Carsharing STRATEGIES International Transportation | Collection 2020 17 A study on free-floating carsharing in Europe Carsharing, Private vehicle holding, Modal shift How many carsharing users sell their vehicle, how many users suppressed a vehicle purchase? This article outlines the impacts of car2go and DriveNow on modal shift, vehicle owner-ship, vehicle kilometers travelled, and CO 2 emissions in 11 European cities. Hansjörg Fromm, Patrick Jochem C arsharing is an important segment of the sharing economy. The sharing economy strives for more efficient use of resources with positive economic, social, and environmental impacts [1]. In a new culture of nonownership, people increasingly prefer temporary access to resources over permanent ownership of resources. In Europe, the number of carsharing users has grown from 200,000 in 2006 to 6.76 million in 2018 [2] and is expected to increase to 15.6 million through to 2020 [3]. Free-floating carsharing, i. e. carsharing that allows pick-up and return of a car anywhere within a specified area in a city, has been on the market for more than 10 years. It is mainly provided by automotive and rental car companies. After a decade of operation and user experience, an evaluation seems appropriate. Car2Go and DriveNow, who merged into Share Now in 2019, are the largest carsharing operators in the world. They are serving over 3 million users, even if they closed their North American operations in late 2019. Share Now commissioned a study to identify the impact of carsharing on vehicle holdings, modal shift, vehicle kilometers traveled, and greenhouse gas emissions. The study was conducted in 2018 (Car2go) and 2019 (DriveNow). It is based on a survey among car2go and DriveNow customers, now customers of Share Now, in 11 European cities. The complete report can be found on our KIT website [4]. A previous study was performed by the University of California, Berkeley, for 5 North American cities in 2016 [5]. Over 10,000 carsharing users regularly using the service participated in the European online survey. We ensured representativeness of the respondents [4]. In the survey, participants were asked detailed questions on how the availability of car2go or DriveNow changed their travel behavior and vehicle ownership. Besides questions regarding the personal demographics of the participants (gender, age, income, education level), a group of questions centered around their change in travel behavior (e. g. trips overall, carpooling) and mode choice (e.g. use of taxis and public transportation). Another group of questions concerned the change in vehicle holdings. Impacts on vehicle holdings To study the impact on private vehicle holdings, participants were asked to specify how many and which cars they had owned before and after subscribing to the carsharing services and if they had sold or acquired cars within this time period. They were asked if they had attributed the sale or acquisition entirely or partially to the services provided by car2go or DriveNow. These answers allow us to determine the absolute number of sold vehicles and the percentage of participants who sold a vehicle. The number of acquired cars that the participants attributed to the carsharing service is negligible. There is a significant other effect of the availability of carsharing: people forego or postpone the acquisition of a car which they otherwise would have purchased. In order to estimate the number of suppressed vehicle purchases, we asked the hypothetical question “Would you acquire a car if car2go or DriveNow disappeared from your region? ”. If the answer is “yes” and this is not simply the replacement of a previously sold car, then we count this as a suppression of a vehicle. Other than the response concerning a vehicle sale, the response concerning vehicle suppression is not verifiable. Nevertheless, we handle the respondents’ information on suppressed vehicles in the same way as we handle the information on sold vehicles although the latter is more reliable. Figure 1: Percentage of participants who sold a vehicle and percentage of participants who suppressed a vehicle purchase STRATEGIES Carsharing International Transportation | Collection 2020 18 Figure 1 presents, for each individual city, the percentage of participants who sold a vehicle and the percentage of participants who suppressed a vehicle purchase. Assuming sample representativeness, these percentages can be applied to the overall population of regular car2go and DriveNow users. This gives us an estimate of the total numbers of sold and suppressed vehicles in each city. Table 1 shows the estimates of sold and suppressed vehicles in absolute numbers and on a per-carsharing-vehicle basis. According to our study, between 3.6 % and 16.1 % of the regular carsharing users in the individual cities claimed to have sold a vehicle due to the carsharing service. Between 13.3 % and 40.7 % claimed to have suppressed the purchase of a vehicle. This would amount to 18,948 vehicles sold and 62,900 vehicles suppressed across all cities. Berlin and Hamburg stand out with 4,616 resp. 3,100 vehicles sold and 11,834 resp. 11,020 vehicles suppressed. Per carsharing vehicle, between 2.1 and 5.3 vehicles have been sold in the individual cities and between 7.8 and 18.6 vehicles have been suppressed. We asked the participants for the reasons why they sold a car. For most participants, one reason was to save costs by getting rid of a car. In Berlin, Vienna, Hamburg, and Helsinki, many participants stated that good public transportation infrastructure was important for their decision. In these cities, a comparably high percentage of sold vehicles can be observed. Other reasons indicated by car2go and DriveNow users are: carsharing sufficiently fulfills their mobility needs, carsharing addresses the scarcity of parking space within the cities, and carsharing contributes to environmental protection. From a demographic point of view, candidates for selling their car or not purchasing a car are especially young people with a relatively high educational background and people who live in small households. This confirms insights from previous studies [6]. Impacts on modal shift Carsharing has a considerable impact on the participants’ choice of other transportation modes. Participants were asked to what extent they had changed their use of urban rail, bus, taxi, intercity rail, bicycles, and motorcycles, and whether they had changed their walking habits. There is a general trend that carsharing users reduce the use of taxis and - to a lower extent - the use of urban rail and busses. Only a few participants report that they have increased the use of public transportation. This means that some participants complement the use of carsharing with other transportation modes, while other participants substitute other transportation modes by carsharing. Table 2 shows the respective numbers exemplarily for Berlin on the left. The results for other cities show noticeable differences in magnitude, even if the general trends are largely the same [4]. This behavioral change is certainly influenced by several factors, such as the individual mobility demand of the participant within the geography, the structure of the public transportation network, and the question if the participant has reduced private vehicles. While the first two factors are difficult to grasp, the latter is known from the participant’s survey response. To assess the impact of private vehicle reduction on modal shift, we separately evaluated the subgroup of users who have sold a car. The results are remarkable. Of the participants who have sold a car, a higher percentage increases the use of public transportation and a lower percentage reduces the use of public transportation. Additionally, a higher percentage of these participants increases the use of bicycles and the quantity of walking. For example in Berlin (table 2 on the right), the share of participants who increased the use of urban rail grows from 9.7 % to 31.8 %, if we consider the subgroup of users who have sold a car. For intercity rail, the percentage increases from 9.9 % to 32%, for bus from 7.1 % to 21.3 %, for bicycle from 13.9 % to 39.8 %, for walking from 16.1 % to 28.8 %. Even if the percentage of participants who sold a car is still moderate (3.6 % to 16.1 % for the different cities, cf. figure 1), the removal leads to considerable reductions of private vehicle kilometers traveled. The relatively high mode shift percentages of participants who sold a car indicate where these saved kilometers might have gone: to public transportation in combination with walking and biking. Impacts on vehicle kilometers traveled The numbers of vehicles sold and suppressed allow us to estimate the impact of carsharing on vehicle kilometers traveled (VKT). From the responses of the participants, neither the VKT of a sold vehicle nor the VKT of a not purchased vehicle can be directly determined. Therefore, we make assumptions for both cases. If participants sold a vehicle, we assume that they sold the vehicle with the lowest VKT in their ownership and count this VKT as a reduction. For a suppressed vehicle purchase, we base our VKT estimate on the average VKT of all participant-owned vehicles in the corresponding city. Our conservative scenario assumes that the suppressed vehicle would have been used with 20 % of this average VKT, and our optimistic scenario assumes that the suppressed vehicle would have been used with 80 % of this average VKT. As a replacement of their private cars, customers will now use - together with other transportation modes - carsharing vehicles. Therefore, the VKT reduced by sold and suppressed vehicles must be netted with the VKT that customers drive with carsharing vehicles. The VKT estimates for sold vehicles are roughly around 12,000 kilometers per year, the VKT estimates for suppressed vehicles roughly between 2,000 (conservative) and 14,000 (optimistic) kilometers per year. This leads to annual VKT reductions e. g. for Helsinki between 9.7 (conservative) and 21.3 (optimistic) million kilometers and for Berlin between 95.1 (conservative) and 190.7 (optimistic) million kilometers. Estimation of total vehicles sold Vehicles sold per SHARE NOW vehicle Estimation of total vehicles suppressed Vehicles suppressed per SHARE NOW vehicle Amsterdam 1.060 3,2 3.340 10,2 Berlin 4.616 4,4 11.834 11,4 Brussels 1.512 5,3 1.481 8,6 Copenhagen 1.367 3,2 6.835 18,6 Hamburg 3.100 3,8 11.020 13,4 Helsinki 424 2,9 1.212 9,0 Lisbon 425 2,1 2.009 10,4 London 868 2,4 3.629 13,3 Madrid 954 2,1 3.846 8,4 Rome 2.388 3,8 8.953 14,4 Vienna 2.234 3,2 5.356 7,8 Table 1: Estimates of sold and suppressed vehicles in absolute numbers and per-carsharingvehicle Carsharing STRATEGIES International Transportation | Collection 2020 19 Impacts on CO 2 emissions The impact on CO 2 emissions is a direct consequence of the VKT reductions. VKT reductions are transformed into CO 2 emission reductions by multiplying them with the official emission factors (g CO 2 per km) for the individual countries obtained from the Eurostat database [7]. Since sold vehicles are reported to be approximately 10 years old, we took the average emission factors for newly registered vehicles in 2008/ 2009. For suppressed vehicles, we used the 2016/ 2017 factors. According to the International Council on Clean Transportation (ICCT), real use phase emissions are on average 40 % higher than the officially reported testing cycle emissions (NEDC) [8]. Therefore, we added 40 % to the Eurostat emissions in our calculations. To obtain the net reduction of CO 2 emissions induced by the carsharing service, we had to balance the emission reductions caused by reduced and suppressed private vehicles with the emission increase caused by the carsharing fleet. For fleet vehicle emissions, we used model-specific factors of the actual fleet, again with the 40 % adjustment as described above. Table 3 shows, for each individual city, the net CO 2 emission reductions in absolute numbers and on a per-carsharingvehicle basis. They range between 1,737 and 17,865 tons/ year for the conservative scenario and between 4,233 and 34,864 tons/ year in the optimistic scenario. Netted CO 2 emission reductions per Share Now vehicle are between 8.7 and 17.2 tons/ year in the conservative scenario and between 21.1 and 33.5 tons/ year in the optimistic scenario. Conclusion After free-floating carsharing offerings have been available in major European cities for more than 10 years, positive impacts on vehicle holdings, modal shift, and CO 2 emission reductions can be observed. Although only a moderate percentage of car2go and DriveNow users have sold a vehicle or suppressed the purchase of a vehicle, the reductions in private vehicle holdings sum up to a considerable number. As a consequence, each fleet vehicle of the carsharing service turns out to remove a multiple of private vehicles. This frees up space in the streets (roadside parking) and parking lots of the cities. and fosters public transportation as a complement to carsharing. Removal of private cars leads to VKT reductions measured in tens or hundreds of millions of kilometers in each city, and, as a consequence, to CO 2 emission reductions in thousands of tons per year. The willingness of carsharing users to sell a car depends on many factors. One factor is certainly the coverage, performance, and ease of use of the public transportation system in the city and its surroundings. Another factor is the diminishing desire for car ownership with its associated fixed costs. But above all, we could observe that the willingness to sell a car increases with the time that the carsharing system has been in operation in the particular city. Therefore, it can be expected that in the coming years the trend to reduce private vehicles with its positive effects will continue. ■ LITERATURE [1] Martin, C. J. (2016): The sharing economy: A pathway to sustainability or a nightmarish form of neoliberal capitalism? . In Ecological economics, 121, pp. 149-159 [2] Shaheen, S. and Cohen, A. (2020): Innovative Mobility: Carsharing Outlook Carsharing Market Overview, Analysis, And Trends [3] Deloitte (2017): Car Sharing in Europe. Business Models, National Variations and Upcoming Disruptions [4] Fromm, H.; Ewald, L.; Frankenhauser, D.; Ensslen, A.; Jochem, P. (2019): A Study on Free-floating Carsharing in Europe: Impacts of car2go and DriveNow on modal shift, vehicle owner-ship, vehicle kilometers travelled, and CO 2 emissions in 11 European cities, Working Paper Series in Production and Energy 36, doi: 10.5445/ IR/ 1000104216, https: / / publikationen.bibliothek.kit.edu/ 1000104216/ 51584214 [5] Martin, E. and Shaheen, S. (2016): Impacts of car2go on vehicle ownership, modal shift, vehicle miles traveled, and greenhouse gas emissions. Working Paper [6] Schmöller, S.; Weikl, S.; Müller, J.; Bogenberger, K. (2015): Empirical analysis of free-floating carsharing usage: The Munich and Berlin case. Transportation Research Part C: Emerging Technologies, 56: 34-51 [7] Eurostat (2018). [online] Available at: http: / / ec.europa.eu/ eurostat/ tgm/ table.do? tab=table&plugin=1&language=en&pcode=t2020_ rk330 [8] Tietge, U.; Zacharov, N.; Mock, P.; Franco, V.; German, J.; Bandivadekar, A.; Ligterink, N.; Lambrecht, U. (2015): From laboratory to road - a 2015 update of official and “real-world” fuel consumption and CO 2 values for passenger cars in Europe, ICCT White Paper Hansjörg Fromm, Prof. Dr.-Ing. Honorary Professor at the Karlsruhe Service Research Institute (KSRI), Karlsruhe Institute of Technology (KIT), Karlsruhe (DE) hansjoerg.fromm@kit.edu Patrick Jochem, PD Dr. rer. pol. Lecturer at the Institute for Industrial Production (IIP), Karlsruhe Institute of Technology (KIT), Karlsruhe (DE) jochem@kit.edu Table 2: Modal shift due to carsharing for all participants and participants who sold a car (Berlin) All Participants Participants who sold a car Increased use No change Decreased use Increased use No change Decreased use Taxi 3,8 % 24,8 % 71,4 % 8,3 % 27,3 % 64,4 % Urban Rail 9,7 % 54,2 % 36,1 % 31,8 % 40,2 % 28,0 % Bus 7,1 % 58,6 % 34,3 % 21,3 % 50,0 % 28,7 % Intercity Rail 9,9 % 85,5 % 4,6 % 32,0 % 61,5 % 6,6 % Bicycle 13,5 % 70,4 % 16,1 % 39,8 % 54,2 % 5,9 % Walking 16,1 % 67,8 % 16,1 % 28,8 % 62,3 % 8,9 % Reduced CO 2 emissions due to SHARE NOW in tons per year (conservative - optimistic scenario) Estimation of total vehicles sold Reduced CO 2 emissions per SHARE NOW vehicle in tons per year (conservative - optimistic scenario) Amsterdam 4,887 - 10,126 14.9 - 30.9 Berlin 17,865 - 34,864 17.2 - 33.5 Brussels 4,343 - 6,423 15.2 - 22.5 Copenhagen 5,857 - 14,915 13.7 - 34.9 Hamburg 11,877 - 27,570 14.5 - 33.6 Helsinki 1,744 - 3,492 11.8 - 23.7 Lisbon 1,737 - 4,233 8.7 - 21.1 London 3,571 - 7,730 10.0 - 21.6 Madrid 4,087 - 9,074 8.9 - 19.9 Rome 7,878 - 18,550 12.7 - 29.8 Vienna 6,604 - 14,013 9.6 - 20.3 Table 3: Net CO 2 emission reductions due to Share Now in absolute numbers and on a percarsharing-vehicle basis