key: cord-0752087-d8fazeu6 authors: Laroche, Florent title: The effect of COVID‐19 on long‐distance transport services in France date: 2022-04-25 journal: Regional Science Policy & Practice DOI: 10.1111/rsp3.12534 sha: 00e5666d49d6b49eb482ae9992e04a357296c0e4 doc_id: 752087 cord_uid: d8fazeu6 This paper explores the effect of COVID‐19 on long‐distance transport services in France. These services have been deeply impacted by the different lockdowns imposed. However, until now, few papers have considered this topic. The research examines the long‐distance transportation market before COVID‐19 to understand how it adapted to the COVID‐19 countermeasures. The empirical research is based on a large pool of data pertaining to four transport routes in France from September 2019 to November 2021 for four modes (coach, carpooling, train, and air). Prices and frequencies are the main variables. The study's results must be considered carefully because the pandemic is ongoing. The first finding is the severe crisis in terms of supply during the first lockdown in March 2020. Throughout the rest of 2020 and 2021, services increased slowly, with rail and carpooling recovering to almost their 2019 levels, though this is not yet the case for coach and air services. The second finding is that the dominant mode, rail, enjoys the greatest market concentration, especially in comparison to air, which is still experiencing difficulties because of the reduction in the business market. The last finding highlights the persistence of conventional services during the different lockdowns and the high variability of low‐cost services, which disappeared during the crisis. However, mainly thanks to state financial support, no operator went bankrupt or left the market. Finally, low‐cost services were recovering faster in autumn 2021 than conventional services, thereby increasing their market share. The crisis in the business market may explain the difficulty conventional services face whereas low‐cost services are better adapted to the leisure market. understand its evolution. The second challenge was to identify the main stakeholders and their strategies to address different COVID-19 countermeasures. The empirical research is based on a large pool of data. The data was collected on four routes in France where competition was expected for train services in 2020, in line with market liberalization. Notably, because the COVID-19 crisis was not anticipated, the initial purpose of the data collection was to assess the effect of train liberalization on the current train market as well as on the other modes of transport in terms of price and frequency. Consequently, the data collection started in September 2019 and was still active in December 2021 for four modes: coach, carpooling, 1 train, and air. The analysis is based on time series from September 2019 to October 2021. Prices and frequencies are discussed as the main variables on an aggregated scale for all modes then mode per mode. Finally, the intensity of the competition is analyzed on two scales: intramodal and intermodal. The Herfindahl-Hirschman Index is used to assess the level of competition through market concentration (Lipczynski et al., 2017) . The main results have to be considered carefully as the pandemic is ongoing as of the end of 2021. However, the progress of the vaccination programs in Europe and France during the second part of 2021 suggests a return to a normal state of the market. The first finding was the severe crisis in terms of supply during the first lockdown in March 2020. Throughout the rest of 2020 and 2021, the COVID-19 countermeasures had progressively less impact, and services increased slowly, with rail and carpooling recovering to a level similar to their 2019 levels. However, this is not yet the case for coach and air services. The second finding is that the market is concentrated to the advantage of the dominant mode, rail services, especially in comparison to air services, which are the main challenger for long-distance trips. Air services still face difficulties recovering, mainly because of reduction in the business market. The third finding highlights the persistence of conventional services during the different lockdowns and the high variability of low-cost services, which disappeared during the crisis. However, no operator went bankrupt or left the market during the crisis, mainly thanks to public support. Moreover, low-cost services recovered faster in September-October 2021 than conventional services, thereby increasing their market share. Consequently, the COVID-19 pandemic has resulted in greater intramodal competition, mainly between the incumbent and challengers in each mode. The last finding reveals the crisis in the business market (Caputo et al., 2021) caused by alternatives to business travel (such as videoconferencing and teleworking). This crisis may explain the challenges conventional services face when confronted by low-cost services better adapted to the leisure market. Additionally, it suggests a new paradigm of work and travel based on the development of alternatives to physical mobility. It confirms the concept of a "new normal" and questions public national incumbents' business models, which are traditionally dependent on business users. There are several practical implications for policymakers and operators. First, the COVID-19 pandemic highlighted the key role public national incumbents play during a crisis in ensuring a minimal service when other operators stop running as well as providing specific services, like sanitary trains in France. Second, there is a risk that the end of public financial support will transform the long-distance transport market into "business for survival" as the sector is highly sensitive to COVID-19 countermeasures and has become more competitive, mainly because of reduced demand (especially from the business market). Finally, most of the market segments are duopolies, so there is a high risk of monopolies developing if challenger operators go bankrupt. Public regulators must remain vigilant in the immediate future regarding the effects of COVID-19 countermeasures and public financial support to avoid greater market concentration. This paper is structured in six sections. Section 2 presents the motivations underlying the French case and the context for long-distance services. Section 3 describes the data collection and method. Section 4 gives the general results, after which Section 5 discusses them mode by mode. Finally, Section 6 concludes the paper. There are several reasons for selecting France as a case study. First, the French market is characterized by a wide variety of services for long-distance travel, from rail services to carpooling, as described in Subsection 2.1. Subsection 2.2 shows that during the pandemic, France, like other European countries, imposed strict sanitary regulations and multiple lockdowns as well as other restrictions, affecting long-distance travel. Finally, Subsection 2.3 highlights the crisis's severe impact on demand for travel and transport companies. The French case is interesting due to the wide choice of transport services available on long-distance routes. Laroche and Lamatkhanova (2021) showed that the diversity of services in the French market is similar to Germany's, Italy's, and the UK's. The services are predominantly train, coach, carpooling, and air. Consequently, analyzing the French market can offer insights into trends in other markets after considering certain specificities of the French market. First, most segments of the long-distance travel market are either duopolies or monopolies. Second, some of them are quite new and were still consolidating when the COVID-19 pandemic hit. Third, low-cost services are well established in the market. More specifically, train services are still a monopoly run by the national incumbent, SNCF. The initial competitors were expected in 2020, but the first lockdown and other countermeasures delayed project until the end of 2021. However, the high-speed market can be characterized as an innovative service, because SNCF was the first European operator to develop the concept of low-cost high-speed services (called Ouigo) in 2013. It is interesting to distinguish low-cost services (Ouigo) from traditional services (inOui) and analyze the internal competition between the two. Air services are a duopoly for long-distance trips (journeys that are three hours longer by rail) and a monopoly for shorter trips. The national incumbent is traditionally Air France, which competes with easyJet. The coach market is relatively new in France in so far as it has only been open to competition since 2015 (Blayac & Bougette, 2017) . One of the first players was Flixbus. There were several competitors in the early years but, confronted by the consolidation of the market, most are now limited to niche markets. There were two main competitors at the beginning of 2020 before the first lockdown: Flixbus and BlaBlaBus. BlaBlaBus can be considered a challenger, and its history is original. Launched by the national train company (SNCF) in 2014 (as iDBUS), the company was bought by BlaBlaCar, the French leader in carpooling services, in 2019 with the aim of becoming a multimodal actor in the passenger transport market. Lastly, BlaBlaCar is a carpooling services success story in France and Europe. Established in 2006, it boasted 100 million members in 2021 (Business Insider France, 2021). The next sections consider the effect of COVID-19 on transport overall and on each of the four modes individually. Another motivating factor to examine the French case is the French government's decision to counter COVID-19 with a wide range of regulations. Figure 1 illustrates the chronology of these regulations from 2020 to 2021. The first lockdown was imposed in March 2020, some days after Italy but in line with the majority of other European countries. For eight weeks-from March 17, 2020, to May 11, 2020-mobility was curtailed to one hour per day within a one kilometer radius of one's residence with a pass. From May 11, 2020, to May 28, 2020, the restrictions on local mobility were loosened but stayed restrictive for long-distance mobility, with journeys limited to 100 kilometers for essential reasons only. Long-distance transport services were particularly impacted by these F I G U R E 1 Timeline of the COVID-19 countermeasures in France. Source: Author restrictions. Summer 2020 was unique for several reasons. On the surface, it looked like a "normal" summer; however, the public were obliged to wear masks in public, including on transport, potentially negatively impacting longdistance services. In addition, the limitation of international trips and the rising "fear of social contact" changed mobility behaviors and habits that summer, especially regarding long-distance journeys. The COVID-19 countermeasures of summer 2020 (a lower level of restriction but permanent sanitary regulations, such as the use of masks) persisted until early September, when universities reopened their doors and invited students to travel long distances to be physically present in the classroom after six months of online courses. Unfortunately, the number of COVID-19 cases increased drastically that autumn, leading the French government to implement the second lockdown from October 30, 2020, to December 15, 2020. Prior to the second lockdown, there was a period of curfews from nine p.m. until six p.m. in the most infected localities. The second lockdown was similar to the first. Trips were limited to one kilometer from one's residence for one hour a day with a pass. Additionally, a curfew throughout mainland France was implemented from six p.m. Once again, long-distance services were deeply impacted. Universities closed their doors until summer 2021, and teleworking became the norm for millions of workers. Although the second lockdown was lifted in December 2020, restrictions remained very strict throughout winter 2021, with the national six p.m. curfew and the closure of restaurants, bars, and ski resorts. The third lockdown was ordered from April 3, 2021, to May 3, 2021. Since then, the restrictions have been progressively alleviated, and the curfew ended on June 30, 2021. Summer 2021 was characterized by the permanence of restrictions, like mask-wearing, and the introduction of a vaccination pass, excluding non-vaccinated people from certain modes of transport (air and train), restaurants, and other places. By October 2021, 70% of the French population had been vaccinated against COVID-19, making it possible to return to "normal" life. The mass teleworking policy was removed, and universities reopened. This research compares autumn 2021 to September-October 2019 to assess the effects of eighteen months of restrictions and verify if a "new normal" has emerged from this period. To conclude, long-distance services have been particularly impacted by the different measures to counter the COVID-19 pandemic. This is because long-distance trips were prohibited for many months for the majority of the French population, plus the closure of universities and mass teleworking resulted in no demand from the student and business markets. This last subsection highlights the effect of COVID-19 on demand for long-distance transport services. It also compares the French trends to the European averages. From a general point of view, the collapse in demand for long-distance transport services in France is similar to the rest of Europe. For France's train market, demand decreased by 41% in 2020 compared to 2019 and by 48% on average in Europe (IRG-Rail, 2021). The situation was similar for air, with a decrease of 67% for Air France and "only" 46% for easyJet. The average in Europe was slightly higher with a decrease of 73% in 2020 compared to 2019 (Eurostat, 2021) . For both modes, teleworking and fewer business travelers deeply impacted the overall dynamic (Caputo et al., 2021) . For coach services, demand decreased by 65% over the period (Autorité de Régulation des Transports, 2021b). Carpooling services were less affected by the collapse in demand than other modes of transport with "only" a 25% reduction in demand (Vérier, 2021) . It is possible that the user profile for carpooling is less sensitive to COVID-19 because of age or another factor. Finally, it is interesting to consider that traffic on France's high-speed roads increased by 7.6% during summer 2021 compared to summer 2019 according to Vinci Autoroute (Investir, 2021) . This finding suggests that long-distance trips experienced a similar behavioral change as urban mobility, where an increase in car use has been observed (Klein et al., 2021; Macharis, 2021) . Subsection 3.1 presents the study's parameters. Details of the data collection are described in Subsection 3.2, and Subsection 3.3 explains the method. The analysis is based on four main routes in France as shown in Figure 2 : Paris-Lyon, Paris-Bordeaux, Paris-Toulouse, and Paris-Nice. The initial research question was about the effect of a new competitor on long-distance train travel; therefore, these four routes were selected because a new train operator was expected to challenge the incumbent in the context of market liberalization in 2020. However, the COVID-19 pandemic delayed the introduction of competition and necessitated changing the research question. The four routes were also chosen in the hope of including routes with different characteristics, like travel time. Lyon and Bordeaux are less than three hours from Paris, while Toulouse and Nice are more than three hours from the capital. Consequently, the Toulouse and Nice are served by more flights but fewer trains and other services, like carpooling and coaches. Moreover, the Paris-Lyon route is one of the most used high-speed lines in Europe, serving more than 44 million travelers in 2019. F I G U R E 2 Study parameters. Source: Author Data were collected every Tuesday since September 2019 on each route and for four modes of transport: train, air, coach, and carpooling. Tuesday was selected for its representativeness during the week. The collection was performed one week before the departure date except for carpooling because of the high variation of supply at the last minute. In this case, collection was performed one day before to consider more than 90% of the supply. The data collection concerned each mode and each service (only direct travel without connections) during the day, and looked at price, schedule, and travel time. In total, the data cover 87 days and comprise more than 14,000 observations 2 between September 2019 and October 2021. The main sources given in Table 1 are the websites dedicated to the different modes. Most of them are comparators because of the competition between operators. The two exceptions are for train and carpooling where the markets are monopolies. For rail, we used the SNCF's commercial website where all the train offerings are presented. For carpooling, there are several operators but only one drives the market; therefore, BlaBlaCar's website was the reference for carpooling. For the other modes, comparators were used. For air, Google Flights gave a full overview of supply, while the European website Omio.fr was used for coach services. The method is based on time series and a descriptive analysis. The times series are given from September 2019 to October 2021. The variations between 2019 and 2021 are calculated on the basis of two reference months: September and October. The main variables used to give an overview of the market evolution are price, number of seats, and number of services (frequency). The price is given as an average price per kilometer for each mode or service. The number of seats is given per day for each mode and service. It is based on the number of seats per service and their frequency during the day. The number of seats can change from one service to another according to its characteristics. This is especially true for train services. For example, a low-cost service proposes 634 seats versus 510 for a conventional high-speed rail service (Autorité de Régulation des Transports, 2021a). There is less heterogeneity for the other modes. The standard number of seats for a coach service is 50, 3 165 for a standard Airbus A320, 4 and four at most per car for a car. Finally, the frequency is based on the number of services a day. Competition is assessed via the Herfindahl-Hirschman Index. It is a well-known index usually used to assess a market's concentration (Lipczynski et al., 2017) . It is based on each operator's market share. In this study, the 2 One observation corresponds to one service one way for one schedule. 3 https://location.flixbus.fr/services/location-bus-places#:$:text=Location%20de%20bus%2036%20%2D%2050,bus%20qu'il%20vous%20faut. 4 https://www.universalis.fr/encyclopedie/airbus-a320/2-les-nombreux-heritiers-de-l-a320/#:$:text=L'A320%2C%20qui%20se%20d%C3%A9cline, hauteur%20de%2011%2C76%20m%C3%A8tres. Herfindahl-Hirschman Index is calculated on the basis of the number of seats for each mode or type of service per day. It is the best indicator to give a realistic view of the level of concentration within a market taking into account the capacity of each mode and frequency of the services. The results are split into two sections. Subsection 4.1 gives an overview of the COVID-19 effects in terms of price and capacity, while Subsection 4.2 focuses on the effects of the pandemic on each mode of transport. The the market (respectively, 0.4% and 2.2% in 2019), rail increased its market share, while air services fell from 17% in 2019 to 13% in 2021. The first lockdown was very severe for air services, when the market share dropped to 4.6% before recovering to around 11% during the two other lockdowns. Nevertheless, the situation remains difficult for air services. For the other modes, it is interesting to consider that carpooling has been more consistent than the other modes with a modal share between 0.4% and 0.2%. However, coach services have been significantly impacted with the stoppage of services during the second and third lockdowns, resulting in rail services taking a greater share of the market. According to Figure 5 , the Herfindahl-Hirschman Index increased by 9.8% over the period, with a peak during the first lockdown, placing the rail incumbent in a dominant position within the market. Figure 6 shows that rail can be observed, especially for coach services, who are well known for their flexibility, but also for rail and air. To conclude, the return of rail services to their initial capacity may hide a lower utilization rate of rail than in 2019. This situation could limit the market power of rail services, particularly in terms of prices. Regarding price, COVID-19 has had unexpected effects. Figure 7 the modes. They decreased during the first and second lockdowns but increased during the third. This can be explained by the significant reduction in supply and state financial support to maintain a minimal service during the first two lockdowns. Concerning train services, it is interesting that the average price per kilometer was similar to that for carpooling. It can provide from the development of the low-cost offer whose the extinction can have a positive effect of prices (increase) as during the three lockdowns. Nevertheless, it is notable that the train incumbent did not benefit from its dominant position within the market by increasing prices. Regarding carpooling prices, the reduction could be related to the price of oil, which dropped in 2020 (INSEE, 2020) . Prices should increase in the next month as oil prices rise. Finally, coach services increased their prices after their services were paused during the lockdowns. Subsection 4.2 highlights the sector's difficulties in finding economic equilibrium. This subsection focuses on each mode's prices and frequency. Rail services can be split into two types: the conventional high-speed rail service (inOui) and the low-cost high-speed rail service (Ouigo). From a general standpoint, Figure 9 shows that the first lockdown dramatically impacted both services. The conventional high-speed rail service almost completely stopped its services, reducing them from an average of 13 trains per day to two, while the low-cost services stopped running altogether. The effect of the subsequent two lockdowns was less severe, although both times traffic was halved, and the low-cost services were more impacted than the conventional high-speed rail trains. However, the low-cost services' market share increased significantly from 2019 to 2021. The average number of conventional high-speed rail trains per day decreased by almost 15% but the number of low-cost high-speed rail trains increased by 30%. The result was a substitution of F I G U R E 8 Detailed price per mode (in € per km). Source: Author conventional high-speed rail services for low-cost high-speed rail services and a bigger market share for low-cost high-speed rail (from 22% in 2019 to 30% in 2021). The Herfindahl-Hirschman Index decreased by 10% during the period, pointing to greater competition between the two services and a form of internal "home-made" competition. Concerning prices for train services, analysis is difficult because of yield management practices. Figure 10 shows that prices seemed to be more sensitive to the summer holidays than the lockdowns. Nevertheless, it is interesting to observe that prices decreased by 17% and 16%, respectively, for conventional high-speed rail and low-cost highspeed rail between 2019 and 2021. This result was unexpected in so far as the concentration of the market increased in favor of rail services during that period. At least two explanations can be given. First, the overall reduced demand generated overcapacity and put pressure on prices. This is clear for the conventional high-speed rail where frequency and prices were reduced during the period. The severe crisis in the business market linked to teleworking and videoconferencing is an explanation. Second, the train market was opened to competition in 2020. With two years of delay, the first competitors should enter the market at the end of 2021. From this perspective, the incumbent has an interest in moderating prices for conventional high-speed rail services and improving its low-cost services in order to be competitive. Air services are split between two operators. The national company, Air France, traditionally runs conventional services in opposition to easyJet, the well-known low-cost company. As shown in Figure 11 , the first lockdown had whereas easyJet's only dropped by 14%. The market share for low-cost services increased from 30% in October 2019 to 40% in October 2021, and the Herfindahl-Hirschman Index fell by 8%, pointing to a higher level of competition within the market. As with conventional rail services, Air France's difficulties in returning to its initial level may be a sign of crisis in the business market, whereas easyJet is in a better situation thanks to its strong position in the leisure market supported by its low-price policy. Regarding prices, as with rail, yield management practices limit the analysis, especially during the lockdowns. However, Figure 12 highlights a strong divergence between Air France's and easyJet's strategies. Between 2019 and 2021, Air France increased its prices by 9%, while easyJet decreased theirs by 50%. As reported in Figure 12 , easyJet has an aggressive strategy of increasing supply and decreasing prices, while Air France seems to have difficulties in keeping its offerings competitive. To conclude, the low-cost services seem to have increased their market share after two years of COVID-19. This last point will be discussed further in Section 5. Figure 12. The coach market is characterized by a duopoly between two companies: Flixbus and BlaBlaBus. Whereas for rail and air services it is possible to distinguish a conventional company from a low-cost one, this is not the case for the Figure 13 presents these characteristics. Both operators adopted a similar strategy during the three lockdowns with a shutdown of their services. Flixbus had a small advantage between the second and the third lockdown as it remained alone on the market. From a general standpoint, Flixbus seems to have performed better between 2019 and 2021 than BlaBlaBus. In 2019, BlaBlaBus had a greater market share than Flixbus (49% versus 38%, respectively), but the situation reversed in September-October 2021 when Flixbus had 62% of the market versus BlaBlaBus's 18%. However, it is important to note that the frequencies are highly variable and sensitive to demand. According to Blayac and Bougette (2020) , this is a competitive advantage inherent in their business model. It can be defined as a two-sided market (Rochet & Tirole, 2003) based on a digital platform where the offer (coach companies' services) matches demand. Both companies have no buses of their own, meaning they are able to adapt their capacity and prices to meet demand. The carpooling market is unique for several reasons. First, there is only one operator for long-distance trips, BlaBlaCar. Other platforms tried to develop alternative services without success. Since 2021, BlaBlaCar can be considered the dominant digital platform for carpooling services. Second, BlaBlaCar is a digital platform where car drivers are free to share their trip and fix their price according to a limit set by the platform (calculated on the distance and based on gasoline price and tolls). Economically speaking, BlaBlaCar's business model can be defined as a two-sided market with a strong position in the long-distance market. Consequently, Figure 15 does not show planned conventional offerings but the results of an aggregate of drivers willing to share their trips. The offerings are very varied, and the peaks are strongly dependent on the holiday periods during winter (Christmas) and summer. The COVID-19 pandemic had two interesting effects on supply. First, traffic stopped during the first lockdown and significantly decreased during the other two. Second, drivers' behavior seems to have changed between 2020 and 2021. Although summer 2020 saw a classic peak, summer 2021 showed below-usual performance levels, and the overall supply fell by 6% between 2019 and 2021. The first explanation could be the simple evolution of the demand for transportation. Considering that the carpooling offer is dependent on the willingness of drivers to make trips, the carpooling supply is certainly the best indicator for assessing the level of demand for long-distance transport. Consequently, it is possible to assume that demand decreased between 2019 and 2021, especially in the leisure market. This could also explain the reduction in supply for other modes. The second explanation can be linked to the fear of social contact. This psychological effect of COVID-19, based on the fear of being infected, has been identified F I G U R E 1 5 Carpooling frequency (average number of cars per day). Source: Author F I G U R E 1 6 Average carpooling price per kilometer (per day). Source: Author regarding other modes of transport (Aaditya & Rahul, 2021) . Finally, the last assumption is based on gasoline prices. They decreased during 2020 before increasing in summer 2021. As the value of drivers' time is very high because of the cost of organizing the services, it is possible that the service has become less attractive to drivers. Figure 16 supports this hypothesis. Prices dropped by 26% throughout the period, before increasing in October 2021. If prices are linked to the price of gasoline, they did not change significantly during the summer when oil prices were already increasing. This last section discusses two further aspects of the COVID-19 pandemic on long-distance services. Subsection 5.1 proposes a spatial approach based on an analysis of seat capacities route by route. Subsection 5.2 further questions the results regarding the increase of low-cost services. In Section 4, the results were analyzed by mode of transport. In this section, the objective is to provide an analysis per route to identify the potential spatial effect of COVID-19 countermeasures in France. We recall that the data is based on the following four routes: Paris-Lyon, Paris-Bordeaux, Paris-Toulouse, and Paris-Nice. Each route has specific characteristics. First, distance is a key parameter for distinguishing them. Lyon and Bordeaux are closer to Paris (466 kilometers and 586 kilometers, respectively) than Toulouse and Nice (677 kilometers and 933 kilometers, respectively). A second characteristic is the metropolitan population. Lyon is the biggest (1.3 million inhabitants), Bordeaux and Toulouse are similar (749,595 and 783,353 inhabitants, respectively), and Nice is the smallest (550,000 inhabitants). Finally, their accessibility is heterogenic. Lyon and Bordeaux are directly linked to Paris by a high-speed line, while Toulouse and Nice are still outside the high-speed rail network. Consequently, the travel time by train to these destinations is longer, and air services to them are more developed. Considering this, it is interesting to compare these routes and identify the potential effect of COVID-19 on seat capacities. Figure 17 highlights the effect of COVID-19 on seat capacities for the four routes. The first comment is the effect of distance on seat capacities. The Paris-Lyon route is the shortest and offers the highest number of seats. The considerable difference in 2019 with Bordeaux can be explained by the fact that Lyon is larger than Bordeaux. Otherwise, Toulouse and Nice are similar in terms of supply. Toulouse has a small advantage due to its closer proximity to Paris and bigger metropolitan area. The results are in line with the standard theory of the gravity model (Hansen, 1959) . Concerning the effect of COVID-19, the reduction in the number of seats was similar for Lyon and Bordeaux (À73% and À72%, respectively) but very different for Toulouse and Nice (À51% and À90%, respectively). In this case, it is difficult to draw a general conclusion. The second lockdown had less of an impact than the first but that is highly heterogenic according to location. The effect of the latter lockdown was unexpected with a stronger effect on short distances (respectively, À54% for Lyon and À52% for Bordeaux) and a lower effect on longer distances (respectively À34% for Toulouse and À26% for Nice). Finally, all the destinations lost seat capacity in September-October 2021 in comparison to 2019. Nevertheless, the two shortest distances lost less capacity than the two others. Lyon and Bordeaux reduced their seat capacities by 5.7% and 9.5%, respectively, while Toulouse and Nice reduced them by 17.5% and 18%, respectively. The higher reduction for the latter two can be attributed to the difficulties air services faced in returning to their initial level of service. Furthermore, the better performance of Lyon and Bordeaux can be linked to the good performance of rail services as discussed in Section 4. To conclude, the cities that are not linked to the high-speed rail network or who are far from the main city (more than three hours away) have suffered more from the effect of COVID-19 than the others in terms of supply. The aforementioned results highlighted a possible increase in low-cost services for rail and air modes between 2019 and 2021. This last section tests this finding. Figure 18 compares the evolution of conventional and low-cost supplies for air and rail in terms of frequency. In September and October 2019, there were on average 120 conventional trains and flights per day for the routes investigated and 30 low-cost trains and flights. The modal share of low-cost services was around 25%. The first lockdown had a dramatic effect on both traditional and low-cost services but impacted low-cost services more. All lowcost air and rail services stopped, whereas conventional services were merely reduced. The two other lockdowns had less impact on overall traffic. However, low-cost services were still more affected than conventional ones. During the second lockdown, frequency decreased by 42% for conventional services versus 82% for low-cost services. During the last lockdown, the decrease was 38% for conventional services versus 52% for low-cost services. This confirms the greater flexibility of low-cost services, although in the case of serious crisis, the pursual of the service offer has been guaranted of service continuity. In September-October 2021, there were on average 34 low-cost services a day versus 87 for the conventional offer. There are several lessons here. First, the number of low-cost services did not increase by much (+5.5%). They regained their pre-COVID-19 levels. Second, the full service offer decreased by 28.5% over the period. Traditional air and rail services remain are far below their initial level with only 87 services per day versus 120 in 2019. F I G U R E 1 8 Average low-cost offer for rail and air per route and day versus the full service offer. Source: Author Consequently, the low-cost market share increased, but the analysis confirms that there has not been an overall increase in low-cost services. There appear to be more low-cost services because the conventional offer is still far below its initial level. To conclude, Figure 19 shows the ratio between low-cost services and conventional services for rail and air to the previous results were driven by one or the other. For both air and rail, the ratio was higher in 2021 than in 2019, which confirms the previous results. If the ratio were higher in 2019 for air, signifying a larger market share for lowcost services than for rail, it was the opposite in 2021. However, the ratio for air almost returned to the difference with rail in 2021, illustrating easyJet's aggressive strategy. It also confirms the severe crisis conventional services, traditionally driven by business trips, face (Caputo et al., 2021 ). An analysis of demand and supply would be an interesting addition to this research to further explore the effect of COVID-19 on the evolution of travel behaviors. The COVID-19 pandemic resulted in the implementation of strict regulations all over the world, an unprecedented development in peacetime. These regulations significantly impacted long-distance transportation by limiting individuals' capacity to travel both between countries and within them. At the same time, behaviors changed to adapt to the lack of physical mobility, such as the rise of teleworking and videoconferencing. This study examined the effects of these regulations on long-distance transport services. France was selected as the case study because of the difficulty in obtaining international data and the country's large variety of services (rail, air, coach, and carpooling) as well as its representativeness in terms of COVID-19 countermeasures (including three lockdowns). There are several main findings, but these must be considered with caution because the pandemic, at the time of writing in November 2021, is ongoing. The first finding is the severe crisis in terms of supply during the first lockdown in March 2020. The subsequent COVID-19 countermeasures and lockdowns during the rest of 2020 and 2021 had progressively less impact on long-distance transport, and services increased slowly until rail and carpooling services were close to their 2019 levels (though coach and air services have not yet recovered to this point). The second finding is that the market is concentrated to the advantage of the dominant mode, rail services. Air services are the main challenger for long-distance trips but face difficulties in recovering their activity mainly because of the reduction in business travel. The last finding highlights the persistence of conventional services during the different lockdowns and the high variability of low-cost services, which disappeared during the crisis. However, it is interesting to consider two facts. F I G U R E 1 9 The ratio of low-cost services to classic services for train and air. Source: Author First, no operator went bankrupt during the crisis or withdrew from a route, mainly thanks to state financial support. Second, low-cost services were recovering faster in September-October 2021 than conventional services, finally increasing their market share in the air, rail, and coach markets. The crisis in the business market can explain the difficulty for conventional services confronted by low-cost ones better adapted to the leisure market. In addition, the effect is higher intramodal competition predominantly between the incumbent and challengers in each mode. Finally, the spatial approach highlights the greater resilience of long-distance travel services serving cities connected to the high-speed rail network that are close to the main cities. The cities that are not linked to the high-speed rail network or that are far from the main city (more than three hours away) have suffered more from the effects of COVID-19 than the others in terms of supply. The results show that the concept of a "new normal" could be applied to the French long-distance travel market considering the structural reduction of the business market. It suggests a new paradigm of work and mobility based on the development of alternatives to physical mobility (such as videoconferencing and teleworking). Moreover, the shrinking business market presents a huge challenge for the public national incumbents (SNCF, Air France) because their business models depend on business users. There are several practical implications for policymakers and operators. First of all, the COVID-19 crisis highlights the key role public national incumbents play during a crisis to ensure a minimum level of service when other operators stop their activities. They also played a key role in providing specific services, like sanitary trains in France. Second, there is a risk that the end of the state financial support will transform long-distance travel business into "business for survival" as the sector is highly sensitive to the effects of COVID-19 countermeasures and is still weakened by the reduction of the business market. Furthermore, most of the market segments are duopolies. There is a risk of monopolies developing should an operator go bankrupt. The public regulators must remain vigilant in the next few months regarding the effect of COVID-19 countermeasures and state financial support to avoid higher concentration within the markets. Finally, the current market evolution with fewer air services could be an interesting trend for the future, considering the objective of carbon neutrality in France and Europe by 2050. The reduction of the business market can be a relevant lever for at least two reasons. First, it may be easier to find alternatives to physical travel (such as videoconferencing) for business trips than for leisure or family trips. Second, business demand prioritizes frequency over price. A reduction of the business segment could favor fewer trains and flights per day but conversely a better utilization rate and commercial capacity thanks to the low-cost model, leading to environmental benefits. https://orcid.org/0000-0001-9961-1159 Psychological impacts of COVID-19 pandemic on the mode choice behaviour: A hybrid choice modelling approach Le marché français du transport ferroviaire en 2020 Autorité de Régulation des Transports, Rapport Rapport annuel sur le marché du transport par autocar et sur les gares routières en France en 2020 Autorité de Régulation des Transports, Rapport Should I go by bus? The liberalization of the long-distance bus industry in France Le marché des bus longue distance en France: Ouverture, consolidation et perspectives intermodales [The long-distance bus market in France: Liberalizaio, consolidation and intermodal perspectives BlaBlaCar passe le cap des 100 millions d'utilisateurs, notamment en raison du prix de l'essence [BlaBlaCar passes the milestone of 100 million users Return to a world transformed: How the pandemic is reshaping corporate travel? Deloitte Insights Comparing French carpoolers and non-carpoolers: Which factors contribute the most to carpooling? Air transport statistics How accessibility shapes land use En juillet, le trafic autoroutier du réseau Vinci est repassé au-dessus de son niveau de 2019. Investir Impacts of the COVID-19 crisis and national responses on European railway markets in 2020 Public transport use in the context of Covid-19 crisis: Evolution and perspectives from Lyon metropolitan area Effects of open access competition on prices and frequencies on the interurban railway market: Evidence from Europe Industrial organization: Competition, strategy and policy November 1-2). The COVID pandemic in Brussels: The necessary boost for a sustainable mobility transition? Why do (or don't) people carpool for long distance trips? A discrete choice experiment in France Platform competition in two-sided markets The influence of COVID-19 on international and long-distance passenger rail transport. The cases of Italy and Poland -The first observations American inflation: Global phenomenon or homegrown headache? The Economist Crise du Covid-19: BlaBlaCar résiste mieux que le train et l'avion Will Covid-19 put the public back in public transport? A UK perspective The effect of COVID-19 on long-distance transport services in France