6, 1452-1469. van Alstyne M.W., Parker,G.G. and Choudary, S.P., Pipelines, platforms and the new rules of strategy, Harvard Business Review, April 2016, 2-9. von Tunzelmann, N., Historical coevolution of governance and technology in the industrial revolutions, Structural Change and Economic Dynamics, 2003, 14, 4, 365-384. <?page no="215"?> 215 6 Costs, restructuring and M&A Decisions on the cost structure, the restructuring of an existing firm or the acquisition of another firm, always aim at a sustainable improvement of the competitiveness of a firm. The cost structure can be influenced by outsourcing parts of fixed costs activities. For example, banks outsource to specialised IT or BPO service providers. Another strategy would be exploiting factor price differences by relocating production to low-wage countries. This is the case in large parts of the textile and clothing industry, and also, increasingly, for accounting and customer care activities. Restructuring aims at a permanent adjustment of costs and the cost structure as a result of a change in factor prices or a permanent change of demand structure. This is accompanied by the creation or reduction of jobs, and an adjustment of the financing structure of a firm and the amount of equity and debt capital. The aim of an acquisition of, or merger with, a competitor, is to generate competitive advantage on the cost side through M&A transactions. Additionallly, other goals for mergers and acquisitions include the development of know-how, access to new markets or risk diversification. These strategic measures intend to increase or restore the profitability of a firm through cost reductions. In order to establish cost leadership as a competitive advantage, firms try to foster economies of scale or economies of scope. Costs of a firm are a significant factor influencing its competitiveness. If a firm competes with products that are, from a customer's perspective, weakly differentiated or not differentiated at all, and if price competition prevails, then the firm’s cost structure may be the decisive factor in its ability to survive. Of course, all the concepts on production decisions from ► Chapter 5 still apply, i.e., decreasing marginal products and varying returns to scale are reflected accordingly in cost functions. There is a duality between production and cost functions: both represent, in the shortand long run, transformation processes of firms. While ► Chapter 5 focusses on input factors and output, this ► Chapter 6 looks at factor prices and costs - in both cases management tries to create efficiency through minimum factor input equivalent with cost minimisation for a given or planned output. Learning Objectives This chapter deals with developing a basic understanding of which costs are relevant to business decisions from a long run and short run perspective; how the change from increasing to decreasing marginal products affects marginal costs, and how increasing and decreasing returns to scale determine the long run costs of firms; the shortand long run effects of wage rate increases and decreases in demand on firms' costs and how this impacts their strategic options; and how concepts of economies of scale and economies of scope can be used for strategic decisions on mergers and acquisitions (M&A) of firms. <?page no="216"?> 6 Costs, restructuring and M&A 216 6.1 Cost functions, decisions and competitiveness From a management perspective, it is crucial to understand the time horizon within which costs can be influenced, and by which actions and how competitive advantage can ultimately be built up. In other words, from a strategic and microeconomic perspective, only future costs are essential - whereas from the perspective of finance and controlling, an analysis or documentation of past costs in a profit and loss account or other external accounting is more important. Short run decisions determine the firm's profit via future costs. For example, a price reduction for a mobile phone offer can lead to higher demand and increasing output, however, a manager must always take into account the development of unit costs with higher output (data usage, telephone minutes, SMS, but especially also increased call centre use for customer care). Long run decisions on cost structures can create competitive advantage. If you look at some successful digital platform business models such as Google, Facebook, Airbnb, WhatsApp, Uber or Alibaba, all of these platforms are based on strategies that involve extremely high fixed costs. As a consequence, the relatively lower variable costs, with marginal costs close to zero thus, create competitive advantage (Ellison and Ellison 2005, Lambrecht et al. 2014, Probst et al. 2015, de Jong and van Dijk 2015). Future costs can always be considered as opportunity costs that are relevant for decisionmaking. These are associated with the implementation of a certain strategy and take into account the foregone opportunities of any best alternative strategy. Future costs are, therefore, used for planning, in decision calculations and in business cases. As an example, opportunity costs arise when some of the best employees from other departments are redeployed to an internal innovation project in order to build up new business. In this case, the associated decline in revenues in the existing business represents opportunity costs. Similarly, the costs of large meetings or off-sites in firms, where most participants are just listening or otherwise engaged in non-business related activities, are opportunity costs. In this instance, those people could just be working instead (Rogelberg et al. 2007). This means that cost functions developed from historical data via regressions can of course only be used for future decisions if they also apply structurally in the future. Fixed costs, sunk costs and variable costs Future costs of a firm can be described in terms of fixed costs and variable costs. Fixed costs (FC) are expenditures that are independent of the output in the short run, i.e., all costs of short run fixed input factors of the firm. Approximately, fixed costs can be equated with the costs of the input factor capital. Fixed costs can only be changed or eliminated by abandoning production and closing down the firm. As such, these costs are relevant for long run decisions, but irrelevant for decisions in the short run. For example, the costs of maintaining the Deutsche Bahn rail network are essentially fixed costs, regardless of the short run number of trains or passengers, and should therefore, be out of scope for short run planning by managers. Fixed costs can at the same time be sunk costs: sunk costs are based on historical costs associated with a decision to enter a market or choosing a specific technology and which cannot be recovered in the event of a market exit or when abandoning the technology. This applies, for example, to firmor industry-specific marketing investments to establish a mobile phone brand or path-dependent research and development expenses for a drug. <?page no="217"?> 6.1 Cost functions, decisions and competitiveness 217 Case Study │ Pharmaceutical firm and project-specific R&D-investments If we look at an example of a pharmaceutical firm’s investment decision, the role of sunk costs becomes clear. Let’s suppose the firm had originally expected possible revenues of EUR 20 billion and R&D expenditure of EUR 15 billion for a new drug. Now a competitor enters the market with a rival product and the revenue forecast is adjusted by the marketing department down to 5 billion EUR, but in the meantime 12 billion EUR have already been invested in R&D - what should the firm do now? The decision must be to continue the R&D investments. Why? Because only future revenues and costs are relevant to the decision. In this case, the EUR 5 billion in revenues that are still possible, minus EUR 3 billion in additional costs, promise a future profit of EUR 2 billion, compared to a strategy of abandoning the product. This is irrespective of the EUR 12 billion already invested, which represent irreversible sunk costs. In such cases, there are irreversible past costs that, although irrelevant to future decisions, significantly limit a firm's decision-making and constrain the firm to a certain strategy ("commitment") and indirectly influence the market structure (Sutton 1991, Münter 1999, Manez 2009 and Sibony et al. 2017). The marketing campaign for the originally planned opening of the BER airport in Berlin in 2011 can serve as another example. On the one hand, the marketing expenditures were fixed costs, independent of the number of passengers handled. On the other hand, due to the cancelled opening of the airport, these costs were irrelevant for decisions regarding subsequent planned opening dates from 2012. This case demonstrates how all past expenditures are irreversibly lost and not decisive for future decisions. However, this is where a sunk cost fallacy sets in for many managers: they try to justify previous wrong decisions through further investments (especially due to the high previous costs) instead of developing a completely new alternative or only considering future costs and revenues (Arkes and Blumer 1985, Pararye 1995, Reinstein et al. 2017, Roeder 2017 as well as ► Chapter 3). Fixed costs and sunk costs are somewhat exogenously determined by industry-specific conditions, legislation and competition policy or technology, but firms can also actively decide on their amount and extent. Substantial endogenous sunk costs are created with the aim of establishing entry barriers, product differentiation, or competitive advantage in order to influence market structure and competitive situation. Research and development and/ or marketing investments are especially good examples for such costs. (► Chapter 4). Variable costs (VC) are expenditures that depend in the short run on cost drivers such as output, number of customers, number of deliveries, or number of projects, i.e., all costs of shortterm variable input factors of production. Depending on the time horizon, all costs for labour, materials or sales commissions can be summarised here in a highly simplified manner. These are relevant to short-term decisions, e.g. for adjusting the allocation of employees to production lines or for controlling a marketing campaign. Short-term total costs (TC) of a firm are then given as (6.1) 𝑇𝑇𝐶𝐶 = 𝑇𝑇𝐶𝐶(𝑞𝑞, 𝐸𝐸, 𝐿𝐿) = 𝐹𝐹𝐶𝐶 + 𝐸𝐸𝐶𝐶 the sum of fixed and variable costs. Marginal costs (MC) correspond to the change in total costs as a result of a marginal change in the output or other cost drivers. The level of marginal <?page no="218"?> 6 Costs, restructuring and M&A 218 costs varies widely across industries. If capital resources or capacity are not fully utilised, then an additional search query at Google or an additional telephone minute at Telefónica cause almost no additional costs, i.e., marginal costs are almost zero. An additional vehicle at BMW or the production of an additional smartphone at Samsung, on the other hand, noticeably increases total costs, i.e., the marginal costs are clearly positive. However, marginal costs also depend on the business model: the marginal costs of a copy of an Australian newspaper Herald Sun in printed and digital form differ significantly due to printing costs, distribution and logistics. Mathematically, the marginal costs are calculated by deriving the total cost function (6.2) 𝑀𝑀𝐶𝐶 = ∆𝑇𝑇𝐶𝐶 ∆𝑞𝑞 for ∆→ 0: 𝑀𝑀𝐶𝐶 = 𝜕𝜕𝑇𝑇𝐶𝐶 𝜕𝜕𝑞𝑞 with respect to output 𝑞𝑞 . If costs defined in this way are related to output, we get (6.3) 𝐴𝐴𝑇𝑇𝐶𝐶 = 𝑇𝑇𝐶𝐶 𝑞𝑞 , 𝐴𝐴𝐸𝐸𝐶𝐶 = 𝑉𝑉𝐶𝐶 𝑞𝑞 , 𝐴𝐴𝐹𝐹𝐶𝐶 = 𝐹𝐹𝐶𝐶 𝑞𝑞 as average total costs (ATC, often also referred to as unit costs), average variable costs (AVC) and average fixed costs (AFC). It is apparent that average total costs can be expressed as (6.4) 𝐴𝐴𝑇𝑇𝐶𝐶 = 𝑇𝑇𝐶𝐶 𝑞𝑞 = 𝑉𝑉𝐶𝐶+𝐹𝐹𝐶𝐶 𝑞𝑞 = 𝐴𝐴𝐸𝐸𝐶𝐶 + 𝐴𝐴𝐹𝐹𝐶𝐶 . Empirical estimation of cost curves and cost function Cost curves or cost functions of individual firms can be identified or reconstructed (in a similar procedure as for production functions) by regressions on empirical cost structures from structured data taken out of profit and loss statements or balance sheets across various time periods. ► Figure 6.1 on the left shows output 𝑞𝑞 and total costs 𝑇𝑇𝐶𝐶 over the period 2002 to 2010, which at first glance do not suggest a pattern. ► Figure 6.1 on the right shows the same data adopting a microeconomic view. A total cost function is depicted that shows the relationship between output 𝑞𝑞 and total costs 𝑇𝑇𝐶𝐶 over the period 2002 to 2010. One often observes nonlinear Sshaped total cost curves, initially disproportionately low, then disproportionately high increases in total costs. Reasons for short-term nonlinear cost curves lie in production: if the production function shows a sequence of first increasing and then decreasing marginal products, then total costs first increase less than proportionally, then more than proportionally. Figure 6.1: Empirical identification of a cost function. <?page no="219"?> 6.1 Cost functions, decisions and competitiveness 219 Of course, every empirical cost function has numerous cost drivers: in addition to the output 𝑞𝑞 , in particular input quantities of capital 𝐸𝐸 and the number of employees 𝐿𝐿 as well as their respective factor prices, the wage rate 𝐻𝐻 and the weighted average cost of capital 𝑟𝑟 . However, the cost situation of each firm can be directly analysed solely through the abstract relationship between total costs 𝑇𝑇𝐶𝐶 and output 𝑞𝑞 as described in ► Figure 6.1. Case Study │ Identifying the cost function of a turbine manufacturer Looking back at the turbine manufacturer from ► Chapter 5, a close relationship between output and total costs can be seen immediately in ► Figure 6.2, top left. By dividing total costs 𝑇𝑇𝐶𝐶 by output 𝑞𝑞 , we get average total costs 𝐴𝐴𝑇𝑇𝐶𝐶 = 𝑇𝑇𝐶𝐶/ 𝑞𝑞 , as can be seen in ► Figure 6.2 top right, which are around 140,000 and are relatively constant over time. Figure 6.2: Production, total cost and average total cost as components of the cost function (top) and regression analysis of total cost (bottom). The total cost function 𝑇𝑇𝐶𝐶(𝑞𝑞) as a function of output can then be empirically estimated by regressing total costs on output. Two cases are sketched in ► Figure 6.2: the lower left linear case is obviously not plausible, despite a relatively high significance level: fixed costs would be negative - the lower right case of a nonlinear total cost function, on the other hand, is typical. <?page no="220"?> 6 Costs, restructuring and M&A 220 If we now analyse the cost situation of this turbine manufacturer using the econometrically estimated total cost function of (6.5) 𝑇𝑇𝐶𝐶(𝑞𝑞) = 𝐸𝐸𝐶𝐶 + 𝐹𝐹𝐶𝐶 = 𝜓𝜓 1 𝑞𝑞 2 + 𝜓𝜓 2 𝑞𝑞 + 𝐹𝐹𝐶𝐶 = = 32.88𝑞𝑞 2 − 31,255.40𝑞𝑞 + 217,711,520.48, then for a range of production quantities considered of approx. 2,200 to 3,600 units, fixed costs 𝐹𝐹𝐶𝐶 amount to approx. 218 million EUR, variable costs are not constant with 𝐸𝐸𝐶𝐶 = 32.88𝑞𝑞 2 − 31,255.40𝑞𝑞 and depend on the output 𝑞𝑞 and marginal costs result as 𝑀𝑀𝐶𝐶 = 𝜕𝜕𝑇𝑇𝐶𝐶/ 𝜕𝜕𝑞𝑞 = 65.76𝑞𝑞 − 31,255.40 and also vary with output. However, this also means that average variable costs are not constant, as can be seen from 𝐴𝐴𝐸𝐸𝐶𝐶 = 𝑇𝑇𝐶𝐶/ 𝑞𝑞 = 32.88𝑞𝑞 − 31,255.40 . If output is expanded, unit costs increase for this firm - contrary to the assumption of constant unit costs often observed in business cases and planning calculations. If freely available data from competitors is used in competitive analysis, their total cost functions as well as associated fixed costs, variable costs and marginal costs can be determined at least approximately by means of simple econometric procedures such as regression analyses. The crucial importance of marginal costs for decisions in strategic competition and the analysis of competitors is considered in ► Chapter 10. 6.2 Short run decisions: Fixed costs and marginal costs In the short run, output can be changed for a given capital endowment by changing labour input. For managers, it is essential to identify the variable costs that are relevant for decision-making in the short run and, in addition, to be able to estimate the change in total costs - marginal costs - when output changes. Any short-term analysis, therefore, concentrates on the question of how total costs change if, for a specific capital endowment, there are short run adjustments made to production output (as a function of marginal output of labour), or a decision can be made to allocate production to factories with different cost functions. In order to demonstrate and work out possible decisions in a short-term perspective, the cost situation of a firm shown in ► Table 6.1 serves as an example. Short-term cost structure output (1) fixed costs (2) variable costs (3) total costs (4) marginal costs (5) average total costs (6) average variable costs (7) average fixed costs (8) Q FC VC TC MC ATC=TC/ q AVC=VC/ q AFC=FC/ q 1 1000 620 1620 620 1620 620 1000 2 1000 920 1920 300 960 460 500 3 1000 1020 2020 100 673 340 333.33 <?page no="221"?> 6.2 Short run decisions: Fixed costs and marginal costs 221 4 1000 1080 2080 60 520 270 250 5 1000 1200 2200 120 440 240 200 6 1000 1440 2440 240 407 240 166.67 7 1000 1847 2847 407 407 264 142.86 8 1000 2527 3527 680 441 316 125 9 1000 3627 4627 1100 514 403 111.11 10 1000 5427 6427 1800 643 543 100 Table 6.1: Short-term cost structure (ATC, AVC and AFC partially rounded). The firm has fixed costs of 𝐹𝐹𝐶𝐶 = 1000 for an output of 1 to 10 and, due to the non-linear increase in variable costs (3), also a nonlinear increase in total costs (4) - the total cost curve is S-shaped, as can be seen in ► Figure 6.3, top left. If marginal and average costs are calculated using equations (6.2) and (6.3), we get e.g. (6.6) 𝑀𝑀𝐶𝐶 = ∆𝑇𝑇𝐶𝐶 ∆𝑞𝑞 = 2440−2200 6−5 = 240, i.e., an increase in total costs by 240 if output is increased from 5 to 6 and average total costs with an output 𝑞𝑞 = 5 of (6.7) 𝐴𝐴𝑇𝑇𝐶𝐶 = 𝑇𝑇𝐶𝐶 𝑞𝑞 = 2200 5 = 440 . If we look at marginal costs (5) as well as variable (7) and average total costs (6) in ► Table 6.1 for an increasing output, we can observe a decrease and a subsequent increase. Obviously, there is a so-called U-shaped cost structure. Average fixed costs (8), on the other hand, decrease continuously with increasing output. In order to gain a more detailed and analytical view of short run cost structures on the one hand, but also to show the typical procedure based on real data of a firm on the other hand, data of output (1) and total costs (4) are checked by means of regression analysis. The estimated equation is (6.8) 𝑇𝑇𝐶𝐶 = 𝐸𝐸𝐶𝐶 + 𝐹𝐹𝐶𝐶 = 16,432𝑞𝑞 3 − 185.95𝑞𝑞 2 + 751.61𝑞𝑞 + 1025.30 with clearly nonlinear variable costs 𝐸𝐸𝐶𝐶 = 16,432𝑞𝑞 3 − 185.95𝑞𝑞 2 + 751.61𝑞𝑞 and fixed costs 𝐹𝐹𝐶𝐶 = 1025.30 with a coefficient of determination of 𝑅𝑅 2 = 0.99 . Obviously, the regression slightly overestimates fixed costs (which are actually 𝐹𝐹𝐶𝐶 = 1000 ) with 1025.30, but nevertheless the regression curve, in ► Figure 6.3 top left, adequately represents the empirical data. To analyse the given data, one can now go two ways based on equation (6.8): one can use (6.2) to calculate the marginal costs (6.9) 𝑀𝑀𝐶𝐶 = 𝜕𝜕𝑇𝑇𝐶𝐶 𝜕𝜕𝑞𝑞 = 49,296𝑞𝑞 2 − 371.90𝑞𝑞 + 751.61 <?page no="222"?> 6 Costs, restructuring and M&A 222 Figure 6.3: Regression analysis of the short-term cost structure from Table 6.1. and, via (6.3), determine the average costs as the (6.10) 𝐴𝐴𝑇𝑇𝐶𝐶 = 𝑇𝑇𝐶𝐶 𝑞𝑞 = 16,432𝑞𝑞 2 − 185.95𝑞𝑞 + 751.61 + 1025.30 𝑞𝑞 and (6.11) 𝐴𝐴𝐸𝐸𝐶𝐶 = 𝑉𝑉𝐶𝐶 𝑞𝑞 = 16,432𝑞𝑞 2 − 185.95𝑞𝑞 + 751.61 . Due to the quadratic terms, all cost curves (6.9) to (6.10) are obviously U-shaped. If equations (6.9) and (6.10) or equations (6.9) and (6.11) are now set equal, we obtain that marginal costs are identical to average total costs for 𝑞𝑞 𝑀𝑀𝐶𝐶=𝐴𝐴𝑇𝑇𝐶𝐶 = 7 and marginal costs are identical to variable average costs for 𝑞𝑞 𝑀𝑀𝐶𝐶=𝐴𝐴𝑉𝑉𝐶𝐶 = 6 . Both values are plotted in ► Figure 6.3 bottom right. Up to these values marginal costs are below variable or average total costs, thereafter, they lie above. The underlying economic reasoning is analogous to the relationship between productivity and marginal products: if marginal costs are lower than average costs, then an expansion of production leads to a less than proportionate increase in total costs and a decrease in average costs, and vice versa. A comparison of marginal costs with average costs, as shown in ► Figure 6.4, enables immediate decision-making from a management perspective: if marginal costs are below average costs, then an increase in production leads to a reduction in average costs. This equals an increase of a firm's competitiveness that can be realised in the short run - an expansion of production at given prices directly increases profits (see also ► Chapter 7 on decisions on output based on marginal costs). <?page no="223"?> 6.2 Short run decisions: Fixed costs and marginal costs 223 Figure 6.4: Marginal costs, average costs and competitiveness. Cost minimisation and location planning at a given production level In addition to analysing whether a change in output increases competitiveness, firms often have to decide how to allocate production across various locations. Examples could be car manufacturers and their international plants, call centre operators and their allocation of employees to service lines, or even employees of a supermarket and their assignment to specific stores across a city. If products or services produced are identical, then this target can be reduced to minimising total costs for a given output (which might be determined by demand side, longterm supply contracts or service level agreements). Typical situations are those in which a given output (6.12) 𝑞𝑞 = 𝑞𝑞 𝐴𝐴 + 𝑞𝑞 𝐵𝐵 = 𝑓𝑓𝐻𝐻𝑥𝑥 is fixed, but can be produced in two locations A or B with the respective production quantities 𝑞𝑞 𝐴𝐴 and 𝑞𝑞 𝐵𝐵 . The objective of management is then a short-term minimisation of total costs (6.13) 𝑇𝑇𝐶𝐶 = 𝑇𝑇𝐶𝐶 𝐴𝐴 + 𝑇𝑇𝐶𝐶 𝐵𝐵 → 𝑆𝑆𝐻𝐻𝑆𝑆! by choosing respective production quantities 𝑞𝑞 𝐴𝐴 and 𝑞𝑞 𝐵𝐵 . Problems of optimising an objective function under a constraint can be solved using a Lagrange function. The Lagrange function is not limited to only one constraint or two production locations (and is also a built-in solution called solver in Excel, for example) and can be formulated as (6.14) 𝑍𝑍 = 𝑇𝑇𝐶𝐶 𝐴𝐴 + 𝑇𝑇𝐶𝐶 𝐵𝐵 + 𝜆𝜆(𝑞𝑞 − 𝑞𝑞 𝐴𝐴 − 𝑞𝑞 𝐵𝐵 ) → 𝑆𝑆𝐻𝐻𝑆𝑆! <?page no="224"?> 6 Costs, restructuring and M&A 224 That is, the Lagrange function combines the objective function 𝑇𝑇𝐶𝐶 𝐴𝐴 + 𝑇𝑇𝐶𝐶 𝐵𝐵 → 𝑆𝑆𝐻𝐻𝑆𝑆! with the constraint 𝑞𝑞 = 𝑞𝑞 𝐴𝐴 + 𝑞𝑞 𝐵𝐵 using the Lagrange multiplier 𝜆𝜆 . The Lagrange function is generally optimised by differentiating the objective function with respect to control variables - these are the variables that can be influenced by management and in this case are the production quantities 𝑞𝑞 𝐴𝐴 and 𝑞𝑞 𝐵𝐵 - as well as the Langrange multiplier λ, so that we obtain (6.15) 𝜕𝜕𝜕𝜕 𝜕𝜕𝑞𝑞𝐴𝐴 = 𝜕𝜕𝑇𝑇𝐶𝐶𝐴𝐴 𝜕𝜕𝑞𝑞𝐴𝐴 − 𝜆𝜆 = 𝑀𝑀𝐶𝐶 𝐴𝐴 − 𝜆𝜆 = 0 and 𝑀𝑀𝐶𝐶 𝐴𝐴 = 𝜆𝜆 (6.16) 𝜕𝜕𝜕𝜕 𝜕𝜕𝑞𝑞𝐵𝐵 = 𝜕𝜕𝑇𝑇𝐶𝐶𝐵𝐵 𝜕𝜕𝑞𝑞𝐵𝐵 − 𝜆𝜆 = 𝑀𝑀𝐶𝐶 𝐵𝐵 − 𝜆𝜆 = 0 and 𝑀𝑀𝐶𝐶 𝐵𝐵 = 𝜆𝜆 and (6.17) 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 = 𝑞𝑞 − 𝑞𝑞 𝐴𝐴 − 𝑞𝑞 𝐵𝐵 = 0 equal to 𝑞𝑞 = 𝑞𝑞 𝐴𝐴 + 𝑞𝑞 𝐵𝐵 . It follows from (6.15) and (6.16) that the Lagrange multiplier is equal to marginal cost MC and when solved for 𝜆𝜆 , it immediately follows 𝑀𝑀𝐶𝐶 𝐴𝐴 = 𝑀𝑀𝐶𝐶 𝐵𝐵 - that is, a firm optimises total costs across two (or more) locations exactly when marginal costs in these locations are equal (see in ► Chapter 5 the equating of marginal products in two bank branches to maximise output). Case Study │ Location planning for an automobile manufacturer An automobile manufacturing supplier is faced with the decision of how to divide the planned total production 𝑞𝑞 = 32,000 in the short run between two factories with different cost functions. The cost functions at these two locations 𝐴𝐴 and 𝐵𝐵 are given by (6.18) 𝑇𝑇𝐶𝐶 𝐴𝐴 = 0.6𝑞𝑞 𝐴𝐴2 + 16,000,000 and 𝑇𝑇𝐶𝐶 𝐵𝐵 = 0.2𝑞𝑞 𝐵𝐵2 + 24,000,000 , the constraint is (6.19) 𝑞𝑞 = 𝑞𝑞 𝐴𝐴 + 𝑞𝑞 𝐵𝐵 = 32,000 . In order to make a decision, the Lagrangian function is defined as (6.20) 𝑍𝑍 = 0.6𝑞𝑞 𝐴𝐴2 + 16,000,000 + 0.2𝑞𝑞 𝐵𝐵2 + 24,000,000 + 𝜆𝜆(32,000 − 𝑞𝑞 𝐴𝐴 − 𝑞𝑞 𝐵𝐵 ) → 𝑆𝑆𝐻𝐻𝑆𝑆! . If we now differentiate (6.20) with respect to the control variables 𝑞𝑞 𝐴𝐴 and 𝑞𝑞 𝐵𝐵 , we get (6.21) 𝜕𝜕𝜕𝜕 𝜕𝜕𝑞𝑞𝐴𝐴 = 1.2𝑞𝑞 𝐴𝐴 − 𝜆𝜆 = 0 with 𝑀𝑀𝐶𝐶 𝐴𝐴 = 1.2𝑞𝑞 𝐴𝐴 (6.22) 𝜕𝜕𝜕𝜕 𝜕𝜕𝑞𝑞𝐵𝐵 = 0.4𝑞𝑞 𝐵𝐵 − 𝜆𝜆 = 0 with 𝑀𝑀𝐶𝐶 𝐵𝐵 = 0.4𝑞𝑞 𝐵𝐵 as well as (6.23) 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 = 32,000 − 𝑞𝑞 𝐴𝐴 − 𝑞𝑞 𝐵𝐵 = 0 . If we now equate 𝑀𝑀𝐶𝐶 𝐴𝐴 = 𝑀𝑀𝐶𝐶 𝐵𝐵 with 1.2𝑞𝑞 𝐴𝐴 = 0.4𝑞𝑞 𝐵𝐵 , using (6.23), we find that output in factory 𝐴𝐴 must be equal to 𝑞𝑞 𝐴𝐴 = 8,000 and 𝑞𝑞 𝐵𝐵 = 24,000 in order to minimise the costs, which are then (6.24) 𝑇𝑇𝐶𝐶 = 𝑇𝑇𝐶𝐶 𝐴𝐴 + 𝑇𝑇𝐶𝐶 𝐵𝐵 = (0.6𝑞𝑞 𝐴𝐴2 + 16,000,000) + (0.2𝑞𝑞 𝐵𝐵2 + 24,000,000) = 193,600,000 . A simple comparison in ► Table 6.2 shows: if production was divided equally between the two factories, total costs would be 245 million - a potential saving of 51 million EUR or about 26% compared to the optimised cost structure based on the Lagrange function. <?page no="225"?> 6.3 Long run decisions: adjusting cost structures 225 Allocation of output and relative cost advantage strategy output fixed costs FC in mill. variable costs VC in mill. TC in mill. TC in mill. cost disadvantage marginal costs MC average total costs ATC marginal costs set equal factory A 8,000 16 38 54 194 0 % 9,600 6,800 factory B 24,000 24 115 139 9,600 5,800 50 % / 50 % allocation factory A 16,000 16 154 170 245 26 % 19,200 10,600 factory B 16,000 24 51 75 6,400 4,700 average costs set equal factory A 3,799 16 9 25 208 7 % 4,558.8 6,490 factory B 28,201 24 159 183 11,280.4 6,490 Table 6.2: Quantity strategies and relative cost advantage. Also, an allocation according to the simple heuristic "equate average costs" - regularly observed in firms (see also ► Chapter 5 on the allocation of employees in branches) - causes higher costs: even with comparatively low marginal cost differences and relatively high fixed cost blocks in this example, the cost disadvantage amounts to 16 million EUR or approx. 7%. 6.3 Long run decisions: adjusting cost structures In the long run, all input factors can be adjusted. The analysis, therefore focuses, for example, on questions of how total costs change if either of the following are taking place: a long run and permanent change in output (in relation to the nature of returns to scale), or a permanent change in absolute or relative factor prices, typically rising wage rates. This brings into view decisions concerning cost structure (and capital intensity), regarding investments in durable capital stocks, or potentially have sunk cost characteristics. All these decisions, in addition to the long-term objective of minimising costs, are suitable for ensuring or increasing the long-term competitiveness of a firm. Based on this, Porter (1980) described the concept of cost leadership in industries with weak product differentiation. In addition, the interaction between competitiveness of countries on the one hand and firms and industries on the other, can be explained based on absolute or relative cost advantages in international trade (Dosi et al. 2015, Krugman 1996 and Porter 1990). In this context, the meaning of "short-term" versus "long-term" again depends on the industry and its capital intensity - short-term is any space of time in which capital input cannot be changed or can only be changed at great expense (Levy 1994). Across industries, the possibilities to influence costs through process innovation, outsourcing, or factor substitution can differ greatly. The long run cost function of a firm is obtained by adding the cost of capital 𝑟𝑟𝐸𝐸 , i.e., <?page no="226"?> 6 Costs, restructuring and M&A 226 capital input 𝐸𝐸 multiplied by the interest rate 𝑟𝑟 , and the cost of labour 𝐻𝐻𝐿𝐿 , i.e., the labour input 𝐿𝐿 multiplied by the wage rate 𝐻𝐻 , as (6.25) 𝑇𝑇𝐶𝐶 = 𝑟𝑟𝐸𝐸 + 𝐻𝐻𝐿𝐿 with 𝐸𝐸𝐶𝐶 = 𝐻𝐻𝐿𝐿 and 𝐹𝐹𝐶𝐶 = 𝑟𝑟𝐸𝐸 . In order to optimise the cost structure, firms must now, for a given production level or planning, include factor prices, i.e., interest rates and wage rates, into consideration - especially in the long run. Changes in factor prices or international factor price differences then lead to an adjustment of the cost structure: the relative shares of labour and capital costs are adjusted. For example, in a 2015 interview, Horst Neumann, the then human resources board member of Volkswagen Group, explained and justified the long-term automation strategy due as follows: an employee working hour in Germany costs 40 EUR, in Eastern Europe 10 EUR, in China below 10 EUR, but a robot working hour including maintenance and energy, costs on average less than 6 EUR (Focus Money 2015). To look at long-term cost decisions, equation (6.25) is transferred to ► Figure 6.5. For given factor prices, long-term total costs can be drawn as an isocost line (analogous to the budget line in ► Chapter 2). The isocost line describes for given total costs 𝑇𝑇𝐶𝐶 alternative possible combinations of input factors 𝐸𝐸 and 𝐿𝐿 at given factor prices 𝐻𝐻 and 𝑟𝑟 . The intersections with the capital and labour axes indicate the maximum amount of capital and labour input if the other factor is not used - that is, the total available budget of the firm is used for only one input factor. The slope of the isocost line is determined by the relative factor prices, the so-called wage/ interest rate ratio −𝐻𝐻/ 𝑟𝑟 : with relatively increasing wage rate 𝐻𝐻 , the slope of the isocost line increases. At the same time, however, by converting (6.25) for 𝐿𝐿 0 = 0 to (6.26) 𝐸𝐸 = 𝑇𝑇𝐶𝐶 𝑝𝑝 − 𝑤𝑤𝑝𝑝 𝐿𝐿 and 𝐸𝐸 0 = 𝑇𝑇𝐶𝐶 𝑝𝑝 the absolute cost level can also be described by the intercept 𝐸𝐸 0 . Figure 6.5: Isocost line and maximum use of input factors. Suppose, for example, current factor prices for labour are 𝐻𝐻 = 12 and capital 𝑟𝑟 = 0.05 and the maximum total costs for a production is 𝑇𝑇𝐶𝐶 = 240,000 . Then, as can be seen in ► Figure 6.5 on <?page no="227"?> 6.3 Long run decisions: adjusting cost structures 227 the right, a maximum of 20,000 hours of labour can be hired or a maximum of 4.8 million EUR of capital can be financed, as well as all combinations of labour and capital at or below the isocost line. If the wage rate increases to 𝐻𝐻 = 16 and the capital cost rate simultaneously increases to 𝑟𝑟 = 0.08 with a constant budget equal to the total cost 𝑇𝑇𝐶𝐶 = 240,000, then obviously the maximum amount of capital and labour decrease. Moreover, the wage/ interest rate ratio changes from −𝐻𝐻 0 / 𝑟𝑟 0 = −240 to −𝐻𝐻 1 / 𝑟𝑟 1 = −200 and the isocost line is now flatter due to the relatively stronger increase in the interest rate. If one brings together these considerations of production (► Chapter 5) and costs and inserts isoquants and indifference curves in ► Figure 6.5, it becomes immediately clear that following an increase in factor prices at a constant budget, production must decline. Figure 6.6: Marginal rate of technical substitution and efficiency. ► Figure 6.6 shows this relationship graphically in detail: if the isoquants are convex to the origin, then two essential relationships can be derived. Firstly, for a given budget there can only be one optimum, i.e., cost-minimising, combination of capital and labour at point A, because this lies on the lowest isocost line. Secondly, at this point A the slope of the isocost line is equal to the slope of the isoquant - so that in general a condition of long-term cost-minimising production is fulfilled if the factor price ratio is equal to the ratio of the marginal products of the input factors, i.e., the so-called marginal rate of substitution. Point B would be inefficient because, on the one hand, higher costs are incurred (as can be seen from the higher isocost line) and, on the other hand, at the given factor price ratio the factor labour is used too much and the factor capital too little. This firm can increase efficiency by restructuring and using more capital and less labour - i.e., by reducing jobs and investing more in automation. In general, the relationships in ► Figures 6.5 and 6.6 can also be described mathematically. If any long-term production function (6.27) 𝑞𝑞 = 𝑇𝑇(𝐸𝐸, 𝐿𝐿) is given, then, based on the total differential (6.28) 𝑑𝑑𝑞𝑞 = 𝜕𝜕𝑇𝑇 𝜕𝜕𝐸𝐸 𝑑𝑑𝐸𝐸 + 𝜕𝜕𝑇𝑇 𝜕𝜕𝐿𝐿 𝑑𝑑𝐿𝐿 <?page no="228"?> 6 Costs, restructuring and M&A 228 the slope along an isoquant on which 𝑑𝑑𝑞𝑞 = 0 applies, as the marginal rate of technical substitution MRT described by the ratio (6.29) 𝑀𝑀𝑅𝑅𝑇𝑇 = − 𝑑𝑑𝐸𝐸 𝑑𝑑𝐿𝐿 = 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 = 𝑀𝑀𝜕𝜕𝜕𝜕 𝑀𝑀𝜕𝜕𝜕𝜕 . The marginal rate of technical substitution describes the current exchange ratio of the input factors measured by the respective marginal products of labour 𝑀𝑀𝑃𝑃 𝐿𝐿 = 𝜕𝜕𝑇𝑇/ 𝜕𝜕𝐿𝐿 and capital 𝑀𝑀𝑃𝑃 𝐸𝐸 = 𝜕𝜕𝐹𝐹/ 𝜕𝜕𝐸𝐸 . It measures what change in capital input 𝑑𝑑𝐸𝐸 is necessary to keep output constant in case of a change in labour input 𝑑𝑑𝐿𝐿 . The ease with which this adjustment can be made depends again on the technology used by the firm. Since the slope of the isocost line −𝐻𝐻/ 𝑟𝑟 is equal to the slope of the isoquant 𝑀𝑀𝑃𝑃 𝐿𝐿 / 𝑀𝑀𝑃𝑃 𝐸𝐸 for efficient and cost-minimising production, because of (6.30) − 𝑤𝑤𝑝𝑝 = 𝑀𝑀𝜕𝜕𝜕𝜕 𝑀𝑀𝜕𝜕𝜕𝜕 at a cost minimum, the ratio of marginal products corresponds to the ratio of factor prices (► Chapter 2 for analogous results concerning the price ratios of products and relative marginal utility). To illustrate these long-term adjustment processes, two typical examples are considered below. First, from a management perspective, it is analysed how firms react to a permanent decline in demand, then possibilities for reaction following a permanent increase in a factor price, typically the wage rate 𝐻𝐻 , are considered. Restructuring given a permanent decrease in production Firms regularly have to take actions to optimise costs when demand declines permanently and in the long run. Osram, as a manufacturer of light bulbs, has been repeatedly affected by new requirements, bans and restrictions on the sale and production of conventional light bulbs since 2012, and has consequently made massive job cuts and closed plants (Augsburger Allgemeine Zeitung 2015 and Handelsblatt 2017). Similarly, firms such as Siemens and General Electric have been forced to cut jobs and restructure their business following various national and international agreements on the energy transition and a sharp decline in demand for gas and steam turbines (Köhn 2017). Restructuring describes a reorganisation of the firm, in which both capital and labour are adapted to the new business environment, often associated with a renewal of the business model, relocation of production, or new shareholders. Firms have to reduce production as a result of a permanent decline in demand, with constant factor prices. According to the decline in production, the quantities of capital and labour used must be reduced in order to cut costs and create efficiency. Alternatively, management can try to enforce a reduction in factor prices - e.g. by renegotiating employee salaries or interest rates for loans or credits. <?page no="229"?> 6.3 Long run decisions: adjusting cost structures 229 Figure 6.7: Long-term perspective in case of production decline. ► Figure 6.7 shows that typically the number of jobs is reduced in the short run, and if demand continues to decline, capital input (equity and debt) is also reduced in the long run: in an initial situation (1) the firm produces an output 𝑞𝑞 0 at a total cost of 𝑇𝑇𝐶𝐶 0 with a capital input 𝐸𝐸 0 and a labour input 𝐿𝐿 0 - production is efficient because the isocost line is tangent to the isoquant at point 𝐴𝐴 ; subsequently, a decline in production from 𝑞𝑞 0 to 𝑞𝑞 1 (2) triggered by a permanent decline in demand, production is no longer efficient - the isoquant 𝑞𝑞 1 intersects the isocost line - and so typically too much capital and labour is employed; in the short run, capital input 𝐸𝐸 0 cannot be reduced, so the firm's short run response is only to cut jobs to 𝐿𝐿 0 ′ (3) - in fact, costs do fall to 𝑇𝑇𝐶𝐶 0′ , but production is inefficient because at point 𝐴𝐴′ the shifted isocost line continues to intersect isoquant 𝑞𝑞 1 ; in the long run, the firm can reduce capital inputs to 𝐸𝐸 1 (4), so that at point 𝐵𝐵 production is now efficient again - this is accompanied by a further reduction in costs from 𝑇𝑇𝐶𝐶 0′ to 𝑇𝑇𝐶𝐶 1 . As a result of a reduced output, input of both input factors 𝐸𝐸 and 𝐿𝐿 is reduced; in any case, at constant factor prices, the total costs are also reduced in the long run. This is usually accompanied by a change in capital intensity, yet direction of change depends on the actual production function and the substitutability of the input factors. <?page no="230"?> 6 Costs, restructuring and M&A 230 Case Study │ Slow-down of production at a light bulb manufacturer The production function of a light bulb manufacturer is described as a (6.31) 𝑞𝑞 = 20𝐸𝐸 0.05 𝐿𝐿 0.9 , the firm produces 𝑞𝑞 0 = 3.2 billion light bulbs. Due to a change in legislation, sales of conventional light bulbs collapse. The firm expects a permanent decline to 𝑞𝑞 1 = 2.0 billion light bulbs at constant factor prices of 𝑟𝑟 = 0.02 and 𝐻𝐻 = 20 . The situation can be depicted analogously to ► Figure 6.6 - but in addition to some qualitative analysis, a quantitative analysis can also be carried out. This is necessary in everyday business for at least two reasons: in order to implement the necessary job cuts, a reconciliation of interests and a social-compensation plan typically have to be negotiated between workers’ representative and employer representative - a robust and comprehensible set of figures can make a significant contribution to objectification here; a reduction of capital employed means a reduction in equity and debt capital - here, too, a reliable set of numbers helps to more accurately inform banks, previous shareholders, and the capital market about the restructuring of the firm. The quantitative analysis is carried out using a Lagrange function, as again an optimisation must be carried out under constraints. First, the cost minimum and the corresponding capital and labour input must be determined for the initial situation, so that under the constraint of a current production of (6.32) 𝑞𝑞 = 20𝐸𝐸 0.05 𝐿𝐿 0.9 with 𝑞𝑞 0 = 3.2 at an interest rate of 𝑟𝑟 = 0.02 and an hourly wage of 𝐻𝐻 = 20 to optimise the cost situation (6.33) 𝑇𝑇𝐶𝐶 = 𝑟𝑟𝐸𝐸 0 + 𝐻𝐻𝐿𝐿 0 = 0.02𝐸𝐸 0 + 20𝐿𝐿 0 using a Lagrange-function (6.34) 𝑍𝑍 0 = 0.02𝐸𝐸 0 + 20𝐿𝐿 0 + 𝜆𝜆�3.2 − 20𝐸𝐸 00.05 𝐿𝐿 00.9 � → 𝑆𝑆𝐻𝐻𝑆𝑆! . This is optimised by choosing the strategy parameters (control variables) 𝐸𝐸 0 and 𝐿𝐿 0 that can be influenced by the firm, so that (6.35) 𝜕𝜕𝜕𝜕0 𝜕𝜕𝐸𝐸0 = 0.02 − 𝜆𝜆�20 ⋅ 0.05𝐸𝐸 0−0.95 𝐿𝐿 00.9 � = 0 → 𝜆𝜆 = 0.02 20⋅0.05𝐸𝐸0−0.95𝐿𝐿00.9 (6.36) 𝜕𝜕𝜕𝜕0 𝜕𝜕𝐿𝐿0 = 20 − 𝜆𝜆�20 ⋅ 0.9𝐸𝐸 00.05 𝐿𝐿 0−0.1 � = 0 → 𝜆𝜆 = 20 20⋅0.9𝐸𝐸00.05𝐿𝐿0−0.1 (6.37) 𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 = 3.2 − 20𝐸𝐸 00.05 𝐿𝐿 00.9 = 0 must be fulfilled. If one now equates 𝜆𝜆 from (6.35) and (6.36), then it follows that (6.38) 0.02 20⋅0.05𝐸𝐸0−0.95𝐿𝐿00.9 = 20 20⋅0.9𝐸𝐸00.05𝐿𝐿0−0.1 → 𝐸𝐸 0 = 55.55𝐿𝐿 0 and (6.39) 𝑞𝑞 0 = 3.2 = 20𝐸𝐸 00.05 𝐿𝐿 00.9 = 20(55.55𝐿𝐿 0 ) 0.05 𝐿𝐿 00.9 the optimal input of labour and capital in the initial situation as (6.40) 𝐿𝐿 0 = � 3.2 20⋅55.550.05 0.95 = 0.1176 and 𝐸𝐸 0 = 55.55𝐿𝐿 0 = 6.5332. so that, at given factor prices, total costs amount to (6.41) 𝑇𝑇𝐶𝐶 0 = 𝑟𝑟𝐸𝐸 0 + 𝐻𝐻𝐿𝐿 0 = (0.02 ⋅ 6.5332) + (20 ⋅ 0.1176) = 2.4826 . In other words: the production of 3.2 billion light bulbs requires a capital input of <?page no="231"?> 6.3 Long run decisions: adjusting cost structures 231 6.5 billion EUR and 0.12 billion hours of labour (at 1,700 working hours per employee per year this equals about 69,000 jobs) and the total costs amount to 2.48 billion EUR - the total average cost of each light bulb is accordingly 0.78 EUR. In the second step, the effect of a decrease in production to 𝑞𝑞 1 = 2.0 on the number of jobs 𝐿𝐿 0 ′ of the firm must be determined - under the condition that capital input is constant in the short run at 𝐸𝐸 = 6.5332 . Using (6.42) 𝑞𝑞 1 = 2.0 = 𝐴𝐴𝐸𝐸 0𝛼𝛼 𝐿𝐿 0 ′ 𝛽𝛽 = 20 ⋅ 6.5332 0.05 𝐿𝐿 0 ′ 0.9 we get that the number of working hours reduces to (6.43) 𝐿𝐿 0 ′ = � 2.0 20⋅6.53320.05 0.9 = 0.0697 . Figure 6.8: Determining the shortand long-term optimization for the light bulb manufacturer. <?page no="232"?> 6 Costs, restructuring and M&A 232 This means, given 1,700 working hours per employee per year, a reduction to a target number of approx. 41,000 jobs, so that the short run reduction in jobs amounts to approx. 28,000 jobs. The total costs in the short run decrease to (6.44) 𝑇𝑇𝐶𝐶 0′ = 0.02𝐸𝐸 0 + 20𝐿𝐿 0 ′ = 1.5258 , compared to 𝑇𝑇𝐶𝐶 0 = 2.4826, a saving of almost 1 billion EUR. In the third and final step, the long-term cost reduction with a reduction in capital input is now determined. The analysis is identical to (6.36) to (6.44) for the now new long-term reduced output 𝑞𝑞 1 = 2.0 , so that via the Lagrange function (6.45) 𝑍𝑍 1 = 0.02𝐸𝐸 1 + 20𝐿𝐿 1 + 𝜆𝜆�𝑞𝑞 1 − 20𝐸𝐸 10.05 𝐿𝐿 10.9 � → 𝑆𝑆𝐻𝐻𝑆𝑆! the long-term optimal use of input factors results as (6.46) 𝐿𝐿 1 = � 2.0 20⋅55.550.05 0.95 = 0.0717 and 𝐸𝐸 1 = 55.55𝐿𝐿 1 = 3.9835 . Compared to the short-term optimisation, the labour input now slightly increases again; the reason for this is the reduction of capital input by about 1.6 billion EUR. This is accompanied by a further - albeit small - reduction in costs, so that the long-term total costs (6.47) 𝑇𝑇𝐶𝐶 1 = 0.02𝐸𝐸 1 + 20𝐿𝐿 1 = 1.5137 would result in an additional cost optimisation of EUR 12 million. Spreadsheet software such as Excel can be used to calculate these optimisations quickly and in various scenarios, as can be seen in ► Figure 6.8. In fact, the 37.50% reduction in production would lead to a permanent reduction of about 26,997 FTE (from 69,175 to 42,179 FTE), about 39.03%. The total costs decrease in the long run from 2.48 billion EUR by approx. 0.97 billion EUR to approx. 1.51 billion EUR, i.e., by approx. 39.03%. This more than proportionate decrease in total costs is due to decreasing returns to scale of the production function with 𝛼𝛼 + 𝛽𝛽 = 0.95. so that labour and capital can be more than proportionately reduced when production decreases. Restructuring following an increase of wages Many firms have to optimise the efficiency of production via substituting capital for labour as a result of an increase in wages (e.g. as a result of new collective labour agreements with the trade unions) at constant output. For example, Deutsche Bahn and its subsidiary DB Schenker Rail announced in 2015, as a direct response to a new collective labour agreement, that it would cut up to 5,000 jobs in the group and replace with higher capital input (Die Zeit 2015). Typically, as a result of a wage rate increase, total costs will rise immediately at constant output levels. According to the adjusted wage/ interest rate ratio, labour has now become more expensive relative to capital - so the wage rate increase must be followed in the long run by an adjustment of capital and labour in order to reduce costs. Since the relative costs of labour have risen in comparison to capital, automation possibilities are sought; in the medium term, the number of jobs is reduced and the use of capital is increased. ► Figure 6.9 shows these relationships schematically: in an initial situation (1), with an initial wage/ interest rate ratio −𝐻𝐻 0 / 𝑟𝑟 0 , the firm produces an output 𝑞𝑞 0 with a capital input 𝐸𝐸 0 and a labour input 𝐿𝐿 0 at a total cost of 𝑇𝑇𝐶𝐶 0 - production is efficient, since the isocost line is tangent to the isoquant at point 𝐴𝐴 ; <?page no="233"?> 6.3 Long run decisions: adjusting cost structures 233 an increase in the wage/ interest rate ratio (2) from −𝐻𝐻 0 / 𝑟𝑟 0 to −𝐻𝐻 1 / 𝑟𝑟 0 for a given output results in a rotation of the isocost line at point 𝐴𝐴 and immediately in an increase in total costs to 𝑇𝑇𝐶𝐶 0 ′ - production is now inefficient since the isocost line intersects the isoquant at point 𝐴𝐴 and the firm cannot react since substitution of labour by capital is impossible in the short run; in the long run (3) the firm can increase the capital stock from 𝐸𝐸 0 to 𝐸𝐸 1 and at the same time reduce the number of jobs from 𝐿𝐿 0 to 𝐿𝐿 1 - production is now efficient again, since at point 𝐵𝐵 the isocost line with slope −𝐻𝐻 1 / 𝑟𝑟 0 , which has been shifted parallel to the left, is now tangent to the isoquant; this is accompanied by a cost reduction (4) from 𝑇𝑇𝐶𝐶 0′ to 𝑇𝑇𝐶𝐶 1 , which can be read from the capital axis intercept. Figure 6.9: Long-term perspective with wage rate increases. Case Study │ Automation following increasing wages at a railroad firm The production function of a railroad transport service provider is given as (6.48) 𝑞𝑞 = 12𝐸𝐸 0.4 𝐿𝐿 0.7 . The firm produces 2.4 billion ton-kilometres of transport based on long-term contracts. The trade unions achieve an increase in the hourly wage from 𝐻𝐻 0 = 22 to 𝐻𝐻 1 = 24, the cost of capital remains constant at 𝑟𝑟 = 0.05 . The firm must now determine shortand long-term actions concerning capital and labour input as well as total costs if output continues to be 2.4 billion ton-kilometres in accordance with the contract. <?page no="234"?> 6 Costs, restructuring and M&A 234 With a Lagrange function one can map the analysis quantitatively, so that given an initial situation of (6.49) 𝑞𝑞 = 𝐴𝐴𝐸𝐸 0𝛼𝛼 𝐿𝐿 0𝛽𝛽 = 12𝐸𝐸 00.4 𝐿𝐿 00.7 = 2.4 and (6.50) 𝑇𝑇𝐶𝐶 0 = 𝑟𝑟𝐸𝐸 0 + 𝐻𝐻 0 𝐿𝐿 0 = 0.05𝐸𝐸 0 + 22𝐿𝐿 0 a Lagrangian function can be developed to be minimised by choosing strategic parameters 𝐸𝐸 and 𝐿𝐿 (6.51) 𝑍𝑍 0 = 0.05𝐸𝐸 + 22𝐿𝐿 + 𝜆𝜆(12𝐸𝐸 0.4 𝐿𝐿 0.7 − 2.4) → 𝑆𝑆𝐻𝐻𝑆𝑆! . Figure 6.10: Determining the shortand long-term optimization for the railway firm. After differentiation of (6.51) and solving for 𝐸𝐸 and 𝐿𝐿 , the original costs and the factor input result as (6.52) 𝐿𝐿 0 ≈ 0.031 and 𝐸𝐸 0 ≈ 7.80 with 𝑇𝑇𝐶𝐶 0 = 0.05𝐸𝐸 0 + 22𝐿𝐿 0 ≈ 1.072 . <?page no="235"?> 6.4 Cost-based competitive advantage and M&A 235 If the wage rate now increases to 𝐻𝐻 1 = 24, then with values of 𝐿𝐿 0 ≈ 0.031 and 𝐸𝐸 0 ≈ 7.80 which continue to be given, the new total costs 𝑇𝑇𝐶𝐶 0′ result as (6.53) 𝑇𝑇𝐶𝐶 0′ = 0.05𝐸𝐸 0 + 24𝐿𝐿 0 ≈ 1.134 this means that the wage rate increase immediately increases total costs by 62 million EUR to 1.134 billion EUR. The firm will now try - for example through further automation - to substitute jobs with higher capital input. The extent of the restructuring can now again be measured by optimising the total costs using a Lagrangian function, so that via (6.54) 𝑍𝑍 1 = 0.05𝐸𝐸 1 + 24𝐿𝐿 1 + 𝜆𝜆(12𝐸𝐸 10.4 𝐿𝐿 10.7 − 2.4) → 𝑆𝑆𝐻𝐻𝑆𝑆! new optimum factor inputs and total costs of (6.55) 𝐿𝐿 1 ≈ 0.0300 and 𝐸𝐸 1 ≈ 8.244 with 𝑇𝑇𝐶𝐶 1 = 0.05𝐸𝐸 1 + 24𝐿𝐿 1 = 1.133 result. The total costs can only be reduced by approx. 10 million EUR with the given production function. ► Figure 6.10 summarises all the results and their effects. In fact, the wage rate increase of about 9.09% would lead to a reduction of about 568 FTE, about -3.11%. Total costs increase in the long run by approx. 61 million EUR, approx. 0.6%. 6.4 Cost-based competitive advantage and M&A Firms can strategically transform cost structures into competitive advantage along two dimensions, both in the long and short run: economies of scale and economies of scope. Both strategies are driving forces in many business models and are integral to multisided digital platforms such as TripAdvisor, LinkedIn and Amazon (Evans and Schmalensee 2016 and ► Chapter 2). They help to understand growth of firms within an industry and how diverse the product portfolio is (Chandler 1990). In addition, both concepts offer possible rationales for mergers or acquisitions. The aim here is to realise synergies from size (economies of scale) or synergies from diversification (economies of scope). Economies of scale are firm-specific competitive advantage in terms of costs - average total costs or unit costs fall with increasing firm size or output: if a firm expands production by 80% but total costs only rise by 60%, we have economies of scale. Causes can be high fixed costs (e.g. due to indivisibilities in the innovation process, in marketing or within the firm’s organisation), increasing returns to scale in production or the use of big data. A particular competitive advantage can arise over time: if a firm's average costs fall as a result of cumulative production quantities over the years, then learning curve effects (experience and learning-by-doing) occur. These competitive advantages based on routines can, for example, lie in higher working speed or reduced waste. A firm with learning curve effects then has a competitive advantage over a newly entering firm without the corresponding experience and cumulative output. A well-established rule of thumb is the 80% rule identified by Dutton and Thomas (1984) over a large number of empirical studies, i.e., when cumulative production doubles, average costs fall to approximately 80% of the previous level. However, these learning curve effects fade if production is suspended or the technology used is changed (Spence 1981, Argote and Epple 1990 and Malerba 1992). <?page no="236"?> 6 Costs, restructuring and M&A 236 Good to know │ Which is better - economies of scale or learning curve effects? When firms want to achieve cost-based competitive advantage, economies of scale and learning curve effects must always be taken into account. From a management perspective, these effects are relevant for new product launches and target cost planning as well as for catching up with cost disadvantages vis-à-vis competitors. Figure 6.11: Economies of scale versus learning curve effects. ► Figure 6.11 on the left shows the decreasing shape of the average total costs of a manufacturer of graphic processing units. With a current output of 𝑞𝑞 = 200, the average cost is 𝐴𝐴𝑇𝑇𝐶𝐶 = 10 . If output is doubled to 𝑞𝑞 = 400 in the coming year, then average cost falls to 𝐴𝐴𝑇𝑇𝐶𝐶 = 9 along the existing cost curve due to economies of scale. Alternatively, if the firm can achieve learning curve effects, the cost curve shifts downwards in parallel. The average costs then fall to 𝐴𝐴𝑇𝑇𝐶𝐶 = 8 in the coming year due to experience even if output remains the same. If both effects come together - the cost curve shifts downwards and the firm becomes larger - then costs are 𝐴𝐴𝑇𝑇𝐶𝐶 = 7, i.e., both effects add up. Which of the two effects reduces costs stronger in reality depends on the actual shape of the cost curve and the firm's growth opportunities. Economies of scope ("advantages from diversification") are also firm-specific competitive advantage on the cost side: they are based on the fact that with increasing diversity of the product portfolio, the average total costs of the individual product types decline due to synergies from diversification. If a consumer goods firm expands production to include toothpaste, and the average cost of producing detergent therefore declines, economies of scope occur. Causes can again be high fixed costs (establishing umbrella brands or leveraging an existing distribution structure); jointly used basic technology (production technology or flexible product platforms such as in the automotive or computer hardware industry); risk diversification through negatively correlated cost development across R&D projects; joint production (for example in the chemical or pharmaceutical industry); or <?page no="237"?> 6.4 Cost-based competitive advantage and M&A 237 know-how or patents applicable to different product types or segments as well as complementary use-cases of the customers. Industries and firms sometimes differ significantly with regard to the importance of economies of scale and scope. These are determined (based on returns of scale) by technological, production-side and cost-side conditions. Thus, numerous business models are fundamentally based on economies of scale and scope. With increasing size, banks can realise fixed-cost-intensive brands, IT platforms or global trading networks at lower average costs; cost-reducing synergies and competitive advantage arise from the combination of lending business (lending activity and corporate finance) and deposit business as well as transaction banking (payment transactions or securities settlement) (Altunbas and Molyneux 1996). In the case of automobile manufacturers - apart from absolute economies of scale - economies of scope are created by the cross-model and cross-series production of engines, powertrains and interiors. At BMW, the share of common parts within diesel or petrol engines is larger than 60 % in each case, while the share of common parts across powertrains is still more than 30 % (BMW 2015). In the consumer goods/ FMCG industry, firms such as Procter&Gamble or Unilever also aim for economies of scale as well as economies of scope. Strong drivers here are fixed marketing/ branding expenses as well as shared B2B sales structures, logistics and production platforms across the products, which ultimately continues into B2C sales at large discounters such as Aldi, Walmart or Carrefour. Alternatively, firms can actively aim to build or scale multisided markets through economies of scale and scope (Hagiu and Wright 2015). Alphabet, the parent holding of Google, addresses both economies of scale (big data in the respective Google or YouTube business models) and economies of scope (linking data across the Android, Google or Chrome business models). Amazon realises its own economies of scale and scope via absolute size and product portfolio and scales these by integrating third-party providers on Amazon Marketplaces and by providing the Amazon Payment and fulfilment platform. LinkedIn, a career platform and subsidiary of Microsoft, achieves economies of scale on its fixed-cost-intensive IT platform from an increasing number of users. In addition, services from partner firms are increasingly achieving economies of scope through portfolio expansions. Economies of scale, minimum efficient size and market structure Economies of scale are not only relevant for individual firms. If firms in an industry use similar technology, then the S-shaped total cost curve described in ► Section 6.3 - as shown schematically in ► Figure 6.12 bottom left - applies to an entire industry and industry cost curves can be identified. Firms then take up positions in different areas of the cost curve according to their output and firm size and can derive strategic implications from their respective positioning. With an expansion of the output, total costs increase. This can occur proportionally, less than or more than proportionally. The effect of a change in output on total costs can be measured by the total cost elasticity <?page no="238"?> 6 Costs, restructuring and M&A 238 (6.56) 𝐴𝐴 𝑇𝑇𝐶𝐶 = ∆𝑇𝑇𝐶𝐶 𝑇𝑇𝐶𝐶 ⁄ ∆𝑞𝑞 𝑞𝑞 ⁄ = ∆𝑇𝑇𝐶𝐶 ∆𝑞𝑞 ⁄ 𝑇𝑇𝐶𝐶/ 𝑞𝑞 = 𝑀𝑀𝐶𝐶 𝐴𝐴𝑇𝑇𝐶𝐶 with 𝐴𝐴 𝑇𝑇𝐶𝐶 ≷ 1 for 𝑀𝑀𝐶𝐶 ≷ 𝐴𝐴𝑇𝑇𝐶𝐶 , i.e., the ratio of marginal costs to average total costs. Economies of scale are present if total costs increase less than proportionately: this is always the case if marginal costs 𝑀𝑀𝐶𝐶 are lower than average total costs 𝐴𝐴𝑇𝑇𝐶𝐶 - firms achieve a competitive advantage from growth and size and 𝐴𝐴 𝑇𝑇𝐶𝐶 < 1 . Diseconomies of scale occur if costs increase disproportionately and 𝐴𝐴 𝑇𝑇𝐶𝐶 > 1, i.e., marginal costs 𝑀𝑀𝐶𝐶 are higher than average total costs 𝐴𝐴𝑇𝑇𝐶𝐶 . Figure 6.12: Schematic industry cost curve and economies of scale. ► Figure 6.12 above shows - analogous to economies of scale from ► Chapter 5 - the effect of a change in output on total costs. If there are economies of scale, an increase in the output leads to a less than proportionate increase, but a decrease in the output also leads to a less than proportionately small decrease in costs. If we have diseconomies of scale, then an increase in output results in a more than proportionate increase in total costs, so that a more than proportionate reduction in costs is also possible in the event of a decrease in output. Scale elasticity can also be seen for the average cost curve in ► Figure 6.12 bottom right. Typically, the total average cost curve for many industries then is bathtub-shaped. Up to a size 𝑞𝑞 0 , average costs decrease and competitive advantage arise from growth; from 𝑞𝑞 1 onwards, cost disadvantages from increasing size can be observed. A size 𝑞𝑞 0 is called the minimum efficient size (MES). A <?page no="239"?> 6.4 Cost-based competitive advantage and M&A 239 firm with a smaller size than 𝑞𝑞 0 thus, has cost disadvantages - in empirical studies, average cost disadvantages of 15% result if a firm only achieves a quarter of the required minimum firm size (Weiss 1975). In addition, statements about the market structure, i.e., the size and size distribution of firms in an industry, can be made via the minimum firm size and economies of scale (Münter 1999). ► Figure 6.13 schematically shows two industries with their respective bathtub-shaped industry cost curves: on the left-hand side there is a relatively large minimum efficient size and a relatively wide range of optimum firm sizes, i.e., the bottom of the bathtub is strongly developed - firms differ strongly in size and market shares; the right-hand side shows a relatively small minimum efficient size and a relatively narrow range of optimum firm sizes, i.e., the bottom of the bathtub is not very pronounced - firms hardly differ in size and market shares. Figure 6.13: Minimum efficient size and market structure. The minimum efficient size in relation to the size of the market, i.e., a minimum market share, also gives an indication of the number of firms to be expected in a market (see also ► Chapter 10 for further details). ► Table 6.3 shows this minimum market share for some US industries. The inverse of the minimum market share provides a rough guide to the real number of firms to be expected in this industry. <?page no="240"?> 6 Costs, restructuring and M&A 240 industry MES as % of market size industry MES as % of market size beet sugar 1.87 breakfast cereals 9.47 cane sugar 12.01 bottled water 0.08 flour 0.68 coffee 5.82 bread 0.12 pet food 3.02 canned vegetables 0.17 baby food 2.59 frozen food 0.92 beer 1.37 Table 6.3: Minimum firm size in relation to total market production. (Source: Sutton 1991, p. 393 ff.). ► Figure 6.14 shows schematically the industry cost curves for the automotive and beer industries in order to identify strategic implications. For the automobile industry, it is obvious that brands such as Tesla or Porsche do not reach a minimum efficient size, but are viable even at high average costs due to high pricing margins. In the beer industry, although the merged AB Inbev / SAB Miller group has grown to about 29% market share by 2020, this has not been associated with a reduction in average costs at a minimum market share of 𝑆𝑆 = 𝑞𝑞/ 𝑄𝑄 = 1.37% - rather, the group now comes close to an upper limit of about 45%, above which average costs could start to rise again (Tremblay and Tremblay 2005 and Barth-Haas Group 2020). Figure 6.14: Market structure and industry cost curves (schematic). <?page no="241"?> 6.4 Cost-based competitive advantage and M&A 241 However, significant economies of scale in an industry do not necessarily mean just one or only a few firms are able to survive. In the global beer industry, there are only a few very large firms with AB InBev, SAB Miller, China Snow and Heineken (all of them with a multitude of associated brands), but in parallel there are also many very small suppliers. In addition to horizontal product differentiation based on marketing and regional customer preferences, coexistence is ensured in particular by sufficiently high prices. Figure 6.15: Economies of scale versus "small is smart". However, besides the typical bathtub-shape of an industry-specific cost curve, the extreme cases of the cost curves in ► Figure 6.15 are also possible. On the left side, the case is sketched that the average total costs 𝐴𝐴𝑇𝑇𝐶𝐶 decrease continuously with increasing output. The larger a firm is, the lower its average costs. If the price now falls from 𝑝𝑝 to 𝑝𝑝′ , only the larger of the two firms with the output 𝑞𝑞 2 is able to cover its average total costs. Therefore, the smaller firm at 𝑞𝑞 1 must either grow or leave the market. This allows the largest firm to strategically force other competitors out of the market, creating a tendency towards a natural monopoly. This option is illustrated in ► Figure 6.16 for the search engine industry: Google has by far the largest market share, e.g. compared to Bing and Yahoo. The cost situation is characterised by high fixed costs and constant marginal costs per search query close to zero, so that the total average cost 𝐴𝐴𝑇𝑇𝐶𝐶 is continuously decreasing. The competitors' business models are based on the sale or auction of advertising space. As can be seen in ► Figure 6.16, Google can set prices 𝑝𝑝 below the level of the competitors' average total costs: for advertisers this means great market access at low prices, for competitors this means continuous losses and, without further measures, exit from the market. <?page no="242"?> 6 Costs, restructuring and M&A 242 Figure 6.16: Search engine market structure (schematic). With permanently falling 𝐴𝐴𝑇𝑇𝐶𝐶 , size drives efficiency and competitive advantage, and industries can become natural monopolies: this is typical in industries with extremely high fixed costs due to the setup and maintenance of infrastructure such as railways, energy, or telecommunications. Competition is impossible and such industries are often regulated by governmental intervention - a challenge that also exists for digital markets (see also ► Chapter 7). Conversely, as output increases, there may be continuously increasing average costs 𝐴𝐴𝑇𝑇𝐶𝐶 : Size has disadvantages ("small is smart") and in these industries there are typically many but in turn small businesses: restaurants, rock bands, or authors benefit only to a limited extent from increasing size and the survivability of organisations decline as firm size increases. Economies of scale and scope as driving forces behind mergers and acquisitions If economies of scale cannot be achieved through organic growth based on marketing or sales strategies, then inorganic growth can alternatively be addressed to achieve economies of scale through mergers with competitors. In many industries, firms try to build competitive advantage through mergers with or acquisitions of competitors (Jansen 2016 and Trautwein 1990). Here, the focus is on three possible sets of activities: an acquisition of new capabilities (e.g. employees, patents or technologies) and core capabilities; building up or expanding market power to control strategic parameters (e.g., pricing power on the sales or procurement side or regional exclusivity); and efficiency gains by achieving economies of scale and economies of scope (essentially the reduction of fixed costs) or by achieving transaction cost advantages (changing firm boundaries, simplifying organisation, and optimising internal coordination and communication). <?page no="243"?> 6.4 Cost-based competitive advantage and M&A 243 Figure 6.17: Types of mergers. Mergers can then be classified as shown in ► Figure 6.17. Horizontal mergers - e.g. the takeover of Tengelmann by Edeka or the acquisition of the E-Plus Group by Telefónica - are based on the consolidation of similar business models or address one or more shared elements of the value chain of an industry in a market. The aim is to reduce the intensity of competition, realise economies of scale to reduce costs and consequently to build up or expand market power. Figure 6.18: Logic of the Telefónica and E-Plus merger. <?page no="244"?> 6 Costs, restructuring and M&A 244 ► Figure 6.18 shows the logic of the merger in the German mobile network operator (MNO) market of Telefónica and the E-Plus Group in simplified form. Due to their relative cost disadvantage compared to the by then market leaders Vodafone and Deutsche Telekom, the profitability of the two small providers was increasingly threatened by decreasing prices. Through the merger and the size thus achieved, the new Telefónica can significantly reduce average total costs. In particular fixed costs can be eliminated by reducing duplicate functions in administrative areas, merging the IT and mobile network, and cutting marketing expenditure as a result of decreased competitive intensity. Vertical mergers take place along the value chain of an industry through strategic forward or backward integration with the aim of improving control of the value chain and reducing transaction costs. In this way, firm boundaries are adjusted, for example through outor insourcing, and internal firm costs are reduced (see also ► Chapter 4). ► Figure 6.19 shows such a vertical merger between a retail bank and a fintech service provider: in 2019, Deutsche Bank acquired a 5% share in the fintech Deposit Solutions. The former supplier relationship for interest rate products was thus, stabilised, in particular in order to acquire know-how, improve purchasing conditions and, if necessary, make it more difficult for Deutsche Bank's competitors to access Deposit Solutions. Figure 6.19: Classification of mergers. Conglomerate mergers take place across industry or market boundaries with the aim of achieving diversification through the expansion of the product portfolio or the establishment of a diversified group and usually a reduction of risk in negatively correlated portfolios (► Chapter 2) as well as the realisation of economies of scope through the joint use of key resources. ► Figure 6.19 shows a conglomerate merger between a billing service provider from the telecommunications industry and a payment service provider from the financial services industry. If one looks at the acquisitions of fiserv since 1990, for example, some of these M&A transactions (acquisition of Information Technology Corp., Check Free, M-Com, CashEdge or First Data) fall into <?page no="245"?> 6.5 Summary and key learnings 245 this group. Through repeated conglomerate mergers, a diversified technology and financial services provider has emerged that addresses B2B customers from numerous industries. Strategic decisions for M&A transactions are often overlaid by general macroeconomic conditions, trends towards mergers ("merger waves"), a self-interest of managers ("empire building") and the influence of the capital market (especially through the availability and relative costs of equity or debt capital to acquire a firm). A merger of firms is always accompanied by a reduction in the number of market participants. The aim is to increase market shares, so that mergers can have a restraining effect on competition and therefore require notification to or approval by national or international competition authorities (Bundeskartellamt in Germany, EU Commission at the level of the European Union or the Federal Trade Commission in the US) (cf. ► Chapter 7 for further details). Regardless of this, many takeovers fail to meet their respective economic targets. The main reasons are to be seen in the implementation of the transaction: the complexity of an integration and merger, a lack of technology fit, differences in corporate culture and overlapping product portfolios as well as a lack of acceptance by customers. In addition, the goals are often set too high and considered as too quickly achievable. Here, an overestimation of possible cost and revenue synergies (see also ► Chapter 3 on the overconfidence of managers), errors in due diligence, a purchase price that is too high, an overestimation of the designability of the future business model, and finally a lack of or insufficient consideration of possible reactions by competitors play essential roles (Gugler et al. 2003, Ferris et al. 2013, Ficery et al. 2007. Malmendier and Tate 2007 and Miles et al. 2014). 6.5 Summary and key learnings Is it a good thing for a firm to have small fixed costs - or is that bad? Is a firm able to improve its competitiveness by growing larger? The main objective of cost decisions is to improve shortand long-term competitiveness of firms. From a management perspective, it is essential to recognise the differences in the decision-making relevance of long-term and short-term costs. In the short run, fixed costs as costs of capital input cannot be changed. Production and cost decisions are based on the analysis of marginal costs to build up competitive advantage. Marginal costs of a firm are typically U-shaped: with an expansion of the output, marginal costs decrease due to increasing marginal products, and with decreasing marginal products, marginal costs increase. The objective of management is primarily to minimise total costs and ensure efficiency for a given output at current prices. Typical decisions concern, for example, the allocation of total production to several locations or branches - total costs are minimised here if production quantities are assigned to individual locations so that marginal costs are identical. A strategically important decision - especially when setting up new digital business models - lies in the relationship between fixed and variable costs: if variable costs are absolutely low, then output can be expanded at marginal costs close to zero and competitive advantage arises based on economies of scale. <?page no="246"?> 6 Costs, restructuring and M&A 246 In the long run, all costs are adjustable: decisions about the cost structure (and capital intensity) depend now on factor prices, the long-term planned production level and the firm's growth ambitions. Relative factor prices (the wage/ interest rate ratio) are a central explanation for rising capital intensity in many industries over time and for the relocation of production to low-wage countries. In the event of a permanent decline in production output, the cost structure is adjusted within the framework of restructuring, i.e., a reduction in the number of jobs and an adjustment in the amount of equity and debt capital. In addition, long-term competitive advantage can exist or be created through economies of scale (costs rise less than proportionately when output is expanded) or economies of scope (synergies from diversification and product portfolio). If there are significant economies of scale and scope, then these firms are larger, ceteris paribus. In addition, market structure and drivers for mergers can be identified via industry cost curves that are usually S-shaped. Recommendations for further reading For an in-depth coverage of cost theory, see Mas-Colell, A. Winston, M.D. and Green, J.R., Microeconomic theory, New York 1995, and Varian, H., Microeconomic analysis, New York 1992. For those who want to understand more about M&A, Pepall, L., Richards, D. and Norman, G., Industrial organization - contemporary theory and empirical applications, Hoboken 2014. Questions for review [1] Describe applications of decisions on costs from a microeconomic perspective as well as their limitations, advantages and disadvantages. [2] Which costs are relevant for decision-making from a microeconomic perspective? Why? [3] Show the relationship between marginal costs and average costs for linear and non-linear cost structures? What can be the reason for non-linear cost patterns? [4] How can a firm determine its own or a competitor's marginal costs? [5] Which rule must managers follow when allocating production to different factories from a cost perspective? What kind of mistakes are often made? [6] What are the shortand long-term consequences of increasing wage rates (at constant interest rates)? [7] What is the effect on demand for capital and labour, if a rapid increase of output is necessary? [8] In the long run, capital intensity increases in many industries - what could be reasons for this? [9] What is the difference between economies of scale and economies of scope? Which role do both concepts play in explaining M&A strategies? [10] How can one infer the number and size of firms in an industry from the shape of the longterm average cost curve? <?page no="247"?> 6.5 Summary and key learnings 247 Literature Altunbas, Y. and Molyneux, P., Economies of scale and scope in European banking, Applied Financial Economics, 1996, 6, 4, 367-375. Argote, L. and Epple, D., Learning curves in manufacturing, Science, 1990, 247, 4945, 920-924. Arkes, H. R. and Blumer, C., The psychology of sunk cost, Organizational Behavior and Human Decision Processes, 1985, 35, 1, 124-140. Barth-Haas Group (eds.), Der neue Barth-Bericht Hopfen 2019/ 2020, Nürnberg 2020. BMW AG, Investor Presentation, München 2015. Chandler, A.D., Scale and scope: the dynamics of industrial capitalism, Cambridge / London 1990. de Jong, M. and van Dijk, M., Disrupting beliefs: a new approach to business-model innovation, McKinsey Quarterly, July 2015. Dosi, G., Grazzi, M. and Moschella, D., Technology and costs in international competitiveness: from countries and sectors to firms, Research Policy, 2015, 44, 1795-1814. Dutton, J.M. and Thomas, A., Treating progress functions as a managerial opportunity, Academy of Management Review, 1984, 9, 2, 235-247. Ellison, G. and Ellison, S.F., Lessons about markets from the internet, Journal of Economic Perspectives, 2005, 19, 2, 139-158. Evans, D.S. and Schmalensee, R., Matchmakers: the new economics of multisided platforms, New York / London 2016. Ferris, S.P., Jayaraman, N. and Sabherwal, S., CEO overconfidence and international merger and acquisition activity, Journal of Financial and Quantitative Analysis, 2013, 48, 1, 137-164. Ficery, K., Herd, T. and Pursche, B., Where has all the synergy gone? The M&A puzzle, Journal of Business Strategy, 2007, 28, 5, 29-35. Gruber, H., The learning curve in the production of semiconductor memory chips, Applied Economics, 1992, 24, 885- 894. Gugler, K., Mueller, D.C., Yurtoglu, B.B. and Zulehner, C., The effects of mergers: an international comparison, International Journal of Industrial Organization, 2003, 21, 5, 625-653. Hagiu, A. and Wright, J., Multisided platforms, International Journal of Industrial Organization, 2015, 43, 162-174. Jansen, S.A., Mergers & Acquisitions: Unternehmensakquisitionen und -kooperationen, 6. Auflage, Wiesbaden 2016. Köhn, R., Stellenabbau bei Siemens - es brennt lichterloh auf den Märkten, Frankfurter Allgemeine Zeitung, 16. November 2017, 13. Krugman, P., Making sense of the competitiveness debate, Oxford Review of Economic Policy, 1996, 12, 3, 17-25. Lambrecht, A., Goldfarb, A., Bonatti, A., Ghose, A., Goldstein, D.G., Lewis, R., Rao, A., Sahni, N. and Yao, S., How do firms make money selling digital goods online? , Marketing Letters, 2014, 25, 3, 331-341. Levy, D., Output, capital, and labor in the short and long run, Southern Economic Journal, 1994, 60, 4, 946-960. Malerba, F., Learning by firms and incremental technical change, Economic Journal, 1992, 102, 413, 845-859. Malmendier, U. and Tate, G., Who makes acquisitions? CEO overconfidence and the market’s reaction, Journal of Financial Economics, 2008, 89, 1, 20-43. Manez, J.A., Rochina-Barrachina, M.E., Sanchis, A. and Sanchis, J.A., The role of sunk costs in the decision to invest in R&D, Journal of Industrial Economics, 2009, 57, 4, 712-735. Miles, L., Borchert, A. and Ramanathan, A., Why some merging companies become synergy overachievers, Bain & Company 2014. <?page no="248"?> 6 Costs, restructuring and M&A 248 Münter, M.T., Wettbewerb und die Evolution von Industrien, Bayreuth 1999. o.V., Bahn will Tausende Stellen streichen, Die Zeit, 18. Oktober 2015. o.V., Lampenhersteller streicht 1300 Jobs, Handelsblatt, 13. November 2017. o.V., Neuer Osram-Chef drückt aufs Tempo - und hält am Stellenabbau fest, Augsburger Allgemeine, 4. Februar 2015. o.V., VW will Menschen durch Roboter ersetzen - Kosten von höchstens sechs Euro pro Stunde, Focus Money, Februar 2015. Parayre, R., The strategic implications of sunk costs: A behavioral perspective, Journal of Economic Behavior and Organization, 1995, 28, 3, 417-442. Porter, M.E., Competitive advantage of nations, New York 1990. Porter, M.E., Competitive strategy: techniques for analyzing industries and competitors, New York 1980. Probst, L., Frideres, L., Pedersen, K., Lide, S., and Kasselstrand, E., New business models - freemium: zero marginal cost, OECD Case Study 49, Paris 2015. Reinstein, A., Bayou, M.E., Williams, P.F. and Grayson, M.M., Resolving the sunk cost conflict, Advances in Management Accounting, 2017, 28, 123-154. Roeder, F.C., Berlin’s Zombie Airport: a textbook example of the sunk-cost fallacy, in: Handelsblatt Global Edition, 20. Juli 2017. Rogelberg, S. G., Scott, C. and Kello, J., The science and fiction of meetings, MIT Sloan Management Review, 2007, 48, 2, 18-21. Sibony, O., Lovallo, D. and Powell, T.C., Behavioral strategy and the strategic decision architecture of the firm, California Management Review, 2017, 59, 3, 5-21. Spence, A.M., The learning curve and competition, Bell Journal of Economics, 1981, 12, 1, 49-70. Sutton, J., Sunk costs and market structure: price competition, advertising, and the evolution of concentration, London 1991. Trautwein, F., Merger motives and merger prescriptions, Strategic Management Journal, 1990, 11, 4, 283-295. Tremblay, V.J. and Tremblay, C.H., The US brewing industry: data and economic analysis, Cambridge/ London 2005. Weiss, L.W., Optimal plant size and the extent of suboptimal capacity, in: Masson, R.T. and Qualls, P.D. (eds.), Essays on Industrial Organization in Honor of Joe S. Bain, Cambridge 1975, 126-134. <?page no="249"?> 249 7 Perfect competition, monopoly and competition policy Firms obviously differ in size, product portfolio, corporate strategy and technology. Yet it is less obvious that firms have very different abilities and possibilities to influence or change other factors such as strategic parameters (prices, quality, type and degree of product differentiation, etc.); the business environment of a competitive situation (entry barriers, industry-specific technology relevant to all competitors, influencing the legislation, etc.); or the characteristics of the competitive process amongst firms (price versus capacity competition, R&D intensity, speed of innovation, etc.). Whether a firm is in a position to influence individual strategic parameters or even the competitive process as such is essentially determined by general conditions such as market structure and competitive intensity (► Chapter 4). Against this background, two opposing cases can be sketched out first, which rarely exist in absolute form in reality, but provide an initial framework for further analysis: perfect competition and monopoly. In the first case, none of the firms can influence strategic parameters; in the second case, one firm alone can control them and has market power. Market power means that a firm can influence a strategic parameter in its own interest without competitors being able to exploit this strategically. For example, a firm with market power can increase the price without losing market share through cheaper offers or price undercuttings by competitors. Similarly, this firm could exploit market power by reducing the quantity produced or limiting the choice of products, by lowering quality or by refraining from innovation efforts. For such strategies to be profitable, reactions of competitors or consumers must be insignificant or in fact non-existent. This can go hand in hand with the fact that firms with market power cannot only determine individual competitive parameters such as price, but can also actively restrict or eliminate competition. For example, the market leader Lufthansa was able to expand its pricing power on domestic German routes by taking over significant parts of the airline Air Berlin (FAZ 2017b). Such a form of restriction of competition, caused by the strategic behaviour of firms, justifies the necessity for competition policy. Thus, a planned acquisition is examined by German and European competition authorities and, in the case of significant restraints of competition, approval within the framework of merger control is only granted subject to conditions (socalled remedies). In the case of Lufthansa/ AirBerlin, for example, this was done by auctioning off some take-off/ landing slots at the airports (FAZ 2017c). Competition policy as part of governmental economic policy has a key objective in creating conditions for workable competition so that innovations and welfare gains are possible. In essence, this is done through national and international legislation and jurisdiction aimed at restrictive practices or market power of firms. From a management perspective, two aspects are crucial: firstly, to assess whether and to what extent strategic degrees of freedom are available to one's own firm, and secondly, whether an exploitation of these degrees of freedom due to market power is in conflict with competition law. <?page no="250"?> 7 Perfect competition, monopoly and competition policy 250 Learning Objectives This chapter deals with: the concepts of perfect competition and monopoly as a framework and yardstick for competition policy; possible causes of market power and dominant behaviour; differences in market outcomes (profits, quantities and prices) in profit maximising behaviour depending on market structure (number and behaviour of firms); and effects on economic welfare in case of market power and the relevance of competition policy, the tasks and roles of German competition authorities as well as some cases on restrictive behaviour of firms. 7.1 Decisions of a firm under perfect competition Many industries are characterised by firms that are small relative to the size of the market, whose products do not differ perceptibly from the customer's perspective, but which have a large number of B2C customers. In these industries firms often have few degrees of freedom since their own strategic behaviour is restricted and predetermined by the competitive situation and the behaviour of their competitors. In addition, all firms often have access to the same resources (e.g., employees, suppliers, or raw materials), there are no significant economies of scale, and none of the firms is in a position to influence the strategic parameters of the industry. Typical examples are free e-mail services, firewood for the fireplace, restaurants in large cities, regional delivery services or container shipping. Markets with perfect competition are characterised by the following features. Market entry and exit without sunk costs - firms can enter or exit the market freely (i.e., without sunk costs). There are no strategic, legal, or structural barriers to entry. Accordingly, there is no minimum efficient size, all firms have access to the best technology and all other resources. Lack of barriers to entry (including branding or customer loyalty) also means that customers have no preference for individual firms or brands. Large number of small firms - there is a large number of firms and customers in the market. All of them are small and strategically insignificant, so everyone has to take the market price as a fact and cannot influence it. In reality, it is already sufficient if there is a potentially large number of small market participants. Homogeneous products and perfect information - the products offered are essentially the same. In reality, it is sufficient that the products are so similar and comparable that customers do not take into account or do not recognise the differences. As a consequence of this only one price can exist, at least in the medium term. All customers and suppliers have perfect information, from which it follows that all relevant changes are immediately known to all market participants. <?page no="251"?> 7.1 Decisions of a firm under perfect competition 251 Figure 7.1: Restaurants in Berlin (Source: Google Maps 2017). One market for which these conditions apply very well is restaurants in larger cities such as Melbourne, London or Madrid, but of course also like Berlin. As shown in ► Figure 7.1, Google Maps and popular portals such as TripAdvisor provide almost perfect information, for both competitors and customers, in real time. The number of firms (i.e., restaurants in Berlin) in 2017 is more than 5,000; each individual firm is very small relative to demand; the products are relatively homogeneous (customers are looking for "a nice evening"); and, with the exception of haute cuisine restaurants and low-cost fast-food chains, prices for a typical dish are very close together. A firm that would try to raise the price above the competitive level while maintaining the usual market quality would immediately lose customers. There are also virtually no sunk costs involved in entering and leaving the market: although restaurants very often go out of business, the premises are typically taken over by the next restaurant operator immediately afterwards, so that there are no sunk costs for kitchen equipment and furnishings (Giersberg 2016). From a management perspective, a firm is very likely to be in perfect competition if discussions at board level are not so much about innovation and business model design as they are about reducing prices in response to competition and if the focus is repeatedly on cost reduction. In this case, a firm is not free to decide on prices, since customers are likely to switch to competitors with lower prices. Where customers do not perceive any difference in product quality between firms, the usual marketing strategies will have no differentiating effect in competition. Profit maximisation by choice of quantity under perfect competition Firms in these industries face perfect competition. From a management perspective, all strategic parameters (especially pricing) must be assumed to be given and determined by the market. In this case, if a firm decides against the usual and established product-price-business <?page no="252"?> 7 Perfect competition, monopoly and competition policy 252 model combination, it will either not find customers or will not be able to cover its costs. In both cases, the existence of the firm is at risk. Individual decisions of the firm, thus, only concern the individual quantity produced and the resulting costs in order to achieve profits. The crucial question for management is then whether, given a market price and a firm-specific cost function, profits can be incurred to ensure the survival of the firm. Figure 7.2: Profits of a firm in perfect competition. ► Figure 7.2 illustrates this situation. If the price of perfect competition is independent of the output, then a revenue curve 𝑅𝑅 , as price 𝑝𝑝 multiplied by the quantity 𝑞𝑞 , increases linearly. Combining this with a firm-specific cost function 𝑇𝑇𝐶𝐶 , the difference directly yields the profit 𝜋𝜋 = 𝑅𝑅 − 𝑇𝑇𝐶𝐶 of a firm. Obviously, profit is maximised at a quantity 𝑞𝑞 2 - here the difference between revenues and total costs is largest. Profits 𝜋𝜋(𝑞𝑞) as a function of the quantity is generally calculated from the difference between revenues 𝑅𝑅 and total costs 𝑇𝑇𝐶𝐶 as (7.1) 𝜋𝜋(𝑞𝑞) = 𝑅𝑅(𝑞𝑞) − 𝑇𝑇𝐶𝐶(𝑞𝑞) → 𝜋𝜋(𝑞𝑞) = 𝑝𝑝(𝑞𝑞)𝑞𝑞 − 𝑇𝑇𝐶𝐶(𝑞𝑞) . If one maximises profits by choosing a quantity 𝑞𝑞 , then via (7.2) 𝜕𝜕𝜋𝜋 𝜕𝜕𝑞𝑞 = 𝜕𝜕𝜕𝜕 𝜕𝜕𝑞𝑞 − 𝜕𝜕𝑇𝑇𝐶𝐶 𝜕𝜕𝑞𝑞 = 𝑝𝑝 + 𝜕𝜕𝑝𝑝 𝜕𝜕𝑞𝑞 𝑞𝑞 − 𝑀𝑀𝐶𝐶 = 𝑀𝑀𝑅𝑅 − 𝑀𝑀𝐶𝐶 = 0, we get that at a profit maximum marginal revenues are equal to marginal costs - in other words: the increase in revenue of the last product sold (the marginal revenue) just covers the additional costs (the marginal costs). This relationship generally applies to every market structure - here in perfect competition just as in monopoly (► Section 7.3) and also in oligopoly (► Chapter 10). <?page no="253"?> 7.1 Decisions of a firm under perfect competition 253 In perfect competition, however, the marginal revenue also corresponds to the market price 𝑝𝑝 . If no firm can influence the price, then this price applies to every quantity produced and so it follows that (7.3) 𝜕𝜕𝑝𝑝 𝜕𝜕𝑞𝑞 = 0 . In other words: a price is determined by competition in the market and no individual firm can influence this price by changing its output. Thus, equation (7.2) simplifies to the following for perfect competition (7.4) 𝜕𝜕𝜋𝜋 𝜕𝜕𝑞𝑞 = 𝜕𝜕𝜕𝜕 𝜕𝜕𝑞𝑞 − 𝜕𝜕𝑇𝑇𝐶𝐶 𝜕𝜕𝑞𝑞 = 𝑝𝑝 − 𝑀𝑀𝐶𝐶 = 0; 𝜋𝜋(𝑞𝑞) → 𝑆𝑆𝑎𝑎𝑥𝑥! ↔ 𝑀𝑀𝑅𝑅 = 𝑝𝑝 = 𝑀𝑀𝐶𝐶 , so that in order to maximise profits each firm must choose a quantity of output 𝑞𝑞 at which marginal costs 𝑀𝑀𝐶𝐶 are equal to the market price 𝑝𝑝 . This abstract relationship can be easily illustrated: restaurant operators observe the market price for 'a dinner of usual quality' in their region, which they cannot influence individually, and now determine an optimum size of the restaurant (tables and chairs) and the number of meals to be produced, knowing their cost situation. ► Figure 7.3 shows this decision graphically. If marginal costs are U-shaped and the market price 𝑝𝑝 0 is constant and independent of output, then there is a unique quantity 𝑞𝑞 ∗ . If average total cost 𝐴𝐴𝑇𝑇𝐶𝐶 is lower than price at this output, the firm makes a profit, which is shown in ► Figure 7.3 on the left by a grey area as 𝜋𝜋(𝑞𝑞) = (𝑝𝑝 − 𝐴𝐴𝑇𝑇𝐶𝐶) ∗ 𝑞𝑞 . This area shows maximum possible profits for this firm: any increase or decrease in output would reduce profits. However, this profit is not necessarily positive, as can be seen from ► Figure 7.3 middle. With an identical cost situation, the market price 𝑝𝑝 1 is now lower. If the firm again chooses an optimum quantity where price equals marginal costs, then this price no longer covers average total costs 𝐴𝐴𝑇𝑇𝐶𝐶 , but still average variable costs 𝐴𝐴𝐸𝐸𝐶𝐶 . The firm makes a loss of 𝜋𝜋(𝑞𝑞) = (𝑝𝑝 − 𝐴𝐴𝑇𝑇𝐶𝐶) ∗ 𝑞𝑞 , but it achieves a positive contribution margin - parts of the fixed costs are covered. The grey area again describes the 'maximum' profit, but in this case as the smallest possible loss. In a situation where, with an optimally chosen quantity 𝑞𝑞 ∗ , 𝑝𝑝 < 𝐴𝐴𝑇𝑇𝐶𝐶 but 𝑝𝑝 > 𝐴𝐴𝐸𝐸𝐶𝐶 applies, a firm should in any case remain in the market in the short run: all variable costs are apparently covered and a positive contribution margin amounting to 𝑝𝑝 − 𝐴𝐴𝐸𝐸𝐶𝐶 on fixed costs is achieved. If this firm were to stop production, the losses increase immediately and then correspond to the amount of fixed costs 𝐹𝐹𝐶𝐶 . This makes it easy to check in reality whether a firm is in a competitive environment characterised by perfect competition: if cost-cutting measures in particular are initiated in a situation of declining or negative profits because there is no strategic space for price increases, then the firm is obviously in perfect competition with high competitive intensity. In ► Figure 7.3 on the right, if the cost situation remains unchanged, the market price is still lower at 𝑝𝑝 2 . Now the rule 𝑝𝑝 = 𝑀𝑀𝐶𝐶 leads to a quantity 𝑞𝑞 ∗ , for which not only an absolute loss occurs, but also 𝑝𝑝 < 𝐴𝐴𝑇𝑇𝐶𝐶 and 𝑝𝑝 < 𝐴𝐴𝐸𝐸𝐶𝐶 applies - a firm should now stop production, because a negative contribution margin is achieved, i.e., with each unit produced, losses increase beyond the level of fixed costs. <?page no="254"?> 7 Perfect competition, monopoly and competition policy 254 Figure 7.3: Short-term profits in perfect competition. <?page no="255"?> 7.1 Decisions of a firm under perfect competition 255 Firm-specific supply curve and market supply curve If we take all observations from ► Figure 7.3 together, then a firm in perfect competition obviously always chooses production quantities for which the market price 𝑝𝑝 corresponds to the firm's marginal costs 𝑀𝑀𝐶𝐶 in order to maximise profit. From this, the firm-specific supply curve sketched in ► Figure 7.4 on the left can be derived directly - it corresponds to the ascending branch of the marginal cost curve starting at the minimum of the average variable costs. The higher the price, the more the firm will offer depending on its marginal cost curve in order to maximise profits. In the area between the curve of total and variable average costs (light dots), a firm achieves a positive contribution to fixed costs, but makes losses in absolute terms. In the area above the average total costs, a firm achieves positive profits. Figure 7.4: Firm-specific supply curve and market supply curve. The market supply curve can then be determined from the horizontal addition of the individual firm supply curves at the respective prices (outlined here for 𝑝𝑝 1 , 𝑝𝑝 2 and 𝑝𝑝 3 ). In ► Figure 7.4 on the right, it can be seen for three firms with different marginal cost curves that the dashed market supply curve becomes flatter with an increasing number of firms and shifts to the right. Consequently, the market supply curve runs horizontally for a large number of firms with identical marginal costs. Competitive strategy and competitive dynamics under perfect competition The interplay between supply and demand on the market and the decisions of the firms can be used to derive some initial statements about typical patterns in perfect competition. ► Figure 7.5 on the left sketches a situation in which, given a demand function 𝐷𝐷 , there is currently a supply curve 𝑆𝑆 1 of all active firms. The firms will be able to produce a total quantity 𝑄𝑄 1 and sell it at a price 𝑝𝑝 1 . In ► Figure 7.5 on the right, this situation is transferred to an individual firm: at price 𝑝𝑝 1 , the firm chooses, according to the rule price equals marginal cost, an individual quantity 𝑞𝑞 1 , so that as a result a profit in the amount of the grey area arises. <?page no="256"?> 7 Perfect competition, monopoly and competition policy 256 Figure 7.5: Profits and market entry of new firms. However, profits of this firm are a signal for new firms to enter the market, because obviously this market is attractive, at least in the short term, and profits can be made. Now, if there are no barriers to entry and potential firms have access to the same technology and resources, this situation will not last. In the long run, firms will enter this market. As a result, as shown in ► Figure 7.5 on the left, the market supply curve will shift to the right and become flatter. In sum, the quantity will increase to 𝑄𝑄 2 , which can only be sold at a reduced market price 𝑝𝑝 2 . ► Figure 7.5 on the right shows the consequences for a firm in the market. Because of the now lower market price 𝑝𝑝 2 this firm will reduce its output to 𝑞𝑞 2 - in the long run the consequence is that identical firms will continue to enter the market until the price falls to the minimum of average total cost. Now 𝑝𝑝 = 𝑀𝑀𝐶𝐶 = 𝐴𝐴𝑇𝑇𝐶𝐶 obviously holds and none of the firms makes a profit. In fact, at this economic profit of zero, all costs are covered, including the cost of capital, so that owners of the firm receive dividends, for example. From a management perspective, a key insight is that although the total market grows from 𝑸𝑸 𝟏𝟏 to 𝑸𝑸 𝟐𝟐 , in perfect competition one's own firm will be forced to reduce production output from 𝒒𝒒 𝟏𝟏 to 𝒒𝒒 𝟐𝟐 , and profits will also fall to zero. The reason lies in the absence of entry barriers: potential new firms will enter the market unhindered, driving price reductions and realising existing profit opportunities in the short term. The competitive process described leads to all firms producing at a minimum of average total costs. In perfect competition, cost-cutting measures happen all the time, as firms with higher average costs are regularly forced out of the market. Perfect competition forces firms to be efficient. This dynamic can be seen in ► Figure 7.6 as a mirror image of the development in ► Figure 7.5: here, in the short term, numerous firms are initially in the market and produce a total quantity 𝑄𝑄 1 at a price 𝑝𝑝 1 . Obviously, however, this market price is so low that each individual firm realises a loss in the amount of the grey area in ► Figure 7.6 on the right. If all firms have identical cost situations, then as a consequence those firms whose equity base cannot cover <?page no="257"?> 7.1 Decisions of a firm under perfect competition 257 these losses are forced out of the market - thus, there are exits of firms leading to a shift in the market supply curve to 𝑆𝑆 2 . The remaining firms are now able to impose a higher price 𝑝𝑝 2 so that the individual firms are again at the minimum average total costs 𝐴𝐴𝑇𝑇𝐶𝐶 and each surviving firm grows from 𝑞𝑞 1 to 𝑞𝑞 2 as a result of the market exits. Figure 7.6: Losses and market exit of firms. In the long run market equilibrium, as described by 𝑝𝑝 2 and 𝑄𝑄 2 in ► Figures 7.5 and 7.6, there is no incentive to enter or exit the industry and the equilibrium price occurs if supply and demand are balanced. In perfect competition, however, competition does not come to a halt. Rather, dynamics in industries are characterised by a repeated interplay of entries and exits and, thus, of profits and losses, but profits realised are typically close to zero. Firms can, for example, realise competitive advantages and increase profits in the short term through innovation. In the long run, all firms will learn and adapt the best available technology so that all firms have identical efficiency and short-term profits erode again. Accordingly, if entry barriers and sunk costs are absent or disappear, there will of course be continuous entry and exit of firms (Petrakis et al. 1997. Jovanovic 1982 and Münter 1999). In empirical studies, the high intensity of competition in perfect competition becomes visible by permanently high entry and exit rates and via firms that are very similar in their strategies due to a lack of market power. A good example of this is restaurants all around the globe: in this industry the highest entry and exit rates are repeatedly observed, while at the same time - aside from system gastronomy, which can achieve economies of scale through branding and a high degree of automation - economic profits are typically close to zero. Thus, such industries clearly belong to the market-based view. These firms do not differ significantly in their strategies, but due to the lack of entry barriers, no lasting profits are achieved. Market dynamics and market results of perfect competition are also tested in economic laboratory experiments. The findings can be assigned to two categories: on the one hand, under <?page no="258"?> 7 Perfect competition, monopoly and competition policy 258 given conditions, the expected price and quantity levels are successively well approximated after several rounds of the game; on the other hand, experience, level of information, and history of the experimental market results play a decisive role (Smith 1962 and 1981 as well as Holt 1995). However, it is also true that even under laboratory conditions a complete manifestation of the results of perfect competition is the exception rather than the rule (► Chapter 10). 7.2 Producer and consumer surplus as a measure of economic welfare The perfect competition model developed in the previous section finds confirmation in empirical reality, especially for industries and markets in which a large number of firms compete but no profits are generated due to high competitive intensity, a lack of product differentiation, and a lack of entry barriers (► Chapter 4). In addition, perfect competition forces firms to be efficient and to price at the level of marginal costs - thus, this model offers a first point of orientation for competition policy analyses or the regulation of markets (Neumann 1999 and Motta 2004). The reason for this is that in perfect competition, especially in comparison with other competitive situations, economic welfare is maximised. Economic welfare Economic welfare describes the benefits provided to society as a result of the existence of a competitive market, i.e., benefits for both firms and customers. ► Figure 7.7 on the left shows individual willingness to pay and the cost situations for a large number of customers and firms. At the intersection of demand and supply curves derived from this, a market price 𝑝𝑝 and an aggregated quantity 𝑄𝑄 result. In this equilibrium, however, there are obviously customers whose individual willingness to pay 𝑧𝑧 𝑖𝑖 exceeds the price - these customers achieve an individual utility and the aggregate utility of all customers who buy at price 𝑝𝑝 corresponds to the consumer surplus 𝐶𝐶𝑆𝑆 . This is also true for firms: at a price 𝑝𝑝 , many firms are able to realize a positive contribution to fixed costs because the price exceeds marginal costs. This aggregate advantage of the firms is called producer surplus 𝑃𝑃𝑆𝑆 . Typically, producer surplus is not equal to profits: the market supply curve corresponds to the marginal cost curves, so that the difference is due to fixed costs, so that 𝑃𝑃𝑆𝑆 − 𝐹𝐹𝐶𝐶 = 𝜋𝜋 applies. A market equilibrium with perfect competition is beneficial for both customers and firms: customers realise utility, firms realise profits. The overall benefit from the existence of a competitive market can be determined by adding consumer and producer surplus and is called economic welfare - shown on the right in ► Figure 7.7 in the form of the two triangular areas. The extent of economic welfare is, on the one hand, the yardstick for measuring the functioning of a market and, on the other hand, all economic policy and competition policy measures must be assessed with regard to their effect on economic welfare. <?page no="259"?> 7.2 Producer and consumer surplus as a measure of economic welfare 259 Figure 7.7: Producer and consumer surplus in a market equilibrium. Quantifying effects of price regulation on economic welfare Economic policy intervenes in markets by regulating prices, for example, setting minimum or maximum prices. Examples of this include rent control, minimum wages, regulating roaming charges in mobile telephony, taxi charges, or the approval of postal tariffs. Often the aim is to put customers or firms in a better position compared to a current competitive situation, for example, to lower consumer prices. Whether an improvement is really achieved depends on the effect of price regulation on consumer and producer surplus. ► Figure 7.8 shows that an introduction of a maximum or minimum price prevents a market equilibrium from being achieved as supply and demand are no longer balanced. A maximum price 𝑝𝑝 𝑚𝑚𝑚𝑚𝑚𝑚 is usually intended to protect customers: the rationale is that with a maximum price (set below the equilibrium price) more customers will be able to purchase the product. However, if a maximum price 𝑝𝑝 𝑚𝑚𝑚𝑚𝑚𝑚 is introduced, as in ► Figure 7.8 on the left, then the quantity offered by firms will actually decrease as a direct consequence. The reason is that at a lower maximum price some firms will no longer be able to cover their marginal costs and hence exit the market. On the other hand, in ► Figure 7.8 on the right, a minimum price 𝑝𝑝 𝑚𝑚𝑖𝑖𝑝𝑝 is introduced to put firms in a better position - but in this case, too, the quantity falls due to reduced demand from customers. In both cases, there is inevitably a redistribution of economic welfare through changes in consumer and producer surplus, but the effects are not clear-cut and need to be quantified. Moreover, price regulation typically leads to a deadweight loss (DWL) - the sum of consumer and producer surplus comparing ► Figure 7.8 to 7.7 is definitely reduced by the white triangle. Economic policy must, therefore, always take into account effects on economic welfare, effects on firms and consumers, as well as consider a potential welfare loss. <?page no="260"?> 7 Perfect competition, monopoly and competition policy 260 Figure 7.8: Welfare loss due to maximum or minimum prices. Figure 7.9: Maximum and minimum prices and transformation of producer surplus into consumer surplus. Welfare effects can be quantified by changes in consumer or producer surplus. In ► Figure 7.9 on the left, a maximum price 𝑝𝑝 1 is drawn, which is below the equilibrium price 𝑝𝑝 0 . The maximum price introduced will cause firms to reduce output from 𝑄𝑄 0 to 𝑄𝑄 1 . This is accompanied by the conversion of part of the previous producer surplus - area 𝐴𝐴 - into consumer surplus, but also by a welfare loss in the form of a decrease in consumer surplus by area 𝐵𝐵 and a decrease in producer surplus by area 𝐶𝐶 - all three areas can be represented by (7.5) 𝐴𝐴 = (𝑝𝑝 0 − 𝑝𝑝 1 )𝑄𝑄 1 (7.6) 𝐵𝐵 = (𝑝𝑝 ′ − 𝑝𝑝 0 ) (𝑄𝑄0−𝑄𝑄1) 2 and (7.7) 𝐶𝐶 = (𝑝𝑝 0 − 𝑝𝑝 1 ) (𝑄𝑄0−𝑄𝑄1) 2 <?page no="261"?> 7.2 Producer and consumer surplus as a measure of economic welfare 261 directly. The absolute welfare loss (also called deadweight loss) results from the addition of (7.6) and (7.7), the change in consumer and producer surplus as (7.8) ∆𝐶𝐶𝑆𝑆 = 𝐴𝐴 − 𝐵𝐵 and (7.9) ∆𝑃𝑃𝑆𝑆 = −𝐴𝐴 − 𝐶𝐶 . The producer surplus decreases in any case with a maximum price. The effect on consumer surplus depends on the size of the two effects 𝐴𝐴 and 𝐵𝐵 and cannot be clearly determined in its direction per se - whether the improvement in consumer surplus intended by the maximum price is achieved depends on the shape of the supply and demand curve and must be determined for each competition policy measure. In the case of minimum prices, the price is raised above the equilibrium level, so that, as sketched in ► Figure 7.9 on the right, the change in consumer and producer surplus as well as in the welfare loss results from (7.10) 𝐴𝐴 = (𝑝𝑝 1 − 𝑝𝑝 0 )𝑄𝑄 1 (7.11) 𝐵𝐵 = (𝑝𝑝 1 − 𝑝𝑝 0 ) (𝑄𝑄0−𝑄𝑄1) 2 and (7.12) 𝐶𝐶 = (𝑝𝑝 0 − 𝑝𝑝 ′ ) (𝑄𝑄0−𝑄𝑄1) 2 as (7.15) ∆𝐶𝐶𝑆𝑆 = −𝐴𝐴 − 𝐵𝐵 (7.16) ∆𝑃𝑃𝑆𝑆 = +𝐴𝐴 − 𝐶𝐶 and (7.17) 𝐷𝐷𝑊𝑊𝐿𝐿 = −𝐵𝐵 − 𝐶𝐶 . In the case of minimum prices, firms can benefit if the effect of a conversion 𝐴𝐴 of consumer surplus into producer surplus exceeds the loss C of producer surplus; consumers are worse off in any case. Case Study │ Regulating the price of milk The European dairy industry has been repeatedly regulated in recent decades through competition policy interventions such as minimum prices or production volume restrictions. The aim is often to support agricultural firms and farmers, most recently with the milk quota regulation that expired in March 2015. With the expiry of this regulation, numerous farms have increased their production volumes, so that as a result price reductions have taken place in perfect competition, threatening the survival of some firms (Grossarth 2016). In the following, we will therefore take a brief look at whether the introduction of a minimum price, as demanded by the milk lobby, can ensure the survivability of the firms. A regional market for milk is characterised by a very large number of small firms, essentially local farms. The demand 𝐷𝐷 for milk is given by 𝑝𝑝 = 5 − 0.001 𝑄𝑄 , the supply 𝑆𝑆 by 𝑝𝑝 = 0.5 + 0.0001 𝑄𝑄 . The farmers have to bear relatively high fixed costs, totalling for all farmers these are 𝐹𝐹𝐶𝐶 = 1,000 . The Minister of Agriculture plans to set a minimum price for milk of 𝑝𝑝 1 = 1.00 in order to secure the farmers' existence. From the perspective of competition policy, the decisive factor is the qualitative and quantitative effects on producer and consumer surplus as well as the welfare of the market. <?page no="262"?> 7 Perfect competition, monopoly and competition policy 262 Figure 7.10: Initial situation and price regulation on the milk market. In order to determine the price 𝑝𝑝 and the total quantity produced 𝑄𝑄 in the initial situation, the supply and demand functions are first equated so that from (7.17) 5 − 0.001𝑄𝑄 = 0.5 + 0.0001𝑄𝑄 and solving for quantity (7.18) 4.5 = 0.0011𝑄𝑄 → 𝑄𝑄 0 = 4090.9 and equilibrium price (7.19) 𝑝𝑝 𝐷𝐷 = 5 − (0.001 ⋅ 4,090.9) = 0.9091 → 𝑝𝑝 𝑆𝑆 = 0.5 + (0.0001 ⋅ 4,090.9) = 0.9091 follows. Thus, consumer and producer surplus as well as the profits of the farmers result as (7.20) 𝐶𝐶𝑆𝑆 0 = (5 − 0.9091) ⋅ 4090.9 2 = 8,367.77 (7.21) 𝑃𝑃𝑆𝑆 0 = (0.9091 − 0.5) ⋅ 4090.9 2 = 836.78 (7.22) 𝜋𝜋 0 = 𝑃𝑃𝑆𝑆 0 − 𝐹𝐹𝐶𝐶 = 836.78 − 1,000 = −163.22 so that, in sum, farmers actually realise a loss of 𝜋𝜋 0 = − 163.22 in the current situation at a price of 𝑝𝑝 = 0.9091 . Due to the rising slope of the supply curve, it may well be that some of the individual firms make positive profits in the short run due to lower marginal costs. If now, as outlined in ► Figure 7.10, a minimum price of 𝑝𝑝 1 = 1.00 above the previous market price is introduced, then the demand for milk is reduced because of (7.23) 𝑝𝑝 1 = 1 = 5 − 0.001 ⋅ 𝑄𝑄 → 0.001 ⋅ 𝑄𝑄 = 4 to 𝑄𝑄 1 = 4,000 . If we first calculate 𝑝𝑝′ based on (7.24) 𝑝𝑝′ = 0.5 + 0.0001𝑄𝑄 1 = 0.5 + 0.0001 ⋅ 4,000 = 0.9 the areas necessary for calculating the welfare effects are as follows (7.25) 𝐴𝐴 = (1 − 0.909) ⋅ 4,000 = 363.64 (7.26) 𝐵𝐵 = (1 − 0.909) ⋅ 4,090.9−4,000 2 = 4.13 (7.27) 𝐶𝐶 = (0.909 − 0.9) ⋅ 4,090.9−4,000 2 = 0.41 . In a second step, the changes in consumer and producer surplus and the effects on profit can now be determined as <?page no="263"?> 7.3 Monopoly and dominant firms 263 (7.28) 𝐶𝐶𝑆𝑆 1 = 𝐶𝐶𝑆𝑆 0 − 𝐴𝐴 − 𝐵𝐵 = 8,367.77 − 363.64 − 4.13 = 8,000 (7.29) ∆𝐶𝐶𝑆𝑆 = −𝐴𝐴 − 𝐵𝐵 = −363.64 − 4.13 = −367.77 (7.30) 𝑃𝑃𝑆𝑆 1 = 𝑃𝑃𝑆𝑆 0 + 𝐴𝐴 − 𝐶𝐶 = 836.78 + 363.64 − 0.41 = 1,200 (7.31) ∆𝑃𝑃𝑆𝑆 = +𝐴𝐴 − 𝐶𝐶 = 363.64 − 0.41 = 363.23 as well as (7.32) 𝜋𝜋 1 = 𝑃𝑃𝑆𝑆 1 − 𝐹𝐹𝐶𝐶 = 1,200 − 1,000 = 200 . The minimum price apparently has the intended effect: the farmers' initial loss 𝜋𝜋 0 = −163.22 turns into a small profit 𝜋𝜋 1 = 200 . Moreover, the welfare loss amounts to be (7.33) 𝐷𝐷𝑊𝑊𝐿𝐿 = −𝐵𝐵 − 𝐶𝐶 = −4.13 − 0.41 = −4.54. This is obviously a small intervention in the market and, depending on one’s basic political orientation, may be acceptable. However, it must be clear: all customers are now paying approximately 10% higher prices, the quantity demanded is declining by 2.2% and with it the number of firms is also declining. This means that despite the intervention in the market by competition policy, firms will also be forced to leave the market without further measures. 7.3 Monopoly and dominant firms Historically, many services in a large number of countries were provided by monopolies - a single firm. Examples include telecommunications, postal services, railways, electricity, or lotteries. Typically, a monopolist has market power and as a result, unlike the firms in perfect competition, can influence strategic competitive parameters and, for example, raise the price above the competitive level. A monopolist, or more generally speaking a dominant firm can also influence product quality, capacity, or technological progress in its own interest. In addition, the lack of competition means that there is no pressure for firms to be efficient - a visit to your local post office may illustrate this. The following origins of monopolies and dominant firms often interact. State licences or state-owned firms - in many industries governments operate firms itself. The reasons can be standardisation of a specific technology, need for investment (e.g., for a nationwide rail infrastructure), price control, quality assurance, or an intended prevention of competition due to potential market failure. However, these historical justifications for monopolies often disappear. Take as an example the still prevalent existence of governmental and regional taxi licences along with the associated investment protection for the acquisition of vehicles, the establishment of a taxi centre with territorial protection (as considered necessary after the Second World War), as well as a local street knowledge test. These licences, and associated conditions, seem outdated in times of navigation devices and taxi apps (Haucap et al. 2015). Economies of scale - extensive economies of scale in relation to absolute market size can create significant cost advantages that give the largest firm a natural monopoly (see also ► Chapter 6). Firm-specific capabilities - one or more firms possess permanently non-imitable capabilities that lead to a unique selling proposition through competitive advantage. If, for example, economies of scale are associated with this, other firms are detained from en- <?page no="264"?> 7 Perfect competition, monopoly and competition policy 264 tering the market due to entry barriers and an unattackable dominant position occurs (► Chapter 4). Strategic effects and roles in an idustry - through long run competitive processes and renewed investment in R&D or marketing, a firm can succeed in building up a dominant role (see ► Chapter 10 on Stackelberg market leadership for further details). This is accompanied in particular by customers’ perception of horizontal product differentiation. Due to high brand loyalty, these customers show a low willingness to switch and will therefore accept higher prices even in the absence of vertical product differentiation. A high market share itself then also justifies market power in the long run (Wernerfelt 1991 and Rhoades 1985). Patents - based on product or process innovations, a firm owns exclusive rights for the production or distribution of certain products or services temporarily. In Germany and the EU these rights are currently conceded for a maximum of 20 years. Essentially, patent protection is intended to change appropriability conditions and, thus, to create incentives for cost-intensive and risky R&D projects (Levin 1986, Levin et al. 1987 and Gilbert and Shapiro 1990). However, firms can also create strategic entry barriers for competitors through preventive patenting (Gilbert and Newbery 1982). Network effects - demand-side direct or indirect network effects (► Chapter 2) can also lead to a dominant position of a firm that is able to build or control a multisided platform (Evans 2003 as well as Haucap and Heimeshoff 2017). Exclusive or extensive control of essential resources - for example, the firm around the de Beers family controlled a significant share of about 85% of all diamond mines in the world until 2005 and had a dominant position. As a result, none of the smaller competitors had any interest in a price war, because all firms in the industry benefited from the very high prices set by de Beers. However, the control of essential resources can also apply to employees. For example, football clubs such as Paris St. Germain FC, FC Barcelona, or FC Chelsea can try to take a dominant position by employing the best players; similarly, this applies to management consultancies and investment banks. Profit maximisation by choice of quantity in monopoly A first key difference between perfect competition and monopoly can be seen in ► Figure 7.11. A monopolist also maximises his profit via the choice of a quantity 𝑞𝑞 ∗ , so that in monopoly too (7.34) 𝜋𝜋(𝑞𝑞) = 𝑅𝑅(𝑞𝑞) − 𝑇𝑇𝐶𝐶(𝑞𝑞) mit 𝜋𝜋(𝑞𝑞) → 𝑆𝑆𝑎𝑎𝑥𝑥! with (7.35) 𝜕𝜕𝜋𝜋 𝜕𝜕𝑞𝑞 = 𝜕𝜕𝜕𝜕 𝜕𝜕𝑞𝑞 − 𝜕𝜕𝑇𝑇𝐶𝐶 𝜕𝜕𝑞𝑞 = 𝑀𝑀𝑅𝑅 − 𝑀𝑀𝐶𝐶 = 0 the condition marginal revenue equals marginal cost is fulfilled. However, the price is not market-determined, but is influenced by the quantity offered by the monopoly. For this reason, the revenue curve 𝑅𝑅 = 𝑝𝑝𝑞𝑞 does not rise linearly but is a downward-open parabola (► Chapter 1). <?page no="265"?> 7.3 Monopoly and dominant firms 265 Figure 7.11: Monopoly and profit maximisation. From a management perspective, it is not so much absolute profit maximisation that is important, but rather the strategies that can be derived and their line of attack: if the current quantity is less than a profit-maximising quantity 𝑞𝑞 ∗ , then the firm should expand the output. The reason is that to the left of 𝑞𝑞 ∗ the revenue curve rises more steeply than the total cost curve - as quantity is increased, revenues rise more steeply than costs, so that marginal costs are below marginal revenue and profits increase. To the right of 𝑞𝑞 ∗ , marginal revenue is less than marginal cost - the firm reduces profits by continuing to grow. The reason for this is not only a more than proportionate increase in costs, but the firm must at the same time reduce prices in order to be able to sell the higher output - both effects taken together explain a decline in profits. Thus, for firms with market power, growth is not always a good strategy. ► Figure 7.12 shows two firms with comparable and short-term stable cost and revenue functions. For firm 𝐴𝐴 , after a loss in 2014, a reduction in production volume apparently makes sense. On the one hand, the costs fall more than proportionately in the area of decreasing economies of scale, and, on the other hand, with lower production volume in 2015, higher prices can be enforced with the customers again, so that revenues rise. <?page no="266"?> 7 Perfect competition, monopoly and competition policy 266 Figure 7.12: Profits and different strategies. Firm 𝐵𝐵 , on the other hand, based on profit and marginal revenue/ marginal cost situation in 2014, has correctly decided in favour of growth - however, the firm has grown too far, resulting in lower profits in 2015 than in 2014. More specifically, decisions by dominant firms can also be explained mathematically analysing explicit demand (7.36) and revenue (7.37) functions and associated cost functions (7.38) such as (7.36) 𝑝𝑝(𝑞𝑞) = 𝑎𝑎 − 𝑏𝑏𝑞𝑞 (7.37) 𝑅𝑅(𝑞𝑞) = 𝑝𝑝𝑞𝑞 (7.38) 𝑇𝑇𝐶𝐶(𝑞𝑞) = 𝑀𝑀𝐶𝐶 ⋅ 𝑞𝑞 + 𝐹𝐹𝐶𝐶 . This results in a profit function (7.39) 𝜋𝜋(𝑞𝑞) = 𝑅𝑅 − 𝑇𝑇𝐶𝐶 = (𝑎𝑎 − 𝑏𝑏𝑞𝑞)𝑞𝑞 − (𝑀𝑀𝐶𝐶 ⋅ 𝑞𝑞 + 𝐹𝐹𝐶𝐶) = 𝑎𝑎𝑞𝑞 − 𝑏𝑏𝑞𝑞 2 − 𝑀𝑀𝐶𝐶 ⋅ 𝑞𝑞 − 𝐹𝐹𝐶𝐶 → 𝑆𝑆𝑎𝑎𝑥𝑥! which, through the choice of output (7.40) 𝜕𝜕𝜋𝜋 𝜕𝜕𝑞𝑞 = 𝑀𝑀𝑅𝑅 − 𝑀𝑀𝐶𝐶 = 𝑎𝑎 − 2𝑏𝑏𝑞𝑞 − 𝑀𝑀𝐶𝐶 = 0 is maximised. Profits are apparently maximised at a quantity (7.41) 𝑞𝑞 ∗ = 𝑚𝑚−𝑀𝑀𝐶𝐶 2𝑏𝑏 . Obviously, there is an economic interpretation for equation (7.41): 𝑎𝑎 describes the maximum willingness to pay of the customers in this market and 𝑀𝑀𝐶𝐶 describes the marginal costs of the firm, so that the difference 𝑎𝑎 − 𝑀𝑀𝐶𝐶 represents a simplified measure of the competitiveness of the firm, where 1/ 𝑏𝑏 is a measure of the size of the market. If one now differentiates equation (7.41) partially with respect to all individual influencing variables in order to determine effects on competitive strategy, then with <?page no="267"?> 7.3 Monopoly and dominant firms 267 (7.42) 𝜕𝜕𝑞𝑞∗ 𝜕𝜕𝑚𝑚 = 1 2𝑏𝑏 > 0; 𝜕𝜕𝑞𝑞∗ 𝜕𝜕𝑀𝑀𝐶𝐶 = − 1 2𝑏𝑏 < 0; 𝜕𝜕𝑞𝑞∗ 𝜕𝜕𝑏𝑏 = − 𝑚𝑚−𝑀𝑀𝐶𝐶 2𝑏𝑏2 < 0 we get, that an increase in willingness to pay 𝑎𝑎 , a decrease in marginal costs 𝑀𝑀𝐶𝐶 and a growth in market size 1/ 𝑏𝑏 each have a positive effect on the optimum firm size 𝑞𝑞 . ► Figure 7.13 shows this abstract relationship in concrete terms for Deutsche Post against the background of current changes in the demand and cost situation. Figure 7.13: Strategy development for the number of post offices in Germany. All three relevant parameters reinforce each other and point in a clear direction with regard to private letter mail: the number of postal branches 𝑞𝑞 will be further reduced in the next few years. In particular, the clear decline in competitiveness 𝑎𝑎 − 𝑀𝑀𝐶𝐶 - the willingness to pay 𝑎𝑎 decreases and marginal costs 𝑀𝑀𝐶𝐶 increase - leads to a downward adjustment of the firm size in order to ensure survivability. If the respective cost function (provided by the controlling department) and the demand function (provided by the marketing department) are explicitly available for a firm, the strategy department can directly determine an optimum number of branches and their future development over time, as is frequently the case in the financial services industry (Schwartz et al. 2017). Measurement of market power and its influencing factors Market power in general enables a firm to influence individual or even all strategic parameters without competitors being able to exploit this. Market power can take many forms: for example, firms restrain the competitive process and prevent market entry; dictate conditions to suppliers; force customers to agree to unfavourable general terms and conditions; set excessive prices compared to competitors; or offer poor quality. All of these forms of market power can exist without customers switching to competitors. Market power often goes hand in hand with a high market share, but is not necessarily connected to a large market share. In the specific case <?page no="268"?> 7 Perfect competition, monopoly and competition policy 268 of pricing, market power can also be measured: market power measures the ability of a firm to raise price above marginal cost. The profit function (7.39) can also be expressed as (7.43) 𝜋𝜋(𝑞𝑞) = 𝑅𝑅(𝑞𝑞) − 𝑇𝑇𝐶𝐶(𝑞𝑞) = 𝑝𝑝(𝑞𝑞)𝑞𝑞 − 𝑇𝑇𝐶𝐶(𝑞𝑞) . If one differentiates (7.43) with respect to the optimal quantity 𝑞𝑞 and rearranges it to (7.44) 𝑝𝑝(𝑞𝑞) + 𝜕𝜕𝑝𝑝(𝑞𝑞) 𝜕𝜕𝑞𝑞 𝑞𝑞 = 𝑀𝑀𝐶𝐶 we again obtain the condition that marginal revenues equal marginal costs. If we now divide (7.44) by price 𝑝𝑝 and rearrange and plug in the price elasticity of demand for 𝑝𝑝/ 𝑞𝑞 𝜕𝜕𝑞𝑞/ 𝜕𝜕𝑝𝑝 = 𝜀𝜀 , we get (7.45) 1 + 𝜕𝜕𝑝𝑝(𝑞𝑞) 𝜕𝜕𝑞𝑞 𝑞𝑞𝑝𝑝 = 𝑀𝑀𝐶𝐶 𝑝𝑝 → 1 + 1𝜀𝜀 = 𝑀𝑀𝐶𝐶 𝑝𝑝 ⇒ 𝑝𝑝−𝑀𝑀𝐶𝐶 𝑝𝑝 = − 1𝜀𝜀 . The price-cost margin (𝑝𝑝 − 𝑀𝑀𝐶𝐶)/ 𝑝𝑝 is determined by the level of price elasticity of demand. The price elasticity is smaller when customers have fewer alternatives to buy a product - so it is clear that prices can be raised by a dominant firm. In addition, the Lerner index (Lerner 1934) (7.46) 𝐿𝐿 𝑖𝑖 = 𝑝𝑝−𝑀𝑀𝐶𝐶 𝑝𝑝 𝜖𝜖[0; 1] as a measure of market power, explains that firms under perfect competition have no market power: output is chosen so that price equals marginal costs, and (7.46) converges to zero. Conversely, the Lerner index for a monopolist with complete market dominance can rise to one. ► Table 7.1 gives the Lerner indices for various industries. Obviously, in some industries such as meat and sausage retailing, coffee roasting, petrol stations, or breweries there is almost perfect competition with price-cost margins close to zero. On the other hand, financial service providers, manufacturers of breakfast cereals, the tobacco industry, or electricity firms sometimes have significant market power and can achieve significant positive price-cost margins. However, market power does not necessarily go hand in hand with high profits for the firms, as the price-cost margin does not take fixed costs into account. Market power in various industries Industry Lerner-Index meat and sausage retailing 0.00 breweries 0.01 coffee roasteries 0.06 petrol stations 0.10 manufacturing 0.18 finance and insurance 0.24 machinery 0.30 <?page no="269"?> 7.3 Monopoly and dominant firms 269 retail trade 0.31 railroad 0.40 financial services and banking 0.40 motor vehicles 0.43 breakfast cereals 0.45 tobacco 0.67 Table 7.1: Market power in selected industries (Source: Tremblay and Tremblay, 2012, p. 322, and research cited therein). The extent to which market power advances and translates into profits depends empirically, in addition to the cost structure and technology of an industry, on the interaction of the following variables (Schmalensee 1989, Caves 2007 and ► Chapter 4). Price elasticity of demand - the stronger and more urgent the customers’ need for a product, the less elastic demand is, and the higher the market power of incumbents tends to be. This is true for one firm as shown above, but also in competition for several firms if they invest heavily in marketing and branding (► Chapter 1). Innovation intensity within the industry - if a firm can use innovation to either offer qualitatively better products (product innovations to increase vertical product differentiation), or produce at lower marginal costs (process innovations), then opportunities emerge to increase profits. These profits can then also be exploited permanently in the case of market power. However, if this innovation advantage ("pioneering profits") is quickly made up for by high innovation intensity of competitors, only temporary profits are achieved. Number of competitors - as the number of firms in an industry increase, market power is typically reduced. Competitors will try to gain market share by lowering prices and tend to reduce the price-cost margin. If there are significant barriers to entry that effectively deter entry, then market power of incumbents is typically significantly higher. Interaction of firms - the intensity and nature of competition (see also ► Chapter 10 on Cournot vs. Bertrand competition) significantly affects the ability of firms to exercise market power in an industry. If firms cooperate implicitly (or even explicitly), behave in a coordinated manner (collusion), or even form a cartel, then market power increases significantly. Pricing strategy of a dominant firm From the considerations above on profit-maximising condition that marginal revenues equals marginal costs, there follows also a simple rule of thumb for pricing strategies of a dominant firm. If one rearranges equation (7.45) to (7.47) 𝑝𝑝 = 𝑀𝑀𝐶𝐶 1+1𝜀𝜀 , then it follows that a dominant firm must make a mark-up on marginal costs 𝑀𝑀𝐶𝐶 depending on the level of price elasticity of demand 𝜀𝜀 (so called mark-up or cost-plus pricing). The lower the price elasticity of demand, the larger the mark-up on marginal costs and vice versa. In a <?page no="270"?> 7 Perfect competition, monopoly and competition policy 270 situation of perfect competition, in which customers can switch to any other firm and the price elasticity is correspondingly very high, it follows that the price cannot exceed the marginal costs. This relationship also applies, for example, across distribution channels: a bottle of Heineken beer costs less in a local supermarket than at night at the petrol station, and again significantly less than in a Heineken airport lounge. While the marginal cost per bottle of beer is almost identical, the price elasticity of demand is significantly different - Heineken and its distribution partners can leverage this for higher prices. From a management perspective, firms with significant market power and facing rather low competitive intensity need to conduct extensive market research to precisely determine the shape of the demand curve and the level of price elasticity in order to determine an optimum pricing strategy. Firms under perfect competition and high competitive intensity do not have to do this: here it is sufficient to accept the competitors' price as the market price and just imitate it. Determinants of pricing strategies in % cost plus pricing competitors‘ pricing other strategies Germany all industries 73.0 17.0 10.0 low intensity of competition 78.9 9.4 11.7 high intensity of competition 69.8 22.5 7.6 Eurozone all industries 54.3 27.1 18.7 low intensity of competition 63.6 14.7 21.7 high intensity of competition 49.8 35.1 15.1 Table 7.2: Determinants of pricing strategy in Germany and the Euroarea. (Data source: Alvarez and Hernando 2006). In fact, this picture is also found in empirical studies: although firms use industryor firm-specific methods or follow seasonal patterns, it can be seen from ► Table 7.2 that if competition is intense, competitors' prices are more likely to be used to determine pricing strategy; if competition is weak, a cost-plus method is more likely to be adopted (Alvarez and Hernando 2006 and Klenow and Malin 2011). <?page no="271"?> 7.4 Restraints of competition, competition policy and competition authorities 271 7.4 Restraints of competition, competition policy and competition authorities Perfect competition maximises economic welfare in a society - this is the key motive for free competition in markets without governmental intervention. Workable competition leads to lower prices for customers and enables innovative firms to enter the market. As a result, customers can freely choose between many offers and decide on the quality they consider to be just right. In addition, less efficient firms are regularly driven out of the market by superior competitors. In some industries, however, competition works only to a limited extent. We see inflated prices; bad products are driving good products out of the market; market entries are not taking place even with high profit incentives; individual customers or customer groups are discriminated against; innovative business models are held back and a lot more. The functioning of competition and markets can be limited by two types of cause: restraints of competition and market failure. Restraints to competition are caused by active behaviour of firms, whereas market failure is due to specific characteristcs of a market. Restraints of competition can arise under free competition if market-dominant firms actively restrict or inhibit the competitive process and, thus, cause welfare losses for society, for example, through cartels or creation of barriers to market entry. Market failure, on the other hand, means that institutional features of a market restrict the functioning of a market or that markets do not come into being - prices, thus, lose their signalling and coordinating function. If competition does not work in an appropriate way, the state can intervene with competition policy (which in the US is called antitrust policy) to ensure that competition and markets work appropriately. Whether and to what extent the state takes action in markets depends on political decision-making and the chosen competition policy framework, which can follow one or a mix of the following four approaches (Shepherd 1991, Neumann 2000 as well as Haucap and Schmidt 2013). In case of a laissez faire approach, the state relies on the self-healing powers of the market and completely refrains from intervening in markets and competition. However, this can again and again lead to strong market concentration processes and a potential abuse of market power. The structure approach uses only weak intervention: here the state generally defines prohibited behaviour through legislation and laws in order to maintain the competitive process. These rules usually apply to all markets and industries in the same way: examples in Germany are the Gesetz gegen Wettbewerbsbeschränkungen (GWB, Act against Restraints of Competition) or the Gesetz gegen unlauteren Wettbewerb (UWG, Act against Unfair Competition). Within the regulation approach, there is stronger intervention: here, the state determines general or industry-specific rules for competition and permissible behaviour. Firms are not allowed to deviate and abuse is monitored. Such regulations are often industry-specific: for example, competitive processes in the German financial services industry are regulated, e.g by the Kreditwesengesetz (KWG, German Banking Act), the Mindestanforderungen an das Risikomanagement (MaRisk, Minimum Requirements for Risk Management) or the Wertpapierhandelsgesetz (WpHG, German Securities Trading Act). <?page no="272"?> 7 Perfect competition, monopoly and competition policy 272 In the most extreme case of an ownership approach, the state severely restricts entrepreneurship and private ownership, operates essential firms itself and puts firms under state control in the event of imminent insolvency or in the supposed interest of customers. The typical problem with this approach is that the state usually proves to be a 'bad entrepreneur', neglecting innovations in favour of maintenance subsidies and increasing bureaucracy. In Germany, this has happened, for example, for Commerzbank (state aid and state participation in 2009 in the course of the sovereign debt and financial crisis) and for Lufthansa (state aid in form of loans in 2020 during the Covid-19 pandemic). Competition policy in market economies is often situated between these extremes: through a regulatory framework, a combined structure and regulation approach attempts to promote private ownership, entrepreneurship and innovation, but at the same time emergence of dominant positions is monitored to prevent the abuse of dominant positions. The effect of competition policy on improving the functioning of markets is empirically reflected in higher macroeconomic growth rates, rising industry-specific productivity and higher innovation rates in the long run (Buccirossi et al. 2013 and Duso 2015). Competition policy comprises legislation and case law directed at restrictive practices, limiting market power and preventing the abuse of a dominant position. We see this in the European Union as part of the antitrust and cartels legislation - especially Articles 101 to 106 of the Treaty on the Functioning of the European Union (TFEU); in the UK through the Competition Act; and in Germany by the Gesetz gegen Wettbewerbsbeschränkungen. Competition policy is implemented by national and international authorities that attempt to uncover restrictive practices, prohibited collusion and cartels, or to prevent market failures due to natural monopolies - in Germany by the Bundeskartellamt (Federal Cartel Office) and the Bundesnetzagentur (Federal Network Agency), among others; at the European level by the Directorate-General for Competition of the EU Commission; in the UK by the Competition and Markets Authority; and by the Federal Trade Commission in the US. Market failures and restrictions of competition can seldom be completely separated empirically due to a multitude of interactions. Thus, the state addresses them with three overlapping groups of policies as follows. Competition policy in a strict sense aims at limiting and preventing restraints of competition in the form of cartels, abuse of market power, or mergers. Regulation is mainly used to limit and straighten market failures, for example, due to incomplete or asymmetrical information (e.g. the control of price mechanisms on German stock exchanges) or in the case of natural monopolies (especially in infrastructure industries such as railways, postal services, electricity or telecommunications). Market design implies the coordination of key market conditions and market mechanisms as well as market participants. Economists and economic authorities act as engineers of markets, so that market failures are proactively reduced and the emergence of restrictions on competition is counteracted (e.g. mobile phone licence auctions, definition of trading models and market supervision in securities exchanges or emissions trading). <?page no="273"?> 7.4 Restraints of competition, competition policy and competition authorities 273 Competition policy and regulation, thus, take two different perspectives. In markets where competition is desirable and seems possible, competition policy basically fulfils the task of preventing firms from behaving in a way that restricts competition. Regulation, on the other hand, is used when, due to market failures per se, competition is either not expected to work or to come about, or when competition between firms is deliberately to be inhibited. The intervention of governmental authorities in markets happens in one of two ways. Firstly, either certain behaviour or strategies are generally prohibited, as in the case of cartels or prohibited agreements. Secondly, competition in a specific market is regulated: examples include the telecommunications industry (e.g., obligatory approval of tariff models in mobile telephony); the financial services industry (e.g., minimum capital requirements for a credit institution); and the pharmaceutical industry (e.g., approval of medicines). Welfare losses and restraints of competition Models of perfect competition and monopoly can serve as an initial yardstick for assessing whether and how well competition works. In perfect competition, market entry by innovative firms is possible due to the lack of market entry barriers; in contrast, a dominant firm can hinder market entry and innovation. As a result, economic welfare is maximised in perfect competition. On the other hand in the case of a monopoly or dominant firm, welfare losses regularly occur. ► Figure 7.14 summarises the key results from ► Sections 7.2 and 7.3 from a competition policy perspective. Figure 7.14: Social costs of competitive restrictions. If one compares monopoly and perfect competition given an identical demand situation and constant marginal costs, it is immediately apparent that prices are higher in a monopoly and the quantities offered are lower than in perfect competition. A part of the consumer surplus is transformed into profits of a monopolist (here to be seen in the producer surplus 𝐴𝐴 ) and in addition welfare is diminished by − 𝐵𝐵 , i.e., there is a static welfare loss. However, this difference <?page no="274"?> 7 Perfect competition, monopoly and competition policy 274 between perfect competition and monopoly also means social costs, among other things (Posner 1975): some customers can no longer afford the product because of the now higher price, as their willingness to pay or income is insufficient; some customers no longer purchase the product because the quantity offered is reduced in order to be able to push through higher prices; and/ or the number of employees required is lower because production is reduced below the level of perfect competition due to existing market power. In particular, there is a dynamic welfare loss: innovation and technical progress slow down due to the lack of profit incentives, because a dominant firm or a monopoly achieves a permanent profit and does not have to fear competitors. One could conclude from the comparison of perfect competition and monopoly that governmental action and legislation in the field of competition policy should aim to bring about a competitive situation equal to perfect competition. However, such a view on competition policy neglects, on the one hand, that economies of scale exist in many industries and lead to minimum efficient sizes. On the other hand, dynamic characteristics of competition are ignored: successful innovation processes enable firms to assume a monopoly or dominant position, at least temporarily, to realise high profits. In some cases firms, especially those based on digitalisation and disruptive innovations, displace previously incumbent firms and hence change the market structure (Schumpeter 1911 and 1950, von Hayek 1969, Geroski 1998 as well as Sidak and Teece 2009). However, innovations and technological progress often lead to price reductions for customers as well as higher product quality and larger variety. It is, thus, clear that competition policy in particular must not limit innovations and opportunities for new firms to enter the market. The main objective of competition policy is then to ensure workable competition (instead of protecting existing firms), to support freedom of entrepreneurial action, which creates incentives for innovation and supports welfare gains. Market failure and regulation Market failure describes a limited or missing functionality of a market due to institutional characteristics or special features of a particular market - as a result, market processes, as described in perfect competition, are not happening. Market failure often is due to lacking or incomplete property rights: if no one owns a product, nobody can buy it. As a consequence, price formation via supply and demand in the market is either impossible or restricted. In addition, products are either not offered or not all product variants or qualities desired by customers are offered. Essentially, market failure can be traced back to four sources: externalities (also called external effects), public goods, asymmetric information, or natural monopolies (see Fritsch 2011, Knieps 2008 and Bator 1958). Market failure is a necessary but not sufficient indication for governmental action in a market by means of direct government intervention in market transactions (e.g., government provision <?page no="275"?> 7.4 Restraints of competition, competition policy and competition authorities 275 of schools), through regulation (e.g., pricing of medicines) or through bans, minimum requirements, or other legislation. Governmental authorities may intervene in a market through regulation - in Germany, for example, through the Bundesnetzagentur or the Bundesaufsicht für Finanzdienstleistung (Federal Financial Supervisory Authority) in certain markets or industries, as well as through general provisions such as the General Data Protection Rule, labour market legislation, or the price indication regulation (Preisangabenverordnung, PAngV). Market failure due to externalities in the absence of property rights In many markets, the behaviour of one market participant has unwanted effects (benefits or costs) on other market participants - these effects are called external effects or externalities. External effects describe mostly indirect positive or negative influences of a market participant on other market participants or even third parties. Typically, there is no price charged or even available for these effects. This is because pricing is rather impossible because of the complexity of identification of cause and effect of positive or negative externalities. If property rights are not clearly defined or ascertainable, production or consumption in a market can cause positive or negative external effects for other market participants or third parties outside the market - loud music can be a joy or a nuisance for the neighbours. Due to the lack of property rights, no price can be established for these externalities and, thus, this market cannot work properly. Negative externalities arise, for example, from environmental pollution or CO2 emission with effects on global warming, i.e., as costs for society that are not, or not fully, borne by the polluter. In addition, it may not be possible to distinguish who is affected by environmental pollution and in what way, since many effects are indirect, global, or time-delayed. Internalisation, i.e., allocation of costs to the polluter, then requires governmental action, e.g. through a tax on CO 2 emissions. Positive external effects are often due to innovations that - if not protected - are adapted free of charge by other firms or in other markets. If property rights to knowledge do not exist or are not secured, there are either no or only reduced incentives for firms to develop new knowledge. Firms will then either not invest at all or invest too little in research and development. This can also lead to a decline in start-ups: if start-ups fear that their innovations will be quickly adapted or imitated by incumbents, either the start-up does not take place, or the fixed costs increase significantly due to the foreclosure of knowledge. Typically, in the case of negative externalities, too much is produced; in the case of positive externalities, too little is produced. In both cases, the individual private costs do not match the collective societal costs. For the correct allocation and attribution (internalisation) of these external effects, the government sets prices or taxes. Another option is to create markets for external effects - examples include certificate trading for pollution rights in case of negative external effects, and patent protection for innovations with the possibility of licensing in case of positive external effects. In both cases, property rights are introduced to reduce the extent of market failure - too much pollution, too little innovation. In the case of a dynamic CO2 tax on carbon dioxide emissions, this also has a coordinating effect on markets, as incentives are created to develop climate-friendly or climate-neutral technologies. In the same way, patent protection creates incentives for R&D investments - although patent protection also leads to at least a temporary monopoly position for innovative firms. In contrast, basic research at uni- <?page no="276"?> 7 Perfect competition, monopoly and competition policy 276 versities in Germany and many other countries is financed by the state: here, too, market failure leads to too little investment in R&D, although the knowledge created here is largely made available to the public through scientific publications. Market failure due to public goods Public goods are products or services that can be used by many customers at the same time without the individual customer suffering any disadvantages or restrictions in use. A lighthouse on the coast serves all captains and ship crews equally, regardless of how many are currently looking at it. These products are non-rival and non-excludable in consumption. Non-excludability means that property rights cannot be enforced - clean air is available to all people in a given region, but neither can individuals be excluded from consumption, nor can a price be charged. Similarly, people do not compete for clean air in public spaces. Since no prices can be charged, no firm has an incentive to offer this product. Similarly, all people in a country use national security and defense or benefit from a functioning legal system. Non-rivalry and non-excludability naturally lead to the fact that no firm has an incentive to produce these products, since no pricing, no exclusion of non-paying customers, and consequently no profits are possible - i.e., a market failure, because no supply is created. Typically, public goods are then financed by taxes or fees and offered or provided by the state. However, true public goods are rare: prices can be determined and charged for numerous services or products, e.g., in the form of motorway tolls or road tolls, for private schools or private universities, or pay-TV and streaming providers. New business models (advertising-financed radio instead of fee-financed public broadcasting stations) or technologies (blockchain or personalisation) make it now possible to establish excludability, set prices and transform supposedly public goods into normal products. Another reason can be that the state offers public goods in poor quality or insufficient quantity, so that firms have an incentive to offer highquality products for a target group with a high willingness to pay. Market failure due to information deficits In many markets there is incomplete or asymmetric information, i.e., some market participants are better informed than others (see also ► Chapter 1 on second-hand markets). Asymmetric information can occur in many situations: insiders in the financial market usually have better information than private investors; firms usually know their own terms and conditions better than their customers; and a university knows the quality of teaching better than potential students. To some extent, such information deficits can be compensated for by signalling on the supply side or information procurement on the demand side, albeit at sometimes considerable transaction costs (see also ► Chapter 2 on search, experience and credence goods). If these marketbased possibilities do not work sufficiently, however, asymmetric information can lead to market failure if uncertainty about the characteristics or quality of products does impede demand, or the willingness to pay is low. As a result of the lack of demand, firms will not offer the product at all (or will only offer it in insufficient quantity or in inferior quality), so that the functioning of the market is limited, and economic welfare is reduced. <?page no="277"?> 7.4 Restraints of competition, competition policy and competition authorities 277 This is true, for example, in the pharmaceutical industry, where high R&D expenditures face an uncertain demand for medicines, but a high quality of the products is necessary. In order to reduce or prevent the possible market failure, firms can engage in signalling (e.g., through marketing, warranties, or quality seals) to attract demand and increase willingness to pay. In addition, the state can reduce customer uncertainty through liability rules, certification or approval procedures for products or services: in the pharmaceutical industry, medicines go through testing and approval procedures and are sometimes subject to prescription requirements or only available via pharmacies. Market failure as a result of information asymmetries is also a possible cause of banking and financial crises: customers are typically unable to assess the characteristics and risks of bank products, such that customers purchase non-transparent products with unknown risks, as in the run up to the subprime and financial crisis of 2007. This asymmetric information in the B2B area (especially securitisation of CDO and CDS instruments) and in the B2C area between banks and their customers ('Lehman certificates' purchased by retail customers at the savings banks), can then lead to a collapse of the banks in the event of a default of the products. As a result, national and international regulation of the financial services industry has been strengthened. Certain products for individual client groups have been banned and better labelling (product information sheet, MiFiD, records of consultations, etc.) has been introduced to protect investors. In addition, the requirements for bank licences as well as for equity capital and liability of managers have been improved. However, market failures can also occur if customers have better information than firms: for example, customers know more about their risks before taking out a contract (e.g., health insurance) than an insurance firm. This can lead to adverse selection (see also ► Chapter 4): only customers with high risks due to previous illnesses ask for insurance, so that insurance is not possible on the basis of the law of large numbers and a pooling of different risks. The state tries to prevent a possible market failure here by making insurance obligatory for everyone. Good to know │ Why does the state support the financing of start-ups? In a perfect capital market with perfect information, all firms and projects are financed that achieve a positive net present value under given financing conditions - accordingly, the financing of a promising start-up can theoretically never pose a problem. From an economic and competition policy perspective, however, financing bottlenecks seem to impede or prevent start-ups (European Commission 2003 and OECD 1998). Empirical evidence often shows that small firms with limited access to finance show reduced survival probabilities and growth (Reid 2003) - but also that financing constraints may play only a subordinate or endogenous role in business failure (Uusitalo 2001). In particular, access to external equity has long been identified as a development barrier and obstacle for young and small firms (Harrison et al. 2004 and Münter 2020). However, start-ups and young firms have numerous peculiarities in corporate finance compared to incumbent firms, which essentially result from imperfect capital markets in addition to limited opportunities for equity financing and borrowing (further Betsch et al. 2016, Fazzari et al. 1988, Myers and Majluf 1984 and Kerr and Nanda 2009). <?page no="278"?> 7 Perfect competition, monopoly and competition policy 278 Typically, the risk of investing in a start-up is higher than investing in an incumbent. Specifically, this means that the probability of a default on equity or debt is higher, while the variance of repayments (profits, dividends, etc.) is also higher. This higher risk of default is due to uncertainties regarding the market acceptance of the new product or service, the technology, the scalability, the legal and political conditions, as well as the performance and survivability of the founding team. In the planning and founding phase, there is incomplete information about the product, the pricing opportunities, the competitive situation and possible customer expectations. In addition, a start-up is characterised by implicit and vague assumptions of the firm founders, which are difficult to describe or grasp as intangible success factors or competitive advantages. As a result, there is a high degree of asymmetric information between founders on the one hand, and potential investors on the other. This asymmetric information is reinforced by the lack of or severely limited availability of planning data and planning quality. The combination of incomplete and asymmetric information with higher risk can lead to market failure in the financing of start-ups - i.e., too low financing volumes or high financing costs. If the state assesses the risk appetite or liquidity of private actors and banks in the capital market for financing the foundation or growth of young firms as low due to this market failure, it can intervene by means of regulation. Governmental or public institutions and development banks, thus, regularly provide access to financial support, for example, by means of venture capital for start-ups in their early stages. This financing support is often aimed at the development or acceleration of start-ups per se. In addition, there are structural support programmes differentiated according to regions, technologies, or the proximity of start-ups to science or universities. Market failure due to natural monopolies Market failure can also be based on market power, in this case essentially due to technological causes that lead to natural monopolies via indivisibilities in the production process, distinct economies of scale, or network effects (see also ► Chapter 6). As a result, a firm can initially drive other firms out of the market by undercutting prices and subsequently deter market entry through structural entry barriers, so that no workable competition emerges. Conversely, it may be that under these circumstances no firm enters investment-intensive industries with high sunk costs at all: all potential firms see the risk of being forced out of the market, so that the necessary investments for market entry do not take place. Regulation here aims to cure a market failure that has arisen or is expected to arise in markets with natural monopoly characteristics. In these markets, competition would lead to inefficiency and therefore dominant firms are regulated. In the 20th century, regulation was often implemented in the form of state-owned firms, especially in infrastructure industries such as railways, postal services, energy supply, or air transport. This left the market to one firm at a time (either by putting an existing firm under state control or by establishing a state monopoly), thus, solving the investment dilemma. However, state owned enterprises have no incentive to innovate and tend to be inefficient in terms of bureaucracy and budget maximisation due to a lack of competition. As a consequence, some <?page no="279"?> 7.4 Restraints of competition, competition policy and competition authorities 279 of these state monopolies were only able to compete to a limited extent when deregulated or privatised (Leibenstein 1966 and Boardman and Vining 1989). However, in the course of coordinated deregulation, a significant increase in welfare often ensues. For example, in the German telecommunications industry, numerous competitors entered the industry after the market was opened in 1998. As a result, production volumes (telephony minutes, data usage, number of connections and contracts) increased significantly, the rate of innovation grew and, last but not least, prices declined continuously (Dewenter and Haucap 2004, Sickmann 2018 as well as Bundesnetzagentur 2019). However, the regulation of a market with characteristics of a natural monopoly can also take place within the market. State authorities regularly choose various forms of price regulation against dominant firms in order to reduce welfare losses compared to a monopoly price. For this purpose, the balance sheet and P&L are analysed to determine the cost function of the firm with market power. Based on this information, a price between the level of perfect competition and the monopoly is set and approved in negotiations - e.g., in Germany between Deutsche Post and the Bundesnetzagentur for letter postage. A regulatory authority can set fixed or dynamic price ceilings (price-cap procedure), establish a return-oriented (rate-of-return procedure), or a cost-oriented (cost-plus procedure) price regulation. The challenge of all three procedures is that the regulated firm has better information than the regulatory authority, so that as a result of asymmetrical information regulation is usually not efficient. In these cases, firms have no incentive to disclose their true cost structures or to implement cost reductions and innovations. As a consequence, the competitiveness of the firm decreases over time. Recent methods try to create incentives for innovation such that the regulated firms themselves optimise their actual costs and remain competitive even in the case of deregulation (Laffont and Tirole 1986 and 1991 as well as Cabral and Riordan 1989). Incentivecompatible regulation is used in Germany by the Bundesnetzagentur, for example in energy markets: revenues and their development are fixed for a period of five years. Firms then have an incentive to reduce costs of electricity generation and distribution and to implement innovation in order to make profits. The cost reductions achieved are then the starting point for determining revenues in the next regulatory period - cost reductions and, if necessary, price reductions for customers are, thus, achieved without the Bundesnetzagentur having to set firmspecific price ceilings or cost structures. Market design and matching-markets Market design is a new approach, which regulates markets by conceptually reversing the logic of a market. Instead of a market (and thus possibly a market failure) emerging in the first instance, which is then monitored or regulated, economists design a market and its mechanisms in advance so that competition can work as desired (Roth 2008). These procedures are used, for example, in auctions of mobile spectrum licences in the telecommunications industry (Binmore and Klemperer 2002 and further ► Chapter 8). The telecommunications industry is characterised by extensive economies of scale, so that the largest firm could potentially achieve a dominant position. In order to prevent this, a sufficient number of firms, each with limited capacity, are licensed via an auction so that competition is facilitated. Market design, thus, al- <?page no="280"?> 7 Perfect competition, monopoly and competition policy 280 ways focuses on the specifics of a certain market - it does not look for general rules that always apply, but for specific solutions to reduce or prevent a possible market failure. Market design is applied in markets that are coordinated via prices, but more often in so-called matching markets in which no prices exist, but precise allocations or assignments are to be made. Examples of matching markets are allocation of university admissions, allocation of social housing, regional distribution of refugees, design of evaluation platforms on the internet, design of dating platforms, or the organisation of organ donation (Roth 2015 and Jackson 2013). Especially in the case of organ donation, market failure can often be observed - a large number of people are waiting for an organ transplant, but the number of potential donors (especially in anonymous relationships without knowing the recipient) is too small; moreover, the donor organs are often incompatible with the recipients. A matching solution here can be based on paired donor willingness. With this approach, patients first bring possible donors, albeit with incompatible organs, into possible exchange relationships bilaterally from their family and friends community. These pairs of patients and donors are then brought into a ring exchange with other pairs via multilateral crossover donations - the more crossover donations are brought in, the larger the possible ring exchange, the more likely the compatibility between some donor and some recipient. In this way, waiting lists and waiting times are significantly shortened and possible black markets or bribery in organ trade are reduced (Roth et al. 2007). Restraints of competition through abuse of dominant position Unlike market failure, which is due to certain characteristics of a market, restraints of competition are caused by firms’ behaviour. They include all kinds of active or implicit behaviour or strategies with the aim of reducing the intensity of competition, reducing the number of competitors, increasing competitors' costs, imposing higher prices and obviously increasing profits. Restrictions of competition can be attributed to three main causes: abuse of a dominant position, coordinated behaviour of firms as a cartel, and mergers and acquisitions. These types of strategies are addressed by international and national legislation, in Germany by the Gesetz gegen Wettbewerbsbeschränkungen (GWB) and are mainly monitored by the Bundeskartellamt. In the following, we focus on German authorities and specifics, however keeping an eye on EU and US authorities as well. Dominant firms can use market power not only to raise prices, but also to restrict or eliminate the competitive process. In most cases, strategic action is applied to block potential competitors from entering the market, to make access to customers impossible, or to force current competitors to leave the market, for example, through exclusive contracts with suppliers. According to Section 18 of the GWB, a firm is dominant if it is without competitors in the relevant market, is not exposed to substantial competition, or has a superior market position in relation to its competitors. Under the GWB, a single firm is presumed to be dominant if it has a market share of 40 percent or more. A group of firms is considered dominant if a maximum of three firms together reach a market share of 50 percent, or if up to five firms have a market share of two thirds (see also ► Chapter 4 on horizontal concentration). In case of a dominant position in multisided markets and platforms several additional factors have to be analysed including direct and indirect network effects; possible switching costs of customers; the role of <?page no="281"?> 7.4 Restraints of competition, competition policy and competition authorities 281 data in competition; as well as the intensity of competition with respect to singlehoming or multihoming (see also ► Chapter 2 on specifics of digital markets and products). Abuse of a dominant position is covered in Sections 19 to 22 of the GWB by criteria for prohibited conduct. These include setting excessive prices, arbitrary pricing below marginal costs (predatory pricing) to drive out competitors, best price clauses, exclusive contracts and supply boycotts, hindering competitors (e.g., by preventing number portability in telecommunications), bundling or tying, denying or limiting access to networks (e.g., network neutrality) or essential facilities. In Germany, the Bundeskartellamt has investigated, amongst many others, the following cases and issued prohibitions in recent years: most-favoured customer clauses and best-price strategies in hotel bookings by the platforms HRS, Expedia and booking.com; ancillary copyrights between Google and German press publishers; ticket distribution by Deutsche Bahn and its competitors; and exclusive agreements for ticket sales by Eventim. In 2015, for example, the Bundeskartellamt prohibited the best-price clauses of the hotel booking platform booking.com. Best-price clauses guarantee customers the cheapest room rate but restrict competition. How is this achieved? On the one hand, strategic market entry barriers for other hotel booking platforms are established through exclusive contracts. On the other hand, indirect network effects force hotels to enter into distribution partnerships with booking.com. Consequently, those hotels who are not listed with booking.com significantly lose customer access. Hotels can no longer decide on their own prices as a result of the price fixing, which in turn leads to an expansion of distribution commissions at the expense of hotels and customers. At the European level, the EU Commission fined Google EUR 4.3 billion in 2018 for abuse of a dominant position in the Android smartphone operating system: Google had imposed unlawful restrictions on both smartphone manufacturers and mobile network operators since 2011 in order to develop a dominant position for general internet search services (ec.europa.eu/ commission/ presscorner/ detail/ en/ IP_18_4581). In 2019, the Bundeskartellamt in Germany prohibited Facebook from combining user data from different Facebook services such as Instagram or Whatsapp due to an abuse of a dominant market position and enforced an adjustment of the general terms and conditions of business and use: Facebook had not given customers a choice as to whether and in what way data from WhatsApp, Instagram, Facebook and other services was collected and linked, thereby, limiting competition with other social media platforms. Restraints of competition through cartels and collusion Sometimes firms realise, that competing on their own against everyone else is difficult - life could be much easier if firms in an industry make agreements on prices or how to divide customers. Restraints of competition with monopoly-like higher prices, lower quantities, and reduced quality can also arise in particular through restrictive agreements or coordinated behaviour by several firms in the form of collective market dominance at the expense of customers or competitors. These agreements can be explicit or implicit. <?page no="282"?> 7 Perfect competition, monopoly and competition policy 282 Explicit collective market dominance - through cartels and collusive behaviour based on contracts or informal agreements, prices or quantities are fixed, regions are allocated. These explicit agreements are often coordinated by informal organisations ('breakfast cartel', 'price reporting centre', 'associations', etc.) and compliance is ensured with sanctions. Implicit collective market dominance - through coordinated behaviour that has proven successful for the firms ('usances', 'recommendations to members', 'agreed market understanding', etc.) without explicit agreements ever having been made. In addition, it could happen that firms can adjust their behaviour to each other, especially in long-term competition, in such a way that all competitors behave as if being in a cartel ('spontaneous parallel behaviour'). Firms, thus, individually forego strategic decision-making freedom because they expect advantages from being part of a collective decision. Section 1 of the GWB effectively prohibits cartels in Germany. All agreements between firms that restrict competition as well as concerted practices that have the purpose or effect of inhibiting, restricting or distorting competition are prohibited. Agreements and collusion usually focus on the coordination of a strategic parameter of competition. The aim is to reduce the intensity of competition and create strategic transparency among the cartel members. The main reason for collusion is higher profits, but an improved basis for long-term planning (resources, employees, investments, etc.), and the creation of entry barriers through collective market dominance also play central roles. These agreements can basically take two forms: horizontal agreements relate to firms at the same stage of the value chain, for example, supermarket chains; vertical agreements are made between firms along the value chain, for example, food producers as suppliers and supermarkets as distribution channels. Horizontal agreements usually cover one or more competitive parameters: prices and conditions, quality, quantities, regions, behaviour in submissions or behaviour in the event of drastic changes in demand. Horizontal agreements are all the more likely and stable the more homogeneous the products and processes, the smaller the number of firms, the more extensive the entry barriers, and if cost structure and technology of the firms is similar. In these cases it is simply easier to monitor compliance with the agreements within the cartel and, if necessary, to impose sanctions. Vertical agreements try to influence competition upstream or downstream in the value chain: price fixing by manufacturers vis-à-vis distributors (resale price maintenance), service and delivery conditions, exclusivity obligations (obligation to have a car being repaired only at an authorised repair shop or exclusive distribution of Apple smartphones), or distribution obligations (quantity restrictions or discrimination of customer groups). The higher the market power in one's own value chain, the easier it is to implement vertical agreements in other elements of the value chain. In Europe, the Bundeskartellamt and the EU Commission have investigated a large number of cases and imposed fines on cartel members in recent years, including the following: truck cartel involving Daimler, Scania, DAF, Renault/ Volvo, Iveco and MAN with a total fine of EUR 3.8 billion; libor cartel with agreements on interest rates involving Deutsche Bank, Société Générale, Royal Bank of Scotland, JPMorgan, Citigroup, RP Martin, Barclays and UBS with a total fine of EUR 1.7 billion; <?page no="283"?> 7.4 Restraints of competition, competition policy and competition authorities 283 cement cartel involving HeidelbergCement, Schwenk Zement, Dyckerhoff, Lafarge, Alsen and Readymix with a total fine of EUR 330 million; beer cartel involving 12 German breweries with a total fine of EUR 338 million; and food cartel involving more than 20 food producers and supermarkets with a total fine of EUR 242 million. However, cartels can also come about without the direct involvement of humans: for example, price algorithms of different firms can develop patterns through mutual imitation or databased learning that in fact amount to an elevated cartel price (Ezrachi and Stucke 2015, Bundeskartellamt 2020 and Monopolkommission 2018 and further ► Chapter 8). Restraints of competition through mergers and acquisitions Restrictions on competition can also arise if the number of firms in a market is reduced as a result of mergers or acquisitions, thus, increasing horizontal concentration or strengthening vertical ties (see also ► Section 6.4). The aim of mergers is often cost synergies from which customers can also benefit through price reductions. However, this effect is often outpaced by an increase of market power of firms due to a simultaneous horizontal increase in market shares or a vertical expansion of product portfolios, but especially because of a smaller number of competitors. Firms can then use this market power to increase prices. Since these effects are not clear-cut and depend on the firms involved and the market situation, Sections 35 ff. of the GWB stipulates that mergers are not per se forbidden, but that mergers are subject to notification and approval. The Bundeskartellamt examines these cases and prohibits mergers if it is to be expected that competition will be significantly restricted, or that an already existing dominant position will be strengthened (bundeskartellamt.de/ DE/ Fusionskontrolle). If this is not the case, or if the firms can prove that the possible welfare gains from cost synergies via price reductions more than compensate for the increase in market power, the merger is approved (Williamson 1968, Neven and Röller 2005 and Neumann 2016). There are usually (internationally varying) criteria and thresholds in terms of revenues or market shares below which mergers are not considered restrictive of competition per se. In Germany mergers are examined under Section 35 of the GWB if one of the following situations apply. Firstly, the firms involved must have joint worldwide revenues of more than EUR 500 million, and at least one firm involved has revenues of more than EUR 25 million in Germany, and another firm involved has revenues of more than EUR 5 million. Secondly, if the previous thresholds are not reached (especially in the case of takeovers of start-ups), mergers are examined if the purchase price or the transaction volume or purchase price is larger than EUR 400 million and the firms together have revenues of more than EUR 25 million. Similarly, mergers and acquisitions with a pan-European restriction of competition are assessed by the EU Commission based on the EU Merger Regulation (Council Regulation (EC) No 139/ 2004 on the Control of Concentrations between Undertakings). Instead of analysing market shares alone (as was previously common in the market dominance test), the SIEC test (Significant Impediment of Effective Competition, in the EU since 2004, in Germany since 2013) is used to examine whether significant restraints of competition are to be expected in the relevant market. Here, microeconomic models are used to examine and estimate how prices, quantities, <?page no="284"?> 7 Perfect competition, monopoly and competition policy 284 and quality will develop in a market after the merger. This could especially be the case if a dominant firm is not involved in the merger, yet can benefit from the merger through indirect (so-called non-coordinated or unilateral) effects and increase its market power. Merger control aims to prevent the creation or strengthening of market power or dominance through mergers, co-operations, or equity investments. Well-known takeovers in Germany were most recently: Kaiser's Tengelmann by Edeka (2016) in the food retail sector, AirBerlin by Lufthansa (2017) among airlines, E-Plus Group by Telefónica (2014) in the telecommunications industry. However, the German Bundeskartellamt or the EU Commission can tie approval to the fulfilment of conditions (so called merger remedies). For example, following the merger of Telefónica and E-Plus in 2014, the new firm had to sell 30 per cent of the network capacity at fixed prices to operators of virtual mobile networks (MVNOs) in Germany, and sell a radio frequency spectrum to a new network operator (ec.europa.eu/ competition/ mergers/ cases/ decisions). Market power versus corporate strategy High profits of individual firms - such as Apple, International & Commercial Bank of China, or Toyota in recent years - can be explained on the one hand by superior products, good corporate governance, or efficiency, but on the other hand, by market power. In addition, innovations can generate profits that can be used to build and strenghten market power. New digital business models in the shape of multisided platforms, such as Uber, Google, or booking.com, are often seen as a particular threat to existing market structures. These platforms can develop monopoly-like structures due to strong indirect network effects and in some cases there are calls for these business models to be broken up (FAZ 2015). Platforms such as Google, Facebook, or ebay do indeed have outstanding market shares and also market power as a result of entry barriers and strong direct or indirect network effects, often in combination with the lack of possibility for multihoming. Against this background, some basic challenges of competition policy have come up that describe the complexity of competition authorities' decisions as follows. Are firms’ profits due to market power or based on efficiency? High firm profits and/ or large firm size, as with Google, Netflix, or Apple, can be a consequence of market power based on barriers to entry (part of the so-called 'Harvard School' approach) - in this case competition policy must intervene accordingly. However, the profits and firm size can also be due to the fact that this firm is more ambitious, has higher efficiency and better products (part of the so-called 'Chicago School' approach) - in this case the profits are from innovations and indicate workable competition. Is there a causal relationship between dominant position and market share? In the past, a dominant position was often equated with a minimum market share of, for example, 40%. In the meantime, the so-called SIEC test (Significant Impediment of Effective Competition) is used, in a more differentiated way, to assess whether a firm significantly impedes effective competition, also independently of its market share (Bundeskartellamt 2012). Apple (with a market share in Germany in 2019 of 14% for smartphones) is a good example of a <?page no="285"?> 7.4 Restraints of competition, competition policy and competition authorities 285 firm that can exercise market power in their iOS ecosystem, even though their market share clearly lags behind competitors. Where are market boundaries and what is the relevant market? Disruptive innovation and digitalisation are shifting industry and market boundaries. Market shares or profits in a specific market segment then only provide limited information about market power or the functioning of competition. In addition, given multisided markets, many products are offered free of charge at least on one side of the market - thus, pricing loses its informative value with regard to market power. Are the restraints of competition based on intention or are they due to general business conditions? The synchronized pricing of firms can be based on intentional behaviour or agreements to restrict competition - but the pricing can also be justified by essential business conditions (for petrol stations, for example, seasonal demand, technology, crude oil prices, etc.). Are digitalisation and network effects leading to natural monopolies and dominant positions or to market failures? Numerous digital business models, such as multisided platforms, have a tendency towards monopolies due to direct and indirect network effects. However, multihoming (members are registered with several competing platforms) or access to data from large platforms (data compatibility, interfaces and access to essential facilities) can also enable and even foster competition. Competition policy will have to continue to take these special features of digital markets into account, especially in order to further promote innovation without enabling permanent monopolies (Wambach 2016, Haucap and Heimeshoff 2017 and Monopolkommission 2014). With the 9th GWB amendment in June 2017, German competition law has been adapted to the increasing digitalisation of markets in some first steps. In 2021, the next steps were taken through the new Section 19 a GWB with respect to platforms: the Bundeskartellamt may now state that a platform firm has a dominant position not only in one market, but also across several markets; or if, by limiting access to competitively relevant data, the firm has an impact on upstream or downstream markets of other firms. Similarly, the European Commission has proposed two legislative initiatives: the Digital Services Act (DSA) and the Digital Markets Act (DMA). The aim is to create a digital market in which the fundamental rights of all users of digital services are protected and a level playing field to promote innovation, growth and competitiveness is achieved. This should be achieved by limiting the market power of so-called gate keepers to platforms and markets, such as Google, Facebook or Amazon. Good to know │ How do Bundeskartellamt and Bundesnetzagentur share their tasks? Competition and competitive strategies of firms in Germany are essentially monitored and observed by the Bundeskartellamt, the Bundesnetzagentur and the Monopolkommission (Monopoly Commission). The Bundeskartellamt with its headquarters in Bonn and about 350 employees, is the main competition authority. The Bundeskartellamt is tasked with applying the GWB: enforcing the ban on cartels; monitoring corporate mergers and acquisitions; observing <?page no="286"?> 7 Perfect competition, monopoly and competition policy 286 abuse of market power by dominant firms; reviewing federal public contracts; and, since 2017, also reviewing consumer protection in terms of competition policy. By enforcing the ban on cartels, the Bundeskartellamt was able to uncover several large cartel agreements, in particular following the leniency programme introduced in 2000. In proceedings against the firms, fines amounting to several billion euros were imposed, peaking at EUR 1.1 billion in 2014 alone. The Bundeskartellamt is also home to the Market Transparency Unit for Fuels. By means of a real-time reporting obligation for fuel prices of all German petrol stations, an attempt is made to prove suspected prohibited price agreements between petrol stations and mineral oil firms. ► Figure 7.15 shows regular price-setting patterns in which Aral and Shell, as market leaders, very often start price increases at 8 p.m., followed by Esso and Total at 9 p.m. and Jet at 11 p.m.. The main driver for these patterns is the shop closing law: when supermarkets have to close, customers buy urgently needed foodstuffs with a high willingness to pay at the petrol station and are also prepared to accept higher fuel prices. So far, it has not been possible to prove that petrol stations collude on fuel prices (Bundeskartellamt 2017 and FAZ 2017a). The data is passed on to customers via apps or navigation devices to support price competition. Figure 7.15: Petrol prices in brand comparison (Data source: Bundeskartellamt (2017), pp. 19 and 20. E5 prices over the course of a day for Frankfurt and Berlin from December 2015 to May 2016). The Bundesnetzagentur (BNetzA) with more than 2,500 employees in Bonn is the German regulatory authority. In addition to the observation and monitoring of firms in network markets, a main task is the promotion and establishment of competition (deregulation) in network markets with previous state monopolies. Firms in the electricity, gas, telecommunications, postal and rail transport sectors are subject to extensive monitoring by Bundesnetzagentur. These firms must, among other things, have pricing models approved, or allow new competitors access to network infrastructure ('last mile') in telecommunications or energy networks, in order to support workable competition. In addition, the Bundesnetzagentur organises the recurrent auctions of mobile spectrum licences and monitors the coverage requirements to be met by the mobile phone firms ('fast internet'). intraday petrol prices in Frankfurt intraday petrol prices in Berlin <?page no="287"?> 7.5 Summary and key learnings 287 The Monopolkommission, with 20 employees in Bonn, is an independent advisory body for the German government in the area of competition policy, competition law and regulation. In addition, an expert report on the status and development of the competitive situation in German industries is published every two years. 7.5 Summary and key learnings Why are firms not allowed to form a cartel, but to do M&A? Should competition authorities try to make each market a perfectly competitive market? Does Facebook have dominant position in social media - and is Facebook abusing its market power? Perfect competition and monopoly serve as benchmarks for assessing the results of competition (prices and quantities). Perfect competition with a potentially large number of firms, with free market entry and exit, and a lack of product differentiation from the customer's perspective, leads to zero economic profits in the long run. Firms have no strategic degrees of freedom and are forced by competition to be efficient - typically firms' profits oscillate around zero in a succession of innovation, imitation and cost reduction. Economic welfare can be measured in competitive markets by adding consumer and producer surplus. Economic welfare is maximised under perfect competition. Dominant firms with significant market power (or even monopolies), on the other hand, sometimes have enormous strategic power and achieve significant profits. However, the absence of competition - especially in case of state-owned monopolies - often leads to inefficiency and a lack of technical progress because existing profits are not threatened by competition or new firms entering the market, so that there are no incentives for innovation. If competition in markets and industries works only to a limited extent, the government intervenes through a legal framework and regulatory actions to support the competitive process. Competition policy aims to monitor the market power of dominant firms and to detect and inhibit anti-competitive behaviour. In Germany, this is enforced by the Bundeskartellamt (Federal Cartel Office) and the Bundenetzagentur (Federal Network Agency) on the basis of the Gesetz gegen Wettbewerbsbeschränkungen (GWB, Act against Restraints of Competition). Digitalisation and new business models in particular require a fundamental review of competition policy regulations in order to support innovation and dynamic competition to increase economic welfare. Recommendations for further reading More in-depth analysis of competitive markets and general equilibrium theory can be found (with a lot of mathematics) in Mas-Colell, A., Winston, M.D. and Green, J.R., Microeconomic theory, New York 1995, with less mathematics in Perloff, J.M., Microeconomics - theory and applications with calculus, Harlow 2018. A comprehensive and sophisticated account of competition policy is provided by Motta, M., Competition policy - theory and practice, London 2004; from a German perspective, Haucap, J. and Schmidt, I., Wettbewerbspolitik und Kartellrecht - eine interdisziplinäre Einführung, München 2013, and Fritsch, M., Marktversagen und Wirtschaftspolitik, München 2018, offer excellent additions. <?page no="288"?> 7 Perfect competition, monopoly and competition policy 288 Questions for review [1] Describe applications of perfect competition and monopoly as well as their limits, advantages and disadvantages. [2] Explain assumptions of perfect competition. Are these assumptions realistic, are there markets in which there really is perfect competition? What is an optimum strategy for a firm in this setting? [3] Explain whether a firm should remain in the market in case of perfect competition even if it incurs losses. [4] Describe the competitive dynamics of perfect competition over time. [5] How can you determine producer and consumer surplus in a market? What does it explain or measure, what happens to consumer surplus and producer surplus if a minimum price is introduced? [6] What is a monopoly, what can be the sources of monopolies? [7] A monopolistic diamond producer knows very well its demand and cost function with 𝑝𝑝 = 2,000 − 0.002𝑞𝑞 and 𝑇𝑇𝐶𝐶 = 20,000,000 + 0.2𝑞𝑞 for the German market. How many diamonds 𝑞𝑞 for engagement rings for the German market should be produced, what is the price, what is the resulting profit? What is the profit margin, measured as price minus marginal cost? [8] How can you measure monopoly power, what does it say? In contrast, what is a welfare loss (deadweight loss) due to a monopoly? [9] Which authorities in Germany are responsible for enforcing functioning competition? What is the focus of their tasks? [10] Explain possible causes of market failure and restraints of competition. Literature Alvarez, L.J. and Hernando, I., Competition and price adjustment in the Euro area, Banco de Espana Research Paper No. WP-0629, 2006. Bator, F.M., The anatomy of market failure, Quarterly Journal of Economics, 1958, 72, 3, 351-379. Betsch, O., Groh, A.P. and Schmidt, K., Gründungs- und Wachstumsfinanzierung innovativer Unternehmen, München 2016. Binmore, K., and Klemperer, P., The biggest auction ever: the sale of the British 3G telecom licences, Economic Journal, 2002, 112, 478, 74-96. Boardman, A.E. and Vining, A.R., Ownership and performance in competitive environments: a comparison of the performance of private, mixed, and state-owned enterprises, Journal of Law and Economics, 1989, 32, 1, 1-33. Buccirossi, P., Ciari, L., Duso, T., Spagnolo, G. and Vitale, C., Competition policy and productivity growth: an empirical assessment, Review of Economics and Statistics, 2013, 95, 4, 1324-1336. Bundeskartellamt (ed.), Das 3. Jahr Markttransparenzstelle für Kraftstoffe (MTS-K), Bonn 2017. Bundeskartellamt (ed.), Leitfaden zur Marktbeherrschung in der Fusionskontrolle, Bonn 2012. Bundeskartellamt (ed.), Algorithmen und Wettbewerb, Bonn 2020. Bundesnetzagentur (ed.), Tätigkeitsbericht Telekommunikation 2018/ 2019, Bonn 2019. <?page no="289"?> 7.5 Summary and key learnings 289 Cabral, L.M. and Riordan, M.H., Incentives for cost reduction under price cap regulation, Journal of Regulatory Economics, 1989, 1, 2, 93-102. Caves, R.E., In praise of the Old IO, International Journal of Industrial Organization, 2007, 25, 1, 1-12. Dewenter, R. and Haucap, J., Grundlagen und Auswirkungen der Liberalisierung in der deutschen Telekommunikationsbranche, in: Ragnitz, J. and Eitner, P. (eds.), Deregulierung in Deutschland - Theoretische und empirische Analysen, Halle 2004, 45-81. Duso, T., Eine bessere Wettbewerbspolitik steigert das Produktivitätswachstum merklich, DIW Wochenbericht 29, 2014, 687-697. European Commission, Green Paper on Entrepreneurship in Europe, Brussels 2003, 27. Evans, D.S., 2003, The antitrust economics of multisided platform markets, Yale Journal on Regulation, 2003, 20, 2, 325-381. Fazzari, S.M., Hubbard, R.G. and Peterson, L., Financing constraints and corporate investment, Brookings Papers on Economic Activity, 1988, 1, 141-195. Fritsch, M., Marktversagen und Wirtschaftspolitik - mikroökonomische Grundlagen staatlichen Handelns, München 2018. Geroski, P.A., Innovation as an engine of competition, in: Mueller, D.C., Haid, A. and Weigand, J. (eds.), Competition, Efficiency and Welfare, Dordrecht 1998, 13-26. Giersberg, G., Pleiten in Deutschland: Die Insolvenzen des Jahres, Frankfurter Allgemeine Zeitung, 30. Dezember 2016, 13. Gilbert, R. and Newbery, D., Preemptive patenting and the persistence of monopoly, American Economic Review, 1982, 72, 3, 514-526. Gilbert, R. and Shapiro, C., Optimal patent length and breadth, RAND Journal of Economics, 1990, 21, 1, 106-112. Grossarth, J., Dramatischer Preisverfall - Deutschland stürzt in die Milchkrise, Frankfurter Allgemeine Zeitung, 17. Mai 2016, 11. Harrison, R., Mason, C. and Girling, P., Financial bootstrapping and venture development in the software industry, Enterprise and Regional Development, 2004, 16, 2, 307-333. Haucap, J. and Heimeshoff, U., Ordnungspolitik in der digitalen Welt, DICE - Ordnungspolitische Perspektiven, Düsseldorf 2017. Haucap, J. and Schmidt, I., Wettbewerbspolitik und Kartellrecht - eine interdisziplinäre Einführung, München 2013. Haucap, J., Pavel, F., Aigner, R., Arnold, M., Hottenrott, M. and Kehder, C., Chancen der Digitalisierung auf Märkten für urbane Mobilität: Das Beispiel Uber, DICE - Ordnungspolitische Perspektiven, Düsseldorf 2015. Holt, C.A., Industrial organization: a survey of laboratory research, in: Kagel, J. and Roth, A. (eds.), Handbook of Experimental Economics, 1993, 1-103. Hüschelrath, K., Competition policy analysis: an integrated approach, Heidelberg 2008. Jackson, M.O., Economic engineering and the design of matching markets: the contributions of Alvin E. Roth, Scandinavian Journal of Economics, 2013, 115, 3, 619-639. Jovanovic, B., Selection and the evolution of an industry, Econometrica, 1982, 50, 659-670. Kerr, W. and Nanda, R., Financing constraints and entrepreneurship, National Bureau of Economic Research, 2009, No. w15498. Klenow, P.J. and Malin, B., Microeconomic evidence on price-setting, in: Friedman, B. and Woodford, M. (eds.), Handbook of Monetary Economics, Amsterdam 2011, 231-284. Knieps, G., Wettbewerbsökonomie - Regulierungstheorie, Industrieökonomie, Wettbewerbspolitik, Heidelberg 2008. <?page no="290"?> 7 Perfect competition, monopoly and competition policy 290 Laffont, J-J. and Tirole, J., The politics of government decision-making: a theory of regulatory capture, Quarterly Journal of Economics, 1991, 106, 1089-1127. Laffont, J-J. and Tirole, J., Using cost observation to regulate firms, Journal of Political Economy, 1986, 94, 614-614. Leibenstein, H., Allocative efficiency vs. X-efficiency, American Economic Review, 1966, 56, 3, 392-415. Lerner, A.P., The concept of monopoly and the measurement of monopoly power, Review of Economic Studies, 1934, 1, 3, 157-175. Levin, R.C., A new look at the patent system, American Economic Review, 1986, 76, 2, 199-202. Levin, R.C., Klevorick, A.K., Nelson, R.R., Winter, S.G., Gilbert, R. and Griliches, Z., Appropriating the returns from industrial research and development, Brookings Papers on Economic Activity, 1987, 3, 783-831. Monopolkommission (ed.), Algorithmen und Kollusion, Hauptgutachten Monopolkommission 2018, Kapitel 1, 62-88, Bonn 2018. Monopolkommission (ed.), Wettbewerbspolitik: Herausforderung digitale Märkte, Sondergutachten 68, Bonn 2014. Motta, M., Competition policy - theory and practice, Cambridge 2004. Münter, M.T., Unternehmensgründung und Finanzierung, in: Betzold, R., Emrich, E., Gassmann, F., Heidenreich, S., Jordanov, S., Koch, M., Münter, M.T. and Nguyen, Q. (eds.), Unternehmensgründungen im Umfeld saarländischer Hochschulen - Empirische Ergebnisse und regionalökonomische Effekte, 2021, 179-220. Myers, S.C. and Majluf, N.S., Corporate financing and investment decisions when firms have information that investors do not have, Journal of Financial Economics, 1984, 13, 2, 187-221. Neumann, M., Efficiency defense in EU and US horizontal merger control if costs are endogenous, German Economic Review, 2016, 17, 1, 48-60. Neumann, M., Wettbewerbspolitik - Geschichte, Theorie und Praxis, Berlin 2000. Neven, D.J. and Röller, L.H., Consumer surplus vs. welfare standard in a political economy model of merger control, International Journal of Industrial Organization, 2005, 23, 829-848. o.V., Die neue Benzinpreistricks der Tankstellen, Frankfurter Allgemeine Zeitung, 28. September 2017. o.V., Digitale Märkte: Monopolkommission will Google nicht regulieren, Frankfurter Allgemeine Zeitung, 1. Juni 2015. o.V., Lufthansa stärkt Marktmacht durch Kauf großer Teile von Air Berlin, Frankfurter Allgemeine Zeitung, 13. Oktober 2017. o.V., Nach Übernahme von Air Berlin: Chef der Monopolkommission rechnet mit Auflagen für Lufthansa, Frankfurter Allgemeine Zeitung, 13. Oktober 2017. Organisation for Economic Cooperation and Development, Fostering entrepreneurship, OECD, Paris 1998. Petrakis, E., Rasmusen, E. and Roy, S., The learning curve in a competitive industry, Rand Journal of Economics, 1997, 28, 248-268. Posner, R.A., The social costs of monopoly and regulation, Journal of Political Economy, 1975, 83, 4, 807-827. Reid, G., Trajectories of small business financial structure, Small Business Economics, 2003, 20, 4, 273-285. Rhoades, S.A., Market share as a source of market power: implications and some evidence, Journal of Economics and Business, 1985, 37, 4, 343-363. Roth, A.E., Sönmez, T. and Ünver, M.U., Efficient kidney exchange: coincidence of wants in markets with compatibilitybased preferences, American Economic Review, 2007, 97, 3, 828-851. Roth, A.E., What have we learned from market design? , Economic Journal, 2008, 118, 527, 285-310. Roth, A.E., Who gets what—and why: the new economics of matchmaking and market design, New York 2015. <?page no="291"?> 7.5 Summary and key learnings 291 Schmalensee, R., Inter-industry studies of structure and performance, in: Schmalensee, R. and Willig, R.D. (eds.), Handbook of Industrial Organization, 1989, 2, 951-1009. Schumpeter, J.A., Capitalism, socialism and democracy, London 1942. Schumpeter, J.A., Theorie der wirtschaftlichen Entwicklung - eine Untersuchung über Unternehmergewinn, Kapital, Kredit, Zins und den Konjunkturzyklus, Berlin 1911. Schwartz, M., Dapp, T.F., Beck, G.W. and Khussainova, A., Deutschlands Banken schalten bei Filialschließungen einen Gang höher, KfW Research, 181, 2017. Shepherd, W.G., Public policies toward business, Homewood 1991. Sickmann, J., Deregulierung und Verbraucherwohlfahrt auf dem deutschen Telekommunikationsmarkt, Bonn 2018. Sidak, J.G. and Teece, D.J., Dynamic competition in antitrust law, Journal of Competition Law and Economics, 2009, 5, 4, 581-631. Smith, V.L., An experimental study of competitive market behavior, Journal of Political Economy, 1962, 70, 2, 111-137. Smith, V.L., Microeconomic systems as an experimental science, American Economic Review, 1982, 72, 5, 923-955. Uusitalo, R., Homo entreprenaurus? , Applied Economics, 2001, 33, 13, 1631-1638. von Hayek, F.A., Der Wettbewerb als Entdeckungsverfahren, Freiburger Studien, Tübingen 1969, 249-265. Wambach, A., Wettbewerbsregeln für das digitale Zeitalter, Wirtschaftsdienst, 2016, 96, 8, 589-593. Wernerfelt, B., Brand loyalty and market equilibrium, Marketing Science, 1991, 10, 3, 229-245. Williamson, O.E., Economies as an antitrust defense: the welfare tradeoffs, American Economic Review, 1968, 58, 1, 18-36. <?page no="293"?> 293 8 Pricing strategies and price discrimination More or less every passenger on the flight of Luxair 4591 on 5 March 2017 at 14: 50 from Luxembourg to London City - on the same day, at the same time, same route, same booking class - paid a different ticket price than the person sitting next to them, with absolutely identical transport services. The obvious reasons for this are differences in date of booking, payment method, timespan to return flight, membership of a frequent flyer programme, or sales channel. Numerous studies have examined the price differences and pricing patterns of airlines. In most cases there is a U-shaped price development over time as shown in ► Figure 8.1. After an initial phase of stable prices over a longer period of time, prices decline slowly but continuously from about 200 days before the flight, then fluctuate irregularly over a period of two to three months, and then rise again about 30 to 50 days before departure, first slowly, then very rapidly. This reflects a decreasing price elasticity of demand and a growing urgency of customers to fly (McAfee and de Velde 2007). However, part of the success of so-called low-cost carriers such as Southwest or RyanAir is based on the fact that these firms sometimes use opposing pricing patterns. Figure 8.1: Airline ticket prices over time (schematic representation based on cheapair.com/ when-to-buy-flights). But even if all the above parameters are still identical, it is possible that two passengers in the same booking class have paid two different prices: the reason lies in differences of willingness to pay of the two customers, which is exploited by the airlines through price discrimination (Busse 2015). Price discrimination summarises all pricing strategies in which customers or customer groups are charged different prices according to their willingness to pay for essentially identical products with the same production costs without horizontal or vertical product differentiation. Firms are in a position to skim off some or even capture all of the consumer surplus and, thus, achieve significant profit increases. <?page no="294"?> 8 Pricing strategies and price discrimination 294 Products in the non-digital world without significant product differentiation with almost identical production costs such as hardor soft-cover books, hotel rooms, rental cars, software as individual products or packages, petrol as Super95 or UltimatePower95 (which is common in Germany, but unknown in the UK), and many others have already been offered in a price-discriminating manner for a long time. However, big data significantly expands the possibilities for price discrimination: through analysing browsing behaviour, the mobile device used, search intensity, and the product alternatives viewed, a firm can very precisely determine a customer's willingness to pay and vary the price on this basis on a customer-specific and algorithm-based over time. This is known as personalised dynamic pricing (Mohammed 2017, Papanastasious and Savva 2017 as well as Levin et al. 2008). From a management perspective, two core issues are associated with price discrimination: on the one hand, individual or customer group-specific willingness to pay must be determined for market segments, on the other hand, possibilities of different concepts of price discrimination as pricing strategies must be worked out and linked to any existing business model. However, it must be kept in mind that price discrimination - especially as personalised dynamic pricing - can also be considered unfair by customers and consumer protection agencies (Bolton et al. 2003 as well as Krämer and Kalka 2016). Learning Objectives This chapter deals with: the possibilities of pricing and price discrimination of firms with market power and the conversion of parts of the consumer surplus into profits; the concepts of nonlinear pricing strategies, two-part tariffs and dynamic pricing; the application of price discrimination to airline tickets, football tickets, package sizes, books, membership fees as well as the possibilities and fields of application of pure bundling and mixed bundling; and auctions as multi-stage, mostly dynamic bidding procedures in B2C and B2B markets. 8.1 Forms and requirements of price discrimination The objective of all pricing strategies is to increase profits. ► Chapter 7 explains that this approach does not work if competition is intense: firms under perfect competition cannot influence the market price. For firms with market power, however, this is possible - they can enforce a price above marginal cost and, thus, convert part of the consumer surplus into profits due to different customers' willingness to pay. ► Figure 8.2 shows the profits of two firms with identical demand and cost situations. The firm on the left has no significant market power and determines the quantity 𝑞𝑞 in such a way that the marginal costs 𝑀𝑀𝐶𝐶 correspond to the market price 𝑝𝑝 : obviously, a profit of 𝜋𝜋 𝐶𝐶 arises in the short run. In contrast, the firm on the right has significant market power and, under the condition that marginal cost 𝑀𝑀𝐶𝐶 equals marginal revenue 𝑀𝑀𝑅𝑅 , can transform a lower quantity at a higher price into a higher profit 𝜋𝜋 𝑀𝑀 - but in particular, in this case, the consumer surplus 𝐶𝐶𝑆𝑆 is significantly lower. <?page no="295"?> 8.1 Forms and requirements of price discrimination 295 Figure 8.2: Capturing consumer surplus. Price discrimination goes even further: a firm sets not only one price for a product, but several staggered prices, so that profits can be increased well above the monopoly level - i.e., the consumer surplus is reduced even further. Requirements for price discrimination (often referred to in marketing as price differentiation or price fencing) are that: firms have at least a certain degree of market power; customers differ in their willingness to pay; customers can be targeted individually or in market segments or show their willingness to pay through self-selection of a price model; and arbitrage (i.e., exploiting price differences by customers trading amongst each other in resales) is impossible or at least can be made costly. For airlines, these conditions are fulfilled: after the acquisition of AirBerlin, Lufthansa has gained a dominant position with market power on individual routes at certain times; customers (business and holiday travellers) clearly have different willingness to pay and also show this in the booking process; and finally arbitrage is ruled out due to the personalisation of airline tickets (Die Zeit 2017). Price discrimination enables three basic pricing strategies, which also could be combined in pricing and business models. Nonlinear pricing through direct price discrimination - different customers or customer groups pay different prices. Typical examples are all forms of price-discriminating market segmentation: student or senior citizen tickets in public transport; bulk discounts in supermarkets such as 'buy four get one free'; auctions such as those on ebay; or dynamic pricing <?page no="296"?> 8 Pricing strategies and price discrimination 296 at Amazon. This option is especially given if differences in demand or willingness to pay of the customers can be determined by market analysis before the purchase. Two-part tariffs and indirect price discrimination - products are offered with a basic or membership fee and additionally a usage fee based on units consumed. Typical examples are: membership cards such as the Bahncard 50 of Deutsche Bahn to reduce regular fares; annual fees in a golf club with an additional green fee; or monthly basic fees for telephony, electricity, or water supply with additional quantity-dependent prices. This option is chosen if one suspects that customers have different willingness to pay and demand, but this only becomes apparent through self-selection during the purchase and use of the product. Bundling - different products, which are also available separately, are offered in packages or in bundles. Typical examples of bundling are: software packages from Microsoft; menus at Burger King or McDonald’s; and programme packages from pay-TV providers such as Sky. This option is used if the willingness to pay for different products is negatively correlated across customers. 8.2 Direct price discrimination and market segmentation Nonlinear pricing can basically take two forms. Either a firm is in a position to: set a price for each individual customer exactly according to his willingness to pay (perfect price discrimination); or set prices for different customer groups according to their group-specific willingness to pay (market segmentation). Perfect price discrimination Perfect price discrimination is always relevant if customers actively express their willingness to pay, or it can be precisely determined via market research. In fact, this is more often the case than it first might appear. For example, customers actively show their respective willingness to pay in the context of B2C and B2B auctions, but buyers and sellers of new cars also go through an auction: first the seller personalises the vehicle according to the extras desired by the customer; then a reverse auction starts based on the list price plus the extras in the course of which the seller successively lowers the price in order to find out the customer's maximum willingness to pay and finally skim it off. In addition, firms sometimes leave it up to customers to set the price individually according to their willingness to pay, often with considerable success (Uken 2013). Examples include restaurants like the Kish in Frankfurt and the Lentil as Anything in Melbourne, the zoo in Münster in Westphalia, or rock bands like Radiohead for the album In Rainbows. However, the "Name Your Own Price" pricing strategy can only work if the firm has at least some market power, and price comparisons are impossible due to customers’ unability to determine reference prices (Kim et al. 2009). In digital business models, market research has gained crucial importance via the use of big data. Not only do customers directly indicate their preferences through likes or ratings in social media, or on marketplaces such as ebay or Amazon, but a customer's maximum willingness to pay can also be determined precisely through observable purchasing behaviour, the devices used and the prices paid for other products (Baker et al. 2014, Tanner 2014 and Rayna et al. <?page no="297"?> 8.2 Direct price discrimination and market segmentation 297 2015). Firms then attempt to skim off higher paying customers through implicit auctions and dynamic pricing, e.g. customers are repeatedly shown new prices online and individually. Figure 8.3: Perfect price discrimination. ► Figure 8.3 shows the effect of perfect price discrimination compared to pricing under perfect competition and with a usual monopoly (only setting one monopoly price). In perfect competition, a quantity 𝑞𝑞 𝑝𝑝 is offered at a price 𝑝𝑝 𝑝𝑝 , so that the areas 𝐴𝐴 + 𝐵𝐵 + 𝐶𝐶 depict consumer surplus. A usual monopolist without the possibility of price discrimination sets a price 𝑝𝑝 𝑀𝑀 and realises a profit of 𝐴𝐴 - area 𝐵𝐵 remains as a consumer surplus, area 𝐶𝐶 represents the welfare loss (► Chapter 7). Perfect price discrimination now means that each customer is charged a price according to his willingness to pay ( 𝑝𝑝 𝑀𝑀1 , 𝑝𝑝 𝑀𝑀2 , . . . 𝑝𝑝 𝑀𝑀𝑝𝑝 ). In this case, a firm is able to convert the entire consumer surplus into profits, so that the profit increases from 𝐴𝐴 to 𝐴𝐴 + 𝐵𝐵 + 𝐶𝐶 . A price-discriminating dominant firm here also sets prices below the previous monopoly price 𝑝𝑝 𝑀𝑀 down to a lower limit of the price of perfect competition 𝑝𝑝 𝑝𝑝 , so that the quantity can be increased above 𝑞𝑞 𝑀𝑀 up to the level of 𝑞𝑞 𝐶𝐶 of perfect competition and, in addition, the welfare loss is eliminated. Good to know │ Dynamic and personal pricing - why does every customer pay a different price? To implement direct price discrimination, firms need precise information on customers’ individual willingness to pay - ideally at any point in time. If you observe, for example, a fishmonger on Winterhude’s market square in Hamburg, you can immediately recognise this personalised and dynamic pricing: some customers are favoured by lower <?page no="298"?> 8 Pricing strategies and price discrimination 298 prices, other customers are disadvantaged by higher prices. The reasons for this are twofold: on the customer side there are different willingness to pay, there are differences between loyal regular customers and walk-in customers, but also the competitive situation and the remaining stocks shortly before the market closes at 1 pm on Saturdays. A fishmonger can, therefore, combine dynamic pricing (time-specific pricing based on collective demand patterns) and personalised pricing (customer-specific pricing based on individual willingness to pay). Personalised pricing requires in particular customer-specific data, while dynamic pricing analyses competitor data and demand patterns over time. With increasing digitalisation and leveraging big data, analysing pricing behaviour based on individual willingness to pay is becoming easier. Firms are able to draw increasingly precise conclusions about individuals’ preferences and willingness to pay: from browser information, devices used, search behaviour on the internet and likes and dislikes from social media accounts. What is new - especially in digital channels - is a greater diffusion of dynamic pricing and personalised pricing based on algorithms. Algorithmic pricing is based on data analysis that enables automated rule-based, dynamic and customer-specific pricing in real time (Bar-Gill 2019, Narahari et al. 2015 and OECD 2018). Empirical studies show that dynamic pricing is already implemented in many online shops - for example, 80% of all shops on Amazon Marketplaces in the UK already use pricing algorithms (Chen and Chen 2015 and Chen et al. 2016). Personal pricing, on the other hand, is rarely observed - and is often disguised or replaced by product ranking or personalisation of the products (Mikians et al. 2012 and Hannak et al. 2014). Figure 8.4: Frequency and volatility of pricing in German online shops. Maximum number of observed adjustments of a product price (frequency) and maximum percentage price difference from the mean product price (volatility) at the 16 online retailers studied in spring 2018. Price difference given as a percentage (source: Dautzenberg et al. 2018a, p. 20). Corresponding observations were also made by Verbraucherzentralen (Germany's consumer protection agency) in two studies (Dautzenberg et al. 2018a and 2018b). As can be seen in ► Figure 8.4, prices in numerous online shops are changed several times a week, sometimes daily, and with high volatility: the price dynamics amount to up to 105% in <?page no="299"?> 8.2 Direct price discrimination and market segmentation 299 relation to some average price. Personalised pricing has been observed - different devices used by customers show different prices, for example - but neither to the same extent as with dynamic pricing, nor is a systematic correlation identifiable. The reason for this is, on the one hand, that instead of personalised prices personalised product rankings with stronger profit leverage are used and, on the other hand, the customers' perception of fairness is disturbed. Numerous empirical studies show that customers perceive dynamic and personalised prices as unfair (Garbarino and Lee 2003, Richards et al. 2017 and Seele 2020 et al.). Fairness regarding price is always given if customers perceive a price-performance ratio as fair, and the actual price does not violate some individually or collectively accepted price. The perception of fairness is the subjective result of a comparison of the exchange relationship between the transaction partners (so-called equity theory). Here, one's own cost/ benefit ratio is compared with that of the transaction partner, or a comparison is made with other customers who are also connected with the transaction partner. Independently of this, transaction partners often concede an appropriate profit or benefit to each other (so-called entitlement theory). Price increases due to increased costs are, therefore, usually accepted as fair, even if they are at the expense of the customer, while price increases that only serve to increase profits and which are not justified by cost increases are seen as unfair. In addition, the perception of fairness depends on reference prices, which are based on current market prices, learned price anchors and past experience. In addition to these cognitive components, there is an emotional component of price fairness - customers react with annoyance and anger if they are disadvantaged and with feelings of guilt if they are favoured. Both advantages and disadvantages from price differentiation result in perceived unfairness (Xia et al. 2004, Reinartz et al. 2017. Jentzsch 2017 and Kahneman et al. 1986). Perceived unfairness increases with the volatility of pricing and the frequency of price changes in the case of dynamic pricing, yet is getting lost over time and is considered a necessary evil in case of cost fluctuations of a firm. With personalised pricing, the perceived unfairness is generally higher because the social norm of inequality aversion is violated here. As a result, customers lose trust in a firm and the willingness to switch increases (Reinartz et al. 2017). In fact, the customers' perception of unfairness can be reduced: Uber informs customers before booking a ride about situational increased prices due to high demand or few available drivers - as a result, higher and dynamically personalised prices are better accepted (Chen 2017 and Dakers 2016). In addition, memberships in loyalty programmes, limited access to exclusive deals, or shopping in different distribution channels lead to reduced perceptions of unfairness. <?page no="300"?> 8 Pricing strategies and price discrimination 300 An assessment of personalised and dynamic pricing from a competition policy perspective is not clear-cut. Consumer surplus can increase, but it can also fall. Moreover, individualised pricing can also enable customers with low income and a low willingness to pay to buy products that would otherwise be too expensive. In addition, producer surplus usually rises, so that economic welfare increases. Essentially, the assessment depends on whether and how well-informed customers are: the more frequent strategic customers are - who inform themselves in detail about prices - the greater the benefits for the customer side. The more myopic customers there are - who tend to buy spontaneously and without seeking information - the more likely it is that consumer surplus will decrease (Heidhues and Köszegi 2010 and Kremer et al. 2017). However, algorithm-based pricing comes at a risk: it may be (intentionally or unintentionally) that algorithms of different firms adaptively behave like a price cartel in their pricing and, thus, enable higher revenues or profits for the firms (Ezrachi and Stucke 2015, Gal 2017 and Schwalbe 2018). Algorithm-based pricing can also lead to curious results: on 18 April 2011, two pricing algorithms repeatedly outbid each other on Amazon Marketplaces until the textbook Making of a Fly by Peter Lawrence was offered, but not sold, at a price of 23,698,655.93 USD (Solon 2011). Market segmentation and marginal revenues Sometimes it appears too costly for firms to precisely determine the willingness to pay of each individual customer. However, willingness to pay instead could be identified very well for specific customer groups or several market segments. If customer or market segments are easily isolated, can be addressed separately and arbitrage can be prevented, then firms apply group pricing. This can be spatial (different prices for new cars in Germany and the US), demographic (students, seniors, etc.), intertemporal (seasonal or day vs. night tariffs) or any combination. Group pricing means that different prices are set for groups of customers whose different willingness to pay can be easily identified and verified. Typically, this happens in museums, cinemas and theatres or in public transport. Reduced ticket prices for students or senior citizens are, thus, not a friendly concession by the firm: rather, the firm sets monopoly prices for each market segment according to their willingness to pay and is able to increase profits. The basic idea can be seen in ► Figure 8.5 for airline tickets, which are differentiated into booking classes such as economy and business. Obviously, the demand curve 𝐷𝐷 2 of business travellers is flatter and the absolute willingness to pay is higher than for the other travellers. <?page no="301"?> 8.2 Direct price discrimination and market segmentation 301 Figure 8.5: Group pricing and market segmentation. For domestic flights or short-haul routes in Europe, marginal costs per passenger are identical, so the firm separates the booking classes just by 'moving the curtain in the cabin'. Product differentiation is achieved solely through the possibility of greater flexibility in changing a reservation. An airline will then apply the rule marginal revenue 𝑀𝑀𝑅𝑅 equals marginal cost 𝑀𝑀𝐶𝐶 in both market segments to determine an optimum number or allocation of tickets per booking class. In ► Figure 8.5 this will result in optimal seat quantities of 𝑞𝑞 1 and 𝑞𝑞 2 at prices of 𝑝𝑝 1 and 𝑝𝑝 2 and the profits indicated by the grey areas will be significantly larger in the business passenger segment. However, the following insight is important: if a firm applies such a group pricing strategy, then, when maximising profits, marginal revenues are the same in all market segments (here 𝑀𝑀𝑅𝑅 1 and 𝑀𝑀𝑅𝑅 2 ) - that is, the same revenue is achieved with the last ticket sold in each booking class (see also analogous results for marginal products in ► Chapter 5 and marginal costs in ► Chapter 6). In ► Chapter 7, it is shown that a firm with market power can maximise profits if prices and quantities are chosen in such a way that marginal costs correspond to marginal revenues. This result also applies to products in several market segments - exemplified for two products whose joint profit is determined by (8.1) 𝜋𝜋(𝑞𝑞 1 , 𝑞𝑞 2 ) = 𝑝𝑝 1 𝑞𝑞 1 + 𝑝𝑝 2 𝑞𝑞 2 − 𝑇𝑇𝐶𝐶(𝑞𝑞 1 + 𝑞𝑞 2 ) . A firm can increase profits by choosing quantities 𝑞𝑞 1 and 𝑞𝑞 2 under the conditions of (8.2) 𝜕𝜕𝜋𝜋 𝜕𝜕𝑞𝑞1 = 𝑝𝑝 1 + 𝜕𝜕𝑝𝑝1 𝜕𝜕𝑞𝑞1 𝑞𝑞 1 − 𝜕𝜕𝑇𝑇𝐶𝐶 𝜕𝜕𝑞𝑞1 = 𝑀𝑀𝑅𝑅 1 − 𝑀𝑀𝐶𝐶 1 = 0 and (8.3) 𝜕𝜕𝜋𝜋 𝜕𝜕𝑞𝑞2 = 𝑝𝑝 2 + 𝜕𝜕𝑝𝑝2 𝜕𝜕𝑞𝑞2 𝑞𝑞 2 − 𝜕𝜕𝑇𝑇𝐶𝐶 𝜕𝜕𝑞𝑞2 = 𝑀𝑀𝑅𝑅 2 − 𝑀𝑀𝐶𝐶 2 = 0 so that, as in ► Figure 8.15, marginal revenue equals marginal cost. If marginal costs are identical, then from (8.2) and (8.3), using the price elasticity of demand 𝜀𝜀 = 𝑝𝑝𝑞𝑞 𝜕𝜕𝑞𝑞 𝜕𝜕𝑝𝑝 , an optimum price ratio for both market segments can be determined as <?page no="302"?> 8 Pricing strategies and price discrimination 302 (8.4) 𝑝𝑝1 𝑝𝑝2 = 1+ 1 𝜀𝜀2 1+ 1 𝜀𝜀1 . So, a simple rule of thumb for managers is to choose a price ratio for the two market segments inversely to the price elasticity of demand: the higher the price elasticity, the lower the relative price and vice versa. This result applies equally to any number of products or market segments. Moreover, from a strategic perspective, it is important to recognise that costs - as can be seen from equation (8.4) - do not play a role in this decision on the pricing structure if they are identical across market segments. Case Study │ Group pricing of airline tickets A product manager of an airline is to set the ticket prices for the route Frankfurt-London in economy and business class. The marginal costs (landing slot, security processing, luggage handling, etc.) are EUR 250. The price elasticity for economy customers is 𝜀𝜀 𝐸𝐸 = −4.0 (very elastic), for business customers 𝜀𝜀 𝐵𝐵 =-1.6 (weakly elastic). To maximise profits, the product manager can refer to equation (8.4). The prices for both customer groups must take into account the respective price elasticities, so that with (8.5) 𝑝𝑝𝐸𝐸 𝑝𝑝𝐵𝐵 = 1+ 1 𝜀𝜀𝐵𝐵 1+ 1 𝜀𝜀𝐸𝐸 = 0.375 0.75 = 12 ⇒ 𝑝𝑝 𝐸𝐸 = 0.5 ⋅ 𝑝𝑝 𝐵𝐵 the price of an economy class ticket being exactly half the business class price. The absolute price per ticket can then be calculated via equation (7.49) with the marginal costs as (8.6) 𝑝𝑝 𝐸𝐸 = 𝑀𝑀𝐶𝐶 1+ 1 𝜀𝜀𝐸𝐸 = 250 0.75 = 333.33 so that the price for a business class ticket is 𝑝𝑝 𝐵𝐵 = 666.67 . In a similar way to airlines also museums, public transport, or firms offering identical products under several brands determine the relative pricing per market segment. These strategies are often based on the fact that marginal costs per customer are only slightly different from or actually zero: this is especially true for sports events or music festivals. Football clubs like FC Bayern München have, therefore, already optimised revenue streams from ticket sales. First of all, more than 150 different ticket variants are sold per match in a personalised manner (online ticketing) in order to prevent arbitrage. Ticket prices are determined by the relative willingness to pay and the price elasticity of demand. In addition, the club has organised an auction platform (secondary ticket market) where tickets can be resold under certain conditions - FC Bayern München benefits once more (see tickets.fcbayern.com). 8.3 Indirect price discrimination and two-part tariffs Often it is impossible for firms to identify customers or customer groups according to their willingness to pay, or arbitrage cannot be effectively prevented. The concepts and strategies described in ► Section 8.2, therefore, only work to a very limited extent. However, in these cases indirect price discrimination can be applied. Here, a firm offers various pricing models to all customers, but leaves the price selection or even the pricing to the customers and can, thus, save parts of the costs of market research: if different price models for identical products <?page no="303"?> 8.3 Indirect price discrimination and two-part tariffs 303 are strategically arranged correctly, customers are addressed according to their willingness to pay and assign themselves to the price model (Bertini and Koenigsberg 2014 as well as Narayanan et al. 2007). This self-selection often leads to significant profit increases for firms. Typical forms of indirect price discrimination are: package pricing and block pricing - quantity-based price discrimination in which customers are offered price reductions (as quantity discounts or rebates) depending on package size, purchase frequency, or demand behaviour; two-part tariffs and flat rates - customers purchase a basic product or membership via a monthly basic fee, and an additional fee is charged depending on usage or consumption; or freemium models and damaged products - a qualitative price discrimination in which different variants of a product are positioned with functional restrictions. Package pricing Typical examples of package pricing are differing package sizes in food retail, coupons for price reductions, buy-one-get-one-free models, or mobile phone tariffs with different data volumes. In most cases, prices per 100 grams or per gigabyte of data become lower the larger the total quantity. This is illustrated in ► Figure 8.6: Lavazza offers identical coffee in different variants under the brand name Lavazza Qualita Oro - the price per kilogram can, thus, be tripled (prices researched in November 2017 at the Woolworth Supermarket on Fitzroy Street in Melbourne). Figure 8.6: Package pricing. <?page no="304"?> 8 Pricing strategies and price discrimination 304 Figure 8.7: Profits from different package sizes. <?page no="305"?> 8.3 Indirect price discrimination and two-part tariffs 305 Package pricing means that customers pay different prices for identical products depending on the quantity selected. A first case is shown in ► Figure 8.7 using package sizes. A firm reckons that there are three customer groups with different willingness to pay. Instead of a single price and a corresponding quantity in all market segments - marked by the grey dot of the monopoly price - the firm addresses these supposed customer groups with three different prices 𝑝𝑝 1 , 𝑝𝑝 2 and 𝑝𝑝 3 for three different package sizes. If these prices are accurately arranged, then profits are equal to the respective grey areas which arise for the individual customer groups. These can in total be larger than monopoly profits based on a single price - this essentially depends on the size of the individual customer segments, their respective willingness to pay and demand behaviour (Cohen 2008). Package pricing works very well in supermarkets when, due to transport costs, flat size or limited shelf life, not all customers buy the bulk packs that are supposedly cheaper per unit or kilogram but choose smaller variants. A special form of package pricing are block prices - prices that apply to all customers in the same way for certain quantity levels. These are often applied by energy suppliers, but the concept can also be found with cumulative quantity discounts in other business models. ► Figure 8.7 shows the cumulative demand for electricity across different customer groups as a function of household size - typically this increases with the number of household members, but the willingness to pay per kilowatt hour of electricity successively decreases. An energy supplier, therefore, often offers block prices in which the prices descend block by block from 𝑝𝑝 1 via 𝑝𝑝 2 to 𝑝𝑝 3 . The firm can, therefore, initially enforce high prices 𝑝𝑝 1 above a uniform monopoly price 𝑝𝑝 𝑀𝑀 for small quantities up to 𝑞𝑞 1 from all customers - single households, married couples, and families - and achieve a profit of 𝐴𝐴 across all customer groups. At this price, however, neither married couples nor families would expand their demand. For this reason, the firm offers the next quantity interval up to 𝑞𝑞 2 at a reduced price of 𝑝𝑝 2 - profits now increase by area 𝐵𝐵 due to customer groups married couples and families. In order to serve the residual demand of quantity 𝑞𝑞 3 of families, the firm offers a price 𝑝𝑝 3 for these quantities. Total profits then increase to 𝐴𝐴 + 𝐵𝐵 + 𝐶𝐶 and clearly exceed profits achievable through a uniform monopoly price as can be seen in ► Figure 8.8. <?page no="306"?> 8 Pricing strategies and price discrimination 306 Figure 8.8: Package pricing and volume discount. Therefore, profits of a package price discriminating monopolist are higher than those of a usual monopolist. An absolute profit level is determined by the respective willingness to pay, different customer group needs and the actual number of addressable customer groups. Two-part tariffs and flat rates Many firms offer products and services based on two-part tariffs: one part of the tariff is a monthly or annual membership fee and consists of a fixed amount, the other part of this tariff is a usage-based fee, or for certain additional services, and is variable. Such pricing models can be found for many leisure activities (golf, tennis, gym, etc.) and banking products (current accounts, credit cards, etc.) but also for mobile phone contracts with a monthly basic fee and an additional usage fee per minute or for data volumes if inclusive volumes are exceeded. A special form of two-part tariffs exists in system markets with indirect network effects: the purchase of a game console or platform is like a membership fee, the purchase of additional games or levels in the game is a usage fee (see also ► Chapter 2). Whether and to what extent membership fees are possible depends essentially on the consumer surplus and the price elasticity of demand. ► Figure 8.9 shows the basic idea behind a membership fee: like all forms of price discrimination, it is based on capturing part or all of customers consumer surplus. A monopolist (under the profit-maximising condition marginal revenue 𝑀𝑀𝑅𝑅 equals marginal cost 𝑀𝑀𝐶𝐶 ) obtains a usage-dependent revenue equal to area 𝐴𝐴 at <?page no="307"?> 8.3 Indirect price discrimination and two-part tariffs 307 a price of 𝑝𝑝 . A two-part tariff is now created by the firm determining the consumer surplus, translating it into a membership fee 𝐵𝐵 and charging it. This two-part tariff can be transformed into a flat rate 𝐶𝐶 if a firm combines the usage-dependent charge 𝐴𝐴 and the membership fee B. Figure 8.9: Monopoly pricing, two-part tariff and flat rate. ► Figure 8.10 shows that both the existence and extent of membership fees depend on the shape of a demand curve: the lower the price elasticity of demand, the lower the possible membership fee. However, even in this case - as is the case with gyms - a firm can sum up the usage-dependent fee directly to the membership fee and in this way only charge a monthly or annual fee as a flat rate. The reason is that an overuse of the gym is impossible. Similarly, flat rate pricing in all-you-can-eat restaurants can be interpreted as a one-day membership fee without a usage fee. Both business models are based on fixed costs that are fully covered by membership fees (Just and Wansink 2011 and DellaVigna and Malmendier 2006). Figure 8.10: Consumer surplus and membership fees. <?page no="308"?> 8 Pricing strategies and price discrimination 308 Whether charging a membership fee and an additional usage fee is effective depends not only on the shape of demand but also on the competitive situation and the level of marginal costs, as can be seen in ► Figure 8.10: depending on the expected usage behaviour and the level of marginal costs, firms can offer flat rates with or without a usage cap. Here, the lower the marginal costs 𝑀𝑀𝐶𝐶 are and the better the usage behaviour 𝑞𝑞 can be predicted, the more likely a flat rate without a cap will be offered and vice versa. Figure 8.11: Membership fees and flat rate. If either marginal costs are relatively high or usage behaviour cannot be predicted precisely, then - as shown in ► Figure 8.11 on the left - a flat rate is offered, but once a quantity 𝑞𝑞 is reached, an additional usage charge of 𝑝𝑝 is levied. This is done to price capacity bottlenecks in particular, which may limit or decrease customer satisfaction. If marginal costs are very low - as in ► Figure 8.11 on the right - then a flat rate can also be offered without a cap. The current tariff schemes in mobile telecommunications are a reflection of limited predictability of data consumption. On the one hand, marginal costs of mobile telecommunications firms are very low, but on the other hand, data volumes are increasing significantly due to new business models and changing customer behaviour, so that flat rates with inclusive minutes, throttling of the data rate, and pricing of additional data packages are very common. <?page no="309"?> 8.3 Indirect price discrimination and two-part tariffs 309 Figure 8.12: Determination of the Bahncard 50 membership fee. Similarly, the price of a Bahncard 50 of Deutsche Bahn - which allows customers to travel at half the normal fare - can be determined. ► Figure 8.12 first shows the revenues of Deutsche Bahn without the Bahncard model. A customer will purchase a quantity of journeys 𝑞𝑞 1 at price 𝑝𝑝 1 , the marginal costs are negligible. If Deutsche Bahn now offers a Bahncard with a 50% discount on the regular fare, then customers will expand their demand to 𝑞𝑞 2 according to their price elasticity at a new usage price of 𝑝𝑝 1 / 2 . This initially reduces revenues of Deutsche Bahn - but in turn, the firm can now charge a membership fee equal to the grey triangular area. ► Figure 8.12 shows that the firm must choose the membership fee as an annual price of the Bahncard in such a way that the sum of the two grey areas on the right is larger than the grey area on the left. Again, customers themselves will do the selection in form of indirect price discrimination - customers with price-elastic demand will buy the Bahncard and travel more, but Deutsche Bahn's profits will increase in any case if the membership fee is chosen appropriately. In advance of introducing two-part tariffs, it is important from a management perspective to analyse whether customers are homogeneous or heterogeneous in terms of their willingness to pay. This question can be understood directly from ► Figure 8.10: if customers differ in willingness to pay and price elasticity, then optimal membership fees will also differ. If only one pricing model can be implemented, then heterogeneity of customer groups typically leads to low membership fees in order not to lose large customer groups - a membership fee may then not make sense against the background of additional administrative and marketing costs. A look at the mobile telecommunications industry shows, however, that a large variety of tariff models (different monthly membership fees, different usage-dependent fees, or three-part tariffs with completely and/ or permanently free components) can be developed in such a way that significant profit increases are still possible through self-selection of tariffs by customers. The <?page no="310"?> 8 Pricing strategies and price discrimination 310 reason is, on the one hand, that customers cannot correctly estimate their future consumption behaviour as this varies over time, and on the other hand, that tariffs are misperceived and customers systematically overestimate their own ability to choose the right tariff (Mitomo 2002, Grubb 2009, Ascarza 2012 as well as Haucap and Heimeshoff 2011). Freemium-models and damaged products strategy A special form of market segmentation through indirect price discrimination are freemium models and the damaged products strategy. Often customer segments differ in their willingness to pay, but it is inefficient to produce several different versions of similar products. In these cases, an existing high-quality product can be renamed, and certain features can be deactivated (damaged in a sense), thus, the lower quality product has higher marginal costs, but the firm's profit still increases. Examples are Intel computer chips, laser printers from Hewlett- Packard and DVD players from Sharp or Panasonic, which only differ due to deactivated buttons on the remote control (McAfee 2007 and Deneckere and McAfee 1996). Freemium models work in a similar way with apps for smartphones or software: in free versions, individual functions and features are locked, only the premium paid version contains full functionality (Kumar 2014). The aim of freemium models is to achieve some critical mass of direct network effects to establish a digital platform - this applies equally to games, music streaming, social media, or professional online career platforms. 8.4 Bundling A multi-product firm can use indirect price discrimination also in the shape of bundling. In this case, different products of a multi-product firm, which could also be sold individually, are sold as a bundle. This is in contrast to package prices, where identical products are sold in different package sizes. The idea of bundling is based on customers differing in their willingness to pay and that the customers' willingness to pay is negatively correlated across products. Bundling is often found with software products (PowerPoint, Excel and Word bundled in Microsoft Office); in services (car sales with financing and warranty services); in fast food chains (burger, fries and soft drink bundled in McDonald’s McMenu or Burger King Value Meals); at pay-TV firms or streaming services (channels or programmes bundled at Netflix or AmazonPrime); or at travel providers that bundle flights, baggage handling, hotels, travel cancellation insurance and rental cars even across firms (Adams and Yellen 1976, McAfee et al. 1989, Bakos and Brynjolfsson 1999, Stremersch and Tellis 2002 and Armstrong 2006). Case Study │ Bundling of software products ► Table 8.1 shows possibilities and effects of bundling using Microsoft as an example. The left part of ► Table 8.1 shows possible results of market research on willingness to pay for various Microsoft products. There are obviously four customer groups 𝐴𝐴 to 𝐷𝐷 with a total of 1,000 potential customers. The customer groups differ clearly from each other: customer group 𝐷𝐷 has a relatively high willingness to pay for PowerPoint at 70 EUR, but the willingness to pay for Excel is very low at 5 EUR - customer group 𝐴𝐴 , on the other hand, apparently has a great interest in Word, but little in Excel. <?page no="311"?> 8.4 Bundling 311 Microsoft as a dominant firm can now, as shown in the right part of the table, first determine separate optimum monopoly prices for each product in order to maximise revenues (costs can be disregarded, since marginal costs of duplicating the software and downloading it are zero). If Microsoft offers the product PowerPoint at a price of 70 EUR, for example, only 100 customers of customer group 𝐷𝐷 would buy - all other customer groups have too low a willingness to pay - so that revenues amount to 7,000 EUR. If Microsoft lowers the price of PowerPoint to 60 EUR, customer group 𝐶𝐶 will also buy. The revenues for a total of 500 customers now amount to 30,000 EUR. A further reduction to 40 EUR leads to revenues of 32,000 EUR - this cannot be increased further, because if the price is reduced further down to 30 EUR, indeed all 1,000 customers will buy, but revenues will now decline due to elastic demand. Table 8.1: Pure bundling using the example of Microsoft Office. Similarly, maximum revenues for the other two products can be determined, resulting in the following optimum pricing strategy for Microsoft: PowerPoint - at a price of EUR 40 to 800 customers of customer groups 𝐵𝐵 , 𝐶𝐶 and 𝐷𝐷 with revenues of EUR 32,000. Excel - at a price of EUR 30 to 500 customers of customer groups 𝐴𝐴 and 𝐵𝐵 with revenues of EUR 15,000. Word - at a price of EUR 40 to 900 customers of customer groups 𝐴𝐴 , 𝐵𝐵 and 𝐶𝐶 with revenues of EUR 36,000. so that total revenues for all products are EUR 83,000. <?page no="312"?> 8 Pricing strategies and price discrimination 312 However, Microsoft can increase total revenues by bundling: if one adds up the willingness to pay for the individual products for each customer group, one obtains the willingness to pay for an Office bundle shown in ► Table 8.1 below. Customer group 𝐴𝐴 would be willing to pay EUR 150 for the bundle, whereas customer group 𝐷𝐷 would only pay EUR 95. If we now apply the same logic as for the individual products to the bundle, we find that at a price of EUR 110, a total of 900 customers from customer groups 𝐴𝐴 , 𝐵𝐵 and 𝐶𝐶 would buy. Total revenues now amount to a maximum of EUR 99,000 - an increase of almost 20% compared to a separate sale of products, and this despite the fact that the bundle, at EUR 110, is just as expensive as the sum of the individual products. Separate pricing versus bundling Bundling works if customers or customer groups differ systematically in their willingness to pay for several products. In case of Microsoft, this is particularly true for customer groups 𝐴𝐴 and 𝐷𝐷 , whose willingness to pay for the three products is strongly divergent; moreover, marginal costs of all products are low in each case. More generally, to inspect how bundling works, the case of two products can be analysed: if customers 𝐴𝐴 , 𝐵𝐵 and 𝐶𝐶 have different willingness to pay 𝑧𝑧 1 and 𝑧𝑧 2 for the two products 1 and 2, and a firm sets separate prices 𝒑𝒑 𝟏𝟏 and 𝒑𝒑 𝟐𝟐 identical to the respective monopoly prices, as on the left in ► Figure 8.13, then customer 𝐶𝐶 in the upper right quadrant I buys both products, because 𝑧𝑧 1𝐶𝐶 > 𝑝𝑝 1 and 𝑧𝑧 2𝐶𝐶 > 𝑝𝑝 2 . Customer 𝐴𝐴 only buys product 2, because 𝑧𝑧 1𝐴𝐴 < 𝑝𝑝 1 but 𝑧𝑧 2𝐴𝐴 > 𝑝𝑝 2 , while customer 𝐵𝐵 only buys product 1 because 𝑧𝑧 1𝐵𝐵 > 𝑝𝑝 1 and 𝑧𝑧 2𝐵𝐵 < 𝑝𝑝 2 . Customers in quadrant III would not buy either product. Figure 8.13: Separate prices, bundling and optimisation of the bundling price. Bundling can be depicted by a bundling price 𝑝𝑝 𝐵𝐵 , which applies to both products in total and separates the customers' willingness to pay as a 45° line in ► Figure 8.13. With given willingness to pay, customers 𝐵𝐵 and 𝐶𝐶 now buy the bundle, because their respective added willingness to pay is obviously larger than the bundling price, i.e., 𝑧𝑧 1𝐵𝐵 + 𝑧𝑧 2𝐵𝐵 > 𝑝𝑝 𝐵𝐵 and 𝑧𝑧 1𝐶𝐶 + 𝑧𝑧 2𝐶𝐶 > 𝑝𝑝 𝐵𝐵 . Customer 𝐴𝐴 , on the other hand, does not buy, because his added willingness to pay for both products 𝑧𝑧 1𝐴𝐴 + 𝑧𝑧 2𝐴𝐴 < 𝑝𝑝 𝐵𝐵 is below the bundling price. z 1 z 2 A B C p 1 p 2 III II I IV p B z 2 A B C 45° z 1 z 2 A B C p 1 p 2 p B1 p B1 p B2 p B2 separate pricing bundling optimization of bundling z 1 p B <?page no="313"?> 8.4 Bundling 313 To optimally determine a bundling price, the 45° line is shifted to the right until the profits are maximised by a combination of number of customers and a bundling price that is then possible - analogous to the previous example with Microsoft Office products: customers in the grey area do not buy, all others buy the bundle. Bundling works well, if there is a negative correlation between the willingness to pay of the customer groups, i.e., they have alternately high and low willingness to pay across products as in ► Figure 8.14 on the left. The customers' willingness to pay is then lined up parallel to the bundling price. Shifting the bundling price from underneath these customers leads to an increase in profits, and at the same time only relatively few customers are deterred from buying the bundle. If, on the other hand, the willingness to pay is strongly positively correlated - some customers on the bottom left have a low willingness to pay for both products, while other customers on the top right have a high willingness to pay for both products - as in ► Figure 8.14 on the right, bundling does not lead to an increase in profits: customers are lost with each increase in the bundling price. Figure 8.14: Negatively and positively correlated willingness-to-pay and the logic of bundling. In order to apply bundling, however, a firm does not need exact knowledge of individual willingness to pay. From a management perspective, it is sufficient to assume or have a conjecture that there is a negative correlation between the willingness to pay across customer groups. Mixed bundling Many firms choose mixed bundling as a pricing strategy, where products are offered separately and at the same time as a bundle. ► Figure 8.15 shows prices of selected individual products and related bundles at the McDonald’s restaurant in Hamburg-Barmbek in May 2017. Apparently, McDonald’s offers these products in different meals at prices 𝑝𝑝 𝐵𝐵 that are on average z 1 z 2 p 2 p B p B p 1 z 1 z 2 p 2 p B p B p 1 45° 45° strong negative correlation of willingness to pay > bundling increases profits strong positive correlation of willingness to pay > bundling will not increase profits <?page no="314"?> 8 Pricing strategies and price discrimination 314 about 10% lower than summing up individual prices 𝑝𝑝 𝐵𝐵′ - behind this lies an analysis of the price elasticity of demand of individual products and of bundles to increase revenues. Pay-TV providers also often use mixed bundling - programmes, channels, or films can be subscribed to individually, or certain packages (sports, action, entertainment, etc.) can be booked exclusively or in addition. In the UK, customers can also specify their individual bundles. Figure 8.15: Mixed bundling at McDonald’s. Figure 8.16: Mixed bundling. <?page no="315"?> 8.4 Bundling 315 To understand how mixed bundling works, ► Figure 8.16 shows separate prices 𝑝𝑝 1 and 𝑝𝑝 2 for two products, a bundling price 𝑝𝑝 𝐵𝐵 and, marked by dots, any number of customers with different willingness to pay 𝑧𝑧 𝑖𝑖 . With two products and prices chosen here, a firm can clearly identify four regions (customer or market segments) as follows. Region 1 - customers will buy neither the bundle nor separate products, because the willingness to pay is too low in each case, so that 𝑧𝑧 1 < 𝑝𝑝 1 , 𝑧𝑧 2 < 𝑝𝑝 2 and 𝑧𝑧 1 + 𝑧𝑧 2 < 𝑝𝑝 𝐵𝐵 apply. Region 2 - these customers will buy the bundle in any case, because with 𝑧𝑧 1 + 𝑧𝑧 2 > 𝑝𝑝 𝐵𝐵 the willingness to pay is above the bundle price. Region 3 - customers in this region will only buy product 2 because of 𝑧𝑧 1 < 𝑝𝑝 1 but 𝑧𝑧 2 > 𝑝𝑝 2 and 𝑧𝑧 1 + 𝑧𝑧 2 < 𝑝𝑝 𝐵𝐵 . Region 4 - customers in this region only buy product 1 because of 𝑧𝑧 1 > 𝑝𝑝 1 but 𝑧𝑧 2 < 𝑝𝑝 2 and 𝑧𝑧 1 + 𝑧𝑧 2 < 𝑝𝑝 𝐵𝐵 . In addition, there are two regions 5 and 6 in which customers could buy the bundle because 𝑧𝑧 1 + 𝑧𝑧 2 > 𝑝𝑝 𝐵𝐵 . However, in each case the relative utility from buying the individual product, for example (𝑧𝑧 1 − 𝑝𝑝 1 ) > (𝑧𝑧 1 + 𝑧𝑧 2 − 𝑝𝑝 𝐵𝐵 ) for region 6, is larger than the utility from buying the bundle - customers in regions 5 and 6 will, therefore, buy the respective individual products. Optimum prices for mixed bundling, especially if more than two products can be combined, are developed and determined with simulation software. Typically, mixed bundling can increase profit compared to pure bundling. ► Table 8.2 demonstrates this for Microsoft Office products. Table 8.2: Mixed bundling using the example of Microsoft Office. <?page no="316"?> 8 Pricing strategies and price discrimination 316 Compared to pure bundling from ► Table 8.1, Microsoft is able to increase profits by a further 11.6% to EUR 110,500 for these four customer groups considered. To achieve this, a firm first addresses customer group 𝐴𝐴 with the highest willingness to pay for Excel and Word with separate prices of EUR 50 and EUR 70 during the market launch of new versions. None of the other customer groups will buy these products at this price. After the market launch, Microsoft positions an Office bundle for the mass market analogous to ► Table 8.1 at a price of EUR 110 - 700 customers in segments 𝐵𝐵 and 𝐶𝐶 will purchase this now. Finally, Microsoft offers the customer group 𝐷𝐷 with the lowest willingness to pay (obviously students who urgently need PowerPoint) a discounted student version with deactivated features at a price of EUR 95 against verification via student ID. From a management perspective, a multi-product firm with market power must, therefore, always decide whether products should be offered separately at respective monopoly prices, as pure bundling at one price, or as mixed bundling in order to maximise profits. The examples of Microsoft or McDonald’s (two of the most profitable firms of the last decades) show that mixed bundling can be a major driver of profitability for multi-product firms with a dominant market position. In addition, bundling can be used as a strategic barrier to entry. A single-product firm can be deterred from entering the market by the bundling prices of a multi-product firm if no profit can be achieved for the single-product firm due to the distribution of customers' willingness to pay (Nalebuff 2004). For this reason, bundling is also a strategy that restrains competition and can strengthen existing market power. Against this background, governmental action by competition authorities to unbundle business models or products is taking place. Examples include the energy industry with a separation of network infrastructure and energy generation; and the intended splitting up of Deutsche Bahn’s rail network, passenger transport and freight transport. However, unbundling can also be an innovative strategy to attack incumbent firms. In the music industry, for example, download and streaming providers have put established providers of the '10 songs on a physical CD' bundle under significant pressure by using new technology and changing customer behaviour (Elberse 2010). 8.5 Auctions In many markets, products or services are offered or demanded via auctions. In B2C, auctions take place at traditional auction firms such as Sotheby's, on TV with 'Antiques Road Trip' in the UK, or 'Bares für Rares' in Germany as well as via platforms such as ebay. In the B2B sector, auctions are common in the context of tenders for projects, in purchasing or procurement processes via electronic auctions and marketplaces such as WaxDigital or marketdojo. Other examples include the pricing of online advertising such as Google ads (so-called positioning auctions); auctioning of government bonds by the Federal Ministry of Finance in Germany; or the allocation of radio frequencies for mobile telecommunications providers by the Bundesnetzagentur in Germany. Auctions are usually multi-stage, dynamic bidding procedures and serve to determine price and/ or allocation and assignment of a product. In 2019, for example, a total of 41 frequency blocks were auctioned off to the four bidders Deutsche Telekom, Voda- <?page no="317"?> 8.5 Auctions 317 fone, Telefónica and 1&1 Drillisch after 497 auction rounds, with a total value of EUR 6.5 billion (Bundesnetzagentur 2019). Whereas, in the past a human auctioneer was usually still necessary to collect the bids and award a contract, most auctions are now conducted by the seller or buyer themselves via electronic platforms. Auctions and asymmetric information Auctions are an alternative for determining prices if a uniform market price as described in ► Chapter 1 or a market equilibrium cannot be expected. On the one hand, there might be information asymmetries on the supply or demand side, which make price formation difficult or impossible due to a lacking market mechanism. On the other hand, it may be a question of one-off projects or rare products for which no information is available in advance to determine a price. A supplier of such a single object could act as a monopolist on the market and set a price. The risk of such a behaviour is, without knowledge of the customers' willingness to pay, this price can be too high, so that no transaction takes place. Alternatively, the price may be too low, so that the maximum willingness to pay is not captured and the firm's profits are not maximised. An auction can also reduce transaction costs - if conducting an auction is cheaper than conducting some otherwise necessary market research to determine customers' willingness to pay (Klemperer 1999, McAfee and McMillan 1987, Varian 2007 as well as Koutroumpis and Cave 2018). The simplest case of a one-sided sales auction is structured as follows: one firm is offering a single item for sale; several potential customers exist for this single object; there is information asymmetry - customers know their respective maximum willingness to pay, yet the firm knows neither the number of customers nor their willingness to pay. This firm could now set a price based on their estimate of the number of customers and their willingness to pay; or conduct an auction to find out information about the number of customers and their willingness to pay. For this purpose, rules for the conduct of an auction (way of submission of bids, continuous submission of bids or a fixed number of auction rounds, information about the bids, mutual acquaintance versus anonymity of the bidders' identity, etc.) as well as for the setting of the price must be determined. The key advantage of an auction is that the price determination takes place on the better-informed customer side and possible disadvantages for a firm due to asymmetric or incomplete information are reduced. If an optimum auction design is chosen, a firm can determine the maximum willingness to pay and, thus, achieve the highest possible price. Auction methods An auction design describes a set of rules by which the purchase price and the winner of an auction are determined. Designs differ with respect to dimensions of bidding (open or sealed) and price determination (highest price bid or second-highest price bid). In sealed bid auctions, each participant may only bid once and the bids of the other participants are unknown. In open bid auctions, bidders can see the bids of other participants and successively adjust their own bids. The fundamental difference is that bidders can estimate the other bidders' will- <?page no="318"?> 8 Pricing strategies and price discrimination 318 ingness to pay when announcing newly received bids in open auctions. In first price auctions, the bidder with the highest bid wins the auction and pays according to his own bid. In second price auctions, the bidder with the highest bid also wins, but only pays the price of the second highest bid. The combination of these auction rules gives rise to the following four main auction designs. English auctions (also called 'ascending-bid auction') are ascending auctions with open bids. The seller has the option to set a minimum bid. Starting from this minimum bid, bidders openly state their bids and have the opportunity to continuously increase their bids. The lot is awarded to the highest bidder, who pays a price corresponding to his last bid. English auctions are common for auctions of works of art, antiques, or real estate. Dutch auctions (also called 'descending-bid auctions') are descending auctions: the seller starts at a high price (starting price) and gradually reduces it until a bidder makes a first bid at the current price. This bidder pays the price that the auctioneer called at the time of bidding. Dutch auctions were made famous by flower auctions in the Netherlands with a 'flower price clock' running backwards, but are also used in US government bond auctions and are a possible method of allocating shares in IPOs, e.g., Google's IPO in 2004. First-price sealed bid auctions are based on sealed bids - bidders submit a sealed bid once within a fixed period of time without knowing other bidders or bids; the highest bidder wins. In a Vickrey auction (second-price sealed bid auction), bidders submit a sealed bid once within a fixed period of time. The highest bidder wins the auction, but - unlike in a firstprice auction - only has to pay the price of the second-highest bid (Vickrey 1961). English and Dutch auctions are dynamic, i.e., the price can change continuously - although English auctions often last longer and may cause bidding wars. First-price and Vickrey auctions, on the other hand, are static - after all bids have been submitted once, the lot can be knocked down immediately. Based on these four basic models, any combination and set of rules can be developed, e.g., to maximise the revenues from a mobile telecommunications licence auction (Haeringer 2018, Myerson 1981 and Jia et al. 2009). Bidding strategies and auction prices Which of these four auction designs yields the highest price from the seller's point of view, which is the optimum bidding strategy for buyers, and whether advantages lie more on the side of sellers or buyers depends on the degree of information asymmetry, the number of bidders, the distribution of willingness to pay as well as the valuation of the auction item. In auctions with private value (also called 'independent private value auction'), each bidder has his own subjective valuation and, thus, an individual willingness to pay for the auction item. These valuations typically differ between bidders and these values are initially independent of each other, such that there is no uniform value for the auction object and there is uncertainty about the achievable price before the auction process. Typical examples of this auction model with preference uncertainty are auctions of art objects (which are not intended for resale) or memorabilia (objects or souvenirs that are very valuable to some people but may be worthless to others), but also bidding processes for contracts if the individual production costs are known only to the respective bidder himself. Auctions with private value are, thus, individual <?page no="319"?> 8.5 Auctions 319 learning processes: each bidder successively learns from the bids of the other bidders about their willingness to pay, but this will not change the willingness to pay of a rational bidder. Private value auctions bidder A bidder B bidder C bidder D bidder E maximum individual willigness to pay 3.30 5.00 2.90 6.50 2.20 English auction highest bid 3.30 5.00 2.90 5.01 2.20 winning bid 5.01 Vickreyauction highest bid 3.30 5.00 2.90 6.50 2.20 winning bid 5.00 Dutch auction highest bid - - - 6.50 winning bid 6.50 first price auction highest bid 3.30 5.00 2.90 6.50 2.20 winning bid 6.50 Table 8.3: Auctions with private values. ► Table 8.3 shows a private value auction in which bids may be submitted in increments of 0.01 EUR, the willingness to pay of five bidders and the highest bids submitted per bidder in each of the four auction designs. Three results are central: (i) bidder D always wins regardless of the auction design chosen, (ii) the Dutch and the highest price auctions lead to the same price of 6.50 EUR, and (iii) the English and the Vickrey auctions lead to (almost) the same price of 5.00 EUR and 5.01 EUR, respectively. Accordingly, it can be shown in general - given a large number of rational and risk-neutral bidders as well as a continuum of willingness-to-pay - that an English auction and a Vickrey auction are strategically equivalent, and that a Dutch auction and a highest-price auction are strategically equivalent (Klemperer 1999 and Haeringer 2018). If there are not only a few bidders, as in ► Table 8.3 with discrete willingness-to-pay, but an infinite number of bidders whose valuation is independent and identically distributed as well as complete (i.e., every possible price is occupied), the bidders are risk-neutral and decide rationally, then the revenue equivalence theorem holds: all auction designs lead to the same price and to the same revenue for the seller. Accordingly, the bidding strategy for rational bidders in English and Vickrey auctions are simple: one bids according to the maximum willingness to pay and obtains a utility according to the difference between the willingness to pay and the actual price. In a Dutch or highest price auction, on the other hand, the maximum bid of each rational bidder should be below the willingness to pay, because only then does some utility arise: this, however, creates the risk that another bidder will win. This risk increases the more bidders participate in the auction and the further the bid is chosen below the willingness to pay. In this case, one's bidding strategy is <?page no="320"?> 8 Pricing strategies and price discrimination 320 essentially determined by risk aversion: the larger the risk aversion, and the larger the number of participants in the auction, the higher the price that can be achieved. If, on the other hand, the bidders decide with bounded rationality, emotion, or risk aversion; or if the number of bidders is small and the willingness to pay is widely spread; then significantly different results are possible in all four auction forms - up to and including bidding wars that can often be shown in experiments (Kagel and Levin 1993 as well as Lusk and Shogren 2007). This is accompanied by the fact that the definition and implementation of bidding strategies becomes significantly more complicated. Moreover, auctions regularly lead to higher prices: a study by Ariely and Simonson (2003) shows higher prices for 98.8% of the analysed online auctions compared to alternative fixed prices with an average price increase of 15.3%. In common value auctions, an auction item has the same objective value for each bidder. The estimation of this value, however, may vary due to different assessments of the quality or due to uncertainty about the auction item. This applies, for example, to auctions of purses in lost property offices, or mining rights of raw materials where an absolute quantity of some natural resource is unknown. Although the value is the same for all bidders ex post, it is only known after all natural resources have been extracted or the purse is opened. Therefore, auctions with a common value are collective learning processes: each bidder learns about the other bidders’ willingness to pay from each other’s successive bids and will adjust their own willingness to pay accordingly. This adjustment of willingness to pay and bids leads to the so-called winners' curse: with an increasing number of bidders and a probability distribution of valuations, bids increase due to the mutual adjustment to the other bidders' valuations - as a result, the highest bidder regularly pays a price above the actual value (Bazerman and Samuelson 1983, Thaler 1988 and Bajari and Hortaçsu 2003). Optimum auction design Similarly, auctions of mobile licences are auctions of common value, although they are mostly designed as combinatorial auctions: mobile telecommunication firms usually need contiguous, complementary frequency blocks to offer services in a 4G or 5G network. However, mobile firms only realise the true value of UMTS or 5G licences over time - after the auction - and can only develop assumptions in advance based on their own business case and on the bids of other firms due to strong technological uncertainty. However, evidence of the winner's curse has not been detected, at least not for the UMTS auction in the UK (Cable et al. 2002). One reason may be that in English auctions, in particular, visible rising bids can be used as signals to other bidders - thus, implicit agreements are possible and reduce the maximum possible auction revenues (Klemperer 2002). Following the first mobile licence auctions during the early 2000’s, numerous competition and regulatory authorities have optimised their auction rules in order to prevent collusion, increase the number of participants in the auctions and, thus, increase revenues. A perfect auction would have many risk-neutral bidders, all bidders having the same level of information, all bidders behaving rationally, and having a clear private or common value for the auctioned item, and the winning bid would be based on these bids alone. The perfect auction, therefore, would not require an auction design: all auction designs would lead to identical re- <?page no="321"?> 8.6 Summary and key learnings 321 sults, i.e., prices for the bidders and revenues for the sellers, as explained by the revenue equivalence theorem (Myerson 1981 and McAffee and McMillan 1987). In contrast, some conditions for optimal auction design can be derived from experiments to maximise auction revenues if the conditions of the revenue equivalence theorem are not fulfilled (Lusk and Shogren 2007, Crawford et al. 2009 and Davis et al. 2014). These conditions would be as follows. Regardless of an auction type, as many bidders as possible should participate - the number of bidders increases competition, the expected highest bid and, thus, the expected second highest bid. In auctions with private value and risk aversion, a Dutch auction usually increases revenues compared to an English auction. In English auctions, a high minimum bid reduces the risk of loss for sellers - especially if it is not clear how many bidders will participate. In common value auctions, a reduction of information asymmetry leads to an increased willingness of risk-averse bidders to place higher bids. English auctions achieve higher auction revenues than first-price auctions if better information becomes available to all auction participants in the course of the auction. Negotiation and cooperation between bidders must be prevented - any form of collusion reduces auction revenues. 8.6 Summary and key learnings How can you find the optimum pricing strategy for a product - do you need to do market research, or will you just auction it off? Are you able to sell identical products at different prices? Price discrimination describes pricing strategies of dominant firms in which different prices are charged for essentially identical products according to the customers' willingness to pay. Firms are then able to transform some or even all of the consumer surplus into significant profit increases. Price discrimination can take place directly through personalised pricing or market segmentation if market research or digital technologies can identify customer groups with sufficient precision. Indirect price discrimination is based on self-selection of customers: according to their willingness to pay, customers assign themselves to mobile telecommunication tariffs or determine their use over time within two-part tariffs such as the Bahncard. Pure and mixed bundling are strategies of multi-product firms such as Microsoft or McDonald’s to achieve higher profits by bundling products. Auctions allow improved pricing in the presence of incomplete information. Auction designs depend on whether bidders have different private values for the auction item or whether there is a common value. The smaller the number and the more different the level of information of the participants, the more important is the choice of auction design. Conversely, the revenue equivalence theorem shows that with many risk-averse participants and the same level of information, there is no difference in the outcome: English and Dutch auctions lead to the same price. <?page no="322"?> 8 Pricing strategies and price discrimination 322 In a narrower sense, the described pricing strategies only apply to firms with market power - it is quite clear that firms also apply price discrimination if competition is intense. However, especially in case of competition in an oligopoly, complex repercussions of competitor behaviour would then have to be taken into account (Stole 2007). Recommendations for further reading Pricing from a business perspective is comprehensively covered by Simon, H. and Fassnacht, M., Preismanagement, Berlin 2016. Extensive analysis from a microeconomic and strategic perspective is provided by Belleflamme, P. and Peitz, M., Industrial organisation: markets and strategies, London 2015. Questions for review [1] Describe applications of price discrimination from a microeconomic perspective as well as their limitations, advantages and disadvantages. [2] Which requirements are needed for a successful application of price discrimination? Describe differences and similarities between direct and indirect price discrimination. Name two cases where the distinction is not clear-cut. [3] Provide two examples each of personalised (perfect) price discrimination, package pricing and group pricing. [4] Describe the basic idea of (pure and mixed) bundling with the help of an illustration. Name and briefly explain three examples from different industries in which (pure or mixed) bundling is typically applied. [5] Mobile phone or smartphone tariffs often include minute or data packages, but minutes or data beyond that cost extra - why? What is the pricing strategy here? Explain the basic considerations and strategies for two-part tariffs. [6] A bowling club offers courts at EUR 20 per hour, currently there is no membership fee. The demand function of amateur team A is 𝑝𝑝 = 50 − 0.025𝑞𝑞 per year - currently this group plays 1200 hours per year, the demand function of a pub group B is 𝑝𝑝 = 40 − 0.025 𝑞𝑞 per year - currently this group plays 800 hours per year. The bowling club is now thinking about introducing a uniform annual membership fee per team, the court rent at EUR 20 per hour should be maintained. Which membership fee should be established? How many hours will these two teams play bowling, once the membership fee is in place? [7] What is the basic idea of Deutsche Bahn offering a Bahncard 50, which conditions have to be fulfilled to make a success? [8] Why does the pay-TV provider Premiere offer individual sports and feature film channels, but also programme packages - which forms of bundling are used here? [9] L'Oréal is planning to launch a new lotion - the firm is thinking about offering this new lotion with a scarf as a bundle during market launch. Marginal costs for the lotion are EUR 3, for the scarf EUR 7. The willingness to pay was determined by market research for five equally sized customer groups, as can been seen from the table below: <?page no="323"?> 8.6 Summary and key learnings 323 customer group willingness to pay lotion scarf A 20 5 B 18 12 C 12 18 D 9 21 E 4 24 At which prices should one sell the products separately? At which price would one sell the bundle (one scarf, one lotion)? Which strategy leads to a higher profit? Why? Cinema chains willingness to pay of cinema chain marginal costs of movie distributor action movie comedy action movie comedy A with 5 cinemas 13,000 6,000 5,000 5,000 B with 5 cinemas 7,000 11,000 5,000 5,000 C with 5 cinemas 15,000 2,000 5,000 5,000 [10] The table shows the willingness to pay for two films (action movie and comedy) for three cinema chains A, B and C. Marginal costs are 5,000 EUR per film per week per cinema. Which prices should the film distributor charge for pure bundling? Is there a possibility to achieve higher profits by mixed bundling? [11] Explain the four common auction methods. [12] Explain differences between auctions with private and common values of bidders. Literature Adams, W.J. and Yellen, J.L., Commodity bundling and the burden of monopoly, Quarterly Journal of Economics, 1976, 90, 3, 475-498. Ariely, D. and Simonson, I., Buying, bidding, playing, or competing? Value assessment and decision dynamics in online auctions, Journal of Consumer Psychology, 2003, 13, 113-123. Armstrong, M., Recent developments in the economics of price discrimination, in: Blundell, R., Newey, W.K. and Persson, T. (eds.), Advances in economics and econometrics: theory and applications, Volume II, Cambridge 2006, 97-141. Ascarza, E., Lambrecht, A. and Vilcassim, N., When talk is “free”: the effect of tariff structure on usage under twoand three-part tariffs, Journal of Marketing Research, 2002, 49, 6, 882-899. Bajari, P. and Hortaçsu, A., The winner's curse, reserve prices, and endogenous entry: Empirical insights from ebay auctions, RAND Journal of Economics, 2003, 34, 2, 329-355. Baker, W., Kiewell, D. and Winkler, G., Using big data to make better pricing decisions, McKinsey Quarterly, July 2014. Bakos, Y. and Brynjolfsson, E., Bundling information goods: pricing, profits, and efficiency, Management Science, 1999, 45, 12, 1613-1630. <?page no="324"?> 8 Pricing strategies and price discrimination 324 Bar-Gill, O., Algorithmic price discrimination when demand is a function of both preferences and (mis)perceptions, University of Chicago Law Review, 2019, 86, 2, 217-254. Bazerman, M.H. and Samuelson, W.F., I won the auction but don't want the prize, Journal of Conflict Resolution, 1983, 27, 4, 618-634. Bertini, M. and Koenigsberg, O., When customers help set prices, Sloan Management Review, 2014, 55, 57-66. Bolton, L.E., Warlop, L. and Alba, J.W., Consumer perceptions of price (un)fairness, Journal of Consumer Research, 2003, 29, 474-491. Bundesnetzagentur (ed.), 5G Frequenzauktion beendet, Pressemeldung 12. Juni 2019. Busse, C., Vorbild Billig-Konkurrenz: Lufthansa startet neues Preismodell, Süddeutsche Zeitung, 24. Juli 2015. Cable, J., Henley, A. and Holland, K., Pot of gold or winner's curse? An event study of the auctions of 3G mobile telephone licences in the UK, Fiscal Studies, 2002, 23, 4, 447-462. Chen, L., Measuring algorithms in online marketplaces, Boston 2017. Chen, L., Mislove, A. and Wilson, C., An empirical analysis of algorithmic pricing on Amazon Marketplace, Proceedings of the 25th International Conference on World Wide Web, 2016, 4, 1339-1349. Chen, M. and Chen, L., Recent developments in dynamic pricing research: multiple products, competition, and limited demand information, Production and Operations Management, 2015, 24, 5, 704-731. Chu, C.S., Leslie, P. and Sorensen, A., Bundle-size pricing as an approximation to mixed bundling, American Economic Review, 2011, 101, 263-303. Cohen, A., Package size and price discrimination in the paper towel market, International Journal of Industrial Organization, 2008, 26, 2, 502-516. Crawford, V.P., Kugler, T., Neeman, Z. and Pauzner, A., Behaviorally optimal auction design: Examples and observations, Journal of the European Economic Association, 2009, 7, 2-33, 377-387. Dakers, M., Uber knows customers with dying batteries are more likely to accept surge pricing, The Telegraph, Oct 2016. Dautzenberg, K., Gaßmann, C., Groß, B., Müller, F., Neukamp, D. and Bodenstein, U., Dynamische Preisdifferenzierung im deutschen Online-Handel - eine Untersuchung der Verbraucherzentralen, Potsdam 2018b. Dautzenberg, K., Gaßmann, C., Groß, B., Müller, F., Neukamp, D., Schmidtke, L. and Bodenstein, U., Individualisierte Preisdifferenzierung im deutschen Online-Handel - eine Untersuchung der Verbraucherzentralen, Potsdam 2018a. Davis, A.M., Katok, E. and Kwasnica, A.M., Should sellers prefer auctions? A laboratory comparison of auctions and sequential mechanisms, Management Science, 2014, 60, 4, 990-1008. DellaVigna, S. and Malmendier, U., Paying not to go to the gym, American Economic Review, 2006, 96, 3, 694-719. Deneckere, R. and McAfee, P., Damaged goods, Journal of Economics and Management Strategy, 1996, 5, 2, 149-174. Elberse, A., Bye-bye bundles: the unbundling of music in digital channels, Journal of Marketing, 2010, 74, 3, 107-123. Ezrachi, A. and Stucke, M.E., Artificial intelligence and collusion: when computers inhibit competition, University of Tennessee, Research Paper #267, May 2015 Gal, A., It’s a feature, not a bug: on learning algorithms and what they teach us, OECD Roundtable on Algorithms and Collusion, Paris 2017. Garbarino, E. and Lee, O.F., Dynamic pricing in internet retail: effects on consumer trust, Psychology and Marketing, 2003, 20, 6, 495-513. Grubb, M.D., Selling to overconfident consumers, American Economic Review, 2009, 99, 5, 1770-1807. Haeringer, G., Market design: auctions and matching, Cambridge 2018. Hannak, A., Soeller, G., Lazer, D., Mislove, A. and Wilson, C., Measuring price discrimination and steering on e-commerce web sites, Proceedings of the 2014 conference on internet measurement conference, 2014, 305-318. Haucap, J. and Heimeshoff, U., Consumer behavior towards on-net/ off-net price differentiation, Telecommunications Policy, 2011, 35, 4, 325-332. <?page no="325"?> 8.6 Summary and key learnings 325 Heidhues, P. and Köszegi, B., Exploiting naïvete about self-control in the credit Market, American Economic Review, 2010, 100, 5, 2279-2303. Jentzsch, N., Wohlfahrts- und Verteilungswirkungen personalisierter Preise und Produkte, Friedrich-Ebert-Stiftung, WP, 2017. Jia, J., Zhang, Q., Zhang, Q. and Liu, M., Revenue generation for truthful spectrum auction in dynamic spectrum access, in: Proceedings of the tenth ACM International Symposium on Mobile ad hoc networking and computing, 2009, 3-12. Just, D.R. and Wansink, B., The flat-rate pricing paradox: conflicting effects of “all-you-can-eat” buffet pricing, Review of Economics and Statistics, 2011, 93, 1, 193-200. Kagel, J.H. and Levin, D., Independent private value auctions: bidder behaviour in first-, secondand third-price auctions with varying numbers of bidders, Economic Journal, 1993, 103, 419, 868-879. Kahneman, D., Knetsch, J.L. and Thaler, R., Fairness as a constraint on profit seeking: entitlements in the market, American Economic Review, 1986, 76, 4, 728-741. Kim, J.-Y., Natter, M. and Spann, M., Pay what you want: a new participative pricing mechanism, Journal of Marketing, 2009, 73, 1, 44-58. Klemperer, P., Auction theory: a guide to the literature, Journal of Economic Surveys, 1999, 13, 3, 227-286. Klemperer, P., How (not) to run auctions: The European 3G telecom auctions, European Economic Review, 2002, 46, 4-5, 829-845. Koutroumpis, P. and Cave, M., Auction design and auction outcomes, Journal of Regulatory Economics, 2018, 53, 275- 297. Krämer, A. and Kalka, R., Dynamic Pricing - Verspielt Amazon das Vertrauen seiner Kunden? , Absatzwirtschaft, 2016, online. Kremer, M., Mantin, B. and Ovchinnikov, A., Dynamic pricing in the presence of myopic and strategic consumers: theory and experiment, Production and Operations Management, 2017, 26, 1, 116-133. Kumar, V., Making “freemium” work: many start-ups fail to recognize the challenges of this popular business model, Harvard Business Review, 2014, 92, 5, 27-29. Levin, Y., McGill, J. and Nediak, M., Dynamic pricing in the presence of strategic consumers and oligopolistic competition, Management Science, 2008, 55, 32-46. Lusk, J.L. and Shogren, J.F., Experimental auctions - methods and applications in economic and marketing research, Cambridge 2007. McAfee, P. and de Velde, V., Dynamic pricing in the airline industry, in: Hendershott, T. (eds.), Handbook on Economics and Information Systems, Amsterdam 2007, 527-570. McAfee, P., McMillan, J. and Whinston, M.D., Multiproduct monopoly, commodity bundling, and correlation of values, Quarterly Journal of Economics, 1989, 104, 2, 371-383. McAfee, P., Pricing damaged goods, Economics: The Open-Access, 2007, 1, 1-19. McAfee, R.P. and McMillan, J., Auctions and bidding, Journal of Economic Literature, 1987, 25, 2, 699-738. Mikians, J., Gyarmati, L., Erramilli, V. and Laoutaris, N., Detecting price and search discrimination on the internet, Proceedings of the 11th ACM workshop on hot topics in networks, 2012, 79-84. Mitomo, H., Heterogeneous subscribers and the optimal two-part tariff of telecommunications service, Journal of the Operations Research Society of Japan, 2002, 35, 2, 194-214. Mohammed, R., How retailers use personalized prices to test what you’re willing to pay, Harvard Business Review, 2017. Myerson, R.B., Optimal auction design, Mathematics of Operations Research, 1981, 6, 1, 58-73. Nalebuff, B., Bundling as an entry barrier, Quarterly Journal of Economics, 2004, 119, 1, 159-187. <?page no="326"?> 8 Pricing strategies and price discrimination 326 Narahari, C., Raju, K., Ravikumar, K. and Shah, S., Dynamic pricing models for electronic business, University of Bangalore WP 23-4.2015, 2015. Narayanan, S., Chintagunta, P.K. and Miravete, E.I., The role of self selection, usage uncertainty and learning in the demand for local telephone service, Quantitative Marketing and Economics, 2007, 5, 1, 1-34. o.V., Lufthansa weist Vorwurf überhöhter Preise zurück, Die Zeit, 26. November 2017. OECD (ed.), Personalized Pricing in the Digital Era, Paris 2018. Papanastasious, Y. and Savva, N., Dynamic pricing in the presence of social learning and strategic consumers, Management Science, 2014, 61, 1-21. Rayna, T., Darlington, J. and Striukova, L., Pricing music using personal data: mutually advantageous first-degree price discrimination, Electronic Markets, 2015, 25, 2, 139-154. Reinartz, W., Haucap, J., Wiegand, N. and Hunold, M., Preisdifferenzierung und-dispersion im Handel, Ausgewählte Schriften der IFH-Förderer, 2017, # 6. Richards, T.J., Liaukonyte, J. and Streletskaya, N.A., Personalized pricing and price fairness, International Journal of Industrial Organization, 2016, 44, 138-153. Schwalbe, U., Algorithms, machine learning, and collusion, Journal of Competition Law and Economics, 2018, 14, 4, 568- 607. Seele, P., Dierksmeier, C., Hofstetter, R. and Schultz, M.D., Mapping the ethicality of algorithmic pricing, Journal of Business Ethics, 2020. Solon, O., How a book about flies came to be priced $24 million on Amazon, wired.com/ 2011/ 04/ Amazon-flies-24-million/ Stole, L., Price discrimination and competition, in: Armstrong, M. and Porter, R., Handbook of Industrial Organization, 2007, Volume 3, 2221-2299. Stremersch, S. and Tellis, G.J., Strategic bundling of products and prices: a new synthesis for marketing, Journal of Marketing, 2002, 66, 1, 55-72. Tanner, A., Different customers, different prices, thanks to big data, Forbes, 26. März 2014. Thaler, R.H., The winner’s curse, Journal of Economic Perspectives, 1988, 2, 1, 191-202. Uken, M., Preissysteme - Zahl so viel du willst., Die Zeit, 14. Januar 2013. Varian, H.R., Position auctions, International Journal of Industrial Organization, 2007, 25, 6, 1163-1178. Vickrey, W., Counterspeculation, auctions, and competitive sealed tenders, Journal of Finance, 1961, 16, 1, 8-37. Xia, L., Monroe, K.B. and Cox, J.L., The price is unfair. A conceptual framework of price fairness perceptions, Journal of Marketing, 2004, 68, 1-15. <?page no="327"?> 327 9 Strategic decision-making and game theory Strategic decisions of a firm must take into account and anticipate possible strategic actions or reactions of other firms - interactions of strategies are analysed using game theory. The following simple example demonstrates some key considerations. A mobile ice cream van named Star-Ice arrives at a linear beach from the right. All potential customers are equally distributed between points A and B and a competitor Triangle-Ice with a likewise mobile ice cream van is already positioned on the beach as can be seen in the upper part of ► Figure 9.1. This firm offers ice cream of the same quality at the same prices and customers have no preference for either ice cream vendor. Where should Star-Ice, which is new to the beach, place its ice cream van in order to maximise its profits? How will its competitor Triangle-Ice react? Figure 9.1: Strategic positioning of ice cream vans on a linear beach. Initially, the new firm Star-Ice can position itself directly to the right of Triangle-Ice - so all customers on the right have a shorter distance to Star-Ice and this firm has a correspondingly large market share. However, Triangle-Ice will react and change its own location to the right - which in turn will cause a reaction from Star-Ice. This dynamic will continue until none of the competitors has an incentive to change position because no further improvements in market share and profits is possible. In order to maximise profits, both ice cream vendors will eventually position themselves exactly in the middle of the beach: each competitor will thus obtain a 50% market share, i.e., each will <?page no="328"?> 9 Strategic decision-making and game theory 328 serve its half of the beach. This result is also known as Hotelling's rule of minimum differentiation (Hotelling 1929). Of course, it does not only apply to ice-cream sellers - firms that offer similar products or products that are weakly differentiated from a customer's perspective choose their location next to or opposite each other: savings banks next to Volksand Raiffeisenbanks in Germany, Zara next to H&M, Aral opposite Shell, Burger King next to McDonald’s, and firms like Aldi and Lidl even share their car park in commercial areas. The central insight, however, is not that firms choose locations next to each other, but that competing firms optimise their respective strategies until these firms no longer have any individual incentive to change their strategy - this is one of the central solution concepts of game theory and a special case of a Nash equilibrium. Playing field and rules of game theory Game theory is a concept for analysing strategic decisions in conflict or cooperation situations - trying to determine optimum decisions while taking into account and anticipating the decisions of all other players (von Neumann 1928). Game theory is applied in fields as diverse as economics, politics, sociology, psychology, biology, but also pure mathematics. Other examples include: competition between firms, but also in auctions, in M&A projects, and in the market design of the auction of mobile spectrum licences; situations in sports such as strategies of goalkeeper and shooter in penalty kicks; political cooperation or coalitions and international institutions; formation and stability of formal and informal organisations; analysis of the emergence and course of political conflicts and wars; general negotiation situations with different objectives of the negotiating partners; child rearing and other social situations; and of course games such as chess, poker or rock, scissors, paper. A basic consideration of game theory - similar to that of ice-cream sellers - is to take into account the possible steps and strategies of all competitors when choosing one's own strategy. Strategy can only be successful if interdepencies with strategies of competitors are taken into account - in other words, one has to get into the thoughts and considerations of all competitors (Courtney et al. 2009). This is usually done informally in the context of so-called war games in management workshops. Increasingly, this is also done in the form of explicit gametheoretical models with the support of specialised management consultancies such as OpenOptions, TWS Partners or Frontier Economics and investment banks (Horn 2011). Google, on the other hand, continues to develop its own search algorithm in competition against other algorithms using evolutionary game theory (Tuyls et al. 2018). For an application of game theory from a theoretical perspective (von Neumann and Morgenstern 1944, and Fudenberg and Tirole 1991) it is necessary that: a clear objective function exists for all players and can be identified; all players act rationally and are mutually aware of the game-like situation; and information is available regarding: <?page no="329"?> 8.6 Summary and key learnings 329 identity and number of players - two or any larger number of participants; level of information of the players - perfect, incomplete, or asymmetric information; duration of the game - one-shot, repeated, or infinite games; structure of the game - simultaneous or sequential decisions and non-cooperative or cooperative search for solutions; and of all potential strategies - pure or mixed strategies and their variety. One cannot expect these conditions to be true for all players at all times in every real-life competitive situation or in all industries. However, similar to a game like rock, scissors, paper (played the same way all over the world), rules of games and sequences of behaviours also emerge in numerous industries that fulfil the above conditions well and can be analysed from a management perspective in terms of game theory (Birshan and Kar 2012 as well as Brandenburger and Nalebuff 1995). Also, due to the findings of bounded rationality (► Chapter 3), a new branch has emerged as behavioural game theory. Here, in particular, strategies based on routines and adherence to beliefs or previous patterns of success are analysed and examined within experiments. Economic laboratories such as VIRTECOLAB or MobLab try to identify patterns of deviations from the results to be expected with perfect rationality, which are highly relevant especially for decisions in firms (Mailath 1998, Camerer 2003, Bonau 2017). From a practical perspective, a list of players (Aldi, Lidl, ...), a description of possible strategies (prices, quality, advertising, ...), an estimate of the payoffs or profits if different strategies interact (if Aldi lowers prices and Lidl does not: what happens to the respective profits? ) and knowledge of the industry-specific rules of a game are often sufficient for an application of game theory. However, due to demanding data requirements, applications in real life settings remain limited to: large firms in oligopolies; strategically important decision-making situations. Yet, in these cases it allows: a plausibility check of possible strategies and their potential repercussions; a far-reaching increase in understanding a competitive situation and development and evaluation of industry-specific strategic competitive advantages, and often turns out to be very effective (Lindstädt and Müller 2009). Learning Objectives This chapter deals with basic ideas of game theory for analysing strategic decision situations; solutions for simultaneous games based on dominant strategies and best responses as Nash equilibria in pure and mixed strategies; reasons for stability and instability of collusion in repeated games; effects of risk aversion, as well as the identification of credible sub-game perfect firstmover strategies in sequential games. <?page no="330"?> 9 Strategic decision-making and game theory 330 9.1 Nash equilibria in simultaneous games Games per se can be described using mathematical equations for any number of strategies and players. However, basic ideas of simultaneous games for two players can also be analysed using the so-called strategic normal form, as shown in ► Figure 9.2. To establish some basic gametheoretical concepts, the well-known game of rock, scissors, and paper between two players is depicted in a matrix. Player 1 is the row player - the strategies rock, scissors or paper are arranged in three rows. Player 2 is the column player, hence strategies are arranged in three columns. Depending on the strategy chosen by both players, a strategy combination occurs, e.g., player 1: rock / player 2: paper - with such a strategy combination, player 1 receives zero points, player 2 earns one point. For all possible strategy combinations, a payoff matrix (in this case with nine possible cells) shows the respective payoffs for both players. Figure 9.2: Rock, scissors, and paper matrix as a game in normal form. The game rock, scissors and paper is a simultaneous game - that means players decide and show their respective strategies at the same time. If rock, scissors, and paper was a sequential game, first one of the players would show their strategy, then the other would react and decide - rock, scissors and paper is obviously easy to win as a sequential game. The more complicated solution in a simultaneous game using mixed strategies is explained in ► Section 9.2. In order to identify solutions for simultaneous games, the payoffs for each strategy combination (typically expressed as profits in static terms as in ► Figure 9.3 on the left or as discounted cash flow in dynamic terms) are determined and then compared in an appropriate way <?page no="331"?> 9.1 Nash equilibria in simultaneous games 331 to identify the optimum strategy depending on the strategy options of the opponent. If both players are able to draw identical versions of a game and mutually assume that the other player can do the same, we have symmetrical perfect information (common knowledge) concerning the game structure and payoff matrix, otherwise we call information asymmetric. Figure 9.3: Payoffs and strategies in normal form games. In order to analyse real-life competitive situations in games, considerable market research and data analysis effort is necessary, but in principle a simple procedure can be followed as outlined in ► Figure 9.3 on the right. First, relevant players are identified, in this case Aldi and Lidl, and the strategies that significantly influence profits (e.g., the number and size of shops in different regions). Second, payoffs (here the projected profits according to the previous P&L and further empirical analyses) are grouped in the respective strategy combinations. For example, profits of small Aldi shops in regions where Lidl operates large shops are 1.1 million EUR per year. In this way strategies can be evaluated, and business cases can be developed to identify optimum strategies. If players select one specific strategy from their possible strategies (from their so-called strategy set), we have pure strategies; if players combine strategies from their strategy set, we call this mixed strategies (► Section 9.2). Dominant strategies ► Figure 9.4, top left, describes a competitive situation between Coke and Pepsi in which both firms only have a choice between two strategies: marketing or no marketing. The payoff matrix shows the profits of both firms for all possible strategy combinations - for example, if both firms choose no advertising, Pepsi's profit is 10, Coke's profit is 2. Both firms can influence their <?page no="332"?> 9 Strategic decision-making and game theory 332 profits significantly through advertising expenditure, but individual profits depend on the strategy of the other player - which is the best strategy for Coke, which for Pepsi? Figure 9.4: Marketing versus no marketing for Coke and Pepsi. Pepsi is the row player - accordingly, in ► Figure 9.4, bottom left, Pepsi compares its own profits row by row depending on possible strategies chosen by Coke. If Coke advertises, the best strategy for Pepsi is also to advertise, because a profit of 12 in the strategy combination Pepsi: marketing / Coke: marketing exceeds the profit of 7 in the strategy combination Pepsi: no marketing / Coke: marketing. If Coke does no marketing, a best strategy for Pepsi is to advertise, because now the profit is 16 in the strategy combination Pepsi: marketing / Coke: no marketing compared to a profit of 10 in the strategy combination Pepsi: no marketing / Coke: no marketing. Regardless of what Coke does, it is obviously always better for Pepsi to advertise - Pepsi will, therefore, advertise in any case, as this strategy is clearly superior. For Pepsi, advertising is thus, a dominant strategy that will be chosen in any case, regardless of the strategic decision of the opponent - the dominant strategy advertising clearly rules out the alternative strategy <?page no="333"?> 9.1 Nash equilibria in simultaneous games 333 no advertising. Coke is the column player - accordingly, in ► Figure 9.4, top right, Coke compares the profits column-wise depending on Pepsi's strategies. The same applies to Coke: marketing is always the superior strategy, i.e., Coke also has a dominant strategy that is always chosen regardless of Pepsi's decision. In this example, marketing is obviously a dominant strategy for both firms. In this game (i.e., this competitive situation), both firms will always choose to advertise. Both players have no incentive to change their selected strategy: if one of the two players deviates from his dominant strategy, given that the other player maintains his dominant strategy, he is worse off. Pepsi's profit decreases from 12 to 7 when switching from marketing to no marketing, Coke's profit would decrease from 5 to 0. The constellation of two dominant strategies represents a stable strategy combination. The result is called an equilibrium in dominant strategies - it is a special case of a Nash equilibrium. Dominant strategies are, if they exist, the simplest gametheoretic strategy choice and solution for simultaneous games. Dominant strategies and best responses The situation between Unilever and L’Oreal is somewhat more complex, as described in ► Figure 9.5 concerning a new product launch. If we first look at the payoff matrix row by row from the perspective of L’Oreal, we see that L’Oreal does not have a dominant strategy. If Unilever were to launch the new product, L’Oreal would also launch the new product (profit of 10 larger than profit of 6) - if Unilever did not launch a new product, L’Oreal would not either (profit of 15 less than profit of 20). L’Oreal's optimum strategy thus, depends on Unilever's strategic choice. Figure 9.5: Dominant strategy and best response. However, to make a decision, L’Oreal can look at the game from Unilever's perspective to determine what its competitor will do. Unilever has a dominant strategy in the form of new pro- <?page no="334"?> 9 Strategic decision-making and game theory 334 duct introduction (which of course can also be identified for L’Oreal from the payoff matrix), so Unilever will choose this strategy in any case regardless of L’Oreal's decision. L’Oreal may align and based on this decide to choose product innovation itself as well. Such a strategic choice is called a best response because the best possible alternative is chosen depending on the decision of the opponent. The combination of a dominant strategy and a best response is again stable - neither player has an incentive to deviate from his strategy given the strategy of the opponent - and represents the second special case of a Nash equilibrium. Nash equilibrium in best responses The general case of a Nash equilibrium is based on mutual best responses of competitors (Nash 1950 and 1951). ► Figure 9.6 describes the competitive situation between Volkswagen and Toyota regarding the set-up of production capacity for a new factory in Mexico. Both firms can choose between a small, a medium and a large factory, the size of which afterwards cannot be adjusted in the short term. Figure 9.6: Mutual best responses and Nash equilibrium. <?page no="335"?> 9.1 Nash equilibria in simultaneous games 335 In ► Figure 9.6, top left, it can be seen that neither firm has a dominant strategy: neither a Toyota row nor a Volkswagen column is unambiguously preferable. In such a situation, each firm must now identify best responses to all conceivable strategies of the competitor. Volkswagen - at the top right of ► Figure 9.6 - identifies the following best responses to each of Toyota's possible factory sizes. If Toyota were to build a small factory, Volkswagen's best response is a medium-sized factory, because the profit of 125 exceeds the profit of either a small or a large factory at 105 each. If Toyota builds a medium-sized factory, Volkswagen's best response is also a mediumsized factory, because the profit of 100 exceeds the profit of a small factory of 85 or a large factory of 70. If Toyota builds a large factory, Volkswagen's best response is a small factory, because the profit of 50 exceeds the profit of a medium factory of 40 or a large factory of 0. Similarly, in the lower left of ► Figure 9.6, the situation can be looked at from Toyota's perspective as follows. If Volkswagen were to build a small factory, Toyota's best response is a medium-sized factory, because the profit of 125 exceeds the profit of a small factory or a large factory with 105 each. If Volkswagen builds a medium-sized factory, Toyota's best response is also a mediumsized factory, because the profit of 100 exceeds the profit of a small factory of 85 or a large factory of 70. If Volkswagen builds a large factory, Toyota's best response is a small factory, because the profit of 50 exceeds the profit of a medium factory of 40 or a large factory of 0. If we combine best responses of the two firms in the bottom right-hand corner of ► Figure 9.6, the intersection of best responses defines a Nash equilibrium: both firms will mutually decide in favour of a medium-sized factory based on the analysis of the game. In a Nash equilibrium, the best responses of all players meet - no player has an incentive to deviate from the chosen strategy, given the strategic choice of its opponents, because he would be worse off. Thus, a strategy that fits a Nash equilibrium is the most advantageous decision an individual rational player can make. The Nash equilibrium is neither limited by the number of players nor the number of possible strategies - however, there can be more than one or even no Nash equilibrium in pure strategies, and furthermore, not all players are or act completely rational. These cases will be considered in the following sections. Multiple Nash equilibria in best responses and necessary selection criteria The competitive situation between Roche and Novartis is shown in ► Figure 9.7. Both firms can decide on the level of R&D expenditure for a new drug. <?page no="336"?> 9 Strategic decision-making and game theory 336 Figure 9.7: Multiple Nash equilibria in best responses. First, it can be seen that neither firm has a dominant strategy, as neither a Novartis row nor a Roche column clearly produces the best results. However, both firms are able to identify best responses to each of their competitor's strategies. At all intersections where best responses meet, Nash equilibria are located. In this case there are two: in the strategy combinations Novartis: medium/ Roche: medium and Novartis: high/ Roche: low. From the perspective of these two firms, the two Nash equilibria differ: Novartis clearly has a preference for the Novartis: high/ Roche: low equilibrium, because here profits for Novartis are 120, whereas Roche prefers the Novartis: medium/ Roche: medium equilibrium, because here profits for Roche are 90. Both equilibria would be stable on their own, but it is unclear which one is chosen. In numerous games, multiple Nash equilibria result - without further assumptions or selection criteria (Harsanyi and Selten 1992 and Cooper et al. 1990), it is initially impossible to determine which of the possible equilibria comes about. Selection criteria that serve to select one of the possible equilibria or to coordinate strategies to one of the equilibria are for example: to continue the behaviour or chosen strategies of the past in the form of path dependencies or to choose certain 'usual strategies' (so called focal points); a generally accepted understanding of roles within an industry; all forms of explicit and implicit cooperation or signalling; and of course risk-averse evaluation of strategies or the choice of mixed strategies; adaptive behaviour or sequential decisions. However, despite the indeterminacy of the two Nash equilibria, the analysis yields significant insights for both firms: ► Figure 9.7 on the right shows that no Nash equilibria exist for the Novartis strategy low and for the Roche strategy high - the firms can completely exclude this row and column from their considerations and strategic planning. From a management perspective, a central aspect of game theory is that by identifying Nash equilibria it is possible to medium medium high low high Novartis Roche 80 125 80 90 75 60 50 20 60 100 90 100 90 50 120 75 30 70 multiple Nashequilibria in best responses medium medium high low high Novartis Roche 80 125 80 90 75 60 50 20 60 100 90 100 90 50 120 75 30 70 initial situation 1 2 low low <?page no="337"?> 9.1 Nash equilibria in simultaneous games 337 exclude conceivable strategies in principle, both one's own and those of competitors (Sutton 1990 and 1992, and Münter 1999), in order to develop robust strategies of one's own. Stability of cooperation and agreements In 1974, the then already successful firm Lego and its new competitor Playmobil, which was entering the toy market, planned to launch toy figures for children. The basic situation was somewhat as shown in ► Figure 9.8. If both firms offer figures of the same size, there is a high intensity of competition, which leads to losses of -5 in each case. If, on the other hand, both firms offer different figure sizes, each can achieve profits of 10 in its market segment based on reduced competitive intensity. There is no dominant strategy for either firm, but both firms can identify two Nash equilibria via mutually best responses: it is obviously advantageous for both firms to choose a strategy opposite to that of the competitor. Figure 9.8: Coordination and signalling. In order to choose a strategy in this context of a one-off decision - with a high investment requirement - that generates some profit, there are three possibilities per se: agreements (i.e., prohibited collusion), sequential decisions or signalling. In fact, both firms opted for signalling and had catalogues produced in the run-up to the 1974 toy fair: in this way, it was clear to both firms in advance what the competitor would do, and profits could be made. This implicit agreement has been stable ever since - both firms have never changed their strategy regarding the size of the toy figures for more than 40 years. The stability of this implicit agreement Lego: small / Playmobil: large is fundamentally based on the fact that this chosen strategy combination is a Nash equilibrium. Instability of agreements and the prisoner’s dilemma ► Figure 9.10 shows the competitive situation and the payoffs of each period of two neighbouring petrol stations Shell and Esso. This situation is described as prisoner's dilemma in the context of economic policy and sociological analyses, but it has a strategic component as well. small figures large figures large figures small figures Playmobil LEGO -5 -5 10 10 -5 -5 10 10 initial situation small figures small figures large figures large figures Playmobil LEGO -5 -5 10 10 -5 -5 10 10 multiple Nash-equilibria 1 2 <?page no="338"?> 9 Strategic decision-making and game theory 338 Both firms have two pricing strategies - high prices and low prices. Obviously, both firms make high profits of 25 if both charge high prices and low profits of 10 if both charge low prices. Analysing the game, it appears that both firms have a dominant strategy of low prices - however, there are only small profits to be made this way (► Figure 9.10 top right). Good to know │ Prisoner's dilemma and collective interest - why do all people hold up their smartphones at concerts? The prisoner's dilemma describes a situation in which two suspects are held separately in custody without the possibility to communicate. Both are accused of having committed a major crime together - however, this crime cannot be proven without a confession. Both have agreed in advance not to confess if they get arrested. In this case, both can only be proven to have committed minor offences. ► Figure 9.9 shows this situation, known to both suspects. The negative payoffs denote the prison sentence in years. If only one player confesses, his sentence is reduced due to the leniency programme, the other player has to go to prison for longer. Figure 9.9: Prisoner's dilemma. In fact, the agreement of these two prisoners is not a stable arrangement: both players have an incentive to switch from the strategy 'do not confess' to 'confess' when questioned - because this allows each player to better himself individually. The prison sentence <?page no="339"?> 9.1 Nash equilibria in simultaneous games 339 is supposedly reduced from three years to one year. However, since this consideration applies to both prisoners in case of rational players, both will switch to 'confess' and thus, remain in prison for six years each. The individually rational strategy combination ‘confess/ confess’ is a Nash equilibrium in dominant strategies in this one-shot game and deters a cooperative agreement ‘not to confess/ not to confess’. The individual interest thus, dominates the collective interest and makes both players worse off. Numerous social and economic situations have structurally similar payoff matrices as shown in ► Figure 9.9: this explains why collectively preferred situations (limiting climate change, reducing resource consumption, stopping overfishing, nuclear disarmament, etc.) often do not come about because individual interests are opposed. In the same way, all people at parties speak too loudly (because some do not want to talk quietly), too many people at concerts hold their smartphones into the visual field of others, and all people on aircrafts jump off their seats at the same time after landing. However, laboratory experiments on the prisoner's dilemma also show that some people have a systematic bias towards cooperative behaviour - albeit to varying degrees. Students in lower semesters appear more willing to cooperate than students in higher semesters, but the willingness to cooperate increases if someone knows teammates or opponents, especially if the duration of the game is changed from a one-shot to a repeated game and trust can develop (Rapoport et al. 1965, Axelrod 1980, Frank et al. 1993 and Brosig 2002). If both petrol stations agree on high prices in a one-shot game, this agreement does not hold: in ► Figure 9.10, bottom left, it can be seen that both firms have an incentive to deviate from this agreement, since profits can be increased from 15 to 20 by lowering prices if the other petrol station maintains high prices. These in principle collectively preferred high prices will not last due to dominant individual interests - the collusion is unstable due to the Nash equilibrium in another strategy combination. This applies to both, one-shot games and to finitely repeated games: if competitors know a final last round of a game in which an incentive to deviate from an agreement arises, they will deviate from a previously existing cooperation in this final round of the game. However, since both players foresee this with perfect information and rationality, both will already deviate in the penultimate round - this iterative backward induction continues accordingly, so that in a finite game, stable cooperation cannot be achieved even in a first round of the game. In order to establish stability of an agreement or cooperation, there are three options in game theory: an infinite game in which there is no final round of the game - this way none of the players has an incentive to deviate in a last round; binding agreements with the possibility of sanctions to ensure compliance with the cooperation - which, however, are typically prohibited by competition policy; or a building of trust concerning a cooperative behaviour of the competitor. <?page no="340"?> 9 Strategic decision-making and game theory 340 Figure 9.10: Instability of agreements. Especially the last option of building trust plays a crucial role. If in a repeated game one of the players deviates from the Nash equilibrium in a way that is visible and understandable to the other, and intentionally puts himself in a worse position, he offers the other player a temporarily higher profit, but in particular the possibility of cooperation (see also ► Chapter 2 on fairness in the ultimatum game). If the other player understands this offer and adapts to that behaviour in the next round of the game, and this trust is not mutually disappointed, cooperation can also be possible in a finite game. This solution of reciprocal behaviour is often observed as a robust strategy in algorithm-based computer simulations and in laboratory experiments with humans and is referred to as tit for tat (Axelrod and Hamilton 1981) and also finds widespread empirical confirmation in biology and politics. <?page no="341"?> 9.1 Nash equilibria in simultaneous games 341 Bounded rationality and the achievement of a Nash equilibrium A Nash equilibrium seems plausible from a strategic and logical perspective but is really challenging regarding the rationality of the players, especially their cognitive abilities. To illustrate this, one can refer to the so-called guessing numbers game (Ledoux 1981, Nagel 1995 and Duffy and Nagel 1997). In this game, the following rules apply to any number of players. Each player can choose an integer number from 0 to 100 as a strategy, so that each player can choose from 101 possible strategies. All strategies are written down secretly and simultaneously, added up and the rounded integer average is calculated. The winner is the player whose strategy is closest to 2/ 3 of this average value. If, for example, the average value of all the strategies mentioned by the players is 45.3, then the winning strategy is the one that is closest to 2/ 3 ⋅ 45.3 = 30 . From a game-theoretic perspective, the game can be solved quiet simply as follows. First, all strategies larger than 66 are ruled out by weak dominance - because even if all players took the number 100, due to 2/ 3 ⋅ 100 = 66.67 the winning strategy is at most 66 - no rational player would choose a number larger than 66. If one were to assume that all players randomly choose any number from 0 to 100, then the mean is 50 and, because 2/ 3 ⋅ 50 = 33.33, the winning strategy would be 33. If - with perfect rationality and foresight of the other players' actions - all players chose the outcome of 33, however, because of 2/ 3 ⋅ 33 = 22 the winning strategy would be 22. This reasoning continues according to iterative elimination of possible strategies until finally all players choose 0 - which is in fact the only Nash equilibrium in this game - and all players also win. In fact, in experiments on the guessing numbers game, almost regardless of the education and age of the participants or prior knowledge of the game, players with strategies between 15 to 25 usually win this game. The reason for this is twofold. On the one hand, some players are not able to recognise the solution to the inherently simple game. On the other hand, clever players recognise the limited cognitive abilities and the insufficient depth of reasoning of the problem of the fellow players - as a consequence, even a fully rational player will then not choose 0, because he knows that he cannot win with 0 due to the "wrong strategies" of the boundedly rational players. Instead, he will make an assumption about the distribution of the degree of rationality of the fellow players and choose his strategy accordingly. This game shows in particular the difference between perfect rationality of a player and the collectively restricted rationality due to the possibly boundedly rational behaviour of one individual or all players (Bosch-Domenech et al. 2002 and Güth et al. 2002). Against this background, from a theoretical point of view, it is understandable why a relatively long and slow convergence to an immediately adaptable Nash equilibrium is often observed in laboratory experiments. It is clear that in empirical studies of real markets and competitive situations - which have a significantly higher complexity and less clear rules than the guessing numbers game - often no stable Nash equilibria can be identified (Aiginger 1998). Thus, from a managerial perspective, although the Nash equilibrium serves as a possible focal point in <?page no="342"?> 9 Strategic decision-making and game theory 342 strategic competition, an achievement of equilibria and a convergence of strategies depends on the degree of rationality and the speed of learning of competitors (Holt 1993, Mailath 1998, Foss 2001, Armstrong and Huck 2010, and Crawford 2013). However, formal game-theoretic training and experience in strategic decision-making situations has been shown to significantly increase the likelihood of making decisions consistent with Nash equilibria (Brandenburger and Nalebuff 1995 and Camerer 2003). 9.2 Risk aversion and mixed strategies In the previous section it became clear that games can have more than one Nash equilibrium. In order to create decidability from a management perspective, it is possible to apply extended concepts based on selection criteria. Figure 9.11: Nash equilibrium and risk aversion. Risk aversion and maximin-strategy In many competitive situations, not all parameters can be captured; moreover, doubts about the complete rationality of an opponent may be appropriate. All these factors contribute to the fact that one or all players make risk-averse decisions - the aim is then no longer to achieve the best possible result but to minimise possible losses (see also ► Chapter 3). ► Figure 9.11 on the left shows a competitive setting between SAP and Oracle based on an investment for a new database system. SAP has a dominant strategy in invest, Oracle has no dominant strategy but can determine invest as a best response, resulting in a Nash equilibrium with Oracle: invest / SAP: invest. <?page no="343"?> 9.2 Risk aversion and mixed strategies 343 The strategy consistent with this Nash equilibrium leads to a profit of 20 for Oracle. However, if SAP - for some reason - does not invest, then Oracle faces a considerable loss of -200. Against this background, Oracle, with justified risk aversion, can now choose a cautious strategy: Oracle determines the worst outcome for each conceivable strategy - and then selects the best from these worst outcomes, i.e., the "lesser evil". This strategy in case of risk aversion is called a maximin strategy: a risk-averse player looks at the minima of his strategies and selects the maximum from these. In ► Figure 9.11 on the right, this maximin strategy is applied for both players: Oracle chooses the strategy not to invest based on selecting the maximum out of the minima -20 and -200 of the rows, SAP accordingly continues to choose the strategy invest in the comparison of 0 and 10. Now, a new solution emerges: Oracle: do not invest/ SAP: invest. Although Oracle now makes a small loss, this is preferred to the potential large loss due to risk aversion - the difference between 20 and -20 can be understood as a risk premium to prevent a possible loss of -200. As board members are generally more risk-averse due to their contract duration and bonus arrangements, such a strategy is not uncommon (Ross 2004 and Bolton et al. 2015). In case of SAP, the strategy does not change - the reason is that risk aversion never changes a dominant strategy. In the second example in ► Figure 9.12, the strategic importance of risk aversion compared to a strategy based on a Nash equilibrium becomes clear: without risk aversion, GSK and Pfizer each invest a high R&D budget; in the Nash equilibrium, Pfizer's profit clearly exceeds GSK's profit. Figure 9.12: Strategic use of risk aversion. What happens if GSK signals to the capital market via some ad hoc announcement - justified and credible from Pfizer’s point of view - that there is massive uncertainty in the markets due to some economic or political situation and both firms then act in a risk-averse manner and <?page no="344"?> 9 Strategic decision-making and game theory 344 apply the maximin rule? The profits will shift in favour of GSK: GSK accordingly has a high strategic interest in risk-averse behaviour. Moreover, as can be seen in ► Figure 9.13, risk aversion can also increase the profits of both firms in absolute terms. In this compete situation between Kellogg’s and Nestlé in breakfast cereals, there is obviously no Nash equilibrium. No matter which starting point is chosen in ► Figure 9.13 on the left, one of the competitors always has an incentive to deviate from the current strategy in order to be in a better position. As a result, the four cells of the matrix are repeatedly passed through anti-clockwise. Figure 9.13: Missing Nash equilibrium and risk aversion. If both firms apply the maximin rule, Kellogg’s strategy will be crispy, Nestlé's strategy is sweet - both firms make a profit of 3 in this case. Moreover, both firms with risk-averse behaviour on the right actually do better than in the situation on the left: when passing through the four matrix cells, the average profit per period for Kellogg’s is (1 + 3 + 2 + 5)/ 4 = 2.75, Nestlé's average profit is (2 + 3 + 4 + 1)/ 4 = 2.5 . Mixed strategies and random decisions The analysis of games between Lego and Playmobil (► Figure 9.8) and competition between Kellogg’s and Nestlé (► Figure 9.13) has shown that there may be several or no Nash equilibrium in pure strategies. A possible solution lies in mixed strategies (Nash 1951) - a player does not decide on a single strategy but chooses a strategy at random based on a probability distribution. Nash equilibria in mixed strategies are thus, a generalisation of Nash equilibria in pure strategies. There are two possibilities. Firstly, if there are no Nash equilibria in pure strategies, then the probability distribution is developed over all available strategies. Secondly, if there are <?page no="345"?> 9.2 Risk aversion and mixed strategies 345 several Nash equilibria, the probability distribution takes into account only those pure strategies that lead to these Nash equilibria. A game-theoretical problem in this context that meanwhile has been extensively studied empirically is penalty kicks in football. ► Figure 9.14 shows the situation between goalkeeper and penalty taker at the penalty kick in a very simplified version: both decide absolutely simultaneously, there are only two strategies right or left. In this example the goalkeeper definitely catches the ball if he is in the correct corner, the penalty taker never shoots at the crossbar or post or even misses. Obviously, none of the opponents has a dominant strategy, but the analysis of the best responses does not lead to a solution either: in no cell of the matrix do best responses meet, so that there is no Nash equilibrium in pure strategies. Figure 9.14: Goalkeeper and penalty taker taking a penalty. However, each of these two players can now randomly choose one of their strategies: the distribution of strategies then represents a mixed strategy across pure strategies. A feasible probability distribution to be chosen in such situations must satisfy three simple conditions: probabilities across all strategies must correspond to their relative probability of success; random strategies chosen must not show any discernible pattern; and the sum of probabilities must add up to one. If, instead of simple probabilities in two possibilities as shown in ► Figure 9.14, more detailed distributions based on previous behaviour of the opponents are known, penalty taker and goalkeeper must take into account the current opponent's strategies weighted with empirical probabilities accordingly. Then they have to decide 'at random' on this basis - of course taking into account that the goalkeeper is subject to a psychological action bias of jumping away from the centre (Chiappori et al. 2002, Bar-Eli et al. 2007. Azar and Bar-Eli 2011 and ► Chapter 3). In order to develop a probability distribution across strategies for a penalty taker, the penalty taker needs to determine an expected value about the goalkeeper's strategy - if the <?page no="346"?> 9 Strategic decision-making and game theory 346 penalty taker knew with certainty where the goalkeeper would jump, there would be no need for a mixed strategy. In the absence of Nash equilibria, all conceivable strategies must be considered for this probability distribution. The expected value depends on the probabilities: 𝑝𝑝𝑟𝑟 𝑇𝑇𝐿𝐿 (the goalkeeper jumps to the left) and 1 − 𝑝𝑝𝑟𝑟 𝑇𝑇𝐿𝐿 (the goalkeeper jumps to the right) describes the probability distribution of the goalkeeper's behaviour - both probabilities add up to 100% because of 𝑝𝑝𝑟𝑟 𝑇𝑇𝐿𝐿 + 1 − 𝑝𝑝𝑟𝑟 𝑇𝑇𝐿𝐿 = 1 . Accordingly, 𝑝𝑝𝑟𝑟 𝑆𝑆𝐿𝐿 (the shooter shoots to the left) and 1 − 𝑝𝑝𝑟𝑟 𝑆𝑆𝐿𝐿 (the shooter shoots to the right) describes the probability distribution of the penalty taker (see ► Figure 9.14 left). The expected value 𝐸𝐸 𝑆𝑆𝐿𝐿 of the utility for the shooter if he shoots to the left is (9.1) 𝐸𝐸 𝑆𝑆𝐿𝐿 = 0 ⋅ 𝑝𝑝𝑟𝑟 𝑇𝑇𝐿𝐿 + 1 ⋅ (1 − 𝑝𝑝𝑟𝑟 𝑇𝑇𝐿𝐿 ) = 1 − 𝑝𝑝𝑟𝑟 𝑇𝑇𝐿𝐿 , the expected value 𝐸𝐸 𝑆𝑆𝜕𝜕 for the shot to the right results as (9.2) 𝐸𝐸 𝑆𝑆𝜕𝜕 = 1 ⋅ 𝑝𝑝𝑟𝑟 𝑇𝑇𝐿𝐿 + 0 ⋅ (1 − 𝑝𝑝𝑟𝑟 𝑇𝑇𝐿𝐿 ) = 𝑝𝑝𝑟𝑟 𝑇𝑇𝐿𝐿 . For the goalkeeper, if he is jumping to the left, results as (9.3) 𝐸𝐸 𝑇𝑇𝐿𝐿 = 1 ⋅ 𝑝𝑝𝑟𝑟 𝑆𝑆𝐿𝐿 + 0 ⋅ (1 − 𝑝𝑝𝑟𝑟 𝑆𝑆𝐿𝐿 ) = 𝑝𝑝𝑟𝑟 𝑆𝑆𝐿𝐿 , as the expected value if jumping to the left and (9.4) 𝐸𝐸 𝑇𝑇𝜕𝜕 = 0 ⋅ 𝑝𝑝𝑟𝑟 𝑆𝑆𝐿𝐿 + 1 ⋅ (1 − 𝑝𝑝𝑟𝑟 𝑆𝑆𝐿𝐿 ) = 1 − 𝑝𝑝𝑟𝑟 𝑆𝑆𝐿𝐿 , when jumping to the right. A strategy can be determined randomly if expected values of both possible strategies are equal - in other words, the goalkeeper and the shooter take into account each other's probabilities of the opponent's strategies and then decide (similar to best responses) on an optimum probability distribution, which is now a Nash equilibrium in mixed strategies. Thus, for the penalty taker's decision, one sets equations (9.1) and (9.2) equal (9.5) 1 − 𝑝𝑝𝑟𝑟 𝑇𝑇𝐿𝐿 = 𝑝𝑝𝑟𝑟 𝑇𝑇𝐿𝐿 , then 𝑝𝑝𝑟𝑟 𝑇𝑇𝐿𝐿 = 12 and 1 − 𝑝𝑝𝑟𝑟 𝑇𝑇𝐿𝐿 = 12 yield the probability distribution for the goalkeeper's behaviour - he jumps to the left with a probability of 50%, and to the right with a probability of also 50%. Analogously, by equating (9.3) and (9.4) with (9.6) 1 − 𝑝𝑝𝑟𝑟 𝑆𝑆𝐿𝐿 = 𝑝𝑝𝑟𝑟 𝑆𝑆𝐿𝐿 , the penalty taker also shoots to the left or right with probabilities 𝑝𝑝𝑟𝑟 𝑆𝑆𝐿𝐿 = 12 and 1 − 𝑝𝑝𝑟𝑟 𝑆𝑆𝜕𝜕 = 12 . If the goalkeeper and penalty taker decide according to these probability distributions, then the ► Figure 9.14 combined probabilities plotted on the right will occur. Each combination of strategies will occur with a frequency of 25%. If the players do not stick to the calculated probability distributions, they worsen their own odds and improve the odds of their opponent: for example, if the penalty taker shoots to the left with too high a probability of 70%, then the goalkeeper should always jump to the left and now catches 70% of all penalties. Mixed strategies are also a feasible solution concept in coordination games. Coordination games are characterised by several Nash equilibria that are asymmetrically preferred by the players: there are obviously conflicts of interest. Such games are often grouped under the header battle of the sexes. ► Figure 9.15 on the left shows such a game: a couple has fallen out over weekend activities - she really wants to go to a rugby match in the evening, he wants <?page no="347"?> 9.2 Risk aversion and mixed strategies 347 to go to the ballet at the same time. Both nevertheless have a preference to spend the evening together with their partner, i.e., the two Nash equilibria lie in strategy combinations ballet/ ballet and rugby/ rugby, but further direct communication or indirect signalling for coordination is impossible. Figure 9.15: Battle of the sexes. Now, in order to determine a probability distribution across their own strategies of ballet and rugby, the husband and wife must mutually determine their expected values of the payoff depending on the partner's behaviour. In addition to these possible strategies and the payoffs, the expected value again depends on the probabilities: 𝑝𝑝𝑟𝑟 𝐹𝐹𝐵𝐵 (the woman goes to ballet) and 1 − 𝑝𝑝𝑟𝑟 𝐹𝐹𝐵𝐵 (the woman goes to the rugby match) describes the probability distribution of the woman's behaviour, 𝑝𝑝𝑟𝑟 𝑀𝑀𝐵𝐵 (the man goes to ballet) and 1 − 𝑝𝑝𝑟𝑟 𝑀𝑀𝐵𝐵 (the man goes to watch rugby) describes the probability distribution of the man's behaviour. Thus, the expected value 𝑢𝑢 𝐹𝐹𝐵𝐵 of the utility for the woman if she goes to the ballet is given as (9.7) 𝑢𝑢 𝐹𝐹𝐵𝐵 = 2 ⋅ 𝑝𝑝𝑟𝑟 𝑀𝑀𝐵𝐵 + 0 ⋅ (1 − 𝑝𝑝𝑟𝑟 𝑀𝑀𝐵𝐵 ) = 2 ⋅ 𝑝𝑝𝑟𝑟 𝑀𝑀𝐵𝐵 , according to her utility as a function of the probability distribution of the man's behaviour. The expected value 𝑢𝑢 𝐹𝐹𝜕𝜕 when attending the rugby game is given as (9.8) 𝑢𝑢 𝐹𝐹𝜕𝜕 = 1 ⋅ 𝑝𝑝𝑟𝑟 𝑀𝑀𝐵𝐵 + 5 ⋅ (1 − 𝑝𝑝𝑟𝑟 𝑀𝑀𝐵𝐵 ) = 5 − 4 ⋅ 𝑝𝑝𝑟𝑟 𝑀𝑀𝐵𝐵 . For the man, the following results accordingly (9.9) 𝑢𝑢 𝑀𝑀𝐵𝐵 = 4 ⋅ 𝑝𝑝𝑟𝑟 𝐹𝐹𝐵𝐵 + 1 ⋅ (1 − 𝑝𝑝𝑟𝑟 𝐹𝐹𝐵𝐵 ) = 1 + 3 ⋅ 𝑝𝑝𝑟𝑟 𝐹𝐹𝐵𝐵 and (9.10) 𝑢𝑢 𝑀𝑀𝜕𝜕 = 0 ⋅ 𝑝𝑝𝑟𝑟 𝐹𝐹𝐵𝐵 + 2 ⋅ (1 − 𝑝𝑝𝑟𝑟 𝐹𝐹𝐵𝐵 ) = 2 − 2 ⋅ 𝑝𝑝𝑟𝑟 𝐹𝐹𝐵𝐵 . The choice of a strategy depends solely on chance if the utility of both possible strategies is equal - i.e., a player is indifferent concerning the choice of strategy. If one sets equations (9.7) and (9.8) for the woman's decision equal with (9.11) 2 ⋅ 𝑝𝑝𝑟𝑟 𝑀𝑀𝐵𝐵 = 5 − 4 ⋅ 𝑝𝑝𝑟𝑟 𝑀𝑀𝐵𝐵 rugby ballet man woman 0 0 2 4 1 1 5 2 ballet rugby rugby ballet man woman 0 0 2 4 1 1 5 2 ballet 1/ 30 4/ 30 5/ 30 20/ 30 pr FB =1/ 5 pr MB =5/ 6 (1-pr FR ) =4/ 5 rugby (1-pr MR ) =1/ 6 <?page no="348"?> 9 Strategic decision-making and game theory 348 then 𝑝𝑝𝑟𝑟 𝑀𝑀𝐵𝐵 = 5/ 6 and 1 − 𝑝𝑝𝑟𝑟 𝑀𝑀𝐵𝐵 = 1/ 6 yield the probability distribution for the man's behaviour. Similarly, by equating (9.9) and (9.10) as (9.12) 1 + 3 ⋅ 𝑝𝑝𝑟𝑟 𝐹𝐹𝐵𝐵 = 2 − 2 ⋅ 𝑝𝑝𝑟𝑟 𝐹𝐹𝐵𝐵 , the woman will go to the ballet with a probability 𝑝𝑝𝑟𝑟 𝐹𝐹𝐵𝐵 = 1/ 5, but will attend the rugby match with a significantly higher probability 1 − 𝑝𝑝𝑟𝑟 𝐹𝐹𝐵𝐵 = 4/ 5 . If both players now decide according to their probability distributions, then combined probabilities given on the right will occur, as shown in ► Figure 9.15: with a probability 𝑝𝑝𝑟𝑟 𝑀𝑀𝐵𝐵 ∙ 𝑝𝑝𝑟𝑟 𝐹𝐹𝐵𝐵 = 5/ 30 they both go to the ballet together, with probability 𝑝𝑝𝑟𝑟 𝑀𝑀𝜕𝜕 ∙ 𝑝𝑝𝑟𝑟 𝐹𝐹𝜕𝜕 = 4/ 30 they both go to the rugby game. However, with 𝑝𝑝𝑟𝑟 𝑀𝑀𝐵𝐵 ∙ 𝑝𝑝𝑟𝑟 𝐹𝐹𝜕𝜕 = 20/ 30, the probability that both spend the evening separately (but at least according to individual preferences) is very high. In contrast, the probability that both spend the evening separately and against their own preferences is relatively low with 𝑝𝑝𝑟𝑟 𝑀𝑀𝜕𝜕 ∙ 𝑝𝑝𝑟𝑟 𝐹𝐹𝐵𝐵 = 1/ 30 . In games based on mixed strategies, a basic requirement is that any choice of strategy is indeed random - to make this clear, just look back at rock, scissors, paper: here, too, the theoretical optimum procedure is to use rock, scissors and paper with a probability of one-third each, based on a mixed strategy, but of course not in this order and alternating. If the opponent is able to recognise a pattern and adjust to it, he or she increases the odds of winning this game. Actually, on the New York Times website (nytimes.com/ interactive/ science/ rock-paperscissors), for example, you can play rock, scissors, and paper against a fair but adaptive algorithm. A human player only has a chance here if he actually decides completely random, i.e., neither considers the current own nor the opponent's strategy nor their history to make a decision. This result seems to limit the possible applications of game theory, but the opposite is true: from the analysis of a game’s structure, randomness can be derived as an optimum strategy for games without Nash equilibria in pure strategies, in which mixed strategies are then chosen. In fact, human players are not very good at applying mixed strategies and making purely random decisions for three main reasons. Firstly, people tend to ascribe meaning and causes to completely random events - such as the strategies of opponents. Secondly, the human brain is not well suited to deal with probabilities. Thirdly, from a management perspective, supposedly complex decisions in firms always require justifications and explanations, even if success or failure are purely random - so that a non-random strategy is unintentionally pursued from then on (Tversky and Kahneman 1974, Kahneman et al. 2011, Allred 2016 and Sibony et al. 2017). However, this already applies to a trivial game such as rock, scissors, and paper: here, too, people are not able to decide purely by chance, but psychological effects and the supposed or actual discovery and understanding of patterns govern decision-making (Wang et al. 2014). 9.3 Sequential decisions and commitment In all the situations considered so far, the players have made decisions simultaneously. In reality, this does not necessarily mean at the very same second, but it means that competitors' decisions develop and become concrete simultaneously over time, so that in fact there is no delayed reaction to a competitor's decision: waiting for a decision would possibly result in strategic <?page no="349"?> 9.3 Sequential decisions and commitment 349 competitive disadvantage. However, there are markets and industries in which competitive advantages can be achieved through sequential decisions - both through the possibility of being the first to decide (first mover advantage) and through the possibility of waiting (second mover advantage) (Kopel and Löffler 2008). The strategic normal form used so far can be used well for one-shot, simultaneous, or static games - but it is impractical for analysing sequential decision-making situations. The extensive form makes it possible to describe and analyse decisions along branches in which players decide once in a sequence or repeatedly in sequences. A normal form game can be transformed into an extensive form game as shown in ► Figure 9.16 - player 1, for example, decides first on his strategy A or B. Player 2 observes this decision and then reacts with one of her strategies C or D in the best possible way. The possible profits of both players depend on the sequence of all decisions made. Figure 9.16: Strategic normal form versus extensive form. In a sequential game, one firm can make a first move. Reasons for this may include: a leadership role in the past; the control of certain competitive parameters (based on market power); or firm size and a dominant position in a market (see also ► Chapter 10 on Stackelberg market leadership). However, a firm can also actively seek strategic leadership if it expects to gain an advantage from it and is able to increase its profits. Sequential games are therefore, solved by iterative backward induction - starting from the desired final result, one checks whether this is actually achievable for the player. In Section 9.2, ► Figure 9.6 shows that in a simultaneous competitive situation between Volkswagen and Toyota, a Nash equilibrium based on best responses exists if both firms build a medium-sized factory - both firms achieve profits of 100. However, the payoff matrix in ► Figure 9.6 shows that profits of 105 or even 125 could be possible. <?page no="350"?> 9 Strategic decision-making and game theory 350 If Toyota now analyses this situation as a sequential game to see if these higher profits are achievable, the extensive form is as shown in ► Figure 9.17. Toyota can first decide on the size of the factory (s=small, m=medium, l=large). Volkswagen can then respond with similar strategies. In fact, there are three possible outcomes for Toyota, indicated by arrows in ► Figure 9.17, which might increase profits compared to the simultaneous game as follows. Path 1: Toyota starts with a small factory - profits will increase to 105 if Volkswagen responds with a small factory. Path 2: Toyota starts with a medium factory - profits will increase to 125 if Volkswagen responds with a small factory. Path 3: Toyota starts with a large factory - profits will increase to 105 if Volkswagen responds with a small factory. Figure 9.17: Competition between Toyota and Volkswagen as a sequential game. To solve this sequential game, consider the possible reactions of Volkswagen to this first move by Toyota. In fact, path 1 and path 2 will not come about in the way Toyota intends: if Toyota starts with a small factory, Volkswagen will react by comparing its own possible profits with a medium-sized factory - so Volkswagen increases profits, but not Toyota. Along path 2, Volkswagen will also not react as Toyota intends: Volkswagen will also build a medium-sized factory in response to a medium-sized Toyota factory and realise the Nash equilibrium known from the simultaneous game. However, if Toyota starts with a large factory, Volkswagen's best response is actually to choose a small factory - here, from Volkswagen's perspective, the profits at 50 are larger than with a response of a medium or large factory. This creates a first mover advantage for Toyota - this first move allows Toyota to increase profits from 100 in a simultaneous game to 105 in a sequential game, because Volkswagen is forced to change its strategy compared to the simultaneous game. <?page no="351"?> 9.3 Sequential decisions and commitment 351 ► Figure 9.18 shows this result. Such an equilibrium is called a sub-game perfect equilibrium because the second-level decisions are consistent (sequentially rational) with the first-level decisions - the first-level player anticipates the second-level player's decision with perfect information (i.e., the second-level player observes and knows the first-level player's decision) and perfect rationality. Figure 9.18: Toyota's sub-game perfect first mover strategy. Strategic barriers to entry and credible threats The analysis of sequential decisions is particularly relevant if one firm is already an incumbent in a market and another firm wants to enter this market. A sequential game can then be used to analyse the possibilities and plausibility of strategic barriers to entry. Tesla is massively trying to enter the German automobile market, while incumbent manufacturers such as BMW are trying to deter them (Süddeutsche Zeitung, 30 June 2017). A strategy often claimed by incumbents in such situations is to expand existing capacity - thus, producing more - in the event of a new competitor entering the market in order to limit the potential market share of the new competitor. In addition, as a consequence of an expansion of supply, prices are lowered so that profits can only be made if a firm realises economies of scale. This is also intended to deter possible new competitors who do not have significant economies of scale when entering the market due to their small size. ► Figure 9.19 shows such a situation schematically as a sequential game. In the initial situation, before Tesla enters the market, BMW apparently incurs a profit of 100; but if Tesla enters the market, BMW’s profits fall to 40 due to the higher intensity of competition. Entry barriers work if an incumbent can prevent a potential entrant from entering because its entry would be unprofitable. In ► Figure 9.19, however, it can be seen that such an entry barrier by BMW is not credible: the incumbent would obviously want to maintain high prices (and not expand volume) even if Tesla entered the market, since its own profits of 40 from maintaining volume exceed the profits of 20 from expanding production. Tesla, on the other hand, would <?page no="352"?> 9 Strategic decision-making and game theory 352 enter the market in any case, since its profits are always positive when entering the market with 20 or 40. In this case, BMW’s announcement is not a credible threat - BMW will not stick to its own announcement - and thus, not effectively prevent Tesla's market entry. Figure 9.19: Incredible strategic barriers to entry - BMW versus Tesla. One way to create credibility to an announcement or threat for a firm is to commit (formally and/ or costly) to some binding behaviour and strategy, i.e., to make a commitment. BMW could do this, for example, by voluntarily investing sunk costs SC that are also relevant for Tesla due to legislation or technological requirements (also called raising rivals' costs strategy). Currently, incumbent German manufacturers are trying to do exactly that: the payoff matrix of the competition for electro mobility is to be changed by costs and conditions for charging infrastructure, or compliance with certain technological standards, in such a way that a market entry of Tesla (or possibly other new firms) is prevented or significantly delayed. Figure 9.20: Credible strategic barriers to entry - BMW versus Tesla. <?page no="353"?> 9.3 Sequential decisions and commitment 353 Whether this works depends on the level of invested sunk costs SC, as can be seen in ► Figure 9.20. With the given payoffs, voluntary sunk costs in a range of 40 < 𝑆𝑆𝐶𝐶 < 60 can effectively deter Tesla from entering the market and protect BMW's profits as follows. If 𝑆𝑆𝐶𝐶 < 40, Tesla will still enter as profits can be incurred - BMW's profits are also 40 − 𝑆𝑆𝐶𝐶 in this case. If 𝑆𝑆𝐶𝐶 > 60, BMW would be worse off compared to a situation when Tesla enters the market - because with sunk costs SC of more than 60. BMW's profits fall below the level of profits that would have been possible if Tesla had entered the market without sunk costs. However, if SC are for example 50, then entry would be credibly and effectively blocked - BMW's profit would still be 100 − 𝑆𝑆𝐶𝐶 = 50 and still larger than in case of Tesla's entry. From a management perspective, three aspects are central: investments in form of sunk costs can create credibility to strategies such as barriers to entry; an upper limit for sunk costs results from a difference between profits achievable with market entry of a competitor and the profit without market entry minus the sunk costs; and this investment must be logical and observable for the competitor. This form of sunk costs also exists if a firm strategically invests in overcapacity that can be used at any time without further investment if potential market entry is imminent. Although such overcapacity initially requires higher costs, it can assure a higher profit level in the long-term. From a management perspective, this is often not something that can be discussed amicably between the strategy department (in favour of spending money on sunk costs) and the finance department (typically against any unnecessary investment), however, it is used in airlines through code-sharing alliances such as OneWorld or Staralliance, in container shipping, in the automobile industry via platform partnerships and in the steel industry (Dixit 1980, Milgrom and Roberts 1982 and Schuler et al. 2014). Chicken game and credible strategies Numerous game and competitive situations have very memorable names due to their references to everyday situations, literature, or movies: for example, the payoff structure of the chicken game refers to situations in which assertiveness or superiority is to be demonstrated through a test of courage. However, the structure of this game also allows for far-reaching insights relevant for competitive strategy. ► Figure 9.21 on the left shows the payoff matrix of two drivers, Redcar and Bluecar, who are racing towards each other at high speed with their brakes removed, similar to the situation in the movie 'Rebel without a Cause'. The payoff matrix of the chicken game is not very attractive - if both go straight, the game ends fatally, if both swerve, both are fainthearted chicken. The game has no dominant strategy, and because of the potentially fatal outcome, it does not lend itself to a mixed strategy to realise one of the two Nash equilibria. One way to bring this game to a positive outcome (at least with a relatively high probability) and to win is to throw one's own steering wheel out of the car, of course visible to the opponent. Thus, in ► Figure 9.21 on the right, Redcar eliminates a possible strategy of his own <?page no="354"?> 9 Strategic decision-making and game theory 354 (the bottom row in the form of evading). But now, he has credibly signalled that he will drive straight ahead - a rational player Blue Car will now evade in any case, because in the simplified matrix in ► Figure 9.21 on the right he has a dominant strategy in the form of evading. At the same time, however, the player Redcar - similar to BMW before - has changed the structure of the game: the simultaneous game has become a sequential game by throwing out the steering wheel, and the payoff matrix has also changed. Figure 9.21: Chicken game - initial situation and eliminating one's own strategy. From a management perspective, it is important to understand that it does not always make sense to have additional strategic options: there may very well be competitive situations in which a firm consciously decides to change the structure of the game and reduce the number of possible strategies in order to force competitors into a situation comparable to the chicken game. By eliminating its own strategic options, a commitment is made, and credibility is established (Courtney 1997 and Bonau 2017). 9.4 Summary and key learnings Is competition just like playing a game? And how should managers think about playing games? Game theory is a concept for analysing strategic decisions to determine optimum decisions while taking into account and anticipating decisions of all other players. Numerous competitive situations in which players, strategies, and payoffs can be identified are strategically thought through in this way. The main solution concept is the Nash equilibrium in pure or mixed strategies: a Nash equilibrium occurs if all players play mutually best responses - in this constellation, no player has an incentive to deviate from his strategy, as she or he would be worse off. <?page no="355"?> 9.4 Summary and key learnings 355 In one-shot and finitely repeated games, a prisoners' dilemma can arise - a collectively preferred situation does not come about because there are individual incentives to deviate from a cooperative solution. However, laboratory experiments show that people sometimes are willing to build mutual trust and establish a collectively preferred solution - at least temporarily. In sequential games, backward induction can be used to examine whether a player has a first mover advantage and can achieve a higher profit compared to a simultaneous game. However, game theory does not only help to identify an optimum strategy under given conditions. Often, the added value from a management perspective lies in better understanding the basic patterns of a competitive situation and, if necessary, changing these rules of competition, be it by changing the payoff matrix, eliminating own strategies to create commitment, building uncertainty, or investing in credibility. In real-life business environments, game theory helps to structure interactive strategic decision-making situations and to illustrate and analyse implicit and explicit dependencies on competitors. Often, this is the only way to justify and review strategies that are currently being used or that are conceivable per se. In addition, possible reactions of competitors can be explicitly played out in scenarios with different environmental conditions. Finally, game theory makes it possible to generate options for one's own strategies, to develop them in a robust manner, and to test them for their feasibility. However, game theory is not a magic bullet for strategy departments: limited or even absent rationality of competitors, indeterminacy in multiple Nash equilibria and arbitrariness of managers often prevent the achievement of a Nash equilibrium. A precise application of game theory also requires comprehensive qualitative and quantitative data on the competitive process - this is often only possible to a limited extent and involves a great deal of effort in creating the model. Recommendations for further reading Game theory with lots of exciting applications can be found in Binmore, K.G., Fun and games - a text on game theory, London 1992, or Riechmann, T., Spieltheorie, München 2014. With a clear application-oriented reference to decision theory, Harrington, J.E., Games, strategies and decision-making, New York 2015, is a good choice. Questions for review [1] Describe applications of decisions-making by means of game theory as well as their limits, advantages and disadvantages. [2] What is a dominant strategy? Why are equilibria based on dominant strategies of all players stable? [3] What is a Nash equilibrium, in what way does it differ from an equilibrium in dominant strategies? What is the difference between Nash equilibria in pure and mixed strategies? [4] How does the so-called maximin rule work, why and when does it apply? Explain why a firm may be interested in creating uncertainty about payoffs or rationality. <?page no="356"?> 9 Strategic decision-making and game theory 356 [5] What is the tit-for-tat strategy? Describe how this tit-for-tat strategy can provide an explanation for high petrol prices. [6] How can a cooperation (even if prohibited) between competitors arise? Under which conditions is it stable? [7] Describe the so-called prisoner's dilemma. Give two examples of economic situations in which a prisoner's dilemma prevents solutions that are otherwise preferred by society. [8] In which situations does a player try to use a first-mover strategy? Under what conditions does this strategy work? How do you analyse a sequential game, how can an optimum strategy be identified? [9] Explain the concept of mixed strategies using an example. [10] Two firms A and B compete in the market for toothpaste. The figure shows the strategies 'marketing' and 'no marketing' and the corresponding payoff matrix of the profits of both firms. Determine which strategies (risk-neutral) firms will choose. What is a Nash equilibrium, is the combination of strategies you have determined a Nash equilibrium? [11] Two tech firms 1 and 2 compete via the size of R&D budgets, the matrix shows the profits of both firms. Determine possible equilibria for sequential and simultaneous games (with pure and mixed strategies) as well as for risk-averse firms and analyse these solutions from the firms' perspectives. Literature Aiginger, K., The use of game theoretical models for empirical industrial organization, in: Mueller, D.C., Haid, A. and Weigand, J. (eds.), Competition, efficiency, and welfare, Amsterdam 1999, 253-277. Allred, S., Duffy, S. and Smith, J., Cognitive load and strategic sophistication, Journal of Economic Behavior and Organization, 2016, 125, 162-178. Armstrong, M. and Huck, S., Behavioral economics as applied to firms: a primer, CESifo Working Paper Series, 2010, No. 2937. <?page no="357"?> 9.4 Summary and key learnings 357 Axelrod, R. and Hamilton, W.D., The evolution of cooperation, Science, 1981, 211, 1390-1396. Axelrod, R., Effective choice in the prisoner's dilemma, Journal of Conflict Resolution, 1980, 24, 1, 3-25. Azar, O.H. and Bar-Eli, M., Do soccer players play the mixed-strategy Nash-equilibrium? , Applied Economics, 2011, 43, 25, 3591-3601. Bar-Eli, M., Azar, O.H., Ritov, I., Keidar-Levin, Y. and Schein, G., Action bias among elite soccer goalkeepers: the case of penalty kicks, Journal of Economic Psychology, 2007, 28, 5, 606-621. Binmore, K. and Samuelson, L., Muddling through: noisy equilibrium selection, Journal of Economic Theory, 1997, 74, 2, 235-265. Birshan, M. and Kar, J., Becoming more strategic: three tips for any executive, McKinsey Quarterly, 2012, 3, 60-66. Bolton, P., Mehran, H. and Shapiro, J., Executive compensation and risk taking, Review of Finance, 2015, 19, 6, 2139- 2181. Bonau, S., A case for behavioral game theory, Journal of Game Theory, 2017, 6, 1, 7-14. Bosch-Domenech, A., Montalvo, J.G., Nagel, R. and Satorra, A., One, two, (three), infinity, ...: Newspaper and lab beauty-contest experiments, American Economic Review, 2002, 92, 5, 1687-1701. Brandenburger, A.M. and Nalebuff, B., The right game: use game theory to shape strategy, Harvard Business Review, 1995, 73, 4, 57-71. Brosig, J., Identifying cooperative behavior: some experimental results in a prisoner’s dilemma game, Journal of Economic Behavior and Organization, 2002, 47, 3, 275-290. Camerer, C.F., Behavioral game theory - experiments in strategic interaction, Princeton 2003. Chiappori, P.A., Levitt, S. and Groseclose, T., Testing mixed-strategy equilibria when players are heterogeneous: the case of penalty kicks in soccer, American Economic Review, 2002, 92, 4, 1138-1151. Cooper, R.W., DeJong, D.V., Forsythe, R. and Ross, T.W., Selection criteria in coordination games: some experimental results, American Economic Review, 1990, 80, 1, 218-233. Courtney, H., Games managers should play, McKinsey Quarterly, 1997, 2, 91-96. Courtney, H., Horn, J.T. and Kar, J., Getting into your competitor’s head, McKinsey Quarterly, 2009, 1, 129-137. Crawford, V.P., Boundedly rational versus optimization-based models of strategic thinking and learning in games, Journal of Economic Literature, 2013, 51, 2, 512-527. Dixit, A., The role of investment in entry-deterrence, Economic Journal, 1980, 90, 357, 95-106. Duffy, J. and Nagel, R., On the robustness of behaviour in experimental ‘Beauty Contest’ games, Economic Journal, 1997, 107, 1684-1700. Foss, N.J., Leadership, beliefs and coordination: an explorative discussion, Industrial and Corporate Change, 2001, 10, 2, 357-388. Frank, R.H., Gilovich, T. and Regan, D.T., Does studying economics inhibit cooperation? , Journal of Economic Perspectives, 1993, 7, 2, 159-171. Fudenberg, D. and Tirole, J., Game theory, New York 1991. Güth, W., Kocher, M. and Sutter, M., Experimental ‘beauty contests’ with homogeneous and heterogeneous players and with interior and boundary equilibria, Economics Letters, 2002, 74, 2, 219-228. Harsanyi, J.C. and Selten, R., A general theory of equilibrium selection in games, Cambridge 1988. Holt, C.A., Industrial organization: a survey of laboratory research, in: Kagel, J. and Roth, A. (eds.), Handbook of Experimental Economics, Princeton 1993, 1-103. Horn, J.T., Playing war games to win, McKinsey Quarterly, 2011, 2, 122-127. <?page no="358"?> 9 Strategic decision-making and game theory 358 Hotelling, H., Stability in competition, Economic Journal, 1929, 39, 153, 41-57. Kahneman, D., Lovallo, D. and Sibony, O., Before you make that big decision, Harvard Business Review, 2011, 89, 6, 50-60. Kopel, M. and Löffler, C., Commitment, first-mover-, and second-mover advantage, Journal of Economics, 2008, 94, 2, 143-166. Ledoux, A., Concours résultats complets - les victimes se sont plu à jouer le 14 d'atout, Jeux et Strategie, 1981, 2, 10, 10-11. Lindstädt, H. and Müller, J., Making game theory work for managers, McKinsey Quarterly, 2009, 1, 1-9. Mailath, G.J., Do people play Nash-equilibrium? Lessons from evolutionary game theory, Journal of Economic Literature, 1998, 36, 3, 1347-1374. Milgrom, P. and Roberts, J., Predation, reputation, and entry deterrence, Journal of Economic Theory, 1982, 27, 2, 280-312. Münter, M.T., Empirical facts, economic theory and everyday experience in industrial organization - three sides of the same story? , in: Mueller, D.C., Haid, A. and Weigand, J. (eds.), Competition, efficiency, and welfare, Amsterdam 1999, 279-297. Nagel, R., Unraveling in guessing games: an experimental study, American Economic Review, 1995, 85, 1313-1326. Nash, J.F., Equilibrium points in n-person games, Proceedings of the National Academy of Sciences, 1950, 36, 48-49. Nash, J.F., Non-cooperative games, The Annals of Mathematics, 1951, 54, 2, 286-295. o.V., BMW will den 3er zum deutschen Tesla Machen, Süddeutsche Zeitung, 30. Juni 2017, 14. Rapoport, A., Chammah, A.M. and Orwant, C.J., Prisoner's dilemma: a study in conflict and cooperation, Ann Arbor 1965. Ross, S.A., Compensation, incentives, and the duality of risk aversion and riskiness, Journal of Finance, 2004, 59, 1, 207-225. Roth, A.E. and Erev, I., Learning in extensive-form games: experimental data and simple dynamic models in the intermediate term, Games and economic behavior, 1995, 8, 1, 164-212. Schuler, F., Veraszto, T., Voigt, N., Schubert, M., Mergelkamp, I. and Singh, T., Coping with overcapacity: navigating steel’s capacity conundrum, BCG Perspectives, 31. October 2014. Sibony, O., Lovallo, D. and Powell, T.C., Behavioral strategy and the strategic decision architecture of the firm, California Management Review, 2017, 59, 3, 5-21. Sutton, J., Explaining everything, explaining nothing? Game theoretic models in industrial economics, European Economic Review, 1990, 34, 505-512. Sutton, J., Implementing game-theoretic models in industrial economics: levels of attack, in: del Monte, A. (eds.), Recent developments in the theory of industrial organization, Houndmills 1992. Tuyls, K., Perolat, J., Lanctot, M., Ostrovski, G., Savani, R., Leibo, J.Z., Ord, T., Graepel, T. and Legg, S., Symmetric decomposition of asymmetric games, Nature - Scientific Reports, 2018, 8, 1015. Tversky, A. and Kahneman, D., Judgment under uncertainty: heuristics and biases, Science, 1974, 185, 4157, 1124- 1131. von Neumann, J. and Morgenstern, O., Theory of games and economic behavior, Princeton 1944. von Neumann, J., Zur Theorie der Gesellschaftsspiele, Mathematische Annalen, 1928, 100, 1, 295-320. Wang, Z., Xu, B. and Zhou, H.-J., Social cycling and conditional responses in the Rock-Paper-Scissors game, Nature - Scientific Reports, 2014, 4, 5830. <?page no="359"?> 359 10 Strategic competition in oligopoly In situations of perfect competition or in monopoly, strategic decisions of firms can be worked out easily and straightforwardly, but repercussions from reactions and strategies of competitors are not taken into consideration. A monopolist only takes potential competitors into account; in the case of perfect competition it is sufficient to look at the market price and its development in order to derive one's own strategy (► Chapter 7). In oligopolies, this is totally different: here, only a few firms compete and interact strategically, i.e., the mutual repercussions of the behaviour of other firms in competition must be taken into account, especially when developing one's own strategy. For example, Airbus will always pay attention to what Boeing is doing; in the German mobile telecommunications industry the strategy departments of Vodafone, Deutsche Telekom, and Telefónica keep a close eye on each other; and of course, Nestlé's headquarter analyses in detail every strategic decision of Procter&Gamble or Unilever - and vice versa. From a management perspective, three interdependent decisions have to be taken in situations of strategic competition in an oligopoly in order to achieve profits and support the survival of the firm. Which strategic parameters should be used to compete against competitors - e.g., should investments be made in production capacity or is it better to start a price war? How and in what combination should the strategic parameters be employed - e.g., which level of marketing investment is necessary to establish horizontal product differentiation or quality leadership? Should the intensity of competition be increased or reduced - i.e., does my firm have competitiveness to displace others, or should it rather escape to a niche, or should it invest in establishing barriers to entry? Of course, every firm will try to leverage its firm-specific capabilities in competition with others in such a way that competitive advantage is created (► Chapter 4): ► Figure 10.1 shows some of the strategic parameters that typically have to be examined to become pillars of a competitive strategy. For example, a firm can develop a strategy based on strong product differentiation, as Nestlé or Apple do, or claim quality leadership in a high-priced niche market, as the guitar manufacturer Suhr does. Of course, firms are not completely free or autonomous in the choice of their strategic parameters, but are affected by market structure, amongst other things. In case of perfect competition, the market price is exogenously determined, so that competition is only about choosing an optimum quantity based on the cost structure (► Chapter 7). A dominant firm, on the other hand, can significantly influence profits through price strategies (► Chapter 8). Competitive situations in an oligopoly are in between. While it is true, that an almost arbitrary variety of strategies is conceivable, most industries have some specifics and some immanent industry-logic of the competitive process based on industry-specific routines and behavioural patterns of firms (Coyne and Horn 2009, Reeves et al. 2012 and Maynkia et al. 2010). <?page no="360"?> 10 Strategic competition in oligopoly 360 Figure 10.1: Strategic parameters. <?page no="361"?> 9.4 Summary and key learnings 361 Perspectives on strategic competition However, the analysis of strategic competitive situations can always be guided by four perspectives - given some general business environment, firms will analyse: market structure (number and size distribution of firms, existence and extent of entry barriers, etc.); demand structure (e.g., willingness to pay, size of the market or degree of vertical and horizontal product differentiation); and technological opportunities (general conditions for product and process innovation or the possibility of imitating technologies from other industries); in order to determine the choice of strategic parameter (capacity and firm size, pricing strategy, R&D or marketing effort, locations, creation of entry barriers, etc.). that will finally drive and explain success of firms (Shapiro 1989 and Corchon and Marini 2018). Typically, a classification of competition in oligopoly is made according to: Cournot competition - i.e., simultaneous decisions on quantities or capacities, Bertrand competition - i.e., simultaneous decisions on prices, and Stackelberg competition - i.e., sequential decisions on prices or capacities. All analyses are centred on game-theoretical models of capacity and price competition, referred to as Cournot and Bertrand competition according to the seminal description (Cournot 1838 and Bertrand 1883). Strategies are depicted as response curves (equivalent to best responses in game theory). The solutions are again Nash equilibria, whereby the success of the firms - in addition to the firm-specific capabilities - is largely determined by the nature of the competitive process. From an empirical perspective, Cournot competition can be observed in markets and industries in which production capacities play a central role and these can be adjusted only in the long run and with high investments - airports, telecommunication networks, or energy networks are typical examples. Bertrand competition, on the other hand, prevails if capacities can be adjusted quickly and with little investment, so that firms influence capacity utilisation through rapid and frequent price changes, such as temporary employment, airlines, or soft drinks (Bresnahan 1989, Haskel and Martin 1994 as well as Aiginger 1999). From a theoretical perspective, it can be argued that competition takes place in two stages. In a first stage, firms decide on long-term investments in capacity building, and in a second stage they decide on prices based on existing capacity. The market outcome is then, however, under certain conditions, identical to the outcome of Cournot competition, so that Cournot competition plays a dominant role from a strategic perspective as well as an analytical tool in competition policy (Kreps and Scheinkman 1983, Shapiro 1989, Bagwell and Wolinsky 2002, Phlips 1995, and Motta 2004). <?page no="362"?> 10 Strategic competition in oligopoly 362 Learning Objectives This chapter deals with the analysis of competitive situations among a small number of firms that interact strategically, and the practical relevance to develop competitive strategies; the main differences between price and quantity/ capacity competition and its impact on market structure and market results; the impact of market entry and increasing competitive intensity on prices, volumes, and profits of incumbent firms; and implications of sequential decisions and incentives for product differentiation in Cournot and Bertrand competition. 10.1 Capacity decisions and strategies in Cournot competition In order to show the close connections of game theoretical analyses with Cournot competition and Nash equilibria, ► Figure 10.2 shows a payoff matrix of Apple's and Samsung's profits for a new smartphone generation of a certain product category. With each new generation of smartphones, firms have to decide on production capacity. Accordingly, both firms have an interest in finding a Nash equilibrium with regard to the strategic parameter of production capacity, which, however, can be chosen between 0 and 'very large' in almost arbitrary steps. Even with a lot more than two strategies of each firm, it is possible to identify potential Nash equilibria. Both firms can in a first step determine the mutually interdependent profits for arbitrarily assumed values of 𝑞𝑞 𝐴𝐴 and 𝑞𝑞 𝑆𝑆 and plot them into the matrix. Light grey cells then show best responses from Samsung, dark grey cells show best responses from Apple. In this example, there are obviously, indicated by two black boxes, two possible Nash equilibria at 𝑞𝑞 𝐴𝐴 = 1000 / 𝑞𝑞 𝑆𝑆 = 2500 and 𝑞𝑞 𝐴𝐴 = 500 / 𝑞𝑞 𝑆𝑆 = 3000, so that due to lower marginal costs, Samsung produces more. It should be obvious that with an increasing number of strategies or more than two competitors, a development and analysis of payoff matrices is only possible using software such as Excel. However, as shown in ► Figure 10.2, the best responses of both firms can be connected by lines. These connecting lines of best responses describe all mutually profit-maximizing responses of these two firms and are called response curves - at the intersection of both response curves the Nash equilibrium can be found. In case of a continuum of possible strategies, this means that Samsung's optimum strategy 𝑞𝑞 𝑆𝑆 ∗ is a function of Apple's possible strategies 𝑞𝑞 𝐴𝐴∗ and vice versa: Cournot competition can be looked at as a game with a continuum of firms' strategies. <?page no="363"?> 10.1 Capacity decisions and strategies in Cournot competition 363 Figure 10.2: Cournot competition as reaction curves (data from the payoff matrix based on 𝑝𝑝 = 1000 − 0.14 ∗ (𝑞𝑞 𝐴𝐴 + 𝑞𝑞 𝑆𝑆 ) , 𝐹𝐹𝐶𝐶 = 100,000, 𝑀𝑀𝐶𝐶 𝐴𝐴 = 400, 𝑀𝑀𝐶𝐶 𝑆𝑆 = 100 .) <?page no="364"?> 10 Strategic competition in oligopoly 364 Cournot competition with two firms In order to describe the general structure of Cournot competition and strategies based on reaction curves, we first consider an industry without product differentiation with two firms deciding simultaneously. The demand function is given by (10.1) 𝑝𝑝(𝑄𝑄) = 𝑎𝑎 − 𝑏𝑏𝑄𝑄 = 𝑎𝑎 − 𝑏𝑏(𝑞𝑞 1 + 𝑞𝑞 2 ) , so that the revenue functions for the two firms can be described as (10.2) 𝑅𝑅 1 = 𝑝𝑝𝑞𝑞 1 = �𝑎𝑎 − 𝑏𝑏(𝑞𝑞 1 + 𝑞𝑞 2 )�𝑞𝑞 1 = 𝑎𝑎𝑞𝑞 1 − 𝑏𝑏(𝑞𝑞 1 + 𝑞𝑞 2 )𝑞𝑞 1 and (10.3) 𝑅𝑅 2 = 𝑝𝑝𝑞𝑞 2 = �𝑎𝑎 − 𝑏𝑏(𝑞𝑞 1 + 𝑞𝑞 2 )�𝑞𝑞 2 = 𝑎𝑎𝑞𝑞 2 − 𝑏𝑏(𝑞𝑞 1 + 𝑞𝑞 2 )𝑞𝑞 2 . The cost functions of the two firms are (10.4) 𝑇𝑇𝐶𝐶 1 = 𝑀𝑀𝐶𝐶 1 ⋅ 𝑞𝑞 1 + 𝐹𝐹𝐶𝐶 (10.5) 𝑇𝑇𝐶𝐶 2 = 𝑀𝑀𝐶𝐶 2 ⋅ 𝑞𝑞 2 + 𝐹𝐹𝐶𝐶 , i.e., the firms have identical industry-specific fixed costs 𝐹𝐹𝐶𝐶 , but differ in their marginal costs 𝑀𝑀𝐶𝐶 1 and 𝑀𝑀𝐶𝐶 2 . The difference between revenues 𝑅𝑅 𝑖𝑖 and costs 𝑇𝑇𝐶𝐶 𝑖𝑖 then yields the profit functions 𝜋𝜋 𝑖𝑖 for both firms in the form of (10.6) 𝜋𝜋 1 = 𝑅𝑅 1 − 𝑇𝑇𝐶𝐶 1 = 𝑎𝑎𝑞𝑞 1 − 𝑏𝑏(𝑞𝑞 1 + 𝑞𝑞 2 )𝑞𝑞 1 − 𝑀𝑀𝐶𝐶 1 ⋅ 𝑞𝑞 1 − 𝐹𝐹𝐶𝐶 and (10.7) 𝜋𝜋 2 = 𝑅𝑅 2 − 𝑇𝑇𝐶𝐶 2 = 𝑎𝑎𝑞𝑞 2 − 𝑏𝑏(𝑞𝑞 1 + 𝑞𝑞 2 )𝑞𝑞 2 − 𝑀𝑀𝐶𝐶 2 ⋅ 𝑞𝑞 2 − 𝐹𝐹𝐶𝐶 . Equations (10.6) and (10.7) provide an initial idea of strategic interaction of the two firms: profits of firm 1 depend not only on its own capacity 𝑞𝑞 1 , but also on the capacity 𝑞𝑞 2 of its competitor - and vice versa. If both firms understand capacities as strategic parameters, then an optimum strategy 𝑞𝑞 1∗ can be determined for firm 1 from the first derivative of the profit function 𝜋𝜋 1 with respect to the strategic parameter 𝑞𝑞 1 as (10.8) 𝜕𝜕𝜋𝜋1 𝜕𝜕𝑞𝑞1 = 𝑎𝑎 − 𝑏𝑏(𝑞𝑞 1 + 𝑞𝑞 2 ) − 𝑏𝑏𝑞𝑞 1 − 𝑀𝑀𝐶𝐶 1 = 𝑎𝑎 − 𝑀𝑀𝐶𝐶 1 − 2𝑏𝑏𝑞𝑞 1 − 𝑏𝑏𝑞𝑞 2 = 0 . If one rearranges this equation, then an optimum strategy for firm 1 (10.9) 𝑞𝑞 1∗ = 𝑚𝑚−𝑀𝑀𝐶𝐶1 2𝑏𝑏 − 𝑞𝑞2 2 , is obtained. Firm 1, according to (10.9), must consider three determinants of an optimum strategy to maximise profits: firm 2's strategic choices and its capacity strategy 𝑞𝑞 2 - the larger the capacity of the competitor, the smaller the optimum own capacity; its own competitiveness (𝑎𝑎 − 𝑀𝑀𝐶𝐶 1 ) - the larger the competitiveness, the larger the optimum capacity; and the size of the market, given by 1/ 𝑏𝑏 - the larger the market, the larger the optimum production capacity. Equation (10.9) corresponds to optimum strategies of a monopolist from ► Chapter 7 in equation (7.40) - with the modification that now the strategy of a competitor must also be taken into account. In the same way, the following results for firm 2 via the derivative (10.10) 𝜕𝜕𝜋𝜋2 𝜕𝜕𝑞𝑞2 = 𝑎𝑎 − 𝑏𝑏(𝑞𝑞 1 + 𝑞𝑞 2 ) − 𝑏𝑏𝑞𝑞 2 − 𝑀𝑀𝐶𝐶 2 = 𝑎𝑎 − 𝑀𝑀𝐶𝐶 2 − 2𝑏𝑏𝑞𝑞 2 − 𝑏𝑏𝑞𝑞 1 = 0 <?page no="365"?> 10.1 Capacity decisions and strategies in Cournot competition 365 as an optimum strategy (10.11) 𝑞𝑞 2∗ = 𝑚𝑚−𝑀𝑀𝐶𝐶2 2𝑏𝑏 − 𝑞𝑞1 2 , so firm 2 has to take into account the decision of firm 1 according to (10.11). ► Figure 10.3 shows the reaction function of both firms based on equations (10.9) and (10.11). Figure 10.3: Reaction curves of firm 1 and 2. Figure 10.4: Cournot-Nash equilibrium and effects on competitive strategies. For each conceivable quantity 𝑞𝑞 2 of firm 2, firm 1 will determine an optimum reaction, i.e., its own quantity 𝑞𝑞 1 (𝑞𝑞 2 ) . A reaction curve describes all profit-maximising reactions to every conceivable capacity strategy of its competitor. Both reaction curves are downward-sloping - i.e., the larger the competitor's capacity, the smaller one's own optimum capacity is chosen. If one now combines the two reaction curves, a Nash equilibrium in best responses results at the <?page no="366"?> 10 Strategic competition in oligopoly 366 intersection on the left in ► Figure 10.4 - each firm behaves optimally depending on the strategic choice of its competitor - and no firm has an incentive to deviate from the selected strategy, because it would be worse-off. An intersection of two reaction curves determines the optimum capacities 𝑞𝑞 1 and 𝑞𝑞 2 of both firms - if one substitutes (10.9) into (10.11), then optimum strategies in a Nash equilibrium in quantities result as (10.12) 𝑞𝑞 1∗ = 𝑚𝑚−2𝑀𝑀𝐶𝐶1+𝑀𝑀𝐶𝐶2 3𝑏𝑏 and (10.13) 𝑞𝑞 2∗ = 𝑚𝑚−2𝑀𝑀𝐶𝐶2+𝑀𝑀𝐶𝐶1 3𝑏𝑏 . Putting these optimum strategies into (10.1), (10.6), and (10.7), prices and profits in the Nash equilibrium are given as (10.14) 𝑝𝑝 ∗ = 𝑚𝑚+𝑀𝑀𝐶𝐶1+𝑀𝑀𝐶𝐶2 3 , (10.15) 𝜋𝜋 1∗ = (𝑚𝑚−2𝑀𝑀𝐶𝐶1+𝑀𝑀𝐶𝐶2)2 9𝑏𝑏 − 𝐹𝐹𝐶𝐶 and (10.16) 𝜋𝜋 2∗ = (𝑚𝑚−2𝑀𝑀𝐶𝐶2+𝑀𝑀𝐶𝐶1)2 9𝑏𝑏 − 𝐹𝐹𝐶𝐶 . It is important to recognise that the shape of these two reaction curves is determined by the intersection with the quantity axis of the competitiveness 𝑎𝑎 − 𝑀𝑀𝐶𝐶 𝑖𝑖 of the firms. If - as shown in ► Figure 10.4 on the right - Firm 1 is able to reduce its marginal costs and, thus, increase its competitiveness, firm 1's reaction curve will shift upwards to the right. Based on this improved competitiveness, firm 1 will grow and make higher profits - conversely, we can see that firm 2 will have to become smaller. The strategies in Cournot competition are strategic substitutes: they always develop reciprocally in opposite directions for both firms. If the competitiveness of both firms is the same, identical production capacities are chosen. Competitiveness and competitive strategy More generally, the effects on the optimum competitive strategy can be determined by expressing the partial derivatives of equations (10.12) or (10.13) as (10.17) 𝜕𝜕𝑞𝑞𝑖𝑖 ∗ 𝜕𝜕𝑚𝑚 = 1 3𝑏𝑏 > 0; 𝜕𝜕𝑞𝑞𝑖𝑖 ∗ 𝜕𝜕𝑀𝑀𝐶𝐶𝑖𝑖 = − 2 3𝑏𝑏 < 0; 𝜕𝜕𝑞𝑞𝑖𝑖 ∗ 𝜕𝜕𝑏𝑏 = − 𝑚𝑚−𝑀𝑀𝐶𝐶𝑖𝑖+2𝑀𝑀𝐶𝐶𝑗𝑗 3𝑏𝑏2 < 0 and 𝜕𝜕𝑞𝑞𝑖𝑖∗ 𝜕𝜕𝑞𝑞𝑗𝑗∗ = − 12 < 0 . The effect of an increase in willingness to pay 𝑎𝑎 or a reduction in marginal costs 𝑀𝑀𝐶𝐶 is scaled by the size of the market 𝑏𝑏 : a firm benefits more from a growth in the market the higher its competitiveness. To make these effects tangible, in ► Table 10.1 market results and market structure under Cournot competition in two different industries are calculated as an example. In industry 1, firms A and B compete with identical marginal costs 𝑀𝑀𝐶𝐶 𝑖𝑖 = 10. In industry 2, firm C has an efficiency advantage over firm D. For both industries, identical business environments initially apply: the maximum willingness to pay is 𝑎𝑎 = 100, the size of the market equals 𝑏𝑏 = 0.02 and fixed costs are 𝐹𝐹𝐶𝐶 = 100 . <?page no="367"?> 10.1 Capacity decisions and strategies in Cournot competition 367 Market structure and market results under Cournot competition industry 1 industry 2 firms with identical marginal costs firms with differences in marginal costs A B C D firm-specifc marginal costs 𝑀𝑀𝐶𝐶 𝑖𝑖 10 10 7 10 (1) initial situation ( a =100, b =0.02, FC =100) 𝑞𝑞 𝑖𝑖 1500 1500 1600 1450 𝑝𝑝 40 39 𝜋𝜋 𝑖𝑖 44900 44900 51100 41950 𝑄𝑄 3000 3050 𝑆𝑆 𝑖𝑖 50 % 50 % 52 % 48 % (2) increasing willingness to pay a ( a =130, b =0.02, FC =100) 𝑞𝑞 𝑖𝑖 2000 2000 2100 1950 𝑝𝑝 50 49 𝜋𝜋 𝑖𝑖 59900 59900 67100 56450 𝑄𝑄 4000 4050 𝑆𝑆 𝑖𝑖 50 % 50 % 52 % 48 % (3) increasing market size b ( a =100, b =0.01, FC =100) 𝑞𝑞 𝑖𝑖 3000 3000 3200 2900 𝑝𝑝 40 39 𝜋𝜋 𝑖𝑖 89900 89900 102300 84000 𝑄𝑄 6000 6100 𝑆𝑆 𝑖𝑖 50 % 50 % 52 % 48 % Table 10.1: Market structure and market outcome with Cournot competition (numbers partially rounded). The effects described in equation (10.17) are confirmed as follows. In an initial situation (1) with identical marginal costs, the firms are of the same size and incur profits of the same amount. With different marginal costs, the firm with higher efficiency is larger in absolute terms and earns higher profits, the other firm is smaller, and profits decrease. If willingness to pay (2) increases (from 𝑎𝑎 = 100 to 𝑎𝑎 = 130 ) in the market, firms choose higher capacity, both can command higher prices and profits increase. This is true regardless of the level of marginal costs, but the relative market shares 𝑆𝑆 𝑖𝑖 remain unchanged. <?page no="368"?> 10 Strategic competition in oligopoly 368 A growth of the market (3) (from 𝑏𝑏 = 0.2 to 𝑏𝑏 = 0.1 ) lets firms grow and profits increase, but prices cannot be raised - this is also true regardless of marginal costs, likewise, market shares remain constant. From a management perspective, it is crucial to recognise that a firm can achieve higher profits per se in case of an increase in willingness to pay or a growth of the market. This effect is stronger the higher the competitiveness of a firm is (here based on lower marginal costs). However, a growth of the market cannot be leveraged to increase prices (see also Neumann et al. 2001 and Münter 2017). Case Study │ Optimum capacity at Airbus versus Boeing With Boeing and Airbus - before a market entry of the new Chinese competitor Comac - two firms are competing in the market for twinjet, medium-sized long-haul aircraft: the Boeing B787 and Airbus A350 product families. From an airlines' perspective, there is no vertical product differentiation: range, seating capacity, variation options, and fuel consumption are within very narrow limits approximately the same; sales prices are almost identical; Airbus' market share in terms of units is about 53%, Boeing's market share is about 47% (Hepher 2017 and Airbus 2017). The key competitive parameter - heavily distorted by subsidies and US and European industrial policy - is production capacity (Irwin and Pavcnik 2004, Esty and Ghemawat 2002 and Neven and Seabright 1995). Both firms know from many years of competition: the demand function is 𝑝𝑝 = 500 - 2 ⋅ (𝑞𝑞 𝐴𝐴 + 𝑞𝑞 𝐵𝐵 ) , marginal costs are 𝑀𝑀𝐶𝐶 𝐴𝐴 = 140 for Airbus and 𝑀𝑀𝐶𝐶 𝐵𝐵 = 180 for Boeing, and industry-specific fixed costs are 𝐹𝐹𝐶𝐶 = 3,000 for both (in millions of EUR each). Currently, the firms have not yet decided what production capacity to build for each of these generations of aircraft. A management consultant now examines the following questions for Airbus. What is the optimum (profit-maximising) production capacity for Airbus? How do the different marginal costs affect firm sizes and market shares? What market price will be achieved at the next International Paris Air Show Le Bourget? What will be profits of the two firms with these aircraft? Should Airbus invest additional fixed costs of 5,000 in process optimisation to reduce the marginal cost per aircraft from 140 to 80? First of all, the profit functions for both firms can be calculated from the demand function and the respective cost structures as (10.18) 𝜋𝜋 𝐴𝐴 = 𝑝𝑝𝑞𝑞 𝐴𝐴 − 𝑀𝑀𝐶𝐶 𝐴𝐴 𝑞𝑞 𝐴𝐴 − 𝐹𝐹𝐶𝐶 = �𝑎𝑎 − 2(𝑞𝑞 𝐴𝐴 + 𝑞𝑞 𝐵𝐵 )�𝑞𝑞 𝐴𝐴 − 𝑀𝑀𝐶𝐶 𝐴𝐴 𝑞𝑞 𝐴𝐴 − 𝐹𝐹𝐶𝐶 = 500𝑞𝑞 𝐴𝐴 − 2𝑞𝑞 𝐴𝐴2 − 2𝑞𝑞 𝐴𝐴 𝑞𝑞 𝐵𝐵 − 140𝑞𝑞 𝐴𝐴 − 3,000 and (10.19) 𝜋𝜋 𝐵𝐵 = 𝑝𝑝𝑞𝑞 𝐵𝐵 − 𝑀𝑀𝐶𝐶 𝐵𝐵 𝑞𝑞 𝐵𝐵 − 𝐹𝐹𝐶𝐶 = �𝑎𝑎 − 2(𝑞𝑞 𝐴𝐴 + 𝑞𝑞 𝐵𝐵 )�𝑞𝑞 𝐵𝐵 − 𝑀𝑀𝐶𝐶 𝐵𝐵 𝑞𝑞 𝐵𝐵 − 𝐹𝐹𝐶𝐶 = 500𝑞𝑞 𝐵𝐵 − 2𝑞𝑞 𝐵𝐵2 − 2𝑞𝑞 𝐴𝐴 𝑞𝑞 𝐵𝐵 − 180𝑞𝑞 𝐵𝐵 − 3,000 . The optimum production capacity based on the reaction curves results from the first derivative of the respective profit functions with (10.20) 𝜕𝜕𝜋𝜋𝐴𝐴 𝜕𝜕𝑞𝑞𝐴𝐴 = 500 − 4𝑞𝑞 𝐴𝐴 − 2𝑞𝑞 𝐵𝐵 − 140 = 0 equal to 𝑞𝑞 𝐴𝐴 = 500−140 4 − 12 𝑞𝑞 𝐵𝐵 and <?page no="369"?> 10.1 Capacity decisions and strategies in Cournot competition 369 (10.21) 𝜕𝜕𝜋𝜋𝐵𝐵 𝜕𝜕𝑞𝑞𝐵𝐵 = 500 − 4𝑞𝑞 𝐵𝐵 − 2𝑞𝑞 𝐴𝐴 − 180 = 0 equal to 𝑞𝑞 𝐵𝐵 = 500−180 4 − 12 𝑞𝑞 𝐴𝐴 . If we now substitute equation (10.16) into (10.15), we get (10.22) 𝑞𝑞 𝐴𝐴 = 500−140 4 − 12 𝑞𝑞 𝐵𝐵 = 500−140 4 − 12 � 500−180 4 − 12 𝑞𝑞 𝐴𝐴 � = 50 + 14 𝑞𝑞 𝐴𝐴 or solved as 𝑞𝑞 𝐴𝐴 = 200/ 3 - Airbus should plan and build an optimum capacity of about 67 aircraft. Conversely, by substituting 𝑞𝑞 𝐴𝐴 = 200/ 3 into (10.21), Boeing's optimum capacity is 𝑞𝑞 𝐵𝐵 = 140/ 3, about 47 aircraft. The different marginal costs of the two firms translate into different optimum capacities - the lower the marginal costs, the larger the optimum capacity. The market price for one of the aircraft is now obtained by plugging the two quantities 𝑞𝑞 𝐴𝐴 and 𝑞𝑞 𝐵𝐵 into the demand function as (10.23) 𝑝𝑝 = 500 - 2 ⋅ (𝑞𝑞 𝐴𝐴 + 𝑞𝑞 𝐵𝐵 ) = 500 - 2 ⋅ � 200 3 + 140 3 � = 273.33, so that profits for the two producers can be calculated from (10.18) and (10.19) as (10.18) 𝜋𝜋 𝐴𝐴 = 𝑝𝑝𝑞𝑞 𝐴𝐴 − 𝑀𝑀𝐶𝐶 𝐴𝐴 𝑞𝑞 𝐴𝐴 − 𝐹𝐹𝐶𝐶 = 273.33 ⋅ 200 3 − 140 ⋅ 200 3 − 3,000 = 5,888.89 and (10.19) 𝜋𝜋 𝐵𝐵 = 𝑝𝑝𝑞𝑞 𝐵𝐵 − 𝑀𝑀𝐶𝐶 𝐵𝐵 𝑞𝑞 𝐵𝐵 − 𝐹𝐹𝐶𝐶 = 273.33 ⋅ 140 3 − 180 ⋅ 140 3 − 3,000 = 1,355.56 . Airbus' higher profit has two reasons: Airbus produces a larger volume and doing so at lower marginal cost. Finally, it must now be checked whether an investment in process optimisation brings strategic advantage and higher profits for Airbus. At first glance, there is not much to suggest that this is the case: Airbus produces 67 aircraft, unit costs fall by 60 - supposedly a one-off cost reduction effect of 60 ⋅ 67=4,020, compared to a one-off investment of 5,000. However, this view overlooks the strategic effects: first, Airbus' response curve shifts due to lower marginal costs to (10.24) 𝜕𝜕𝜋𝜋𝐴𝐴 𝜕𝜕𝑞𝑞𝐴𝐴 = 500 − 4𝑞𝑞 𝐴𝐴 − 2𝑞𝑞 𝐵𝐵 − 80 = 0 or 𝑞𝑞 𝐴𝐴 = 500−80 4 − 12 𝑞𝑞 𝐵𝐵 so that optimum quantity of Airbus becomes (10.25) 𝑞𝑞 𝐴𝐴 = 500−80 4 − 12 𝑞𝑞 𝐵𝐵 = 500−80 4 − 12 � 500−180 4 − 12 𝑞𝑞 𝐴𝐴 � = 65 + 14 𝑞𝑞 𝐴𝐴 and, thus, increases to 𝑞𝑞 𝐴𝐴 = 260/ 3, about 87 aircraft. Boeing must now also factor this in, so that their optimum capacity is reduced because of (10.26) 𝑞𝑞 𝐵𝐵 = 500−180 4 − 12 𝑞𝑞 𝐴𝐴 = 36.67 . With an increase in total quantity of both manufacturers, the price naturally drops to (10.27) 𝑝𝑝 = 500 - 2 ⋅ (𝑞𝑞 𝐴𝐴 + 𝑞𝑞 𝐵𝐵 ) = 500 - 2 ⋅ � 260 3 + 110 3 � = 253.33 . At this reduced price and adjusted market shares - Airbus has grown significantly to 70%, Boeing has shrunk to 30% - profits are now also changing. Airbus profits (10.28) 𝜋𝜋 𝐴𝐴 ′ = 𝑝𝑝𝑞𝑞 𝐴𝐴 − 𝑀𝑀𝐶𝐶 𝐴𝐴 𝑞𝑞 𝐴𝐴 − 𝐹𝐹𝐶𝐶 = 253.33 ⋅ 260 3 − 80 ⋅ 260 3 − 8,000 = 7,022.22 rising sharply despite significant increase in fixed costs, Boeing's profits (10.29) 𝜋𝜋 𝐵𝐵′ = 𝑝𝑝𝑞𝑞 𝐵𝐵 − 𝑀𝑀𝐶𝐶 𝐵𝐵 𝑞𝑞 𝐵𝐵 − 𝐹𝐹𝐶𝐶 = 253.33 ⋅ 110 3 − 180 ⋅ 110 3 − 3,000 = −311.11 are now negative. However, with a unit price of 𝑝𝑝 = 253.33 and marginal costs of 𝑀𝑀𝐶𝐶 𝐵𝐵 = 180, Boeing realises a clearly positive amount of contribution to fixed costs, which, however, is not sufficient to fully cover the fixed costs. If Boeing had overlooked the strategic effect of the Airbus investment in process optimisation and established the originally <?page no="370"?> 10 Strategic competition in oligopoly 370 planned production capacity of 𝑞𝑞 𝐵𝐵 = 140/ 3, losses would have been even larger. Because of a total capacity of 𝑞𝑞 𝐴𝐴 + 𝑞𝑞 𝐵𝐵 = 260/ 3 + 140/ 3 = 400/ 3 then available, price per aircraft would have fallen to 𝑝𝑝 = 500 - 2 ⋅ (260/ 3 + 140/ 3) = 233.33, so Boeing's loss would now have been higher with 𝜋𝜋 𝐵𝐵′ = 233.33 ⋅ 110/ 3 − 180 ⋅ 110/ 3 − 3,000 = −511.11 : calculating the Cournot-Nash equilibrium would have prevented 200 million EUR loss. Cournot competition, number of firms and intensity of competition The Cournot model can also be extended to situations of competition between more than two firms to analyse strategic decisions and assess the intensity of competition. For 𝑆𝑆 firms competing, which in case of a demand function without product differentiation means (10.30) 𝑝𝑝 = 𝑎𝑎 − 𝑏𝑏𝑄𝑄 = 𝑎𝑎 − 𝑏𝑏(𝑞𝑞 1 + 𝑞𝑞 2 + ⋯ + ⋯ + 𝑞𝑞 𝑝𝑝 ) = 𝑎𝑎 − 𝑏𝑏(𝑞𝑞 𝑖𝑖 + 𝑄𝑄 −𝑖𝑖 ) with 𝑄𝑄 = 𝑞𝑞 𝑖𝑖 + 𝑄𝑄 −𝑖𝑖 and a total quantity of (10.31) 𝑄𝑄 = 𝑞𝑞 1 + 𝑞𝑞 2 + ⋯ + ⋯ + 𝑞𝑞 𝑝𝑝 with for simplicity assumed identical marginal costs 𝑀𝑀𝐶𝐶 1 = 𝑀𝑀𝐶𝐶 2 = ⋯ = 𝑀𝑀𝐶𝐶 given cost functions (10.32) 𝑇𝑇𝐶𝐶 𝑖𝑖 = 𝑀𝑀𝐶𝐶 ⋅ 𝑞𝑞 𝑖𝑖 + 𝐹𝐹𝐶𝐶 , a profit function results for each firm 𝐻𝐻 (10.33) 𝜋𝜋 𝑖𝑖 = 𝑅𝑅 𝑖𝑖 − 𝑇𝑇𝐶𝐶 𝑖𝑖 = 𝑝𝑝𝑞𝑞 𝑖𝑖 − 𝑀𝑀𝐶𝐶 ⋅ 𝑞𝑞 𝑖𝑖 − 𝐹𝐹𝐶𝐶 = [𝑎𝑎 − 𝑏𝑏(𝑞𝑞 𝑖𝑖 + 𝑄𝑄 −𝑖𝑖 )]𝑞𝑞 𝑖𝑖 − 𝑀𝑀𝐶𝐶 ⋅ 𝑞𝑞 𝑖𝑖 − 𝐹𝐹𝐶𝐶 . If each firm maximises profits by choosing an optimum production capacity 𝑞𝑞 𝑖𝑖 , then differentiation of profit functions yields (10.34) 𝜕𝜕𝜋𝜋𝑖𝑖 𝜕𝜕𝑞𝑞𝑖𝑖 = 𝑎𝑎 − 2𝑏𝑏𝑞𝑞 𝑖𝑖 − 𝑏𝑏𝑄𝑄 −𝑖𝑖 − 𝑀𝑀𝐶𝐶 = 0 . Assuming all firms choose identical strategies and, thus, 𝑄𝑄 −𝑖𝑖 = (𝑆𝑆 − 1)𝑞𝑞 𝑖𝑖 holds, so that optimum strategies are given by (10.35) 𝑞𝑞 = 1 𝑝𝑝+1 𝑚𝑚−𝑀𝑀𝐶𝐶 𝑏𝑏 , then it follows that the larger the number 𝑆𝑆 of competitors, the smaller the individual capacity. In addition, the number of competitors also has an effect on the intensity of competition and on prices and profits: via the number of firms 𝑆𝑆 and the optimum strategies 𝑞𝑞 it follows with (10.36) 𝑄𝑄 = 𝑆𝑆𝑞𝑞 = 𝑝𝑝 𝑝𝑝+1 𝑚𝑚−𝑀𝑀𝐶𝐶 𝑏𝑏 as total output Q of an industry, so that price 𝑝𝑝 and profit 𝜋𝜋 can be calculated as (10.37) 𝑝𝑝 = 𝑎𝑎 − 𝑏𝑏𝑄𝑄 = 𝑚𝑚−𝑀𝑀𝐶𝐶 𝑝𝑝+1 + 𝑀𝑀𝐶𝐶 and (10.38) 𝜋𝜋 = 𝑅𝑅 − 𝑇𝑇𝐶𝐶 = 1 𝑏𝑏 � 𝑚𝑚−𝑀𝑀𝐶𝐶 𝑝𝑝+1 � 2 − 𝐹𝐹𝐶𝐶 . In order to better understand these results, ► Figure 10.5 shows graphically for several firms from 𝑆𝑆 = 1 to 𝑆𝑆 = 100 the effects on prices 𝑝𝑝 , the optimum firm-specific quantities 𝑞𝑞 , the resulting profits 𝜋𝜋 , and the total quantity 𝑄𝑄 of an industry for equations (10.35) to (10.38). If we <?page no="371"?> 10.1 Capacity decisions and strategies in Cournot competition 371 look at the level of profits as a function of the number of firms, it becomes clear that high profits are only possible with a very small number of firms - as the number of firms grows and the intensity of competition increases, profits decrease significantly. Prices 𝑝𝑝 successively approach marginal costs, and in the same way individual firms become continuously smaller - even though the total market, measured in production 𝑄𝑄, is growing. Figure 10.5: Cournot competition and number of firms (given business environment: maximum willingness to pay a=100, size of market b=0.02, identical marginal costs MC=10, industry specific fixed costs FC=0. Profits and individual output log scale). The Cournot model can also explain - for homogeneous firms competing in capacity - the well-known results of monopoly and perfect competition. For 𝑆𝑆 = 1 we obtain monopoly results known from ► Chapter 7, for large numbers 𝑆𝑆 of firms an equilibrium results where price equals marginal cost, each firm is very small and there are no profits - identical to a situation under perfect competition. Market entry and Cournot-Nash equilibria From a management perspective, these results first underline the importance of entry barriers - significant profits are only possible with a small number of competitors, all other things being equal. To illustrate this, ► Table 10.2 shows how market structure and market results change in case an additional firm enters the market. In the initial situation (4) there are two incumbents C and D (► Table 10.1), which differ in size and profitability under given conditions ( 𝑎𝑎 = 100, 𝑏𝑏 = 0.02, 𝐹𝐹𝐶𝐶 = 100 ) due to different marginal costs. If firm E now enters the market, albeit with significantly higher marginal costs, due to missing or ineffective entry barriers, all firms will adapt to the new Nash equilibrium (5) over time. In particular, the two incumbent firms will adjust their strategies, each reducing output and losing market share. This is accompanied by a decline in profits of slightly more than 35%. If, however, the two incumbents were to maintain their current strategies (6) and deviate from the <?page no="372"?> 10 Strategic competition in oligopoly 372 Nash equilibrium, both would be able to push through higher market shares. However, due to higher production volume in total, prices would have to be adjusted significantly downwards, so that as a result profits would drop significantly by more than 50%. If no entry barriers can be established, then giving up market shares and adjusting to the new Nash equilibrium in the wake of market entries stabilises the profitability of incumbents. Impact of market entry on market structure and market results under Cournot competition firms with differences in marginal costs C D E firm-specific marginal costs 𝑀𝑀𝐶𝐶 𝑖𝑖 7 10 15 (4) initial situation with two incumbents C and D ( a =100, b =0.02, FC =100) 𝑞𝑞 𝑖𝑖 1600 1450 - 𝑝𝑝 39 𝜋𝜋 𝑖𝑖 51100 41950 - 𝑄𝑄 3050 𝑆𝑆 𝑖𝑖 52 % 48 % - (5) market entry and incumbents adjust to new Cournot- Nash equilibrium ( a =100, b =0.02, FC =100) 𝑞𝑞 𝑖𝑖 1300 1150 900 𝑝𝑝 33 𝜋𝜋 𝑖𝑖 33700 26350 16100 𝑄𝑄 3350 𝑆𝑆 𝑖𝑖 39 % 34 % 27 % (6) market entry and incumbents stick with current strategies ( a =100, b =0.02, FC =100) 𝑞𝑞 𝑖𝑖 1600 1450 900 𝑝𝑝 21 𝜋𝜋 𝑖𝑖 22300 15850 5300 𝑄𝑄 3950 𝑆𝑆 𝑖𝑖 41 % 37 % 23 % Table 10.2: Influence of market entries on market structure and market outcome in the case of Cournot competition (numbers partially rounded). Good to know │ Endogenous market structure - how many firms can enter a market? Market entry barriers limit or impede market entry of new competitors (see also ► Chapter 4). Therefore, it is necessary to assess from three perspectives whether market entry is still possible: competition authorities want to check whether competition is still functioning; start-ups want to find out whether there is still space for them in the market; incumbent firms want to know whether they are threatened by market entry. So, the question is how many firms can exist in a market without causing losses for individual firms. Assuming that the firms have access to the same technology and cost structure, it is possible to use the above profit function <?page no="373"?> 10.1 Capacity decisions and strategies in Cournot competition 373 (10.39) 𝜋𝜋 = 𝑅𝑅 − 𝑇𝑇𝐶𝐶 = 1 𝑏𝑏 � 𝑚𝑚−𝑀𝑀𝐶𝐶 𝑝𝑝+1 � 2 − 𝐹𝐹𝐶𝐶 as a starting point. Setting equation (10.39) equal to zero and then solving for a maximum number of firms 𝑆𝑆 , we get (10.40) 𝑆𝑆 = 𝑚𝑚−𝑀𝑀𝐶𝐶 √𝑏𝑏⋅𝐹𝐹𝐶𝐶 − 1, which is obviously determined by the customers' willingness to pay 𝑎𝑎 , the size of the market 1/ 𝑏𝑏 , and the firms' cost function ( 𝑀𝑀𝐶𝐶, 𝐹𝐹𝐶𝐶 ). The smaller the fixed costs 𝐹𝐹𝐶𝐶 or the marginal costs 𝑀𝑀𝐶𝐶 , and the larger the willingness to pay 𝑎𝑎 and the size of the market 1/ 𝑏𝑏 , the larger the endogenous number of firms 𝑆𝑆 . This inverse relationship between fixed costs and the number of firms is also clearly visible empirically. For example, as a result of increasing regulation following the financial crisis of 2007 onwards, fixed costs are rising for all banks (supervisory law, compliance, equity capital regulations, etc.), so that in Germany numerous Volksand Raiffeisenbanken (as well as savings banks) are merging, i.e., the number of firms is declining as fixed costs rise. How can a start-up now decide whether or not to enter a specific market? Suppose a start-up in an early stage observes its target market - nine competitors are active already, all of them achieve a profit of 𝜋𝜋 = 35 . Market research has shown that the business environment is best described by (𝑎𝑎, 𝑏𝑏, 𝑀𝑀𝐶𝐶, 𝐹𝐹𝐶𝐶) = (50; 0.15; 5; 100) - if one puts these values into (10.40), then with (10.41) 𝑆𝑆 = 𝑚𝑚−𝑀𝑀𝐶𝐶 √𝑏𝑏⋅𝐹𝐹𝐶𝐶 − 1 = 50−5 √0.15⋅100 − 1 = 10.62 a maximum number of ten firms would be able to survive in this market with non-negative profits - i.e., market entry is possible, but has the consequence that all firms become smaller, the price must be reduced, and profits decline. If now some of the incumbent competitors already in the market increase the industry-specific fixed costs to 𝐹𝐹𝐶𝐶 = 150 via a raising rivals' costs strategy (Salop and Scheffman 1983), then, because of (10.42) 𝑆𝑆 = 𝑚𝑚−𝑀𝑀𝐶𝐶 √𝑏𝑏⋅𝐹𝐹𝐶𝐶 − 1 = 50−5 √0.15⋅150 − 1 = 7.21 the number of viable firms is now reduced to 7. Market entry for new firms is now blocked, moreover, two of the incumbents (with the weakest equity base) are forced out of the market. In fact, above a certain level, fixed costs are of course barriers to entry, be it in the form of legal requirements (banking licence, compliance with regulatory requirements, etc.), technological conditions (as a minimum efficient size due to economies of scale) or strategic initiatives by competitors (e.g., by increasing sunk costs for marketing or R&D). If one assumes that a fraction 𝜑𝜑 < 1 of fixed costs only has to be borne by new firms entering the market, but that already established firms can consider this fixed cost fraction as sunk costs and, thus, only have to bear fixed costs of production to the amount of (1 − 𝜑𝜑)𝐹𝐹𝐶𝐶 , no unambiguous Cournot equilibrium results: rather, 𝑆𝑆� is now the upper limit of the number of firms in Cournot competition. The lower limit 𝑆𝑆 is determined by the extent of the sunk costs within the fixed costs. If the actual number of firms is between 𝑆𝑆� and 𝑆𝑆 , no entry takes place because entry is blocked by the sunk costs (Martin 1993 and Münter 1999). Similarly, a maximum number of firms can be determined if technologies of the firms and, thus, the marginal costs are different (Münter 2017). <?page no="374"?> 10 Strategic competition in oligopoly 374 10.2 Sequential decisions and strategies under Stackelberg competition In the previous section, market structure and market results are described in case of simultaneous decisions by firms in the form of capacity competition. In many industries, however, there is a dominant firm that adopts a first mover role as market leader: for example, Deutsche Telekom, as a former state monopoly, continues to play a market-leading role in the German telecommunications industry with a large number of market followers. This does not necessarily imply a dominant market share, as market leadership can also be reflected in technological or strategic decisions that set or change business environment for other firms. Industries of this type are often analysed using the Stackelberg model (von Stackelberg 1934, Amir and Grilo 1999, Spence 1977. Dixit 1979 and Dowrick 1986): one firm decides first on its capacity, all others observe this decision and adapt themselves to this fixed capacity. Roles, commitment, and transparency Competitive processes and the assertion of a market-leading role are well described by the Stackelberg model if the following three conditions are met. Clear assignment of roles - in Stackelberg competition there is a clear assignment of roles between the market leader and one or more (not necessarily smaller) market followers, i.e., none of the market followers claims market leadership. The leadership role can, for example, be based on: a leading role in the past (former monopoly such as Deutsche Telekom or Deutsche Bahn); control of certain competitive parameters (e.g., Intel as technology leader in the development of computer chips and specification of compatibility); or absolute size (e.g., Amazon, which due to its market share - especially via Amazon Marketplace - decisively determines payment and fulfilment procedures for other online shops, even outside the Amazon Marketplace platform). Commitment - leadership can only work if the market-leading firm can credibly commit to a certain strategy, i.e., the strategy must be cost-intensive and irreversible - it must, therefore, involve significant sunk costs on the part of the market leader as in the two examples below. Pharmaceutical firms such as Pfizer or Merck continuously invest sunk costs in R&D competition in order to establish certain technological paths also for the market follower - the higher the cumulative investment in a certain path, the more likely innovations are and the less a market follower can ignore it. The investment in the new Istanbul Yeni Havalimani airport and the development of enormous transfer capacity between North America and Asia in particular is an attempt to establish a market-leading role in order to take international passengers and market share away from the airports Rhein-Main International Frankfurt, Charles de Gaulle Paris, and London Heathrow, which are unable to expand their capacity. Transparency - moreover, this strategy and its implementation must be comprehensible and adaptable for all market followers, so that they can follow easily and without making errors for example: technology firms like Intel, Linux, or Tesla ensure this by disclosing certain product parameters; <?page no="375"?> 10.2 Sequential decisions and strategies under Stackelberg competition 375 specification and publication of industry standards for the production of compatibility, such as for CDs through the so-called Redbook and subsequently the Bluebook by Sony and Philips. Decisions of the market follower and the market leader Market structure and market results under Stackelberg competition can most easily be determined by backward induction, i.e., one starts to analyse the decision situation from the perspective of the market follower. An industry with two firms is considered, firm 1 is the market leader, firm 2 is the market follower and both do not question the assignment of roles and sell homogeneous products. If the same business environment applies as in ► Section 10.1, then the profit function of firm 2 (10.43) 𝜋𝜋 2 = 𝑅𝑅 2 − 𝑇𝑇𝐶𝐶 2 = 𝑎𝑎𝑞𝑞 2 − 𝑏𝑏(𝑞𝑞 1 + 𝑞𝑞 2 )𝑞𝑞 2 − 𝑀𝑀𝐶𝐶 ⋅ 𝑞𝑞 2 − 𝐹𝐹𝐶𝐶 . can be used as a starting point. This firm maximises its profits by choosing production capacity via (10.44) 𝜕𝜕𝜋𝜋2 𝜕𝜕𝑞𝑞2 = 𝑎𝑎 − 𝑏𝑏(𝑞𝑞 1 + 𝑞𝑞 2 ) − 𝑏𝑏𝑞𝑞 2 − 𝑀𝑀𝐶𝐶 2 = 𝑎𝑎 − 𝑀𝑀𝐶𝐶 2 − 2𝑏𝑏𝑞𝑞 2 − 𝑏𝑏𝑞𝑞 1 = 0, so that, after rearrangement, an optimum strategy is (10.45) 𝑞𝑞 2∗ = 𝑚𝑚−𝑀𝑀𝐶𝐶2 2𝑏𝑏 − 𝑞𝑞1 2 . A comparison with equation (10.11) shows: the optimum strategy of a Stackelberg market follower is identical to the strategy of a Cournot competitor. However, if, as described above, there is a clear understanding of roles in this industry, then the market leader can and will anticipate the market follower's decision and take it into account in his choice of strategy. The profit function of the market leader is given as (10.46) 𝜋𝜋 1 = 𝑅𝑅 1 − 𝑇𝑇𝐶𝐶 1 = 𝑎𝑎𝑞𝑞 1 − 𝑏𝑏(𝑞𝑞 1 + 𝑞𝑞 2 )𝑞𝑞 1 − 𝑀𝑀𝐶𝐶 1 ⋅ 𝑞𝑞 1 − 𝐹𝐹𝐶𝐶 , and, by inserting (10.41), becomes (10.47) 𝜋𝜋 1 = 𝑅𝑅 1 − 𝑇𝑇𝐶𝐶 1 = 𝑎𝑎𝑞𝑞 1 − 𝑏𝑏 �𝑞𝑞 1 + 𝑚𝑚−𝑀𝑀𝐶𝐶2 2𝑏𝑏 − 𝑞𝑞1 2 � 𝑞𝑞 1 − 𝑀𝑀𝐶𝐶 1 ⋅ 𝑞𝑞 1 − 𝐹𝐹𝐶𝐶 . Obviously, the profit function of firm 1 now depends only on its own strategy 𝑞𝑞 1 , since 𝑞𝑞 2 is no longer explicitly included in function (10.47). If one differentiates the profit function (10.47) with respect to the production capacity (10.48) 𝜕𝜕𝜋𝜋1 𝜕𝜕𝑞𝑞1 = 𝑎𝑎 − 𝑀𝑀𝐶𝐶 1 − 2𝑏𝑏𝑞𝑞 1 − 𝑚𝑚−𝑀𝑀𝐶𝐶2 2 + 𝑏𝑏𝑞𝑞 1 = 𝑎𝑎 − 𝑀𝑀𝐶𝐶 1 − 𝑚𝑚−𝑀𝑀𝐶𝐶2 2 − 𝑏𝑏𝑞𝑞 1 = 0, then, after rearrangement, an optimum strategy of firm 1 is (10.49) 𝑞𝑞 1∗ = 𝑚𝑚−𝑀𝑀𝐶𝐶1 𝑏𝑏 − 𝑚𝑚−𝑀𝑀𝐶𝐶2 2𝑏𝑏 , which is obviously determined by differences in marginal costs of the two firms. If the follower behaves as expected by the leader and the leader anticipates this in his strategy, then the result in Stackelberg competition in equations (10.45) and (10.49) is a sub-game perfect Nash equilibrium, i.e., the second-stage decision is made consistently with the first-stage considerations. <?page no="376"?> 10 Strategic competition in oligopoly 376 In situations where both firms have equal marginal costs such that 𝑀𝑀𝐶𝐶 1 = 𝑀𝑀𝐶𝐶 2 , (10.49) simplifies to (10.50) 𝑞𝑞 1∗ = 𝑚𝑚−𝑀𝑀𝐶𝐶1 2𝑏𝑏 , i.e., the strategy of a Stackelberg market leader is identical to the strategy of a monopolist (► Chapter 7 equation (7.43)). ► Table 10.3 shows (for the same business environment as in ► Table 10.1 for Cournot competition) the market structure and market results under Stackelberg competition. Market structure and market results under Stackelberg competition industry 1 industry 2 firms with identical marginal costs firms with differences in marginal costs market leader A market follower B market leader C market follower D firm-specific marginal costs MC i 10 10 7 10 initial situation ( a =100, b =0.02, FC =100) 𝑞𝑞 𝑖𝑖 2250 1125 2400 1050 𝑝𝑝 32.5 31 𝜋𝜋 𝑖𝑖 50525 25213 57500 21950 𝑄𝑄 3375 3450 𝑆𝑆 𝑖𝑖 67 % 33 % 70 % 30 % increasing willingness to pay a ( a =130, b =0.02, FC =100 ) 𝑞𝑞 𝑖𝑖 3000 1500 3150 1425 𝑝𝑝 40 38.5 𝜋𝜋 𝑖𝑖 67400 33650 75500 29825 𝑄𝑄 4500 4575 𝑆𝑆 𝑖𝑖 67 % 33 % 69 % 31 % increasing market size ( a =100, b =0.01, FC =100) 𝑞𝑞 𝑖𝑖 4500 2250 4800 2100 𝑝𝑝 32.5 31 𝜋𝜋 𝑖𝑖 101150 50525 115100 44000 𝑄𝑄 6750 6900 𝑆𝑆 𝑖𝑖 67 % 33 % 70 % 30 % Table 10.3: Market structure and market outcome in Stackelberg competition (numbers partially rounded). The market-leading firm can obviously achieve a first mover advantage in every case: with identical marginal costs, the market leader chooses a production capacity that is exactly twice as large as that of the market follower - the market share is accordingly 2/ 3 for the market leader, 1/ 3 for the market follower; with different marginal costs and efficiency advantages for the market leader, the latter can even extend its lead in market share; <?page no="377"?> 10.2 Sequential decisions and strategies under Stackelberg competition 377 the operating profits (contribution margin before fixed costs) of the market leader are twice as high as those of the market follower; the profits of both firms rise symmetrically as willingness to pay increases and if the market grows, both benefit in the same way. Stackelberg market leadership versus Cournot competition Comparing the results of Stackelberg competition in ► Table 10.3 with the results of Cournot competition in ► Table 10.1, the following differences become visible: under Stackelberg competition, a higher total production capacity is established, more is produced, and prices are lower - customers in particular have advantages, as consumer surplus increases; profits of the market leader are typically larger in Stackelberg competition than in Cournot competition, all other things being equal - so that ceteris paribus the producer surplus also increases. From a management perspective, the question arises whether Stackelberg market leadership should generally be pursued - even without the conditions mentioned above. In ► Figure 10.6, the three strategies Cournot capacity 𝐶𝐶𝐸𝐸 , Stackelberg leader capacity 𝑆𝑆𝑀𝑀𝐸𝐸 , and Stackelberg follower capacity 𝑆𝑆𝑀𝑀𝐹𝐹 are combined under the same business environment in a simultaneous strategy selection game, on the left for the case of equal marginal costs, on the right for the case of different marginal costs. Figure 10.6: Game-theoretic test of Cournot vs. Stackelberg competition (strategies: Cournot capacity CK, Stackelberg leader capacity SMK, Stackelberg follower capacity SMF, payoff matrix values rounded in thousands). If two firms consider the three possible strategies discussed here in a higher-order game to choose the right type of competition, it follows that with identical marginal costs the com- <?page no="378"?> 10 Strategic competition in oligopoly 378 bination of two Cournot strategies always represents a stable Nash equilibrium in best responses. This is also true - as in ► Figure 10.6 on the right - if marginal cost differences are not too large: a firm C with marginal cost advantages would choose a Cournot strategy in a simultaneous game; firm D has a best response also in choosing a Cournot strategy. This proves the plausibility and great importance of the Cournot model in simultaneous decisions - Stackelberg leadership typically requires a clear understanding of roles, commitment to a strategy through sunk costs and transparency. 10.3 Pricing decisions and strategies in Bertrand competition Cournot and Stackelberg models in the previous two sections describe competition via capacity decisions. Firms typically achieve positive profits here. In price competition - originally described by Bertrand (1883) - prices are used as a strategic parameter: as a result, firms cannot realise significant profits due to mutual price undercutting. Bertrand competition without product differentiation An industry that competes on prices without vertical product differentiation is airlines. Over the past decades, there is robust empirical evidence that competition between airlines, especially on domestic routes, regularly leads to profits close to zero and to losses when demand drops (IATA 2013, Morrison and Winston 2010, Borenstein 1992 and Oum et al. 2004, see also ► Chapter 4). Therefore, on the one hand, entry barriers, e.g., in the form of strategic code-sharing alliances such as Oneworld, Star Alliance, or Skyteam, are of significant importance in this industry (they may protect profitable routes); whereas on the other hand, marketing plays a major role in establishing horizontal product differentiation. Analogous to the consideration in ► Section 10.1 of viewing Cournot competition as a game with a continuum of capacity strategies, one can now view Bertrand competition as a game with a continuum of pricing strategies. Each firm tries to find an optimum pricing strategy against the background of possible pricing strategies of all competitors, i.e., the resulting strategy combination is again a Nash equilibrium in best responses - here in prices. ► Figure 10.7 shows the competitive situation between Lufthansa and AirBerlin for a domestic route until 2017 as a game in prices. Apparently, there are, marked by two black boxes, two Nash equilibria, however, the firms make losses in both strategy combinations. The cause of the losses lies in the combination of lacking product differentiation, overcapacity, and price competition. If products are not sufficiently differentiated, customers have no preference for one of the airlines, so that these undercut each other in prices in order to leverage their respective capacity. If marginal costs are identical, as assumed here, firms will lower prices down to the level of marginal costs and realise losses equal to fixed costs. It was precisely this relationship that contributed significantly to AirBerlin's insolvency in 2017: the losses on domestic routes could not be sufficiently offset by profits on international routes, and the equity base could not cover the losses in the long run. Bertrand competition between two firms in an industry without product differentiation with identical marginal costs and no capacity constraints can generally be described as follows. A demand function <?page no="379"?> 10.3 Pricing decisions and strategies in Bertrand competition 379 (10.51) 𝑝𝑝 = 𝑎𝑎 − 𝑏𝑏𝑄𝑄 with = 𝑞𝑞 1 + 𝑞𝑞 2 , could also be written as (10.52) 𝑞𝑞 1 + 𝑞𝑞 2 = 𝑚𝑚−𝑝𝑝 𝑏𝑏 . Figure 10.7: Pricing strategies of Lufthansa and AirBerlin (data of payoff matrix based on 𝑝𝑝 = 400- 0.2 ⋅ 𝑄𝑄, 𝑄𝑄 = 2000- 5 ⋅ 𝑝𝑝, 𝑀𝑀𝐶𝐶 𝐿𝐿𝐻𝐻 = 𝑀𝑀𝐶𝐶 𝐴𝐴𝐵𝐵 = 50, 𝐹𝐹𝐶𝐶 = 40,000 ). If firms have identical marginal costs 𝑀𝑀𝐶𝐶 1 = 𝑀𝑀𝐶𝐶 2 , then depending on the pricing strategy, the following quantities will result: (10.53) 𝑞𝑞 1 = � 0 𝐻𝐻𝑓𝑓 𝑝𝑝 1 > 𝑝𝑝 2 𝑚𝑚−𝑝𝑝 2𝑏𝑏 𝐻𝐻𝑓𝑓 𝑝𝑝 1 = 𝑝𝑝 2 𝑚𝑚−𝑝𝑝 𝑏𝑏 𝐻𝐻𝑓𝑓 𝑝𝑝 1 < 𝑝𝑝 2 In the first case 𝑝𝑝 1 > 𝑝𝑝 2 - given unlimited capacity of each competitor and homogeneous products - all customers will buy from firm 2, in the opposite case of 𝑝𝑝 1 < 𝑝𝑝 2 all customers will buy from firm 1. Both firms have a reciprocal incentive to lower their prices below those of the competitor. Only in the case of equal prices 𝑝𝑝 1 = 𝑝𝑝 2 customers will randomly buy from firm 1 or 2 and the expected value of market shares is 50%. In case of simultaneous decisions, both firms will anticipate the pricing strategy of the respective competitor, so that already for two firms because of (10.54) 𝑝𝑝 1 = 𝑝𝑝 2 = 𝑀𝑀𝐶𝐶 1 = 𝑀𝑀𝐶𝐶 2 <?page no="380"?> 10 Strategic competition in oligopoly 380 profits are (10.55) 𝜋𝜋 1 = 𝜋𝜋 2 = 0 . Profits are at the level of perfect competition with only two firms in the market, a larger number of firms does not lead to a change in prices, quantities or profits - only market shares become smaller. This result is a Nash equilibrium in prices - none of the competitors has an incentive to change its pricing strategy: a price cut below marginal cost expands losses, a price increase causes all customers to switch to the competitor. ► Figure 10.8 on the left shows the reaction curves for Bertrand competition without product differentiation. Both firms will repeatedly undercut each other in prices along their reaction curves in the direction of the dashed line until finally both firms set prices 𝑝𝑝 1∗ = 𝑝𝑝 2∗ equal to marginal cost. Figure 10.8: Bertrand competition with and without product differentiation. Under the given conditions, profits in Bertrand competition are only achievable from a management perspective if one of the following approaches is chosen: collusion and agreement on prices at a level above marginal costs - however, this is typically prohibited from a competition policy perspective and the agreement may not be stable (► Chapters 7 and 9); capacity restrictions of all competitors, so that incentives for price undercutting are reduced or eliminated - capacity restrictions, however, are a form of capacity competition from a strategic perspective (see also ► Section 10.1); or development of vertical or horizontal product differentiation - in case of horizontal product differentiation customers have preferences for one of the suppliers, so that even with lower prices not all customers switch to competitors (see also ► Chapter 2). Bertrand competition under horizontal product differentiation Profits in Bertrand competition can occur if firms are able to establish and maintain a sufficient degree of product differentiation. Products can be differentiated vertically (in terms of quality) <?page no="381"?> 10.3 Pricing decisions and strategies in Bertrand competition 381 or horizontally (in terms of branding or taste). Horizontal product differentiation is mostly achieved through marketing, e.g., branding, colouring, the 'look and feel' of the products, or by attributing or highlighting certain characteristics or use cases. For example, current accounts of savings banks, Volks- und Raiffeisenbanken, or Commerzbank and Deutsche Bank are functionally identical - the differentiation is essentially through marketing, which tries to create some level of customer loyalty. Price competition in heterogeneous products often takes place in horizontal product differentiation. Figure 10.9: Degree of product differentiation. The heterogeneity of the products, in this case due to horizontal product differentiation, can be described by the relative proximity of the products to each other. The industry-specific degree of product differentiation 𝜸𝜸 then determines whether and to what extent consumers see both products as substitutes (in a joint market), or whether the two products are seen as completely different (totally separate markets) (► Chapter 2): 𝛾𝛾 = 1 : perfect substitutes (homogeneous products) - customers have no preferences for brands or firms and switch to the cheapest offer at will and at any time; 𝛾𝛾 = 0 : perfect heterogeneity - each firm has a monopoly or dominant position for its market, so that customers have extremely strong preferences for brands or firms and do not switch to a competitor in any case due to this commitment or high switching costs; 1 > 𝛾𝛾 > 0 : horizontal product differentiation - each firm has achieved a certain firm-, brandor product-positioning through marketing investments and has established customer loyalty (regular customers), but these customers would switch to competitors if the price differences were sufficiently large. In case of heterogeneous products, demand for a firm’s product of course continues to depend on one's own product quality (which triggers a certain willingness to pay 𝑎𝑎 with the customer) <?page no="382"?> 10 Strategic competition in oligopoly 382 and one's own cost structure, but indirectly also on the quality and cost structure of the competitor. If 𝑎𝑎 1 and 𝑎𝑎 2 differ, then there is vertical product differentiation; if, in addition, 1 > 𝛾𝛾 > 0, then there is also horizontal product differentiation, so that the demand functions for two firms 1 and 2 are given by (10.56) 𝑝𝑝 1 = 𝑎𝑎 1 − 𝑏𝑏 1 𝑞𝑞 1 − 𝛾𝛾𝑏𝑏 2 𝑞𝑞 2 and (10.57) 𝑝𝑝 2 = 𝑎𝑎 2 − 𝑏𝑏 2 𝑞𝑞 2 − 𝛾𝛾𝑏𝑏 1 𝑞𝑞 1 , where 𝑏𝑏 1 and 𝑏𝑏 2 describe the size of the respective market segments of the two firms. The closer 𝛾𝛾 is to zero, the larger is the share of regular customers of a firm, the smaller are the repercussions of competition and vice versa, so that for firm 1 𝑏𝑏 1 𝑞𝑞 1 describes the direct effect of its own quantity and 𝛾𝛾𝑏𝑏 2 𝑞𝑞 2 the indirect effect of firm 2's output. If 𝛾𝛾 tends towards 1, the two demand functions (10.56) and (10.57) obviously merge, so that - in case of an eradication of market segments due to a lack of differentiation - the demand function (10.51) would emerge again in a situation of homogeneous products. In order to analyse price competition in case of horizontal product differentiation, we assume that the respective market segments are of equal size and that 𝑏𝑏 1 = 𝑏𝑏 2 = 𝑏𝑏 and the customers' willingness to pay for the products of both firms is identical, so that 𝑎𝑎 1 = 𝑎𝑎 2 = 𝑎𝑎 . Then demand functions (10.56) and (10.57) can be rearranged as (10.58) 𝑞𝑞 1 = 1−𝛾𝛾 (1−𝛾𝛾2)𝑏𝑏 𝑎𝑎 + 𝛾𝛾 (1−𝛾𝛾2)𝑏𝑏 𝑝𝑝 2 − 1 (1−𝛾𝛾2)𝑏𝑏 𝑝𝑝 1 and (10.59) 𝑞𝑞 2 = 1−𝛾𝛾 (1−𝛾𝛾2)𝑏𝑏 𝑎𝑎 + 𝛾𝛾 (1−𝛾𝛾2)𝑏𝑏 𝑝𝑝 1 − 1 (1−𝛾𝛾2)𝑏𝑏 𝑝𝑝 2 . With heterogeneous products and price competition, the demand 𝑞𝑞 𝑖𝑖 for one's own product depends on one's own price and the competitor's price. Product differentiation implies that firms do not have a joint but individual (yet linked) demand function. The higher one's own price, the lower is demand, but, the higher the price of the competitor's product, the higher the demand for one's own product. The parameter 1 > 𝛾𝛾 > 0 still maps horizontal product differentiation under price competition. If firms have identical cost functions of 𝑇𝑇𝐶𝐶 𝑖𝑖 = 𝑀𝑀𝐶𝐶 ⋅ 𝑞𝑞 𝑖𝑖 + 𝐹𝐹𝐶𝐶 , then the profit functions for both firms 𝐻𝐻 and 𝑗𝑗 are given as (10.60) 𝜋𝜋 𝑖𝑖 = 𝑝𝑝 𝑖𝑖 ⋅ 𝑞𝑞 𝑖𝑖 − 𝑀𝑀𝐶𝐶 ⋅ 𝑞𝑞 𝑖𝑖 − 𝐹𝐹𝐶𝐶 = 𝑝𝑝 𝑖𝑖 � 1−𝛾𝛾 (1−𝛾𝛾2)𝑏𝑏 𝑎𝑎 + 𝛾𝛾 (1−𝛾𝛾2)𝑏𝑏 𝑝𝑝 𝑗𝑗 − 1 (1−𝛾𝛾2)𝑏𝑏 𝑝𝑝 𝑖𝑖 � − 𝑀𝑀𝐶𝐶 � 1−𝛾𝛾 (1−𝛾𝛾2)𝑏𝑏 𝑎𝑎 + 𝛾𝛾 (1−𝛾𝛾2)𝑏𝑏 𝑝𝑝 𝑗𝑗 − 1 (1−𝛾𝛾2)𝑏𝑏 𝑝𝑝 𝑖𝑖 � − 𝐹𝐹𝐶𝐶 . In order to determine an optimum pricing strategy of a firm, one differentiates (10.60) with respect to the strategic parameter price 𝑝𝑝 𝑖𝑖 , so that via (10.61) 𝜕𝜕𝜋𝜋𝑖𝑖 𝜕𝜕𝑝𝑝𝑖𝑖 = 1−𝛾𝛾 (1−𝛾𝛾2)𝑏𝑏 𝑎𝑎 + 𝛾𝛾 (1−𝛾𝛾2)𝑏𝑏 𝑝𝑝 𝑗𝑗 − 2 (1−𝛾𝛾2)𝑏𝑏 𝑝𝑝 𝑖𝑖 + 1 (1−𝛾𝛾2)𝑏𝑏 𝑀𝑀𝐶𝐶 = 0, after rearrangement, the optimum price is (10.62) 𝑝𝑝 𝑖𝑖 = (1−𝛾𝛾)𝑚𝑚+𝑀𝑀𝐶𝐶 2 + 12 𝛾𝛾𝑝𝑝 𝑗𝑗 for 𝐻𝐻, 𝑗𝑗 = 1.2 <?page no="383"?> 10.3 Pricing decisions and strategies in Bertrand competition 383 for both firms. Analogous to the procedure under Cournot competition, equation (10.62) can be understood as a reaction function - for each price 𝑝𝑝 𝑗𝑗 , firm 𝐻𝐻 will identify and implement an optimum (profit-maximising) reaction 𝑝𝑝 𝑖𝑖 . ► Figure 10.8 on the right shows these reaction curves: with product differentiation the intercept shifts and becomes larger, at the same time the reaction curves rotate and become flatter for firm 1 and steeper for firm 2. As a result, the new intersection is at prices for both firms that are above marginal costs. Firm 𝐻𝐻 , according to (10.62), must consider three determinants of an optimum strategy in order to maximise profits: strategic decisions of firm 𝒋𝒋 and its pricing strategy 𝑝𝑝 𝑗𝑗 - the higher the competitor's price, the higher its own price can be set; customers' willingness to pay - the larger 𝑎𝑎 is, the higher the optimum price can be chosen; and degree of product differentiation, given by 𝛾𝛾 - the larger the product differentiation (the closer 𝛾𝛾 is to zero), the stronger a higher customer willingness to pay can be exploited for higher prices. Horizontal product differentiation reduces the intensity of competition and allows firms to incur profits under Bertrand competition. The same is true, of course, for firm 𝑗𝑗 , so that mutual best responses of both firms to the competitor's pricing strategy represent a Bertrand-Nash equilibrium in prices. ► Figure 10.10 shows the response curves for two firms. Figure 10.10: Bertrand-Nash equilibria and effects on competitive strategies. In contrast to Cournot competition, the reaction curves are now upward-sloping, i.e., a price increase by the competitor leads (e.g., as shown in ► Figure 10.10 on the right, as a result of an increase in the customers' willingness to pay or also due to higher marginal costs) to an own price increase and vice versa. This effect is stronger, the higher the degree of product differentiation, the smaller 𝛾𝛾 . Conversely, price reductions are of course still countered with price reductions: if a firm can reduce its marginal costs, then according to equation (10.62) its own <?page no="384"?> 10 Strategic competition in oligopoly 384 price will be reduced, and the other firm will also reduce its prices. Strategies in Bertrand competition are strategic complements: they always run in the same direction for both firms. Solving equation (10.62) for two firms 𝐻𝐻 and 𝑗𝑗 gives the Nash equilibrium in prices as (10.63) 𝑝𝑝 𝑖𝑖 = 𝑝𝑝 𝑗𝑗 = (2+𝛾𝛾)�(1−𝛾𝛾)𝑚𝑚+𝑀𝑀𝐶𝐶� 4−𝛾𝛾2 , so that by inserting into (10.58) and (10.60) symmetrical quantities and profits of both firms due to equal marginal costs can be calculated as (10.64) 𝑞𝑞 𝑖𝑖 = 𝑞𝑞 𝑗𝑗 = 𝑚𝑚−𝑀𝑀𝐶𝐶 2+𝛾𝛾+𝛾𝛾2 and (10.65) 𝜋𝜋 𝑖𝑖 = 𝜋𝜋 𝑗𝑗 = (1−𝛾𝛾)(𝑚𝑚−𝑀𝑀𝐶𝐶)2 (1+𝛾𝛾)(2−𝛾𝛾)2 − 𝐹𝐹𝐶𝐶 . If the degree of product differentiation increases, i.e., 𝛾𝛾 becomes smaller and tends towards 0, then the reaction curves shift outwards and become flatter. Prices rise if the degree of product differentiation increases because of (10.63), and firms also grow because of (10.64). In any case, firms can make profits in Bertrand competition if product differentiation is sufficiently large (► Section 10.4). If, in the absence of product differentiation, 𝛾𝛾 would tend towards 1, it follows from (10.63) initially 𝑝𝑝 𝑖𝑖 = 𝑝𝑝 𝑗𝑗 = 𝑀𝑀𝐶𝐶 and from (10.65) a loss in the amount of fixed costs. Case Study │ Price wars between Air France and British Airways Air France and British Airways compete on the Paris CDG - London Heathrow route. As both airlines invest heavily in marketing on this international route, there is horizontal product differentiation from the customer's point of view. Both firms compete on price, as the respective capacities can be expanded at any time to cover the entire demand. Business class demand is given by 𝑞𝑞 𝐴𝐴 = 2000 − 2 𝑝𝑝 𝐴𝐴 + 𝑝𝑝 𝐵𝐵 for Air France, and 𝑞𝑞 𝐵𝐵 = 2000 − 2 𝑝𝑝 𝐵𝐵 + 𝑝𝑝 𝐴𝐴 for British Airways, where 𝑞𝑞 𝐴𝐴 is the quantity (number of seats) of Air France, 𝑞𝑞 𝐵𝐵 is the quantity of British Airways, 𝑝𝑝 𝐴𝐴 is the ticket price of Air France and 𝑝𝑝 𝐵𝐵 is the ticket price of British Airways. Marginal costs of both airlines on this route are 𝑀𝑀𝐶𝐶 = 20 per seat, and fixed costs are 𝐹𝐹𝐶𝐶 = 250,000 each. A management consultant should now answer the following questions for British Airways. Which prices should Air France and British Airways charge, what quantities will result, what is Air France's profit, what is British Airways' profit? Air France decides to set the price 𝑝𝑝 𝐴𝐴 = 400 . What is British Airways’ optimum response to this pricing strategy in order to maximise its own profit? To determine the optimum prices of both firms, one first sets up profit functions (10.66) 𝜋𝜋 𝑖𝑖 = 𝑝𝑝 𝑖𝑖 ⋅ 𝑞𝑞 𝑖𝑖 − 𝑀𝑀𝐶𝐶 ⋅ 𝑞𝑞 𝑖𝑖 − 𝐹𝐹𝐶𝐶 of both firms and plugs in the respective demand functions, so that (10.67) 𝜋𝜋 𝐴𝐴 = 𝑝𝑝 𝐴𝐴 ⋅ (2000- 2 𝑝𝑝 𝐴𝐴 + 𝑝𝑝 𝐵𝐵 ) − 𝑀𝑀𝐶𝐶 ⋅ 𝑞𝑞 𝐴𝐴 − 𝐹𝐹𝐶𝐶 for Air France and (10.68) 𝜋𝜋 𝐵𝐵 = 𝑝𝑝 𝐵𝐵 ⋅ (2000- 2 𝑝𝑝 𝐵𝐵 + 𝑝𝑝 𝐴𝐴 ) − 𝑀𝑀𝐶𝐶 ⋅ 𝑞𝑞 𝐵𝐵 − 𝐹𝐹𝐶𝐶 for British Airways. The optimum prices 𝑝𝑝 𝐴𝐴 and 𝑝𝑝 𝐵𝐵 are determined by taking derivatives of profit functions with respect to 𝑝𝑝 𝐴𝐴 and 𝑝𝑝 𝐵𝐵 , so that <?page no="385"?> 10.4 Strategic competition with product differentiation 385 (10.69) 𝜕𝜕𝜋𝜋𝐴𝐴 𝜕𝜕𝑝𝑝𝐴𝐴 = 2000 − 4𝑝𝑝 𝐴𝐴 + 𝑝𝑝 𝐵𝐵 + 12 𝑀𝑀𝐶𝐶 = 0 or 𝑝𝑝 𝐴𝐴 = 500 + 14 𝑝𝑝 𝐵𝐵 + 12 𝑀𝑀𝐶𝐶 and (10.70) 𝜕𝜕𝜋𝜋𝐵𝐵 𝜕𝜕𝑝𝑝𝐵𝐵 = 2000 − 4𝑝𝑝 𝐵𝐵 + 𝑝𝑝 𝐴𝐴 + 12 𝑀𝑀𝐶𝐶 = 0 or 𝑝𝑝 𝐵𝐵 = 500 + 14 𝑝𝑝 𝐴𝐴 + 12 𝑀𝑀𝐶𝐶 results. If one now substitutes (10.70) into (10.69), then after solving for (10.71) 𝑝𝑝 𝐴𝐴 = 500 + 14 𝑝𝑝 𝐵𝐵 = 500 + 14 �500 + 14 𝑝𝑝 𝐴𝐴 � , the optimum price of Air France equals 𝑝𝑝 𝐴𝐴 = 680 . Due to the symmetry of these two firms in demand and cost structure, the optimum price of British Airways is also 𝑝𝑝 𝐵𝐵 = 680 . Both firms can, because of (10.72) 𝑞𝑞 𝐴𝐴 = 2000- 2 𝑝𝑝 𝐴𝐴 + 𝑝𝑝 𝐵𝐵 = 2000 − 2 ⋅ 680 + 680 = 1,320 and (10.73) 𝑞𝑞 𝐵𝐵 = 2000- 2 𝑝𝑝 𝐵𝐵 + 𝑝𝑝 𝐴𝐴 = 2000 − 2 ⋅ 680 + 680 = 1,320, sell 1,320 business class tickets per day on this route. The profits of both firms are calculated as follows (10.74) 𝜋𝜋 𝐴𝐴 = 𝑝𝑝 𝐴𝐴 ⋅ 𝑞𝑞 𝐴𝐴 − 𝑀𝑀𝐶𝐶 ⋅ 𝑞𝑞 𝐴𝐴 − 𝐹𝐹𝐶𝐶 = 680 ⋅ 1320 − 20 ⋅ 1320 − 250,000 = 621,200 and (10.75) 𝜋𝜋 𝐵𝐵 = 𝑝𝑝 𝐵𝐵 ⋅ 𝑞𝑞 𝐵𝐵 − 𝑀𝑀𝐶𝐶 ⋅ 𝑞𝑞 𝐵𝐵 − 𝐹𝐹𝐶𝐶 = 680 ⋅ 1320 − 20 ⋅ 1320 − 250,000 = 621,200, also of equal amount. If Air France tries to gain more customers and change the competitive process by setting a price of 𝑝𝑝 𝐴𝐴 = 400, British Airways cannot stick to the previous price. However, directly from the reaction function (10.66) taking into account the price arbitrarily set by Air France, British Airways can find the optimum new pricing strategy as (10.76) 𝑝𝑝 𝐵𝐵 = 500 + 14 𝑝𝑝 𝐴𝐴 = 500 + 14 100 = 600 . With the new prices, the quantities shift to (10.77) 𝑞𝑞 𝐴𝐴 = 2000- 2 𝑝𝑝 𝐴𝐴 + 𝑝𝑝 𝐵𝐵 = 2000 − 2 ⋅ 400 + 600 = 1,800 and (10.78) 𝑞𝑞 𝐵𝐵 = 2000- 2 𝑝𝑝 𝐵𝐵 + 𝑝𝑝 𝐴𝐴 = 2000 − 2 ⋅ 600 + 400 = 1,200, of course, in favour of Air France. However, if one considers profits of (10.79) 𝜋𝜋 𝐴𝐴 = 𝑝𝑝 𝐴𝐴 ⋅ 𝑞𝑞 𝐴𝐴 − 𝑀𝑀𝐶𝐶 ⋅ 𝑞𝑞 𝐴𝐴 − 𝐹𝐹𝐶𝐶 = 400 ⋅ 1800 − 20 ⋅ 1800 − 250,000 = 434,000 and (10.80) 𝜋𝜋 𝐵𝐵 = 𝑝𝑝 𝐵𝐵 ⋅ 𝑞𝑞 𝐵𝐵 − 𝑀𝑀𝐶𝐶 ⋅ 𝑞𝑞 𝐵𝐵 − 𝐹𝐹𝐶𝐶 = 600 ⋅ 1200 − 20 ⋅ 1200 − 250,000 = 446,000, then Air France's arbitrary deviation from the Bertrand-Nash equilibrium leads to reduced profits, but due to high fixed costs with simultaneous product differentiation having a stronger negative effect on Air France. Again, a deviation from the Nash equilibrium reduces profits. 10.4 Strategic competition with product differentiation Strategically, product differentiation is used, on the one hand, to create firm-specific market segments, and, on the other hand, to increase customers' willingness to pay. If there is also vertical product differentiation, the effects of qualitative differences in the products can be strengthened by horizontal product differentiation. In case of homogeneous products in markets without vertical or horizontal product differentiation, competition among a fixed number of firms, as shown in the previous sections, leads to very different results. Without limiting factors such as capacity restrictions, customeror contract-ties, or sequential decisions, firms typically do not achieve a profit in Bertrand competition as a result of reciprocal price undercutting. In Cournot competition, however, positive profits are not only possible, but typical. From a <?page no="386"?> 10 Strategic competition in oligopoly 386 management perspective, a crucial task is to work out in which type of competition there are greater incentives for product differentiation in order to achieve profits for a firm. In the following, optimum strategies are developed depending on the degree of product differentiation for firms with heterogeneous competitiveness (differences in customers' willingness to pay and in marginal costs) (Singh and Vives 1984, Tremblay and Tremblay 2011 and Münter 2017). Cournot competition and product differentiation Consider two firms 𝐻𝐻 = 1; 2 that compete in capacities 𝑞𝑞 𝑖𝑖 , whose marginal costs 𝑀𝑀𝐶𝐶 𝑖𝑖 are different, which have industry-specific fixed costs 𝐹𝐹𝐶𝐶 of equal amount and whose products are differentiated both vertically and horizontally. Thus, by equations (10.81) and (10.82), the demand function and the cost function of the two firms can be described as (10.81) 𝑝𝑝 1 = 𝑎𝑎 1 − 𝑏𝑏 1 𝑞𝑞 1 − 𝛾𝛾𝑏𝑏 2 𝑞𝑞 2 𝑝𝑝 2 = 𝑎𝑎 2 − 𝑏𝑏 2 𝑞𝑞 2 − 𝛾𝛾𝑏𝑏 1 𝑞𝑞 1 and (10.82) 𝑇𝑇𝐶𝐶 1 = 𝑀𝑀𝐶𝐶 1 𝑞𝑞 1 + 𝐹𝐹𝐶𝐶 𝑇𝑇𝐶𝐶 2 = 𝑀𝑀𝐶𝐶 2 𝑞𝑞 2 + 𝐹𝐹𝐶𝐶 with 1 > 𝛾𝛾 > 0 , 𝑎𝑎 1 > 𝑀𝑀𝐶𝐶 1 and 𝑎𝑎 2 > 𝑀𝑀𝐶𝐶 2 . Here, 𝑏𝑏 𝑖𝑖 is an indicator for the size of the firm-specific market segment, 𝑎𝑎 𝑖𝑖 is a measure for the maximum firm-specific willingness to pay of customers to illustrate quality-related vertical product differentiation - the greater the product quality perceived by customers (e.g., as a result of product innovations), the higher the value of 𝑎𝑎 𝑖𝑖 . The parameter 𝛾𝛾 also measures the industry-specific degree of horizontal product differentiation. For each firm this results in a profit function (10.83) 𝜋𝜋 𝑖𝑖 = 𝑅𝑅 𝑖𝑖 − 𝑇𝑇𝐶𝐶 𝑖𝑖 = 𝑎𝑎 𝑖𝑖 𝑞𝑞 𝑖𝑖 − 𝑏𝑏 𝑖𝑖 𝑞𝑞 𝑖𝑖 2 − 𝛾𝛾𝑏𝑏 𝑗𝑗 𝑞𝑞 𝑗𝑗 𝑞𝑞 𝑖𝑖 − 𝑀𝑀𝐶𝐶 𝑖𝑖 𝑞𝑞 𝑖𝑖 − 𝐹𝐹𝐶𝐶 , which can be maximised by choosing the production capacity 𝑞𝑞 𝑖𝑖 , (10.84) 𝜕𝜕𝜋𝜋𝑖𝑖 𝜕𝜕𝑞𝑞𝑖𝑖 = 𝑎𝑎 𝑖𝑖 − 2𝑏𝑏 𝑖𝑖 𝑞𝑞 𝑖𝑖 − 𝛾𝛾𝑏𝑏 𝑗𝑗 𝑞𝑞 𝑗𝑗 − 𝑀𝑀𝐶𝐶 𝑖𝑖 = 0 . This provides optimum strategies for both firms (10.85) 𝑞𝑞 1∗ = 𝑚𝑚1−𝑀𝑀𝐶𝐶1 2𝑏𝑏1 − 𝛾𝛾𝑏𝑏2𝑞𝑞2 2𝑏𝑏1 𝑞𝑞 2∗ = 𝑚𝑚2−𝑀𝑀𝐶𝐶2 2𝑏𝑏2 − 𝛾𝛾𝑏𝑏1𝑞𝑞1 2𝑏𝑏2 . These strategies are again reaction curves depending on the production capacity of the competitors, capacities are still strategic substitutes, but due to horizontal product differentiation and 1 > 𝛾𝛾 > 0, these curves are flatter than in the case of homogeneous products (see also ► Section 10.1). The intersection again determines a Nash equilibrium via the resulting strategy combination, but this is typically not symmetric if there are differences in marginal costs, willingness to pay, or size of the market segments. Without vertical or horizontal product differentiation and market segments of equal size, optimum strategies for 𝑎𝑎 1 = 𝑎𝑎 2 , 𝑏𝑏 1 = 𝑏𝑏 2 and 𝛾𝛾 → 1 follow as (10.86) 𝑞𝑞 1∗ = 𝑚𝑚−𝑀𝑀𝐶𝐶1 2𝑏𝑏 − 𝑞𝑞2 2 𝑞𝑞 2∗ = 𝑚𝑚−𝑀𝑀𝐶𝐶2 2𝑏𝑏 − 𝑞𝑞1 2 , which are identical to the strategies developed in ► Section 10.1 in equations (10.9) and (10.11). <?page no="387"?> 10.4 Strategic competition with product differentiation 387 Vertical product differentiation under Cournot competition If we look at firm 1, for example, with the partial derivative of equation (10.85) (10.87) 𝜕𝜕𝑞𝑞1 𝜕𝜕𝑚𝑚1 = 1 2𝑏𝑏1 > 0, it is obvious that an increase in vertical product differentiation (i.e., 𝑎𝑎 1 becomes larger) causes the optimum firm size to increase - this effect is all the stronger the larger the firm's own market segment (the smaller 𝑏𝑏 1 ). The effects for vertical product differentiation are asymmetrical: if firm 1 can increase the firm-specific willingness to pay 𝑎𝑎 1 , e.g., through product innovations, this direct positive effect on its own firm size is accompanied by an indirect negative effect on the size of the other firm 2, as can be seen from (10.88) 𝜕𝜕𝑞𝑞2 𝜕𝜕𝑞𝑞1 = − 𝛾𝛾𝑏𝑏2 2𝑏𝑏1 < 0 . The negative effect of increasing vertical product differentiation on the other firm is stronger the larger the firm's own market segment is, but is weakened with increasing degrees of horizontal product differentiation, i.e., absolutely smaller values of 𝛾𝛾 . If equations (10.87) and (10.88) are taken together, the overall effect of increasing vertical product differentiation is clear: firm 1 grows directly because of equation (10.87), this growth is reinforced by the indirect effect (10.88) of the reduced size of firm 2. Horizontal product differentiation under Cournot competition If we look at the effect of horizontal product differentiation for firm 1 in isolation, from the partial derivative of (10.89) 𝜕𝜕𝑞𝑞1 𝜕𝜕𝛾𝛾 = − 𝑏𝑏2𝑞𝑞2 2𝑏𝑏1 < 0, it is clear that an increase in horizontal product differentiation (i.e., 𝛾𝛾 becomes smaller and tends towards 0) has a positive direct effect on the size of both firms and vice versa. The effect is stronger the larger the own market segment is. However, if we formulate (10.85), e.g., for firm 1 as (10.90) 𝑞𝑞 1 = 2(𝑚𝑚1−𝑀𝑀𝐶𝐶1)−𝛾𝛾(𝑚𝑚2−𝑀𝑀𝐶𝐶2) 𝑏𝑏1(4−𝛾𝛾2) and now differentiate with respect to 𝛾𝛾 , in order to also capture the indirect repercussions from a change in 𝑞𝑞 2 , we get (10.91) 𝜕𝜕𝑞𝑞1 𝜕𝜕𝛾𝛾 = 4𝛾𝛾(𝑚𝑚1−𝑀𝑀𝐶𝐶1)−�4+𝛾𝛾2�(𝑚𝑚2−𝑀𝑀𝐶𝐶2) 𝑏𝑏1(4−𝛾𝛾2) . This effect can obviously be positive or negative - firm 1 grows as the degree of product differentiation decreases (i.e., 𝛾𝛾 increases and tends towards 1) precisely if (10.92) 4𝛾𝛾(𝑎𝑎 1 − 𝑀𝑀𝐶𝐶 1 ) > (4 + 𝛾𝛾 2 )(𝑎𝑎 2 − 𝑀𝑀𝐶𝐶 2 ) , that is, the degree of horizontal product differentiation in relation to the relative competitive capabilities of the two firms is (10.93) 𝛾𝛾 > 2�(𝑚𝑚1−𝑀𝑀𝐶𝐶1)−�(𝑚𝑚1−𝑀𝑀𝐶𝐶1)2−(𝑚𝑚2−𝑀𝑀𝐶𝐶2)2� 𝑚𝑚2−𝑀𝑀𝐶𝐶2 . <?page no="388"?> 10 Strategic competition in oligopoly 388 The direction and strength of the effect cannot be determined without knowing or looking at the relative firm-specific competitiveness (𝑎𝑎 1 − 𝑀𝑀𝐶𝐶 1 ) and (𝑎𝑎 2 − 𝑀𝑀𝐶𝐶 2 ) . In any case, a firm will only grow with a declining degree of horizontal product differentiation if its own competitiveness is relatively large compared to that of its competitor - the size of the other firm will definitely decline in this case. From a management perspective, a detailed analysis of the competitiveness of competitors is obviously necessary. Bertrand competition and product differentiation The two firms 𝐻𝐻 = 1; 2 are further considered in the same business environment as outlined in the previous section, i.e., as described via the system of equations (10.81) and (10.82), with product differentiation under the assumption that prices are now used as strategic parameters. First, the inverse demand functions (10.81) are reformulated into a system of demand functions (10.94) 𝑞𝑞 1 = 1 (1−𝛾𝛾2)𝑏𝑏1 𝑎𝑎 1 − 𝛾𝛾 (1−𝛾𝛾2)𝑏𝑏1 𝑎𝑎 2 + 𝛾𝛾 (1−𝛾𝛾2)𝑏𝑏1 𝑝𝑝 2 − 1 (1−𝛾𝛾2)𝑏𝑏1 𝑝𝑝 1 𝑞𝑞 2 = 1 (1−𝛾𝛾2)𝑏𝑏2 𝑎𝑎 2 − 𝛾𝛾 (1−𝛾𝛾2)𝑏𝑏2 𝑎𝑎 1 + 𝛾𝛾 (1−𝛾𝛾2)𝑏𝑏2 𝑝𝑝 1 − 1 (1−𝛾𝛾2)𝑏𝑏2 𝑝𝑝 2 , so that the respective profit functions can be seen as (10.95) 𝜋𝜋 𝑖𝑖 =𝑝𝑝 𝑖𝑖 � 𝑚𝑚𝑖𝑖−𝛾𝛾𝑚𝑚𝑗𝑗+𝛾𝛾𝑝𝑝𝑗𝑗 (1−𝛾𝛾2)𝑏𝑏𝑖𝑖 − 1 (1−𝛾𝛾2)𝑏𝑏𝑖𝑖 𝑝𝑝 𝑖𝑖 � − 𝑀𝑀𝐶𝐶 𝑖𝑖 � 𝑚𝑚𝑖𝑖−𝛾𝛾𝑚𝑚𝑗𝑗+𝛾𝛾𝑝𝑝𝑗𝑗 (1−𝛾𝛾2)𝑏𝑏𝑖𝑖 − 1 (1−𝛾𝛾2)𝑏𝑏𝑖𝑖 𝑝𝑝 𝑖𝑖 � − 𝐹𝐹𝐶𝐶 . These differ compared to ► Section 10.3 because vertical product differentiation (𝑎𝑎 𝑖𝑖 ≠ 𝑎𝑎 𝑗𝑗 ) and market segments of different sizes (𝑏𝑏 𝑖𝑖 ≠ 𝑏𝑏 𝑗𝑗 ) are now also included. If one differentiates these profit functions with respect to the strategic parameter price, (10.96) 𝜕𝜕𝜋𝜋𝑖𝑖 𝜕𝜕𝑝𝑝𝑖𝑖 = 𝑚𝑚𝑖𝑖−𝛾𝛾𝑚𝑚𝑗𝑗+𝛾𝛾𝑝𝑝𝑗𝑗 (1−𝛾𝛾2)𝑏𝑏𝑖𝑖 − 2 (1−𝛾𝛾2)𝑏𝑏𝑖𝑖 𝑝𝑝 𝑖𝑖 + 1 (1−𝛾𝛾2)𝑏𝑏𝑖𝑖 𝑀𝑀𝐶𝐶 𝑖𝑖 = 0, then optimum pricing strategies of both firms result as (10.97) 𝑝𝑝 1∗ = 𝑚𝑚1−𝛾𝛾𝑚𝑚2+𝑀𝑀𝐶𝐶1 2 + 12 𝛾𝛾𝑝𝑝 2 𝑝𝑝 2∗ = 𝑚𝑚2−𝛾𝛾𝑚𝑚1+𝑀𝑀𝐶𝐶2 2 + 12 𝛾𝛾𝑝𝑝 1 . Thus, even in the case of product differentiation in the context of price competition, the strategies of both firms are strategic complements, and the reaction curves are increasing. Equation (10.97) can be rearranged as follows (10.98) 𝑝𝑝 1 = �2−𝛾𝛾2�𝑚𝑚1−𝛾𝛾𝑚𝑚2+2𝑀𝑀𝐶𝐶1+𝛾𝛾𝑀𝑀𝐶𝐶2 4−𝛾𝛾2 𝑝𝑝 2 = �2−𝛾𝛾2�𝑚𝑚2−𝛾𝛾𝑚𝑚1+2𝑀𝑀𝐶𝐶2+𝛾𝛾𝑀𝑀𝐶𝐶1 4−𝛾𝛾2 to determine the effects of vertical and horizontal product differentiation. Vertical and horizontal product differenation under Bertrand competition If we look at firm 1, for example, the partial derivative of (10.98) with (10.99) 𝜕𝜕𝑝𝑝1 𝜕𝜕𝑚𝑚1 = 2−𝛾𝛾2 4−𝛾𝛾2 > 0, it is clear that with product innovations that increase vertical product differentiation <?page no="389"?> 10.4 Strategic competition with product differentiation 389 ( 𝑎𝑎 1 increases), firm 1 can push through higher prices, and that this effect is stronger the higher the degree of horizontal product differentiation (the smaller 𝛾𝛾 is). The other firm 2, due to (10.100) 𝜕𝜕𝑝𝑝2 𝜕𝜕𝑚𝑚1 = − 𝛾𝛾 4−𝛾𝛾2 < 0, has to reduce prices accordingly, although the strength of this effect is determined by the degree of horizontal product differentiation - the less horizontally differentiated the products are (i.e., 𝛾𝛾 becomes larger and converges towards 1), the more strongly the prices of firm 2 are reduced as a result of higher vertical product differentiation of firm 1. If we now look at the effect of horizontal product differentiation for firm 1, then with the partial derivative (10.101) 𝜕𝜕𝑝𝑝1 𝜕𝜕𝛾𝛾 = − (𝑚𝑚1−𝑀𝑀𝐶𝐶1)4𝛾𝛾+(𝑚𝑚2−𝑀𝑀𝐶𝐶2)�4+𝛾𝛾2� (4−𝛾𝛾2)2 < 0 , it is clear that increasing horizontal product differentiation makes it possible to increase the respective prices. These effects are going in the same direction for both firms due to the complementary nature of the strategic variable price. The size of these effects is determined by the respective competitiveness (𝑎𝑎 𝑖𝑖 − 𝑀𝑀𝐶𝐶 𝑖𝑖 ) . Competitiveness, market structure and profits under Bertrand and Cournot competition In fact, however, it can be shown that, depending on whether Cournot or Bertrand competition is present, the degree of horizontal product differentiation has both quantitatively and qualitatively different effects. In order to quantify the differences of Cournot versus Bertrand competition, the following simulation shows the respective firm strategies and the resulting market structure depending on the degree of horizontal product differentiation. First, a situation is considered in which two firms compete that differ only slightly in their competitiveness. For demand and cost functions, the following parameters from ► Table 10.4 apply first. The firms have minor differences in competitiveness with (𝑎𝑎 1 − 𝑀𝑀𝐶𝐶 1 ) = 21.0 and (𝑎𝑎 2 − 𝑀𝑀𝐶𝐶 2 ) = 22.5 . Firm 1 has lower marginal costs per se, firm 2 has a higher willingness to pay of customers in its own market segment due to vertical product differentiation. parameter firm 1 firm 2 willingness to pay 𝑎𝑎 𝑖𝑖 of firm i 23.0 25.0 market size 𝑏𝑏 𝑖𝑖 of firm i 0.02 marginal costs 𝑀𝑀𝐶𝐶 𝑖𝑖 of firm i 2.0 2.5 industry specific fixed costs 𝐹𝐹𝐶𝐶 2000 competitiveness 𝑎𝑎 𝑖𝑖 - 𝑀𝑀𝐶𝐶 𝑖𝑖 21.0 22.5 Table 10.4: Small differences in the competitiveness of the firms. <?page no="390"?> 10 Strategic competition in oligopoly 390 The respective market segments and the industry-specific fixed costs are equal with 𝑏𝑏 1 = 𝑏𝑏 2 = 0.020 as well as with 𝐹𝐹𝐶𝐶 = 2000 . The only exogenous variable varied is the degree of horizontal product differentiation 𝛾𝛾 ∈]0; 1[ , which in reality is of course not fully exogenous as described above: rather, firms can invest in horizontal product differentiation via branding and marketing. Figure 10.11: Prices, capacities and profits given small differences in competitiveness depending on the degree of horizontal product differentiation. ► Figure 10.11 shows that, with a higher degree of horizontal product differentiation ( 𝛾𝛾 → 0 ), both firms successively choose pricing strategies and associated capacities that correspond to the monopoly outcome based on increasing market power. The more homogeneous the products are ( 𝛾𝛾 → 1 ), the lower the firms set prices in Cournot and Bertrand competition due to higher competitive intensity. For both firms in both models of competition profits decline as horizontal product differentiation decreases. However, this trend is convex in Cournot competition and S-shaped in Bertrand competition, and the decline in profits accelerates if product homogeneity is large. Similarly, price-cost margins PCM decline with decreasing horizontal product differentiation in Cournot competition but remain at a relatively high level. This is different in Bertrand competition: the PCM falls more than proportionately and eventually becomes negative due to the industry-specific fixed costs for the less competitive firm 1. The capacities or production quantities 𝑞𝑞 𝑖𝑖 of both firms show the expected effect in the case of Cournot competition. With decreasing horizontal product differentiation, the respective production quantities decrease less than proportionately, but in fact variance of market shares of both firms increases due to the assumed differences in competitiveness. A firm with <?page no="391"?> 10.4 Strategic competition with product differentiation 391 high competitiveness benefits more from a decrease in horizontal product differentiation. This is different with Bertrand competition where differences in competitiveness lead to the fact that, with a decreasing degree of horizontal product differentiation, the relatively more competitive firm 2 grows, the other firm 1 becomes smaller. The results show that if product differentiation is high (e.g., 𝛾𝛾 < 0.2 ), the competitive intensity and the market result differ only slightly between Cournot and Bertrand competition. As product differentiation decreases (e.g., 𝛾𝛾 > 0.4 ), it becomes apparent that, especially in Bertrand competition, there are greater incentives for product differentiation in order to achieve profits, since low competitiveness increases the risk of being forced out of the market. Now we consider the case in which the two competing firms show significant differences in their competitiveness. The parameters shown in ► Table 10.5 now apply to the demand and cost functions. parameter firm 1 firm 2 willingness to pay 𝑎𝑎 𝑖𝑖 of firm i 30.0 25.0 market size 𝑏𝑏 𝑖𝑖 of firm i 0.02 marginal costs 𝑀𝑀𝐶𝐶 𝑖𝑖 of firm i 22.0 2.5 industry specific fixed costs 𝐹𝐹𝐶𝐶 2000 competitiveness 𝑎𝑎 𝑖𝑖 - 𝑀𝑀𝐶𝐶 𝑖𝑖 8.0 22.5 Table 10.5: Large differences in the competitiveness of firms. The firms now show large differences in competitiveness with (𝑎𝑎 1 − 𝑀𝑀𝐶𝐶 1 ) = 8.0 and (𝑎𝑎 2 − 𝑀𝑀𝐶𝐶 2 ) = 22.5, with firm 1 achieving a higher willingness to pay among customers, however at significantly higher marginal costs. <?page no="392"?> 10 Strategic competition in oligopoly 392 Figure 10.12: Prices, capacities and profits given large differences in competitiveness depending on the degree of horizontal product differentiation. Again, it is obvious that the differences in market results due to having either Cournot or Bertrand competition are small from a firms’ perspective as long as there is a high degree of horizontal product differentiation: here the shape of strategies of prices and quantities and the resulting profits and profit margins run very closely together. Only when product differentiation decreases significantly - in the example chosen here, with a given parameter constellation, starting approximately from 𝛾𝛾 > 0.4 - do these curves diverge and the respective competitive strategies differ due to different competitive capabilities depending on the type of competition. It is also confirmed, that in case of Bertrand competition and a decreasing degree of horizontal product differentiation, firms react strategically with price reductions. As a result, PCMs decline, but firm 2 - after firm 1 successively reduces output to zero due to high marginal costs - is able to expand output and achieve higher absolute profits even with declining profit margins. However, from the shape of the curves in ► Figure 10.12, significant quantitative and qualitative differences are also evident. In the case of Cournot competition, the strategies of the firms clearly depend on the differences in competitiveness as the degree of horizontal product differentiation changes. Firm 2 exhibits a U-shaped pattern for all variables as the degree of product differentiation decreases. Initially prices and quantities are reduced from a monopoly position at 𝛾𝛾 → 0, however, as the intensity of competition increases due to an increase in 𝛾𝛾 , firm 2 is able to use its relatively higher competitiveness to increase both prices and quantities and also profits and profit margins. Substituting the parameters of competitiveness from ► Table 10.5 into equation (10.93), we obtain <?page no="393"?> 10.4 Strategic competition with product differentiation 393 (10.102) 𝛾𝛾 > 2�(𝑚𝑚1−𝑀𝑀𝐶𝐶1)−�(𝑚𝑚1−𝑀𝑀𝐶𝐶1)2−(𝑚𝑚2−𝑀𝑀𝐶𝐶2)2� 𝑚𝑚2−𝑀𝑀𝐶𝐶2 ≈ 0.3676 . At a value of 𝛾𝛾 ≈ 0.3676, prices, quantities and profits of firm 2 reach their minimum under the conditions given here, but from a value of 𝛾𝛾 > 0.3676, firm 2 can grow, increase prices and increase profits due to its relatively higher competitiveness, despite a decline in horizontal product differentiation (equivalent to an increase in 𝛾𝛾 ). Firm 1, on the other hand, must reduce prices and quantities at an increasing rate as competitive intensity increases due to a decline in horizontal product differentiation (equivalent to an increase in 𝛾𝛾 ), and has to accept a decline in the profit margin PCM. Absolute profits also decline continuously to a level of 𝜋𝜋 1 = −𝐹𝐹𝐶𝐶 in case where 𝑞𝑞 1 = 0 - in the example chosen here, this happens for Bertrand competition for values of 𝛾𝛾 > 0.58 . Firm 1 clearly has an incentive to maintain a high degree of horizontal product differentiation. From a management perspective, the incentives for marketing and branding investments to build or expand horizontal product differentiation are clearly stronger for firms with low relative competitiveness. Competitiveness and horizontal product differentiation are strategic substitutes: if a firm has low competitiveness, it may try to evade competition by investing heavily in horizontal product differentiation. Moreover, less competitive firms have a clear incentive to compete on capacity rather than on pricing. This is not necessarily the case for firms with higher competitiveness, as can be seen in ► Figure 10.11 for firm 2 but depends on the degree of horizontal product differentiation. Firms choose very different strategies with regard to horizontal product differentiation depending on Cournot or Bertrand competition. Prices in Bertrand competition are lower than in the case of Cournot competition under otherwise identical conditions; conversely, quantities are higher in the case of price competition than in the case of capacity competition. This can be explained via the (relative) competitiveness of the firms. Firms with high competitiveness - relative to competitors - are typically able to survive even with a low degree of product differentiation. Firms with relatively low competitiveness have strong incentives for horizontal product differentiation - with low horizontal product differentiation their ability to survive is reduced. Price competition à la Bertrand reinforces these effects - capacity competition à la Cournot weakens them. If firms are sufficiently heterogeneous, Cournot and Bertrand competition give rise to asymmetric effects on price and capacity strategy. Firms with low relative competitiveness have a strong interest in establishing or expanding horizontal product differentiation. In empirical studies, one can expect that the firm-specific marketing expenditures for the establishment of horizontal product differentiation will be higher. Firms with strong capabilities on the product and/ or process side do not have this interest - they may try to force other firms to exit the market through reduced horizontal product differentiation and expand their own profits. <?page no="394"?> 10 Strategic competition in oligopoly 394 10.5 Relevance for competitive strategies It has become clear during the previous sections that any development of a competitive strategy must take into account numerous parameters which, moreover, impact each other. What type of competition is at hand: Cournot or Bertrand? Do firms take decisions simultaneously, or is there an implicit understanding of roles with a Stackelberg market leader and sequential decisions? Is the number of firms limited by entry barriers, or does the intensity of competition change as a result of market entry? Are products homogeneous, or is there vertical and/ or horizontal product differentiation? Do all firms have similar competitiveness, or is it significantly different? In fact, answering these questions first requires a comprehensive market and competition analysis, which is either carried out by a firm's own strategy department or with the support of management consultants. Despite the usefulness of the application of Cournot and Bertrand models, this is not daily practice in firms. However, the findings from ► Sections 10.1 to 10.3 show how they can indirectly guide decision-making in strategic competitive situations. Basic ideas of Cournot and Bertrand models for strategy implementation Many of the basic strategic ideas of Cournot and Bertrand competition have - not least as a result of the work of Porter (1980 and 1985) - meanwhile found their way into many strategy workshops and discussions, however, in a significantly simplified form, and often without the essential quantitative implications. Degree of product differentiation - product differentiation is desirable because the repercussions of competitors' strategies on one's own profits are weakened (in Porter's words as a differentiation strategy). The basic questions are: how easy (or at what cost) can permanent product differentiation be built up, and can product differentiation with high sunk costs also act as a barrier to entry? Competitiveness - differences in competitiveness (product quality or efficiency) can lead to the shake-out of firms (Porter calls it cost leadership). The basic questions are: can one's own competitiveness be permanently increased, how quickly do competitors catch-up, and can cost leadership create some sustainable competitive advantage? Moreover, the two-stage nature of decision-making is implicit in decisions of firms, so that in many cases competition can actually be interpreted as a two-stage game: during a first stage, medium and long-term capacity decisions are made and determined by irreversible investments; so that in a second stage price competition takes place to signal competitiveness to customers and indirectly to competitors. In addition, guidance for strategic decisions follows from the analysis of whether Cournot or Bertrand competition prevails. As shown in ► Figure 10.13, the strategic parameters differ in their interplay. In Cournot competition, capacities are strategic substitutes - if one firm can implement a growth strategy due to improved competitiveness, the other firm will have to become smaller: the reaction curves are downward-sloping, and the strategies develop in opposite directions. <?page no="395"?> 10.5 Relevance for competitive strategies 395 Figure 10.13: Strategic substitutes and complements. In Bertrand competition, prices are strategic complements - if one firm can implement higher prices due to improved competitiveness, the other firm will also increase its prices: the reaction curves are upward-sloping, and the strategies develop in the same direction. From a management perspective, however, there is - as with many methods for strategic decision-making - an implementation hurdle. As was made clear in Sections ► 10.1 to 10.4, the quantitative strategies developed on Cournot and Bertrand competition require an understanding and comprehensibility of the mathematics used - formulas alone often imply a natural barrier here. However, this can be significantly reduced by using heat maps to explain results, as can be seen in ► Figure 10.14. Heat maps clearly show both the current competitive situation and the feasible strategic options by means of colouring or structure, analogous to a traffic light system. In this way, alternative strategies can be evaluated and analysed in a highly objectified and structured manner in discussions with management during strategy workshops, and strategies that increase profits can easily be separated from those that reduce profits. Empirical evidence, real behaviour of managers, and situations of strategic competition in experiments Empirical studies (Aiginger 1996a and 1996b as well as Cherchye et al. 2013) directly or indirectly investigate, whether, firms actually use Bertrand or Cournot strategies. The following two findings are regularly observed. Market structures and competitive processes point to two-staged competition. The market result in market shares and profits is partly in line with Cournot competition - however, positve profits are often observed even in Bertrand competition without product differentiation, which may indicate either collusion or limited aggressive behaviour in competition. <?page no="396"?> 10 Strategic competition in oligopoly 396 Figure 10.14: Heat maps of Cournot and Stackelberg competition. Cournot-competition with three firms („heat map“ of capacities and profits) q 1 profits of firm 1 q 2 + q 3 = <?page no="397"?> 10.6 Summary and key learnings 397 In a direct survey of 930 top managers, 38% say they 'play Cournot', 62% rather see competition according to the Bertrand model (Aiginger 1999) - which may be due to a twostage game and rare long-term capacity decisions as opposed to regular discussions on pricing strategies. In order to test the effectiveness and plausibility of the strategies, laboratory experiments are being conducted. The aim here is to monitor actual competitive behaviour under controlled conditions and to test whether, and how quickly, for example, Cournot-, Bertrandor Stackelberg-Nash equilibria are reached (Armstrong and Huck 2010, Holt 1993, Huck et al. 1999 and Huck et al. 2000). The results of these experiments largely confirm boundedly rational behaviour and insufficient depth of reasoning in strategic decision situations as shown below. Bertrand competition - typically players do not determine price floors in the experiments but test different prices by trial and error. Nevertheless, profits are made, albeit in small amounts - accordingly, a Nash equilibrium does not come about. In addition, an extended tit-for-tat behaviour emerges: attempts are made to punish other players' profits by undercutting prices; spontaneous parallel behaviour or collusion that emerges is usually destroyed again in the next period. Cournot competition - satisficing prevails, i.e., one's own strategies are changed only little if positive and deemed sufficient profits repeatedly occur, even if the profits of other players turn out to be higher. Again, players do not calculate optimum strategies although all relevant information is available. In fully transparent games (with evidence of competitors' profits and disclosure of others' chosen strategies), markets are more competitive - i.e., lower prices and lower profits tend to emerge. Stackelberg competition - although the roles are clearly dedicated in the laboratory experiment, the asymmetry of market shares is less pronounced than predicted by the various models, and over time battles between market followers with too large quantities arise, so that Stackelberg market leaders partially give up their leading role. Often in these laboratory experiments - in addition to bounded rationality and satisficing (► Chapter 3) - revenge, belief in a strategy, and not taking competitors into account play a much greater role than strategic rationality and typically reduce profits significantly. 10.6 Summary and key learnings Strategic competition in an oligopoly implies that firms mutually analyse the strategies of their competitors and take these into account for the development of their own strategy. In addition, the business environment of market structure, demand structure, and technological opportunities must be taken into account for the choice of strategic parameters. Competition can take many different forms, but three types of competition have proven to be at the core of the analysis of strategic situations: Cournot competition, Bertrand competition, and Stackelberg competition. In Cournot competition, firms decide on quantities or capacities. Cournot competition is generally to be expected if capacities can be changed only at great cost and in the long run. The resulting Nash equilibrium depends on marginal costs, the number of firms, and <?page no="398"?> 10 Strategic competition in oligopoly 398 demand structure. Firms typically earn positive profits; the higher the relative competitiveness of a firm, the larger its market share and profit. In Bertrand competition, firms decide on prices. Bertrand competition can be expected if capacities can be changed at low cost and quickly. The resulting Nash equilibrium again depends on marginal costs: however, firms typically make profits only with a sufficient degree of product differentiation; without product differentiation, mutual price undercutting leads to a result similar to perfect competition. Stackelberg competition describes competitive situations in which, due to an accepted understanding of roles and credible strategies, a market leader can achieve a first mover advantage. The profits of a Stackelberg market leader based on a sub-game perfect Nash equilibrium exceed the profits in Cournot competition. The applicability and success of the strategies are significantly influenced by a firm's competitiveness and the degree of product differentiation in an industry. Horizontal and vertical product differentiation can significantly reduce the intensity of competition, so that especially in Bertrand competition and generally for less competitive firms there are incentives to ensure survival through strategic investments in marketing. Irrespective of this, the implementation of these strategies requires far-reaching market and competitive analyses; in addition, acceptance for the chosen methods must be established in a suitable manner - e.g., in the form of heat maps, instead of showing equations and formulas to C-level managers. Recommendations for further reading A full range of models covering strategic competition in oligopoly is presented in Pepall, L., Richards, D. and Norman, G., Industrial organization - contemporary theory and empirical applications, Hoboken 2014, and in Belleflamme, P. and Peitz, M., Industrial organization: markets and strategies, London 2015, or in strong combination with game theory in Pfähler, W. and Wiese, H., Unternehmensstrategie im Wettbewerb, Berlin 2008. Questions for review [1] Describe applications of the analysis of strategic competition in oligopoly from a microeconomic perspective as well as their limits, advantages and disadvantages. [2] Briefly explain the basic considerations of Cournot, Stackelberg and Bertrand competition and the main results with regard to quantities, price and profit. Give two examples each from different industries. Explain the practical relevance and areas of application of Cournot, Bertrand and Stackelberg competition for the development of competitive strategies. [3] Explain the different market outcomes of Bertrand competition with and without product differentiation. [4] Two firms in a market without product differentiation have a free choice to compete either on quantities or on prices or to invest in product differentiation - what would you recommend they do? <?page no="399"?> 10.6 Summary and key learnings 399 [5] A firm in a market without product differentiation is considering a Stackelberg market leadership - what are necessary requirements to make this a successful effort? [6] Describe and explain the relationship between the number of firms and the market outcome (prices, quantities, profits) in Cournot competition? [7] There are two firms competing in credit card payment systems: M-Card and V-Card. The firms definitely compete in quantities of connected terminals, i.e., production capacity. From experience, both know: (1) the demand function is 𝑝𝑝 = 5,000 − 2 ⋅ (𝑞𝑞 𝑀𝑀 + 𝑞𝑞 𝑉𝑉 ) , (2) marginal cost is 𝑀𝑀𝐶𝐶 𝑀𝑀 = 180 and 𝑀𝑀𝐶𝐶 𝑉𝑉 = 140, (3) fixed costs are 𝐹𝐹𝐶𝐶 = 100,000 for each firm. You are acting as a consultant for M-Card, currently the firms are competing simultaneously. Consider whether M-Card should strive for market leadership from an economic perspective (compared to simultaneous competition), and if so, how much should M-Card invest (at a maximum) in market leadership? [8] Describe and explain the effects of product differentiation in Cournot and Bertrand competition. [9] Two firms A and B compete in NFC readers in quantities, i.e., production capacity. From experience, both know: (1) the demand function is 𝑝𝑝 = 1000 − 2 ⋅ (𝑞𝑞 𝐴𝐴 + 𝑞𝑞 𝐵𝐵 ) , (2) marginal costs are 𝑀𝑀𝐶𝐶 𝐴𝐴 = 240 and 𝑀𝑀𝐶𝐶 𝐵𝐵 = 280, (3) fixed costs are 𝐹𝐹𝐶𝐶 = 100,000 for each firm. For Cournot competition, find the market shares and profits of both firms. Show and explain with the help of a figure how the Cournot-Nash equilibrium changes if one of the two firms is able to reduce its marginal costs. Using this figure, show and explain the effects of changes in marginal costs, willingness to pay and market size on the output quantities of two firms under Cournot competition. [10] In the credit card industry, two firms A and B compete, the main competition parameters are the respective quantities 𝑞𝑞 𝐴𝐴 and 𝑞𝑞 𝐵𝐵 measured in novel connected terminals. The firms know from experience that (1) the demand function for both firms is given 𝑝𝑝 = 2,000 − 2 ⋅ (𝑞𝑞 𝐴𝐴 + 𝑞𝑞 𝐵𝐵 ) , (2) marginal costs are 𝑀𝑀𝐶𝐶 𝐴𝐴 = 280 and 𝑀𝑀𝐶𝐶 𝐵𝐵 = 360, (3) fixed costs for each of the firms are 𝐹𝐹𝐶𝐶 = 10,000 . (a) Determine the optimum output for firm A using a Cournot model. What is the consequence of the different marginal costs for the size and profits of the firms? (b) You are considering proposing a first mover strategy to firm A. What conditions must be fulfilled? What quantity and profits result from applying the Stackelberg model? (c) Finally, you consider a merger of the firms into a monopoly - marginal costs of the merged firms are 𝑀𝑀𝐶𝐶 𝐴𝐴𝐵𝐵 = 𝑀𝑀𝐶𝐶 𝐴𝐴 = 280, fixed costs are 20,000 . Does this merger make economic sense for the two firms, which effects will occur?