Car-sharing, an incomplete response to the need to reduce car traffic and the challenge of the ecological transition in mobility

Nolwen Biard, septembre 2023

Forum Vies Mobiles

Public policy has been focusing on car sharing for a number of years, but it has only recently been put on the mass agenda. Legislative changes relating to the LOM (orientation law on mobility – 2019 have led to a whole series of reconfigurations in the car-sharing ecosystem. They have also helped to reconfigure practices. The public policies deployed and their impact can be analysed in terms of the areas of relevance defined in the previous section. Beyond the immediate effects observed, we will also take into account the role of car sharing in a longer-term perspective.

À télécharger : 2023.09.11_vf_etude_covoiturage.pdf (5 Mio)

Although boosted by public funding, car-sharing practices encouraged by public policies are still very much in the minority

Intermediated car sharing still marginal despite growing commitment from public authorities

The rise of platform-based car sharing has been particularly strong, especially since 2022: an inflationary context has combined with the increased implementation of a variety of schemes by local authorities across France. The number of journeys rose from over 220,000 in January 2022 to over 780,000 in January 2023. In 2023, the car-sharing bonus boosted the momentum and the average number of journeys per working day doubled 1. The « boom » in car sharing is reported in the media and in the communications of operators and local authorities by three-figure growth rates. For some local authorities, such as Rouen Métropole, the number of journeys recorded by the RPC has increased tenfold in one year.

The role of the financial incentives distributed by the AOMs to carpoolers is decisive in the growth of carpooling journeys via platforms. In 2022, 95% of the journeys recorded by the RPC were incentivised 2. Furthermore, in the ranking of areas recording the most journeys published on the Carpooling Observatory, in 2022 we find almost exclusively areas that have set up local financial incentive schemes, in partnership with one or more operators. These operators note a strong correlation between the local authority’s contribution and carpooling volumes.

The AOMs pay drivers different levels of incentives, depending on the number of passengers and the number of kilometres travelled. For passengers, the journey may be free (with the AOM paying the driver in full) or paid for. Drivers are paid according to the number of passengers, generally at a rate of 10 centimes per km and per passenger. Thresholds and ceilings mean that average incentive levels vary considerably from one local authority to another. Since the introduction of the LOM, local authorities have been authorised to provide incentives that exceed the cost of the journey 3. The incentive lever for drivers can be particularly strong: in some incentive campaigns launched by AOMs, the cost per passenger has been as high as €6, for example.

According to this operator, car-sharing is now a « drip-feed » practice designed to « bring users onto the platforms »: « [Car-sharing] is « booming » today because local authorities have agreed to give drivers three to five euros per journey ». The Plan covoiturage bonus scheme has accentuated this effect, awarding €25 for the first journey, then a further €75 once 10 journeys have been made.

The financial incentives distributed to drivers, as well as the commissions distributed to certain operators for each registered journey, result in budgets that increase with the number of journeys, which is difficult for the local authority to manage. With such a model, the increase in the number of journeys can quickly become a « bottomless pit », as a technician from Tisseo collectivités describes it: « Unlike public transport, where you control the situation, you have a service offer, and when there are too many people you add a public transport offer, for car sharing it can be exponential, because the practice multiplies and you end up paying a new driver, a new commission to the operator. The more carpoolers there are, which is a good thing, the more the local authority will pay ». Some partnerships are even counterproductive for car-sharing policy, because the amount of commission paid to operators increases as the number of journeys increases.

Based on the current model of incentivised car sharing, the budget that would have to be allocated to incentives alone would be substantial if we were to follow the objectives set by the Car Sharing Plan. In fact, to increase the number of daily journeys made using this method from 0.9 to 3 million, at the level of remuneration recommended by the CPP (2 euros per passenger + 10 ct per km over 20 km), we would arrive at a budget of 5 million euros per day and 100.8 million euros per month for financial incentives to drivers alone4. To this sum should also be added the commission per journey paid to the operators and/or the operating costs of the car-sharing services. In the end, this would represent very high sums to finance intermediation. Several of the local authorities we met expressed their desire to withdraw from the financial incentive in the long term, but they feel that it is still too early to consider doing so, as they feel that the car-sharing policy is still in the development phase. For this technician from the Pôle Métropolitain du Genevois français (PMGF), financial incentives should be used as « a sort of bait to redirect them towards carpooling, and if possible on a long-term basis ».

In fact, despite the exponential growth in carpooling journeys via platforms, carpooling is still very much in the minority compared with other daily journeys. Nationally, journeys recorded by the RPC in 2022 represented an average of 0.04% of home-to-work journeys and 0.09% in the first quarter of 2023. This represents only about 1.5% of the number of journeys targeted by the Carpooling Plan (3 million daily journeys). Among the local authorities studied in this study, the modal share per kilometre of carpooling encouraged by local schemes varies, but is very low overall. On average, it represents 1.1% of the kilometres travelled for home-to-work journeys in the Rouen metropolitan area and the Seine-Eure agglomeration, in 2022, and 3% in February 2023; 0.71% of the kilometres travelled for home-to-work journeys targeted by the SMMAG at the beginning of February; or 0.18% of the kilometres travelled for journeys targeted by the Covoiturage Pays de la Loire scheme. In addition, these differences observed between local authorities do not systematically indicate differences in the effectiveness of public policies in increasing overall car sharing, but rather differences in the level of development of car sharing as measured by the RPC.

Platform carpooling is widely promoted, even though it remains much less widespread than informal carpooling

The « boom » in platform car sharing must also be re-examined in the light of informal car sharing practices. In the first quarter of 2023, platform carpooling represented only 3% of total carpooling 5. In the local authorities studied for this study, informal carpooling is still far superior. In the Toulouse conurbation, the journeys recorded during the Commute scheme are five times lower than the level of carpooling declared by employees in the mobility surveys carried out before the project; the journeys recorded by Nantes métropole represented an average of 1% of informal home-work carpooling journeys; in Rouen métropole, the levels of informal carpooling to work measured by the EMD 2017 are five times higher than the carpooling recorded by the RPC in 2022.

Informal carpooling, although largely in the majority, is invisible to the benefit of intermediated carpooling. This can be explained in part by a lack of knowledge of informal carpooling practices on the part of local players. For example, a technician from a metropolitan area said: « We’re starting from such a long way back, from zero, 0% carpooling before. Carpooling between home and work doesn’t exist, apart from a few colleagues who live nearby. That’s why we wanted to make the most of it ». However, in the interviews, this attitude was in the minority; the majority of respondents mentioned the existence of informal car-pooling practices, even though they did not know the profile of the car-poolers or how many there were. For this technician from a mixed syndicate: « It’s a real issue. We know that informal car-pooling exists, we see it at our park-and-ride sites. But it’s difficult to grasp, and we can’t manage to size it up at the moment. Various people we met, working for local authorities or the State, also expressed their desire to take better account of informal practices, even if no satisfactory solution seems to have emerged yet.

Despite this shared observation, elected representatives and technicians are placing more emphasis on intermediated car sharing: solutions linked to digital platforms are technologically « easier to promote » than informal car sharing (Frétigny, 2022). Digital platforms record journeys on a daily basis and provide « proof of carpooling ». This evidence can be used, particularly in the media or in public communications, as it provides proof of the effectiveness of the public policies implemented by local authorities. Communication based on the RPC figures can be used to promote the areas that « carpool the most », particularly through rankings.

In this way, measured carpooling journeys are used to legitimise their actions and can be promoted, whereas informal carpooling, organised without the help of public authorities, is largely « invisible ». As a result, car-sharing « suffers from less political support » when it is « self-organised or non-technological » (Frétigny, 2022). However, if the justification for supporting intermediated car-sharing (and the distribution of financial incentives attracting users to the applications) is based solely on the production of regular and valuable evidence, this amounts to « paying the thermometer » 6.

Car sharing is invisible to public and private players in their discourse, but it is also invisible in the public policies put in place. To date, the majority of funding has been earmarked to support car sharing as measured by the RPC, in particular for the distribution of financial incentives requiring the production of proof of car sharing to avoid fraud. These journeys must also be systematically recorded and geolocated via partner applications. Platforms that do not comply with this level of verification cannot offer subsidies to carpoolers, such as the regional public platform Ouestgo, or the Covoiteo platform of Tisseo collectivités.

The government’s car-sharing plan and the 3 measures announced for 2023 also concentrate two-thirds of the announced funding (€150 million) on financial incentives, which only concern intermediated car-sharers, and in particular drivers. The new feature of the bonus scheme is that new platforms can distribute it, such as OuestGo or the France Covoit scheme proposed by Ecov 7. The last measure announced in the Plan can benefit both informal and intermediated carpoolers, as it concerns the development of infrastructure to facilitate carpooling, which does not necessarily require the use of an application to use it.

The only other scheme that provides a financial incentive for informal car-pooling is the Forfait mobilité durable, if the employer includes car-pooling among the reimbursable modes of transport and does not make it conditional on the use of a platform 8. Lastly, the trial of a « positive toll » in Lille Métropole, due to be introduced in 2023, could also provide financial rewards for informal car-sharing practices 9.

Are public policies creating new practices?

In the absence of a complete overview enabling both informal and platform-based practices to be measured, it is difficult to know precisely the effects of public policies on levels of car sharing. There is in fact a risk of a deadweight loss effect due to financial incentives: people who were already carpooling informally may switch to the platforms in order to benefit from subsidies. While this may be seen as rewarding virtuous practices, it does not mean that public policies are creating new practices. « The proportion of additional practices and the proportion of the revelation of an informal practice are not clear from the CPP data » 10. By way of example, the Greater Geneva 2021 mobility panel shows equivalent levels of car-sharing between French and Swiss workers in the area (9% of respondents say they car-share occasionally), but a much higher level of registration on car-sharing platforms among French residents (15%) than among Swiss residents (3%). The risk of a windfall effect is known to the local authorities we met, and is often pointed out as a limitation when interpreting the journeys recorded by the RPC on their territory. Similarly, the payment of a commission to operators raises questions about journeys that were already being organised in the past, without their services.

The lack of qualitative survey data makes it difficult to assess the creation of new car-sharing practices. A recent survey carried out by the Pays de la Loire Region and Cerema showed that 53% of carpooling scheme users were new carpoolers, while 47% were already carpooling before. Of these, 11% were already carpooling for other reasons, but are now also carpooling for work or study. 21% of users of the scheme say they carpool daily and 44% several times a week.

The presence of subsidies can also lead to financial opportunism and fraud. Last February, for example, an article in the Journal d’Elbeuf revealed major fraudulent practices among Klaxit users in the Rouen metropolitan area 11. This « glove-box car-sharing », as the article put it, is carried out by drivers who are on their own for their journeys, but who use two smartphones to meet the application’s authentication criteria, or by public transport users who activate the application several times during their bus journeys. None of these practices is carpooling, although they have been recorded as such by Klaxit and paid for by the Rouen metropolitan authority. In the Nantes metropolitan area, data from the Covoit’tan system revealed « macro-users » making an abnormally high number of journeys per day, raising suspicions of possible fraud.

While fraud is detrimental to public transport usage figures (since the journey is not recorded), it artificially inflates the results of car-sharing services. The same problem can be found with the « free seats » remuneration offered by Ecov on the various carpooling lines deployed. While this system should make it possible to create the service and reach a sufficient critical mass of drivers who are integrated into the system and ready to take passengers in their vehicles with the least possible waiting time, this Ecov employee describes how certain drivers tried to hijack this system and how Ecov and the local authority integrated new conditions to avoid such practices:

« We became aware of a few abuses by drivers who only registered their journey on the application at Sainte-Julie [the penultimate stop on one of the Plaine de l’Ain Covoit’Ici routes], just before arriving, in order to collect the 50 cents for the free seat. As far as we’re concerned, it’s not useful for running the service. So from October [2022] there has been a change: the 50 cents can only be collected from the terminuses of the lines, on departure ». This change enables Ecov and the local authority to ensure that the drivers who are paid for the use of their seats cover the entire route and are really ready to take passengers. The local authority states that it is authorised to blacklist people who never take any passengers, even if the penalty has not yet been applied.

Carpooling thus represents a small proportion of daily journeys, and even more so when it comes to carpooling registered by the RPC. However, financial incentives have boosted this form of car sharing and created windfall effects and even fraud. Car-sharing policies have been characterised by a lack of targeting and steering on the part of the AOMs. Some of them are now including more methods in their new incentive campaigns to redirect use towards areas where car sharing is appropriate.

Inadequate public policies and lack of targeting limit the real impact on decarbonisation

A car-sharing network that is mainly created by users, with no public authority oversight, leading to a large number of irrelevant uses.

The vast majority of journeys recorded by the RPC in 2022 were made with planned car-sharing operators (Blablacar Daily, Karos, Klaxit). The network is thus created directly by users, who choose the pick-up and drop-off locations. As this operator said in an interview, « We don’t have any control over the definition of journeys ». In most of the financial incentive campaigns introduced since the LOM, there are very few conditions attached to the distribution of subsidies to carpoolers. Once again, this reflects the lack of consideration given to the relevance of carpooling. Car-pooling is deemed beneficial in any case, without considering the competition with already existing or relevant alternatives.

The absence of specific conditions on the journeys made is also due to the fact that some carpooling is encouraged without a local strategy, through non-targeted schemes. Prior to the Carpooling Plan, operators benefited in particular from the financing of their schemes and incentives for carpoolers via Energy Saving Certificates (CEE). Fuel or gift vouchers were offered by operators, who were able to encourage journeys without imposing any special conditions. The Carpooling Plan has added to this momentum by offering a bonus to first-time drivers, with no conditions attached to the type of journeys made. Journeys of up to 80 km benefit from the short-distance carpooling bonus, and those over 80 km can benefit from a bonus for long-distance carpooling. EECs also offer operators significant financial margins. With the new bonus for first-time drivers under the Carpooling Plan, the levels of energy savings associated with each new driver have been raised. Each EEC issued has a value greater than the amount of the bonus, which enables operators to earn a significant margin once the 100 euros has been paid to the driver 12. The operators’ financial margins enable them to encourage journeys outside the targets predefined by the AOMs. For example, the Syndicat mixte des mobilités de l’aire Grenobloise (SMMAG) concentrates its financial incentives on business parks or predefined routes. However, of all the journeys recorded in its area in February 2023, half (48%) were not incentivised by the AOM, but were the subject of subsidies via the driver bonus or incentives distributed directly by private operators.

Given the lack of steering, the majority of journeys were recorded in the densest areas where incentives had been put in place.

To analyse the deployment of carpooling journeys by platform, we used the typology of eight EPCI classes 13. There are significant differences between the EPCI classes, with 222 journeys per 1,000 inhabitants recorded over one year in the inner suburbs of Paris, compared with just 23 journeys per 1,000 inhabitants in the wider suburbs and just 5 journeys per 1,000 inhabitants in the rural EPCIs. Thus, the areas where most platform-based car sharing has been deployed are the most densely populated, where the modal share of public transport is higher.

This development of platform car sharing in densely populated areas can also be seen in the proportion of journeys made within the same municipality, in the heart of a conurbation. While in France as a whole, these internal journeys will account for 7.3% of journeys made in 2022, on a regional scale, they can account for a much larger proportion of total journeys. For example, 23% of journeys recorded in the Pays de la Loire region between September 2021 and March 2022 are internal to Nantes or Angers, due to local schemes set up by the conurbations. Within Tisseo collectivités, the AOM for the Toulouse conurbation, 10% of journeys will be within Toulouse in November 2022, and 10% will be between Toulouse and Blagnac, a neighbouring town that is also very densely populated. Although this phenomenon is mainly observed in metropolitan areas, it is also taking place in smaller towns that are developing partnerships with carpooling operators and financial incentive policies: in November 2022, the internal journey to Lannion (20,000 inhabitants, 466 inhab. /km2) was one of the 10 most car-pooled journeys in France, and in February 2023, the journey within Beauvais (a medium-sized town with 56,000 inhabitants and a population density of 1,708 inhabitants/km2) was the third most car-pooled journey in France.

Journeys within densely populated areas run counter to the decarbonisation objective described in the previous section, because within these areas, other alternatives exist or are relevant. Intra-communal journeys are on average very short (7.3 km in 2022) 14, even though the average distance of informal car-sharing journeys is 20 km 15, and 23.7 km for car-sharing registered by the RPC on average in 2022.

Over 2022 as a whole, a significant proportion of journeys are short or even very short distances: a quarter (23.2%) of journeys are less than 9 km and 7.6% less than 4 km. These journeys are therefore in direct competition with active modes 16 or urban public transport. Incentivised car-sharing adds to a range of solutions that are already available or relevant, calling into question in part the observation that it complements public transport.

In addition, long journeys have also been observed: in 2022, 8.7% of journeys are over 50 km and 16% are over 40 km. These are particularly long distances, especially for carpooling to and from work. The Angers metropolitan area, for example, recorded a quarter of journeys over 50 km in November 2022, with a large proportion of journeys going to Cholet and Le Mans. Between dense conurbations, public transport exists or would be relevant. While pooling journeys of such distances has positive effects on decarbonisation, the AOM has little interest in encouraging them. On the one hand, the longer the distance, the greater the interest in carpooling, as the time lost by carpooling is offset by the economic gain from sharing costs. On the other hand, providing financial incentives for journeys of more than 50 km, particularly if they are made on a daily basis, can encourage households to live very far from their place of work, with the local authority offsetting the cost of such journeys.

The challenge of data access and evaluation

The effectiveness of car-sharing policies cannot therefore be measured solely in terms of the number of journeys made. Access to and qualification of data are crucial to understanding usage and determining whether car-sharing is used as an addition to existing services, or whether it is a real alternative to car-pooling. Bruno Tisserand, Director of Mobility and Transport Operations at the Metropole, describes the difficulty of accessing the data: « We had to fight hard to get access to the data of the car-sharing operator in our area. It’s a start-up that’s just getting off the ground and doesn’t have a culture of working with local authorities. We had to demand that they qualify the origin-destination data, and we wanted to know more precisely what journeys were being made, so that we would know which links to oppose or mitigate ». The local authority found that on its most car-pooled route, between Rouen and Saint Etienne du Rouvray, only the university campus actually recorded car-pooling journeys, meaning that the system was used exclusively by students, and moreover on a route identical to one of the tramway lines serving the campus.

The production of data qualifying practices should enable a more detailed analysis of the environmental and social impact of carpooling incentive schemes. The recent evaluation by Cerema and the Pays de la Loire Region of the Covoiturage Pays de la Loire incentive scheme revealed, following a questionnaire distributed to users of the scheme, that 53% had not previously carpooled, and that 88% were former car-poolers 17.

In addition, some of the schemes do not involve registering or geolocating carpoolers, but organise, secure or promote carpooling or hitchhiking. This is the case, for example, with Rezo Pouce’s organised hitchhiking or simplified carpooling routes (simple illuminated signs with push-buttons, with no recording or geolocation of journeys). Some of these schemes are very simple, focusing on making carpooling more visible in the public space, using poles or illuminated signs. Jacques Toulemonde, chairman of the Autosbus association, discusses the conclusions of a survey carried out by his association comparing the effectiveness of organised hitchhiking and « free » hitchhiking: « During the first 2 years [of the association], we did a lot of research, including a tour of European experiences of organised hitchhiking. We were disappointed with the results when we went out and tried it out in the field: organised hitchhiking was no better than free hitchhiking in terms of waiting time, which averaged 6 minutes. However, he adds: « Rezo Pouce gives a better image of hitchhiking. Free hitchhiking means a 5-minute wait, but nobody does it. If, thanks to organised hitchhiking or a carpooling service, we also have a 5-minute wait, but a lot more people use it, it will be a success.

The development of car sharing areas is also a public policy whose effectiveness is more difficult to measure. In the ADEME 2017 guide, the development of carpooling areas is an investment deemed to be « particularly effective », based on the following estimate: « On the basis of an average implementation cost of €2,000 excluding VAT and taking into account the specific features of carpooling via a dedicated area, the amount of public money invested thus amounts to around €0.50 per journey carpooled and €0.01 per kilometre carpooled. » This particularly optimistic assessment, showing a cost per journey well below that currently practised by local authorities in their various partnerships with car-sharing operators, seems questionable. Other uses can be observed at car sharing areas: meeting points for exchanges, break areas for motorists, traditional parking for carrying out a nearby activity, etc. It is particularly difficult to control the uses induced by a car sharing area. First of all, part of the development policy for these areas is based on labelling car parks where car sharing is already taking place. Accreditation often involves putting up a carpooling sign and carrying out a few refurbishments. Surveys of car park users report general satisfaction with the facilities, which suggests that car parks promote car sharing in the public space and make it safer. On the other hand, to our knowledge, no survey has shown that the presence of a car park encourages new carpooling practices. It should also be noted that the construction of car-sharing areas can involve the artificialisation of land, despite the fact that there is an average of 1.85 parking spaces per car in France 18.

As a departmental officer summed up in an interview, « I like to talk about supporting the practice and putting in place the conditions to encourage people to carpool, but this is only one of the building blocks of a carpooling policy that the partners should have.

The evaluation of car-sharing policies must therefore be extended beyond recorded journeys to include facilities designed to encourage car-sharing, by making it safer or enhancing its value, although these effects may be more diffuse and therefore more difficult to quantify.

The necessary - and difficult - targeting of car-sharing policies to areas of relevance

The AOMs that have implemented incentive policies with few allocation criteria have gradually introduced new procedures to mitigate the irrelevant use observed. Floor thresholds have been introduced or raised to limit very short journeys. Ceiling thresholds have also been introduced, to limit long journeys and enable the local authority to control its budget. For example, the Rouen Metropolitan Area encourages journeys of between 2 and 30 km; the French Genevois Metropolitan Area between 4 and 40 km; and the Pays de la Loire Region between 5 and 30 km. In Rouen, these new arrangements have reduced the average distance travelled (from 26km to 17km) and cut the number of journeys over 30km (6% of the total volume, compared with 28% before).

To avoid encouraging journeys on existing public transport lines, some AOMs have stopped offering incentives in certain areas. The Rouen Metropolitan Area now excludes from financial incentives journeys whose origin and destination lie within a 400 metre zone on either side of the lines of its public transport network. The Angers metropolitan area has done the same for its tram lines.

From the outset, other local authorities have targeted routes to meet their objectives. The Syndicat mixte des mobilités de l’aire grenobloise (SMMAG), the organising authority for mobility in the greater Grenoble area, has decided to organise its various carpooling services according to their area of relevance, targeting as a priority journeys between the Grenoble metropolitan area and the suburban EPCIs, where the challenge of decarbonisation is greatest (60% of greenhouse gas emissions for 19% of journeys).

With more restrictive criteria for distributing incentives, there are fewer carpooling journeys by platform. For the Rouen metropolitan area, the new arrangements have reduced the number of journeys by 37%. The SMMAG is recording fewer journeys than other equivalent areas, despite a substantial budget for the operation of the various M covoit’ services. To have a real impact on decarbonisation, while limiting rebound effects, car-sharing policies require more effort. The players involved in car-sharing policies share the same observation: car-sharing requires a continuous approach and a great deal of effort to make it work. According to Mathilde Remuaux and Nadège Peteuil, who are working to set up car-sharing routes in the Plaine de l’Ain industrial park, « You have to keep harping on about it, because there’s a lot of turnover, and you have to be present and visible ». It’s a more complicated service to explain than a bus route. There’s an application, you have to fill in a bank card (even though there’s also an SMS route), and you have to explain the principle of the guaranteed journey. When you say carpooling, people think Blablacar ». The complexity also comes from the different timetables: there are times when the line can be used, but when the guaranteed journey is not guaranteed, then there are times when the guaranteed journey applies, but only in the specific direction of the line (corresponding to commuters coming to the Park from outside).

Contrary to what had been hoped, car-sharing does not appear to be an easily activated lever for decarbonisation. However, this is what a number of local authorities and transport operators are talking about: an immediately available service that can be deployed more quickly than public transport. In suburban and rural areas in particular, organised carpooling is slow to be rolled out, if at all. Motorisation (or even dual motorisation) is almost compulsory in the absence of structuring alternatives, and there are few constraints on car-soloing; these are all reasons that do not encourage people to accept the constraints inherent in car-sharing. Similarly, outside metropolitan areas and heavily congested roads, the negative effects of the car (congestion, air pollution) are less visible, as Chrystelle Beurrier, an elected member of the Pôle métropolitain du Genevois français, explains: « In the least densely populated areas, where there are the most benefits to be gained from the ecological transition [with car sharing], acculturation is the most complicated to implement. People are so used to using the car in these areas that they use it for very short journeys. These are areas that are apparently less affected by the impact of pollution ». What’s more, the lack of alternatives to carpooling makes the risk and uncertainty for the passenger even greater.

A number of journeys have been observed that compete directly or indirectly with more relevant alternatives within densely populated areas. Intermediated car-sharing develops much more easily where density is highest and the AOMs have little room for manoeuvre to control the network created by users. These practices are also due to advantageous financial incentives, which create windfall effects and do not guarantee sustainable use of car sharing. Targeting relevant uses that are genuinely effective from the point of view of decarbonisation is difficult and results in a smaller number of journeys than in communities without targeting. From a longer-term perspective, what are the opportunities and limitations of car-sharing as a means of initiating the ecological transition, given our ever-increasing need for mobility and the growing distance between home and work?

Opportunities and limitations of car-sharing policies

Encouraging the formation of car-sharing communities to make car-sharing practices more secure

Car-sharing platforms, supported by public policies, have deployed a range of functions to make car-sharing by platform more secure. Having a verified profile, having a departure guarantee, having access to assistance on the application in the event of a problem… All of this helps to make it easier to share a vehicle with strangers. While the policy of financial incentives has played a significant role in attracting new carpoolers, it is not the only factor. Financial transactions in informal car-sharing relationships are absent for nine out of ten journeys, according to data from the EMP 2019. Among the users questioned during the evaluation of the Pays de la Loire Carpooling scheme, 60% stated that they would continue to carpool despite the discontinuation of the incentive scheme introduced by the Region. Making car sharing more secure and familiar for everyday journeys, thanks to various incentives, can help to develop the shared use of the car.

However, there are limits to platform-based car-sharing that technological advances cannot overcome: in sparsely populated areas, critical mass remains an obstacle to the development of car-sharing, whereas in densely populated areas, intermediated car-sharing has developed more easily. However, the critical mass issue is not the only obstacle: in SMMAG’s EMC 2020, only 1% of respondents who never or exceptionally use car-sharing justify their non-practice by the fact that they have not found anyone interested in car-sharing. On the other hand, the reason that came out on top of all the others (cited by more than 30% of non-users of car sharing) was that they did not want to travel with a stranger. As mentioned in part 2, there is strong resistance to the need to meet new people on a daily basis, in a form of closeness and sociability that is uncommon in our modes of transport. Almost 10% of respondents to the above-mentioned survey said that they do not carpool because they prefer to be alone in the car. Time constraints (inflexible working hours or not wanting to waste time) are also frequently mentioned.

This constraint on sociability, which is renewed on a daily basis, can be partly overcome by organising carpooling within circles of carpoolers, with a certain degree of mutual acquaintance and geographical proximity (from home: neighbours, friends, family, or at destination: workers in the same area of activity or colleagues, members of the same association, etc.). Geographical proximity on departure and/or arrival reduces the time wasted by detours. The fact that the members of the carpooling circle know each other means that trust is less of a barrier. It also activates a social lever, that of being of service, an essential element in carpooling and hitchhiking. Doing a favour within circles of acquaintances brings social recognition of the practice, and enables people to enter into a system of « give and take » 19. The formation of circles of regular carpoolers is one of the proposals put forward by the AutosBus association to massively increase carpooling. For a decade, this association has been researching and experimenting with alternatives to the private car in the suburbs of Bourg-en-Bresse, a medium-sized town of 40,000 inhabitants in the Ain département. The association has noted the limitations of carpooling intermediation systems and advocates a collective model based on small circles of regular carpoolers, built up by individuals: « You need a highly motivated driving force, who talks to their colleagues and neighbours to form a small group, which facilitates pooling and makes a change in mobility possible ». In a broader format, « ecomobile clubs » can bring together one or more dozen people in a company or village. Jacques Toulemonde has witnessed this informal car-sharing: « We assumed that we couldn’t develop informal car-sharing, so we had to go through organisations and intermediaries. However, we see it forming and operating on a daily basis. But how can we develop it?

There is therefore a need to encourage and promote carpooling and hitchhiking, without creating a new digital dependency or the daily constraints of sharing a car with strangers.

Offering a marginal solution to long-term trends

Decarbonising the transport sector is a long road, strewn with obstacles and resistance; emissions from the sector have risen by 2% in 2022 compared with 2021 20. Changes in mobility practices are particularly slow in coming. The car remains unthinkable, even for people who are committed to ecology (Demoli, Y., Sorin, M. & Villaereal, A., 2020), and its use is often the daily practice perceived as the most difficult to change (OpinionWay poll for ADEME, 2021) 21.

Mobility policies are affected by a form of inertia, in which they continue to give a predominant place to the car, even though alternatives exist and the challenges of decarbonisation are known and urgent. This ‘path dependency’ is illustrated by an increase of 12.6% in the length of roads between 1997 and 2017, 8% in road traffic and 2 million vehicles in the private car fleet (Sajous, 2020).

Sajous et al (2020) show the contradiction that exists between measures that restrict car use and measures designed to make the fleet less polluting or optimise its use (via car-sharing in particular), which amounts to confirming the role of the car in the organisation of mobility. The growing involvement of the car industry in the development of car-sharing is significant. The desire to turn the millions of empty seats in circulation into a new public transport network sums up this inertia: it adapts public action to a state of affairs in which 80% of the kilometres travelled are by car, in a vehicle that is oversized for the majority of everyday uses. Everyday mobility is fragmented and flexible, encouraging a form of individual mobility that the development of car-sharing clashes with.

Feedback from experience and survey results show that, for individuals and local authorities, car sharing has been a marginal solution in the face of long-term trends. Firstly, the particularity of some of the local authorities we met for the study, which have been involved in developing car sharing for more than a decade or which have developed proactive and ambitious policies, is that they are experiencing sustained demographic growth, higher than the national average. This growth is concentrated in certain areas of the region, and is visibly putting pressure on the mobility system, with the creation of major congestion on flows entering the metropolis, bringing together dispersed flows from the suburbs. The pressure can also be seen on public transport infrastructure, which is often described as saturated on certain routes, and the difficulty for the AOMs in rapidly deploying additional services to meet mobility needs. Car-sharing is therefore seen as part of a « package » to be offered, capable of absorbing some of the pressure from the growth in demand for mobility concentrated in certain areas. Car-sharing was seen as an easy solution to deploy, with significant potential that could easily be activated to tackle the problems of congestion and infrastructure saturation.

The example of the PMGF illustrates the imbalances created by metropolisation, between the concentration of flows in certain areas (to or from the Geneva metropolis, for example), and the continuing urban sprawl in areas of more diffuse density (in this case, in the outlying areas of France). A technician from a community of agglomerations on the outskirts of a metropolis argues: « If we want to invest in major projects, we can’t put in lines that don’t work. It’s also a question of density, and people understand that. Today, the region is already quite well covered by Transport on Demand, and the cost per passenger is close to that of a taxi. At a time when major transport projects are being developed or are set to be developed around the main metropolises 22, the counterpart to these structuring flows is less dense, dispersed flows, for which local authorities have few alternatives or resources. The dispersal of these flows has also been made possible by the car. The carpooling journeys recorded by the RPC showed a great diversity of origin-destination, reflecting the dispersal of flows and the territorial coverage of carpooling.

The development of public transport can also be slowed down by political and administrative difficulties; some flows cross administrative boundaries, and these complicate the implementation of public transport. As this car-sharing operator explains: « We deploy our services outside city centres, where there is no public transport. We go beyond the very borders of metropolises, and generally operate links from outside the metropolis to the metropolis. But in these areas, we don’t fall within the administrative perimeter ». Car-sharing is a mobility skill that is more easily delegated to an inter-territorial level than the organisation of public transport. Such an inter-territorial scale makes it possible to cover this type of travel in exchange between several administrative levels. Of the seven local authorities interviewed for the study, four have developed car-sharing projects on an inter-territorial scale, and of these, three are inter-territorial structures: the Syndicat mixte des mobilités de l’aire grenobloise (SMMAG), the Pôle métropolitain du Genevois français (PMGF), and Tisseo collectivités. The EPCI members of these three bodies have all delegated shared mobility powers relating to car sharing. Only Tisseo collectivités exercises all mobility powers for all its members.

Finally, the concentration of activities in metropolitan areas has been one of the causes of the increase in commuting distances, which have risen by half in rural areas over the last twenty years 23. Half of all working people who live in rural areas but work in towns travel at least 20 km to get to work, compared with 5 km for those who stay and work in rural areas. Faced with these longer distances, in areas where there is no alternative to the private car, the burden of the transport budget is increasingly heavy. In the Plaine de l’Ain, between the suburbs and the countryside, the attractiveness of the Lyon metropolitan area, thirty kilometres away, is having an impact on house prices, forcing people on the lowest incomes to live further and further from their place of work. Such long distances make some of the jobs in the Plaine de l’Ain Industrial Park unattractive. In the short term, mobility policies, which include car-sharing, can compensate for the fact that people are dependent on cars and live and work further away.

Without stronger constraints on car-pooling, car-sharing policies have a reduced impact on decarbonisation.

A modal shift, from car-pool driver to car-pool passenger, is necessary to have a real impact on reducing car traffic and decarbonisation. However, car-poolers are the hardest to convince, as shown by a survey by G. Monchambert and A. le Goff: some car-poolers who said they were prepared to car-share would only choose the driver’s seat in order to retain their autonomy 24. In addition to the loss of autonomy and the uncertainties associated with the practice (return journey, possible driver delays), carpooling also involves a loss of time, linked to the time taken to pick up or drop off the car, as well as the possible break in the load if the journey is not « door-to-door ». However, this loss of time is only economically compensated for by households with a low value of time.

Car sharing is therefore in direct competition with the ease of being a car owner and the preference for autonomy, freedom and time savings. To reduce the use of private cars, we need to encourage car-sharing by reducing the speed of car-pooling and improving the speed of car-pooling. Public policies can encourage the restriction of car-pooling by setting up reserved lanes for car-pooling, reducing the number of lanes open to motorists, or allocating reserved parking spaces, particularly at company sites. Constraints on car-pooling (tolls that increase the cost of car-pooling, parking difficulties) improve car-pooling levels. This is the conclusion of the diagnosis of home-work commuting in Belgium 2018 (see box p.36), which analyses the effect of employers’ mobility policies on employee practices: « Parking spaces reserved for carpooling greatly increase the number of carpoolers, whereas the number of workers who come alone with their car is much lower if they have to pay for their parking space ».

In addition to encouraging car-pooling, constraints on car-pooling can also help to reduce car traffic. This is the concept of evaporated traffic, developed in particular in P. Hosotte’s thesis, which uses several examples of traffic lane restrictions or closures to show that some road traffic disappears and is not transferred to other routes 25.

The car-sharing strategy for the Lake Geneva region 26 has made constraints on car-pooling the main lever for encouraging the growth of car-sharing. These constraints imposed on car-poolers must become opportunities for car-poolers, based on a « push-pull » principle. The constraints imposed on car-poolers (parking constraints, bottlenecks) should give car-poolers advantages in terms of comfort and savings in time and/or money (pull) while reducing these advantages for car-poolers (push). SMMAG has adopted a similar stance with the carpool lane introduced in 2020, seeking to make this lane a constraint for car drivers, while offering a carpool service as compensation. The aim is therefore to use car sharing as a means of restricting car-pooling, while offering an alternative service.

For the moment, reserved lanes in France are still at the experimental stage, since no fines are issued for misuse of the lanes on the roads where they are introduced. So, « at the beginning, the use of the lane was respected to some extent, but today, less and less » reports this elected representative of the SMMAG. As the rate of compliance is very low, and there are no penalties, residents are not yet complaining about the lane. If penalties are gradually to be introduced 27, the question of the feasibility of the constraint arises: should the use of video tagging be generalised?

These restrictions on car use are likely to provoke strong opposition 28, at a time when cities are preparing to introduce Low Emission Zones (LEZ). Dedicated lanes can in fact create two-speed mobility: the lane reserved for carpooling in the SMMAG area is also accessible to very low-emission vehicles with a Crit’Air zero emission sticker, i.e. 100% electric or hydrogen vehicles, regardless of the number of people on board, as well as taxis. On the one hand, there are constraints on car-pooling (which itself is a source of organisational constraints and uncertainty) and, on the other, unregulated individual mobility is possible, but in low-emission vehicles, even though these are still too expensive for many households. The constraints posed by car-sharing raise questions of social and territorial justice in the absence of alternatives to the car, particularly if the incentives are restricted to the use of platforms and create a new digital dependency or an excessive level of uncertainty for car-sharing passengers.

At this stage, car sharing has been proposed more as an additional service, rather than as a real alternative to the private car. To date, constraints on car-pooling have been difficult to implement politically, as a local authority technician explained: « So far, we’ve been proposing additional services and not too many restrictions. It’s not easy. Aren’t we putting the region at a disadvantage economically? There are other issues [besides the ecological transition]. It also depends on the development model that elected representatives want to bring to the region.

However, if there are no constraints on car-pooling, sharing the costs of car-pooling is only economically attractive to low-income individuals. The constraints inherent in carpooling, particularly as a passenger, give rise to strong resistance. To overcome this, local authorities have introduced policies that provide strong incentives for drivers. The costs of such policies, with a view to decarbonisation, result in high CO2 abatement costs per tonne. For example, for a 20 km journey paid for at 10 ct per kilometre, with a commission to the operator of 50 ct per journey (i.e. a cost of €2.50 to avoid driving 20 km), the cost per tonne of CO2 avoided comes to €725. If we add the car-pooling bonus of €100 for 10 journeys, the cost per tonne of CO2 avoided comes to €3,623 29. In the various local authorities studied for this study, the cost per tonne of CO2 avoided varies according to the period, the car-sharing service or the type of cost considered. On average in 2022, it varies between €580 and €9,700 for the most expensive services, and between €520 and €3,500 at the beginning of 2023. These different levels are well above the value of climate action recommended by the Quinet commission, which in a report published in 2019 proposed raising the value of climate action to €250 per tonne of CO2 saved by 2030. This benchmark value is an incentive to give priority to decarbonisation sources whose abatement costs are and will be lower than this 30. The costs of decarbonisation made possible by car-sharing are higher in sparsely populated areas, where various schemes seek to overcome the obstacle of critical mass and the ease of use of the car. However, these schemes have very high operating costs.

Other mechanisms can also reduce the incentive factor of subsidies, by lowering the overall cost of car use: the State has spent 7.5 billion euros to reduce fuel prices 31, which amounts to keeping fuel prices artificially low, while local authorities fund car-sharing, which amounts to artificially lowering its cost. These two policies have opposite effects.

If there are no stronger constraints on car-pooling, car-pooling is only attractive to people in situations of restricted mobility (using a car is impossible or too expensive), linked to the absence (temporary or structural) of alternatives to the private car. A number of mobility surveys have revealed that car-sharing is more common among young people, who generally have fewer motorised vehicles: up to 20% of young people aged between 18 and 24 on work placements, training courses or apprenticeships car-share several times a day in the SMMAG area, for example (EMC2 2020). A number of qualitative feedbacks show that students have also taken up carpooling via the platform: the most carpooling journeys in the Rouen metropolitan area are between Rouen and two neighbouring communes, each of which has a university campus. In the Angers metropolitan area, a large proportion of journeys within Angers were made from or to the Belle-Beille Angers university campus. This is an area served only by buses, where the tram is still in the construction phase, which may explain the use of car-sharing platforms. Similarly, among the most car-pooled journeys in the Île-de-France region, many municipalities are home to university campuses (Evry-Courcouronnes, Massy, Palaiseau), where the lack of public transport to cope with the crowds is regularly highlighted 32. The social implications of car-sharing policies should therefore be emphasised, particularly in a context of fuel price inflation. On the other hand, the impact on decarbonisation is low or very costly.

In the longer term, the challenge is not to maintain the largely dominant and inescapable position of the car in travel practices. However, by reducing the cost of a car journey, car-sharing could encourage households to live further away from employment centres, « reinforcing the phenomena of urban sprawl and peri-urbanisation, and consequently the car dependency of households that are most often modest » (Coulombel, Delaunay, 2019). This would contribute to reinforcing the structural problems of mobility and regional planning. Faced with these problems, which public transport is finding it increasingly difficult to solve, the car would continue to compete with public transport that is considered less efficient, less flexible or less comfortable. The risk is of entering a vicious circle, where poor public transport services lead to low ridership, which in turn leads to a reduction in services. In the Shift Project’s Guide to Low-Carbon Everyday Mobility (2021), the authors suggest « taking the problem in the other direction »: « Instead of ‘simply’ adapting the [public transport] service to an existing urban area with low frequency and reduced opening hours, we need to […] create a structuring transport line and densify at the same time ». To get out of the vicious circle of low density and low ridership, we need to bet on the developmental nature of public public transport. To encourage a modal shift from car to public transport or to active modes of transport, however, we need to question the dynamics of longer distances and dispersed traffic flows.

  • 1 On average 21,139 per working day in 2022 compared with 43,799 in the 1st quarter of 2023.

  • 2 Author’s calculation based on raw CPP data.

  • 3 Excess travel expenses incurred by the driver are allowed for journeys of less than 15 km.

  • 4 The calculation was based on an average journey distance of 24 km, i.e. a payment of €2.40 per journey (one journey = 1 driver and 1 passenger), assuming that the journey is free for the passenger. That’s 2.1 million journeys per day multiplied by €2.40, and 42 million journeys per month (20 working days) multiplied by €2.40.

  • 5 In the first half of 2023, there were 27,000 car-sharing journeys made via platforms, according to the National Car-Sharing Observatory. According to the national carpooling plan, 900,000 journeys are carpooled every day. Journeys made via platforms therefore represent 3% of carpooling journeys.

  • 6 This expression was used by one of the participants in the round tables on local car sharing organised in Rosporden in November 2022 by the Ehop association.

  • 7 With France Covoit, carpoolers do not use the Ecov carpooling lines and meet informally. Passengers have to register and provide proof of identity, then scan the QR code generated by the driver. The driver receives up to €150 and the passenger €16.

  • 8 To award the Forfait mobilité durable, employers can choose to ask employees to provide proof that they have actually made the journeys, via the RPC, or ask them to fill in a sworn statement (which can then relate to journeys made via a platform or outside a platform). According to the Forfait mobilité durable 2022 Barometer, companies are not very familiar with the Carpooling Proof Register, and most of them use the attestation on honour.

  • 9 The positive toll is an « ecobonus » of two euros per journey for motorists registered with the scheme who change their mode of transport, those who start carpooling, and those who shift their journeys to less intensive hours. The scheme is designed to relieve congestion on the city’s busiest roads. (

  • 10 Ecov.

  • 11 Le journal d’Elbeuf, « Métropole de Rouen : quand l’argent public passe par les fenêtres de certains covoitureurs », 05/03/2023. (

  • 12 See the Alternatives économiques article, « Le boom du covoiturage est-il lié à l’action du gouvernement?", Alternatives Economiques no. 434 - 05/2023.

  • 13 These eight classes are : the 22 metropolises (class M); the inner suburbs of Paris (class A); the suburbs of the Paris region and certain metropolises or cross-border areas (class B); the wider suburbs and the countryside around metropolises or medium-sized towns (class C); the EPCIs of medium-sized towns (class D); the EPCIs of medium-sized or small towns, the countryside, EPCI XXL (class E); the rural and suburban EPCIs of medium-sized towns (class F); the rural EPCIs (class G).

  • 14 This average and the next calculations were based on all the raw CPP data for 2022 (January to December).

  • 15 For extra-family carpooling, based on data from ENTD 2008 (ADEME-Inddigo, 2015).

  • 16 In less than fifteen minutes, journeys of 4 km can be made by conventional bicycle and those of less than 9 km by electrically assisted bicycle.

  • 17 Of the 88% of self-drivers, the survey does not indicate the proportion of passengers and drivers who claim to be former self-drivers. However, while drivers were probably car-poolers before car-pooling, passengers may be non-motorists and former public transport users.

  • 18 According to estimates from a study by the Fédération nationale des associations d’usagers des transports (Fnaut)

  • 19 The give-and-take system was put forward by the anthropologist Marcel Mauss as « a service that mutually obliges giver and receiver and which, in fact, unites them through a form of social contract ». It is based on reciprocity: even if counter-giving is not contractually obligatory, it is socially encouraged. Finally, giving and receiving allow people to belong to society. (

  • 20

21 Of the fifteen actions proposed in the Barometer of social representations of climate change, the three actions receiving the most unfavourable responses ("you can't do it" or « you can do it but with difficulty ».


En savoir plus