Intermodal competition in passenger transport services: empirical evidence for Spain
Javier Campos / Manuel FernándezDepartment of Applied Economics
University of Las Palmas
/Department of Economics
Universitat Pompeu Fabra
FIRST DRAFT – COMMENTS WELCOME
This version is dated: February, 2004
Abstract. When competition in the market is not feasible or desirable, as it happens in many transport markets, it has been usually considered that competition for the market and/or between markets yields similar results. In this paper we consider that interurban transport services in Spain could possibly provide a good laboratory to empirically test some of the implications of these ideas. The radial structure of the Spanish road network has traditionally favored a concession system for bus routes that was only slightly adjusted after the privatization of ENATCAR in 1999. Nowadays there coexist a number of private (and some public) concessionaries organized on a for-the-market competition scheme with large duration contracts and limited rivalry among them. The major source of competition on selected routes comes from rail services, operated by the national company RENFE. Therefore, by empirically analyzing price and frequency decisions of a sample of Spanish firms providing coach services during the 2000-2003 period, this paper shows that this expected result does not always arise. This provoking result still deserves some qualifications, but it could also yield relevant new policy recommendations for the future of transport concession markets in Spain.
Keywords:Intermodal competition, price and quality competition in transport services.
JEL Nos:L1, L42, D44.
1Introduction
In many passenger transport markets competition in the market – among a large number of firms providing competing services to the same potential users – is not feasible due to the capacity restrictions associated to their supporting infrastructures (roads, rail lines…) Sometimes it is the non-storable nature of these services and their irregular demand what makes entry by too many firms unprofitable. It has been also argued that – even if feasible – this type of competition might be not desirable in cases where externalities (such as congestion or pollution) are prone to emerge, or when there exist equity concerns (regarding tariffs, frequencies, access by the poor, etc.) that might not be properly addressed by pure market rules.
In most of these circumstances it has been often presumed that competition for the market – in the form of concession contracts competitively awarded to exclusive providers – and competition between markets – in the form of intermodal rivalry for the same or parallel routes – should suffice in passenger transport markets to play the disciplinary role that actual competitors cannot play (Baumol et al., 1988). Recent transport liberalization and deregulation experiences across the world seem to suggest that this approach yields better results than the traditional one, which consisted in the straight price and quality regulation of a (usually, state-owned) monopolistic firm providing transport services under the strict supervision of a regulator.[1]
Interurban bus services fit particularly well in the above description. With the exception of corridors between major cities or within areas of relevant economic activity, market size is relatively small to support a large number of transport operators, and even if intercity roads are not excessively congested, travelers always demand convenient and safe connections among a number of origins and destinations that might be not profitable to serve in a pure profit-maximization strategy by a private firms (Morrison and Winston, 1985). In addition, the generalization of the use of private cars in the last decades, have left this market mainly to customers with lower-than-average income and low price-elasticity.
All these features have dominated the (re-)organization of bus transport markets in the world during the last three decades, although the depth of the reforms varies widely across countries mostly depending on the local tolerance to the introduction of private participation. Within the European Union, for example, only the United Kingdom and Ireland have achieved a complete deregulation with free entry of private operators, whereas the dominant market structure elsewhere is the concession system with exclusive rights for a number of years ranging from 10 to 30.[2] The future European Transport policy (see COM, 2001) suggests a consolidation of this ‘controlled competition’ mechanism, together with a support of intermodal competition as a disciplinary device with potential positive effects on the existing externalities.
In this paper we consider that interurban transport services in Spain could possibly provide a good laboratory to empirically test some of the implications of these polices. In line with the GDP, the Spanish transport market has been steadily growing since the beginning of the 1980s, at annual rates well over 4-5%. This growth has mainly benefited to private transport, domestic flights and high-speed train corridors, whereas traditional rail services have lost market share and interurban bus services have grown at a lower pace. The radial structure of the road network – centered on the capital, Madrid – has traditionally favored a concession system for bus routes that was only slightly adjusted after the privatization of the major state-owned operator (ENATCAR) in 1999. Nowadays there coexist a number of private (and some public) concessionaries organized on a for-the-market competition scheme with large duration contracts and limited rivalry among them. The major source of competition on selected routes comes from rail services, operated by the national company RENFE.
This structure gives rise to several questions, two of which constitute the basic motivation of this paper. First, we will investigate to what extent is intermodal competition driving the price and quality (e.g., frequencies, fleet improvements, etc.) decisions in this market; and second, we are interested in knowing whether the concession system used in Spain is effectively promoting the efficiency results it is intended for.
Previous works have already addressed these issues but from a more general perspective. For example, De Rus and Herce (1996) and De Rus and Nombela (1997) describe the characteristics of the Spanish bus market and provide some results on urban concessions. De Rus (2000) criticizes the design of the concession system and its perverse effects on firms’ costs and requested subsidies, whereas Bel (1994) provides one of the best accounts on the demand for intermodal competition and Inglada (1992) specifically focuses on their effects on estimated price elasticities. Our approach differs from previous studies in the use of actual price and frequency data (obtained from the internet and directly from the companies)[3] to construct an structural model that allows the testing of how are the companies’ decisions affected by the presence/absence of intermodal competitors (namely, railways services) on selected routes. Despite the intrinsic difficulty of capturing all the differences among routes, services and concessionaires, and the fact that this paper only provides an empirical analysis of some stylized facts without further theoretical development, our results show that the extent of intermodal competition in passenger services in Spain is lower than expected, thus suggesting that there remain unexploited opportunities for liberalization to explore in the future.
To carry out this introductory analysis, this paper has been structured as follows. After this brief introduction Section 2 has been devoted to the empirical work. After describing the database and its limitations, we summarize the results of our estimations.Finally, Section 3 concludes discussing the main results and the policy implications of this paper.
2The empirical model
Economic theory predicts that competition for the market – that is, a concession system which allows the winner of a bid provide in a monopolistic way transport services on one or more routes – should yield the same price as competition in the market if the concession system is well designed and transfers most of the concessionaire’s market power in favour of its customers. Since, as mentioned above, the legal framework of Spanish passenger transport serviceschooses the concession system as the mechanism to organize the market, our null hypothesis will be that the degree of market power that the winner of the competition will possess does not affect the price, which is fixed throughout the bidding process.[4]
However it is possible that, either because of lack of perfect competition in the tender process or because of later adjustments in prices (that can be negotiated ex-post), bus firms manage to exercise their monopoly power. In this case a profit maximizer firm will tend to charge larger prices whenever the consumers’ best alternative happens to be a bad substitute.
As it was argued above we will consider the quality (mainly, the frequency) of train passenger services (on the same routes) as a proxy for the market power of the bus monopolistic operators, thus resulting that our empirical analysis will be then summarized into the contrast of the following hypothesis:
H0: Pbus*=f[qualitybus*, demandbus*]
H1: Pbus*= f[qualitybus*, demandbus*, qualitytren*]
2.1Data
Bus
In Spain the inter-provincial road transportation of passengers is organised in a number of lines that communicate different areas of the country. In principle, the monopolistic exploitation of each line is granted periodically by the Ministerio de Fomento through an open auction process, and the firm offering the best (economic and technical) conditions is granted monopolistic rights over the line for a long (and renewable) period of time.
According to data provided by the Ministerio de Fomento, in September 2002 there existed around 120 different bus routes connecting one or more Autonomous Communities in Spain[5]. This group of lines is very heterogeneous in size, geography, demand conditions, etc. There exist routes from 22kms up to 3367kms. Some lines, with 65kms of distance on average, providedrevenuesover 1,400 millions pesetas a year (8,4€ million) while some others, with 45kms of distance, make little more than 200.000 pesetas (1,200€). There exists a line with 900kmswith revenuesof 74 millions and another of 1,200km that produces 1500 millions of pesetas.[6]
Out of these 120 lines we only considered those ones that linked at least two provincial capitals, which reduces the population size to 65. Out of these lines data was gathered for those cases where the operating company offered online information. Data collection took place in September 2002 .This procedure allowed to cover a 35 out of the 60 lines. These lines include the majority of the traffic and the possible connections.
Information includes the price in 2002 of the trip between any two provincial capitals in the line, the time required to complete the trip, the daily frequency of the service on a Monday, and the number of stops between the origin and the destination.
Additionally information was also gathered relative to the shortest road distance between any two cities,[7] the total number of travellers/km that used the line in 2001 (“Plan Infraestructuras 2000-2007”, Ministerio de Fomento) and the index of gasoline prices by state ( September 2002).
Train
In Spain train services are provided exclusively by the Red Nacional de los Ferrocarriles Españoles (RENFE), a public monopoly.In September 2002 we gathered data from the RENFE website ( relative to all routes between provincial capitals which can be also covered directly by bus.
When there was no direct train connection[8], the fastest[9] connection was considered. For every route information included the price of a 2nd Class ticket, the time required to do the trip taking the earliest train on a Monday, the daily train frequency and the number of stops. A dummy variable was constructed indicating whether the trip was direct. Additionally information was obtained directly from RENFE relative to the train distance between any two destinations and the number of passengers/km between any two destinations in the year 2000.
Control variables
To control for possible factor affecting the demand of transport services a series of variables was selected from the Spanish National Institute of Statistics INE corresponding to year 2001: the aggregate GDP of the departure and the arrival provinces, the number of inhabitants, the average per capita GDP, and the average price of gasoline for the provinces of departure and arrival.
< Include Table I here —Description of BUS, TRAIN and GENERAL data >
2.2Results
First we estimate the determinant of bus prices under hypothesis I, this is, assuming that it they are only determined by costs and that no monopoly power can be exercised.We study both the global variation in prices and that variation that occurs within each line.
< Include Table II here —Determinants of Bus prices >
< Include Table III here —Determinants of Bus prices within each line >
First we consider data directly related to bus transportation.Although within a line distance does not affect the price, across lines those ones with longer distances have lower prices per km. The speed of the bus affects negatively the prices both within and between lines. This effect however is non linear, decreasing its importance for larger velocities. One possible interpretation of this result would be that although a higher speed may increase the willingness to pay by consumers, this is offset by the cost-reducing effect. Note that the number of daily services increases the price between lines, but not within.
Data relative to the number of passengers that use each line is also available (although not for each route). Lines that transport larger numbers of passengers have higher prices.Other controls have been also included. The price of gasoline appears to be positive related to the price in each line. Prices also increase with the amount of population (proxy for potential demand). However they decrease with wealth, both in total and in per capita terms. This last result could be due to the fact that the use of bus transportation services tends to be decreasing with wealth.
Similarly, we study the determinants of the price of the train. As in the case of the bus price/km is decreasing with distance. However, controlling for the type of train used, the price is increasing in the time which required to travel. Again as in the case of the bus the frequency of trains it is also positively related to the price. The number of passengers however is not related. In contrast to the bus, RENFE does not seem to charge larger prices in those routes with higher demand. The total GDPs of the origin and the destiny increase also the price, but when we consider it in per capita terms the result is the opposite: in richer areas the train is cheaper.
< Insert Table IV here –Determinants of Train Prices >
2.3The quality/price of the train services and the price of the bus services
As a first approach one could consider the case where both the bus and the train take each others price and services as taken (see Table above).We find that although the price of the train does not affect the price of the bus significantly, the quality of the train as an alternative transportation method does matter. If the train is relatively fast bus prices are lower.
Why does this happen? Within a perfect competition framework this result is difficult to interpret. It is consistent however with the existence of a monopoly by the bus transportation provider whose closer competitor will be the train.A possible criticism to this interpretation of the result is the fact that speeds of buses and trains have been taking as exogenous. This assumption is based on the independence of the companies’ decision to built new infrastructures.
We can also control for those routes where there is no direct train. In this case the effect is negative. No direct connection implies relatively lower prices. This may be related to the higher incommodity for the consumer that it means, or alternatively, to the traditional existence of a compensation tax that bus lines where there was a trade line had to pay.
However this analysis may overlook the fact that maybe train and bus decisions are taken considering each others situation. In order to solve this potential endogeneity problem we implement a simultaneous equation estimation of the decision both of bus and train prices. The fact that distances between any two destinations differ for the train and for the bus allows us to identify the system of two equations. Note that these infrastructures are typically old and can be taken as exogenously given.
< Include Table V here -- Simultaneous equation estimation>
It can be seen that the results are not significantly different from the previous estimations.The relative price of the bus and train services is negatively affected by the distance of the route and positively by the size of the population. The price of the bus is also higher with the time it requires. The effect of the amount of people transported is however opposite. While a larger amount of passengers increases the price of the bus, a larger amount of users of the train has a negative effect on its price. Finally, and consistent with the previous results, again the quality of the train (measured in terms of time) affects positively the price of the bus.