The Governance Model of the Dutch main railway market for passengers, a failure in itself?

(1)

The Governance Model of the Dutch main railway market

for passengers, a failure in itself?

Univers ity of Amsterdam, Faculty of Economics and Bus iness

Master Thesis in Industria l Organisation, Regulation and Competition Policy

2016, August 15

Authors:

Simone de Boer 11090219 Joep Hoefs loot 6168752

Supervisor:

(2)

Abstract

The aim of this thesis is to determine the optimal governance model for the Dutch main passenger railway market so the welfare is maximized, and to provide insights if and how the current institutional framework can be improved. One of the key characteristics of the ra ilway network is the imbalance of power between the different actors on the market: the government, the infrastructure manage r and the railway operator. To argue how the governance could optimally be structured, a theoretical framework is constructed, which is based on economic criteria and tested by interviews and existing empirical results. This research concludes that the first best option is an integrated railway market, where the railway organizations are publicly owned and pos itioned at distance from the government. Though, this research argues that separation of the operations and infrastructure manager is social optimal, s ince reintegration could result in high costs. Moreover, the governance model fa ils, and has to be improved.

(3)

Table of Contents

FOREWORD ... 4

1. INTRODUCTION ... 5

2. INSTITUTIONAL FRAMEWORK: MAIN RAILWAY MARKET FOR PASSENGERS ... 8

2.1 Legal framework ... 8

2.1.1 European law ... 8

2.1.2 National Law ... 9

2.2 Historical analys is ... 9

2.2.1 Institutional changes in the market organization ... 10

2.2.2 Concluding remarks ... 12

2.3 Government - organization relations ... 13

2.3.1 Public shareholding ... 13

2.3.2 Concession ... 14

2.3.3 Subs idy provis ion ... 15

3. PUBLIC INTEREST ... 18

3.1 Market power ... 19

3.1.1 Ways to intervene and possible government fa ilure ... 21

3.2 Externalities ... 23

3.2.1 Ways to intervene and government failure ... 24

3.3 Information asymmetry ... 26

3.4 Public goods ... 29

3.5 Transaction costs ... 31

3.6 Non-economic reasons: paternalism and redistribution of income ... 32

3.7 Summary and concluding remarks ... 33

4. THEORETICAL FRAMEWORK AND THE WORKING PROPOSITIONS ... 37

4.1 Theoretical framework ... 37

4.1.1 Interviews ... 38

(4)

4.2.2 Government production – Private production ... 41

4.2.3 Competition - Monopoly ... 42

5. RESULTS AND ANALYS IS ... 44

5.1 Integration or separation? ... 44

5.1.1 Government control of the coordination between the infrastructure manager and the operator ... 45

5.2 Government production or private production? ... 46

5.2.1 Government control and information exchange issues... 48

5.3 Competition for the market ... 53

5.4 Summary and concluding remarks ... 54

6. POLICY IMPLICATIONS ... 56

6.1 The Railway agency ... 56

6.2 Improved current governance structure (third best option) ... 59

7. CONCLUSION ... 61

LITERATURE ... 63

(5)

Foreword

This thesis focuses on the governance of the Dutch main railway market (Hoofdrailnet), and analyses where government intervention fa ils, and how it can be improved. To come to our results, we have used literature and interviews with the representatives of ProRail, the Nederlandse Spoorwgeen and civel servants of the ministries involved. This research has been conducted to finish our master’s degree in the fie ld of Economics at the Univers ity of Amsterdam. We started with our orienta tion in Januari, 2016, under supervis ion of prof. dr. B.E. Baarsma, and finished our thesis in August, 2016.

During this time, we have learned a lot about how complex the (Dutch) railway market is. This means that changes on the ra ilway market have differentiated effects, depending on the measure and on the environment of the market (politics, demographics, country specific economics, etc.). To master this complexity, we had to invest a large amount of time and effort. To guide us, prof. dr. ir. A.M. Boersma has introduced us to the fie ld of governance, which has given us valuable ins ights about the Dutch railway market. Prof. dr. ir. A.M. Boersma has also provided us the contact information. Therefore, we would like to thank prof. dr. B.E. Baarsma and Prof. dr. ir. A.M. Boersma, for helping us conducting our research, which could not have been done without their help, as we started analys ing this complex market from scratch.

Furthermore, we would like to thank the following people, for taking part in the interviews: H. Boussen (concession issuer); H. Borburgh (shareholder ProRail); J. Durrer (shareholder ProRail); A. Es-Saghir (shareholder NS); M. Burggraaf (shareholder NS); P. Eringa (CEO ProRail); R. van Boxtel (CEO NS) and J. Bierman (public transport expert). All these people were really helpful and put a fa ir amount of effort in giving useful ans wers, for which we are really thankful.

Enjoy reading,

(6)

1. Introduction

Although the reform of the Dutch railway market started in 1995, consensus about the institutional framework does still not exist (Kamerstukken II, 2011; Kamerstukken I, 2012; Veraart, 2013).1_{The main}

objective of the reform was to introduce competition, however, the complexity of the market makes this difficult. The infrastructure entity is separated from the company that operates the trains. However, accurate coordination between the infrastructure manager and the operator is required because of the technical interdependence between the entities. Furthermore, the independent pos ition of the railway organizations seems to conflict with how the government intervenes. The government directly intervenes to gain more control over the organization. However, too much political interference with the operational environment can result in inefficient decis ion-making (Peschelt, 2010; Beck, Bente & Schilling, 2013; Desmaris, 2013). Discontent by the passengers and the media strengthens the discussion; overcrowded trains2_{, delays and unclean tra ins are the order of the day (Rijden de treinen?, 2016).}

Railway networks seem to work we ll ins ide a framework of clear rules, clear strategies and a clear financia l constraint (Peschelt, 2010; Beck et al., 2013; Desmaris, 2013). One of the key characteristics of the railway network is the imbalance of power between the different actors on the market: the government, the infrastructure manager and the railway operator. The government sets the railway policy and rules and is respons ible for the functioning of the railway market. The railway organizations are the executing organizations and account to the government. The train operator is dependent on the infrastructure manager since the operator has to use the infrastructure. However, to structure the different principal-agent relations in a clear and transparent way is complicated. Not only because all actors have their own interests, but the complex operational and dynamic environment of the market is also challenging. Governance theory can shed light on key elements of these principal-agent relations and the underlying functioning of the railway network.

1_{Other sources: http://tinyurl.com/hjdd7nf; http://tinyurl.com/gno39n9} 2_{Source: http://tinyurl.com/zjlem3k}

(7)

1. Introduction

Consequently, our research question is: How should the governance of the Dutch main railway

market for passengers be structured so that welfare is maximized? This thesis focuses on the Dutch main

ra ilway passenger market for passengers 3_{because this is the most important part of the market for train} passengers and because competition issues on the regional railway market can be excluded. Based on the discussion in Spanhove & Verhoest (2007), governance is defined as how the government steers, controls and monitors the railway organizations, focussing on an efficient and effective realizations of the objectives, as well as on how the railway organizations account to the government (and other stakeholders). This is a relevant research question s ince another governance model can improve the railway services and reduces costs. This increases the welfare of train passengers and taxpayers. Furthermore, a lot of research has been done trying to determine the optima l institutional framework for the Dutch railway market. Bekkers (2008) gained some pros and cons of the chosen governance model, and in particular into investments, performance and innovation. Veraart (2013) suggests improvements in the government role given the particular market organization. Furthermore, there have been a lot of observations on the functioning of the institutional context; however, economic analys is of improvements in the Dutch market organization has little attention (Kamerstukken II, 2005; Kamerstukken II, 2011). We try to contribute to the debate by constructing a theoretical framework, which is based on economic criteria and tested by interviews and existing empirical results, to argue how the governance could optimally be structured.

To answer the research question, it is discussed how the current Dutch main railway market is organized. This part is based on an analys is of the institutional changes, which determines the current institutional framework. Thereafter, it is analysed how the different government-organization relations are structured. Subsequently, the question is whether government intervention in the market is justified and how the government could intervene from economic perspective. Thereafter, it is argued how the governance model in the Dutch main railway market for passengers should be structured based on a theoretical framework. The purpose of this research is to determine the optimal governance model for the Dutch main railway passenger market, and to provide ins ights in if and how the current institutional framework can be improved.

(8)

1. Introduction

This thesis is structured as follows. Section 2 discusses the current market organization based on a historical analys is. Subsequently the government-organization relations are examined. Section 3 discusses the public interests to analyse whether and how the government can intervene according to economic theory. In section 4, the theoretical framework is constructed, which is used to answer the research question. Section 5 analyses the results of the interviews and existing empirical results to determine whic h economic criteria apply and to examine how the governance in the Dutch main railway market for passengers should be structured. Section 6 provides ins ights how the results can enhance the current governance model of the Dutch main railway market for passengers.

(9)

2. Institutional framework: main railway market for passengers

This section briefly discusses the current institutional framework of the Dutch main railway market for passengers. In order to situate the subject, the section starts with a brie f overview of the legal context. Section 2.2 analyses the institutional changes after the reform in 1995. Subsequently, the current market organization is presented. Section 2.3 discusses the current relations between the government, infrastructure manager and operator.

2.1 Legal framework

2.1.1 European law

Since the nineties, the European Commission (EC) is gradually trying to liberalize the railway markets in Europe. It started with the first railway directive in 1991 (91/440/EC), which is fo llowed by three other railway packages4_{. Ma in objective of the first railway directive was to separate infrastructure}

managers and transport operators by keeping at least separate profit and loss accounts and balance sheets and publish them individually.5

Most recently, in April 2016, the European Parliament and the Council of the European Union reached an agreement on the fourth railway package. The aim of this package is to revita lise the rail sector and deliver better quality of service and more choice to passengers. This package contains two major changes. Firstly, it aims for a single European market for rail services. The package will establish the right for European undertakings to operate everywhere in the EU from 2019 onwards and set the rules for competition more stringently so that train companies get non-discriminatory access to the network in all EU countries for the purpose of running domestic passenger services. Secondly, the tendering of public service contracts will ensure that authorities get the best possible value for public money while retaining an

4_{The second and the third railway package mainly concern the freight market.} 5_{Source: http://tinyurl.com/jdu3rql}

(10)

optimal level of public transport services for citizens. In order to not disrupt the current railway organisation in the EU countries, the Commission added, “Member states authorities would be able to opt for direct award if justified by the structure and geographical characteristics of the market and network and if it would improve the quality of services and/or cost efficiency”. In The Netherlands, the EC has acknowledged that direct award is justified for the main railway market for passengers. However, regional areas have been subject to tendering.6

2.1.2 National Law

The ministry of Infrastructure and Environment (hereafter: Ministry of I&E) is the legis lator of the Dutch railway market. The two most important laws concern the Railway Act (Spoorwegwet) and the Concession Act (Concess iewet personenvervoer per trein). The (revised) Railway act and the Concess ion act came into force in 2003. The purpose of these laws is to guarantee an efficient and safe railway market and to give the government possibilities to control the railway market. The Railway Act determines the roles, respons ibilities and authorities of the market participants and the government. It includes for example how the infrastructure is provided, how the capacity is divided and the amount of user fee that the operator has to pay to the infrastructure manager. The principles behind the concessions used in this market can be found in the Concess ion Act. The legis lation also ass igns specific ministries to have the final responsibility for a well-functioning railway market (Blank et al., 2013).

2.2 Historical analysis

A historical analys is is given to discuss the institutional context. The institutional changes in the period after the reform in 1995 are analysed because these have determined the current institutional framework.

(11)

2.2.1 Institutional changes in the market organization

Market organization of the Dutch railway market can be set against a long history; especially the period between 1990 and 2016 is characterized by important institutional changes. Before the reform in 1995 the Nederlandse Spoorwegen NV (hereafter: NS) was the owner, manager and operator of the whole Dutch railway infrastructure (Kamerstukken I, 2012). The NS was clarified to be of major importance for society to justify yearly subs idies from the government. However, the open-ended high subs idies caused discontent and the idea existed that the subs idies were inefficiently used. Discontent and the introduction of the new European Directive 91/440/EC gave rise to reform the Dutch railway market started in 1995 (Blank et al., 2013). Ma in objectives of the reform were separation of the railway company, introduction of competition and independency of the railwa y company.

The separation of the railway company concerns the separation of the entity respons ible for owning and managing the railway infrastructure from the entity that operates trains upon it (Blank et al., 2013). Although separation was intended, at the time of the preparation of the agreement the NS was not prepared to completely separate the infrastructure entity from operations (Kamerstukken II, 1999). Reason was the technical interdependency between the activities of infrastructure and operations, which is very complex and is crucia l for the functioning of the market. Therefore, the infrastructure and the operation part were only judicia lly separated. The NS remains the operator, while the ownership and management of the infrastructure were transferred to three so-called task organizations. This judicia lly separation model resulted in unclearness of control s ince both the government and the Board of Directors of the NS controlled the infrastructure organizations. This unclearness caused a stepwise organizational procedure started in 2001 to separate the infrastructure part completely from the operation part. This resulted in a merger of the three task organizations into one infrastructure organization named ProRail in 2005 (Kamerstukken I, 2012). However, what are the effects of the separation of such a highly interdependent technical organization and what is the role of the government?

Separation was seen as an important institutional arrangement for introducing competition (Blank et. al, 2013). The entrance of Lovers Rail introduced competition on the railway market in 1996. However,

(12)

le ft the market in 1999. Since competition on the market was not a success, a concession model was introduced with the purpose to stimulate competition for the market. In 2005, the government directly awarded the concessions to both the NS and Prorail, for respectively the operation and the infrastructure. In 2015, the concessions were again directly awarded to both the companies for a period of ten years.7_{In both}

concession periods, the NS and ProRail have a monopoly pos ition in the main railway market. However, the direct awarding procedure does not allow for competition for the market. Hence, whether competition in this market is feasible can be questioned.

Both the operator and the infrastructure manager have been always in public hands, even though the operations were seen as a commercia l activity and an initia l public offering of the NS was proposed in 1993 (Blank et. al, 2013). According to the government, the NS should remain in public hands because of the complexity of the market (Kamerstukken II, 2014). Infrastructure is seen as a public task because of the significant subs idy, the spatia l planning of the country, the des ire of regulating the use of scarce infrastructure capacity and the general concern for safety (Kamerstukken II, 2000). However, as long as law and regulation can guarantee the interests, private firms operate in general more efficiently (Boot, 2006). Moreover, are subs idies a justification for government production?

Although the railway organizations remain in public hands, independency of the railway organizations was also an important objective of the reform (Kamerstukken I, 2012). Independency is defined as the distance between the government and the organization, which determines to which extent the ra ilway organizations can function independently. However, independency referred more to reducing the subsidies because the NS was already an independent public organization because of the legal form (NV).

At the time ProRail was founded, the idea was that the infrastructure manager should positioned closer to the government than before. However, such a change could have resulted in too much turmoil, therefore, ProRail remained an organization at distance from the government (Kamerstukken II, 2011). Despite formal independency of both organizations, the government had significant influence on the po licy of both organizations. However, the operational environment and the monopoly pos itions, makes it difficult to

(13)

determine how much intervention the government should exercise. The operational environment may require more independency, while monopoly power may require the opposite policy.

All in all, the reform that started in 1995 has resulted in the organization of the Dutch main railway market for passengers as presented in figure 1. ProRail B.V. is the infrastructure manager of the Dutch ra ilway market and the Ministry of I&E is the owner of the organization via Railway Infratrust BV. The Nederlandse Spoorwegen N.V. (the NS) is the operator of the Dutch main railway for passengers and the

Dutch Ministry of Finance is the owner of the NS.

2.2.2 Concluding remarks

The institutional context has been under discussion for years. Not only because of the separation of the railway organization and the attempts to introduce competition, but also government control over the independent publicly owned firms has been questioned. The following section further discusses how the re lations between the government and the railway organizations are structured.

Figure 1. Market organization of the Dutch main railway market of passengers

Figure 1. Authority for Consumers & Markets and Human Environment and Transport Inspectorate

(14)

2.3 Government - organization relations

The government owns and controls both railway organizations. Next to law and regulation, the government uses different instruments to control the railway organizations , which are public shareholding, a concession and subs idy provis ion (Ministerie van Financien, 2013). The following paragraphs discuss the government – organization re lations by elaborating on these control instruments.

2.3.1 Public shareholding

The Dutch government has been always the only shareholder of both ProRail and the NS. Before 2007, the idea was that public shareholding should only be implemented if law and regulation would not guarantee the public interests. In 2007, a new ownership policy was introduced and this trend reversed to “public unless”. In 2013, the current policy has been introduced which states that the government should investigate what the optimal organization is per case. According to this policy, the government as a shareholder should ensure that the public money is managed in a respons ible way within the public organization (Ministerie van Financien, 2013).8

Since 2007 active public shareholding is implemented, which means that the public interests are actively pursued by the ministries. The legal powers of the shareholder depend on the chosen company regime, namely, a full regime or a weakened regime. At a full regime, the shareholder is responsible for, inter alia, appointment and dismissal of the Supervisory Board, approval of major investments, approval of major changes in the strategy of the organization and approval of the financial statements. A weakened regime gives the shareholder even more possibilities to influence the organization. The Civil Code part two and the statutes are the judicia l frameworks . The Ministry of I&E and the Ministry of Finance are the shareholders of ProRail and the NS, respective ly (Ministerie van Financien, 2013).

Public shareholding is based on private law and the organizations concerned are often two -tier companies. A two-tier company is required to have next to the Board of Directors, a Supervisory Board.

8_{The policy of public shareholding in the Netherlands can be found in Nota Deelnemingenbeleid}

(15)

The Board of Directors is responsible for the management of the company and is accountable to the Supervisory Board. The legal powers of the Supervisory Board depend on the chosen company regime, namely, a full regime or a weakened regime. At a full regime, the Supervisory Boa rd approves important decis ions (e.g. investments or changes in the statutes), appoints directors of the company and approves the financia l statements. A weakened regime gives the Supervisory Board fewer possibilities. The NS is characterized by a full regime, while ProRail is characterized by a weakened regime9_{(Kamerstukken II,}

2005).

Public ownership is preferred whether independent functioning of organization is desired so that it can make efficient decis ions without too much political involvement (i.e . at distance from the government). Although public shareholding seems beneficia l, public ownership has its limitations. Not only because private shareholding has the desired discipline in most cases, but also s ince the distance could result in less transparency and less clear information exchange between the government and the organization. As a result, the organization could behave in its own interest and the government could have the tendency to interfere in a more direct way (Kamerstukken II, 2005).

2.3.2 Concession

A concession is a contract, which grants the concess ionaire the right to reap the benefits from a certain (part of a) market for a certain period (OECD, 2007). ProRail and the NS are the concessionaires of the infrastructure and the operations respectively. As an issuer of the concessions, the Ministry of I&E is respons ible for enforcement of the concessions.

9_{ProRail is a policy holding and operates under a weakened regime. The organization has therefore}

less business autonomy than a “normal” public holding. This could result in a high degree of operational control. A policy holding is chosen whether it is inefficient to separate the policy maker role and the shareholder role into different ministries (i.e. if it is difficult to distinguish policy aspects from the autonomous functioning of the organization) (Ministerie van Financien, 2013).

(16)

The concession is an important control instrument s ince it contains measurable key performance indicators (KPIs) (Ministerie van Infrastructuur en Milieu, 2015). Performance indicators for the infrastructure manager concern, inter alia, customer judgement, passenger punctuality, rails delivered and customer inconvenience caused by infrastructure distortions. Performance indicators for the operator concern, inter alia, customer judgement, travel information, transport capacity during the peak hours, social safety and passenger punctuality. For each indicator, target values are set. The organizations account for whether or not these target values are achieved each year. Although a concession model seems beneficia l, the success of a concession depends on the design and the awarding procedure (OECD, 2007). For example, when a KPI is not correctly defined, the railway agent could behave in its own interest at the expense of the purpose of the particular KPI.

2.3.3 Subs idy provis ion

Another important control instrument is the provis ion of subs idy (Ministerie van Infrastructuur en Milieu, 2015). ProRail is highly dependent on the subsidy. The costs are for eighty percent subs idized by the State and the users’ fees paid by the train operators concern the other twenty percent. Based on the management plan, ProRail composes a subs idy application each year. Based on the application, the Ministry composes a subs idy provis ion. The Ministry of I&E can exert pressure on which aspects ProRa il should invest. However, this could result in inefficient decis ions. On the contrary, the infrastructure manager can abuse its monopoly power to obtain a higher amount of subs idy than needed. The NS is not (formally10_{) subs idized.}

10_{The NS is indirectly subs idized s ince the organization pays a lower user fee than whether the}

(17)

2.3.4 Concluding remarks

The government – organization relations are structured based on different control instruments, which are summarized in figure 2. Although the control instruments seem to have the desired effects on paper, the implemented instruments could be ineffective when the information exchange is non-transparent and the government’s interests and organizations’ interests are not aligned. The organization can behave in an opportunistic way s ince the organizations (agents) have an information advantage over the government (principal). Next to potentia l government – organization problems, issues could also exist between the infrastructure manager and the operator. The operator (principal) is highly dependent on the infrastructure manager (agent) s ince they are technical interconnected. Another important relation is the relation between the Parliament and the ministries. The Parliament can exert pressure on the ministries, which can strengthen government control over the railway organizations. Table 1 summarizes these principal-agent re lations.

To analyse the effectiveness of the government control, crucia l is whether the control instruments are optimal chosen and implemented. Therefore, the next section discusses the reasons for government intervention and analyses which instruments are potentia l for the intervention reasons according to economic theory.

Figure 2. Government – organization relations in the Dutch main railway market for passengers

(18)

Table 1

Principal – agent relations in the Dutch main railway market for passengers

Principal Agent Citizen Parliament Parliament Parliament Ministry of I&E Ministry of Finance

Ministry of I&E* ProRail BV **

Ministry of I&E* *** NS NV **

Ministry of Finance* *** NS NV **

NS NV ProRail BV

Note. * = European Union sets the prerequisites for the national railway regulation and policy; ** = Supervisory Board has a supervisory role on the Executive Board; *** = Ministry of I&E and Ministry of Finance could have a principal – principal relation.

(19)

3. Public interest

Ideally, the government does not have to intervene when the market facilitates a socially acceptable level of we lfare by itself. However, in a s ituation without government intervention, markets can fa il. A market fa ils whether the price mechanism fa ils to optimally allocate resources. Market fa ilure is a reason for the identification of public interests and can justify government intervention. Government intervention, in its broadest sense, is defined as government influence on the market outcome. In addition to market fa ilure, social and political preferences can be arguments for public interests , which can justify government intervention. Consequently, the government can have a role to correct market fa ilures and (political) undesired effects to guarantee public interests (Baarsma & Theeuwes, 2009).

Economic theory indicates the welfare effects of the different instruments, which the government can use to correct market failures. The optimal choice depends on the nature of the market problem(s). A discrepancy between the theoretical optimal way to intervene and how the government intervenes reveals government failure. Government fa ilure can be higher than the social costs of the market fa ilure. Therefore, although market failures and public interests justify government intervention, the government should only intervene in case the benefits of government intervention outweigh the costs of government intervention, i.e. the socia l we lfare increases if the government intervenes (Baarsma & Theeuwes, 2009).

This section discusses whether the Dutch main railway market for passengers fails to satis fy the public interests and argues how the government could intervene. This section analyses the standard market fa ilures, market power, externalities, information asymmetry and public goods as well as transaction costs. Each market failure is examined separately by analyzing the Dutch main railway market for passengers without government intervention. For each market failure, it is also discussed how the government could intervene and how the government could fa il according to economic theory. Thereafter, the non-economic reasons for government intervention are discussed.

(20)

3. Public interest

3.1 Market power

In case of an integrated railway market (the infrastructure and operations are integrated within one company), economies of density cause a monopoly to end up with the lowest costs. Economies of dens ity are defined as the number of train kilometres per kilometre of railway track. Thus, the higher the density, the more efficiently the infrastructure is being used. A high dens ity is important in the railway market because the railway company has to earn back the very high fixed costs of constructing the infrastructure (see figure 3). 11_{Therefore, the long run average costs decrease if the density on the railway market}

increases. Furthermore, figure 3 shows that the fixed costs realize a high share of the total costs, which means that the costs for an extra unit of output (marginal costs) are relative ly low. This means that a ra ilway company incurs relative ly small marginal costs for increas ing the dens ity. Since this point is reached at a high density due to the high share of fixed costs, the integrated railway market is naturally served by a monopoly (The World Bank, 2011). 12

The analys is of a separated railway market, when the operations are separated from the infrastructure, concludes otherwise. The reason for this is the difference in cost structure between the infrastructure market and the operator market. The infrastructure market concerns high fixed costs, while the fixed costs for the operator market are much lower. Since the monopoly power of the integrated railway company comes from the cost structure of the infrastructure only, the infrastructure market in a separated market will a lso be optimally served by a monopoly. As can be seen in figure 4, the operator market is concerned with lower fixed costs and a steep s lope of the variable costs line in comparison with the infrastructure market. This means that marginal costs increase faster on the operator market following from

11_{These figures are for railway markets in general. The World Bank does not differentiate between}

countries. The accuracy of the analys is may increase when there is country specific data available.

12_{Economies of scale do not play a s ignificant role. As the World Bank (2011) summarizes from}

different studies, economies of scale may arise when railway networks are very small. Realizing economies of scale with larger railway networks is harder because of increas ing management complexity and loss of corporate agility.

(21)

3. Public interest

an increase in density. For example, an increase in density would cause that the in frastructure has to be inspected more often, while the operator has to hire more employees and use more fuel. The latter has a higher correlation with dens ity, which means a higher increase of the costs in case the density increases (The World Bank, 2011).

Following The World Bank (2011), what is important for the cost structure of an operator is the amount of passengers in a train, the amount of personnel on a train and the s ize of a train. Passengers pay for tickets and therefore generate income, while personnel results in high costs. Furthermore, the s ize of the train determines the costs. For example, two trains are more expensive than one train that can carry the same amount of passenger because building one train is cheaper and one train uses less ene rgy13_{. These}

variables do not hold back entrants from entering the market because they are easily copied. Therefore, it can be concluded that the cost structure does not result in market power on the operator market.

However, when several firms provide the network, coordination costs can be high s ince it is difficult to optimally interconnect all ra ilway tracks. The monopolistic characteris tic of the infrastructure market gives not a problem between infrastructure managers. Interconnection of the operator(s ) and the infrastructure manager can be complicated, which may negatively impact the functioning of the railway market. Consequently, high coordination costs can allow for market power, which can harm we lfare.

When the government does not intervene, the railway organizations will have the incentive to increase price or to reduce quality to maximize profit at the expense of quality. Quality of the railway service concerns reliability, safety and comfort. The infrastructure manager for example does not have direct incentives to invest in new technology, even when this makes the train passages significantly safer. Hence, the following public interests are derived a reliable, safe and comfortable railway service against a competitive price.

13_{Tra ins can easily be connected on to each other, however, building two levels instead of one will}

(22)

3. Public interest

Figure 3. Rail infrastructure costs

Source: The World Bank, 2011.

Figure 4. Railway operating costs

Source: The World Bank, 2011.

3.1.1 Ways to intervene and possible government fa ilure

From the previous section, it should be clear that the government should ascertain the benefits of the infrastructure market with natural monopoly characteristics, while restraining the negative effects of market power on both markets. To overcome the undes irable effects of market power and maintain the benefits of a natural monopoly, the government can use different regulatory instruments. Price and quality regulation are common instruments to incentivize a monopoly to operate in an optimal way. Quality

(23)

3. Public interest

regulation concerns regulation of reliability, safety and comfort. However, if regulation is not designed properly, the monopoly will manoeuvre in such a way that it maximizes profit, possibly at the expense of we lfare. For example, if a price cap is set, a subtle cut in quality can be a very tempting way of cutting costs. Moreover, measurement difficulties in quality and costs provide a regulated firm with an advantage re lative to the regulator (Viscusi, Harrington & Vernon, 2005).

Alternative for regulation is introduction of competition for the market (OECD, 2007). Introduction of competition for the market forces the firm to lower their price / increase quality as in the case of a competitive environment. The argument goes that the pressure from other (potentia l) bidders will induce efficient decis ions by the incumbent. For the railwa y market, concess ions appear to be useful s ince they can both introduce competition and regulate at the same time. The government can issue concessions, incorporating performance indicators and quality standards. In this way, the government retains control over the company, while the organization carries out the day-to-day operations. However, information asymmetry between the government and the organization can make it difficult to enforce the concession. Information asymmetry between the incumbent and an entrant could impede entry. The asymmetric information problems are further explained in section 3.3. Another way of introducing competition for the market can be to split the infrastructure market into a construction part and a managing part. This is bene ficia l s ince the construction of infrastructure is easy to contract opposed to the complex markets of managing and operating the infrastructure. However, separation may increase the coordination costs.

At last, the government may produce the network by itself in case the coordination costs between the different operator(s) and the infrastructure manager are high. The government can also regulate the interconnection between rail operators with compensating measures if the firms cannot recoup the coordination investments. The coordination costs between the infrastructure manager and the operator(s) can be internalized by regulation and/or contracts, which is expla ined in section 3.3.

(24)

3. Public interest

3.2 Externalities

This paragraph discusses the negative, pos itive and network externalities. Positive and negative externalities are often defined as the effects of economic activities on unrelated third parties (CPB, 2004). Network externalities are defined as the effects that an individual us ing the network has on the total value of the network, i.e. the more who use the service the higher the value of the network (Katz & Shapiro, 1985). Since those effects do not directly influence the parties involved in the transaction, the parties will not take the externalities into consideration. Externalities cause market failure when the social cost (or benefit) of such transactions differs from their private cost (or benefit). However, externalities will not always result in a market failure. In case of a manageable number of stakeholders and well-defined property rights, private negotiations will result in a socia l desirable outcome (Baarsma & Theeuwes, 2009). Moreover, if the network externalities are measurable, the incumbent can set the price according to the value added to the network, internalizing the effect (Katz & Shapiro, 1985).

The external effects can be divided in effects caused by construction or maintenance of the infrastructure and the operations. In addition, the external effects can have effect on the passengers and/or

non-passengers. Construction of the infrastructure causes negative external effects when it produces

inconveniences like noise, roadblocks, and pollution, which can have effect on both the passengers and non-passengers. The existing infrastructure can dis tort the spatial planning because of the rail track itself and the cut through to the landscape. The latter especially has effect on non-users. The operations result in more noise and emissions than is socially optimal, which has negative impact on both the passengers and non-passengers. Furthermore, accidents have negative impacts on the passengers, the local residents and the passers-by. However, although the railway network negative ly impact the environment, the railway service is seen as environmentally friendly in comparison with other transport modes (Givoni, Brand &

(25)

3. Public interest

Watkiss, 2009). 14_{Consequently, one could argue that the railway service causes “pos itive” external effects}

in terms of environmental quality relative to other transport modes.

The number of connections and the number of tra ins per hour between particular areas benefits the passenger and increases the value of the network.When the number of passengers increases, the probability that the network will grow increases. This will attract other consumers and the network will become even more attractive to other consumers. This will further enhance the value of the network, which (potentia lly) benefits the other passengers. However, the passenger may not take into account this benefit on other passengers, resulting in a lower willingness to pay than the marginal costs, which can impede to achieve the optimal s ize of the railway network. For example, railway services may not be provided within a specific time s lot or some rural areas may not be reachable by train.

The reachability of (economic) areas is also beneficia l for non-passengers. The reachability benefits the employees and established companies and attracts new companies, which improves the equilibrium of labour demand and supply. Furthermore, easiness in reachability increases the attractiveness of a particular city, which in the end benefits the hospita lity and the retail sector for example.

Negative network externalities arise if the passengers’ demand is higher than the available capacity, resulting in negative effects of a passenger on another passenger. This kind of negative externalities can arise especia lly during peak hours when demand exceeds supply.

Consequently, the main ra ilway market fails because of negative and pos itive (network) externalities. A privately owned operator and infrastructure manager will not have the incentive to internalize the external effects on non-passengers (and passengers) in the prices. Hence, the following public interests can be derived: environmental quality concerning noise, pollution, spatia l distorting and safety, reachability of economic areas, and coordination of the use of the network. The next paragraph discusses how the government can satis fy these interests.

14_{Tra ins in The Netherlands are the most environmentally friendly transport mode in comparison}

with the bus, metro, tram, plane and car. However, the case that trains are more environmentally friendly than cars is not the case in each country (Givoni et al., 2009).

(26)

3. Public interest

3.2.1 Ways to intervene and government failure

The government can ensure internalisation of the external costs until a socia l optimum through regulation and standards, or by providing the right incentives to railway organizations . Environmental standards can set the prerequis ites for the fuel quality and noise disturbance. Taxes can also be used since key figures of the costs of a ir pollution and noise of railway transport exist (Korzhenevych, Dehnen, Brocker, Holtkamp, Me ier, Gibson, Varma & Cox, 2014). However, uncertainties may exist in the cost estimations for each railway track s ince the cost calculation is based on different variables as density of the ra ilway track and the population. This can impede optimal internalization of the external costs. In the end, the tax revenue can be used to reduce the damage on non-users, e.g. noise walls. Furthermore, a liability regime and safety standards can reduce the probability of accidents. Spatial planning can also be regulated. The government may fail, however, when difficulties exist in measuring the external effects of spatia l distorting on res idents, which may be influenced by lobby groups (e.g. neighbourhood committees or ra ilway stakeholders).

Underinvestment in R&D to reduce pollution and noise may justify subs idy. However, the innovation costs for infrastructure (maintenance and construction) are low (Swier, 2014). Moreover, the train is environmentally friendly and the operations of the main railway market are mostly profitable, which seems also not to justify provis ion of subs idies. One could also argue to subs idize the railway transport s ince the railway service is more environmental friendly than other transport modes. When the price of a train ticket decreases, the demand of the train may increase and the demand of the other transport modes may decrease. However, the barriers faced by travellers considering a shift of mode to railway transport are many and varied (Bla iney, Hickford & Preston, 2012). The consumer choice not only depends on prices, but also on quality aspects as reliability, comfort and safety. Therefore, the government may fa il in influencing the behaviour of the passenger. Moreover, stimulating the demand seems to be based on political arguments (see section 3.6).

When the positive network externalities are not internalized, the private benefit is lower than the social benefit. The government can use price regulation based on departure point and arrival point (determined by public transport chip card data) to internalize the benefit of the use of the railway network

(27)

3. Public interest

on other passengers. However, if these benefits are measurable, a privately owned firm can also set such a price scheme. When a privately owned railway organization expects not to recoup an investment, a railway track may not be constructed or operated. In this case, the government can regulate which tracks are socially desirable, taking into account the network externalities . This may justify provis ion of subs idies (financed by the taxpayer). The government can fa il when difficulties exists in measuring the socia l benefit of a particular railway track, which may be influenced by lobby groups (e.g. neighborhood committees). Moreover, the arguments for subs idizing a railway track can be questioned from economic perspective and seems to be more a political decis ion (see also section 3.1.6).15

To overcome the negative externalities of a passenger on another passenger, the government can introduce peak-load pricing by charging higher prices during peak hours to shift away passengers from the peak demand towards lower demand time s lots. However, a privately owned firm will set the same pricing scheme. Though a firm with market power will set a higher tariff, which will result in a lower demand (than social optimal).

3.3 Information asymmetry

As ymmetric information referred to a situation in which one party in a transaction has more information compared to another party concerning the quality, the price/quality combination and/or the characteristics of the exchanged good or service. The information advantage could be a harmful s ituation when one party can take advantage of the other party’s lack of knowledge. As a result, the less informed party would make suboptimal choices (CPB, 2004).

15_{The operations of the main railway market are mostly profitable because of the high density in}

the main railway network, which means that the market will provide these railway tracks without subs idies (the prices may be higher, however). Furthermore, other transport modes can also be chosen to guarantee the reachability of (economic) areas (e.g. buses or metros).

(28)

3. Public interest

As ymmetric information may hit railway consumers since the railway operator has an information advantage over the consumers. The railway operator can exploit the information advantage since the ra ilway consumer has little (technical) knowledge about the railway service. However, the consumer can immediately judge the price and quality of the service s ince she/he is part of the service; a clean and safe ra ilway service is observable and a clear timetable is a tool to observe punctuality. Moreover, repeat of the purchase of railway services can have the countervailing power to the railway organization not to exploit this advantage. Reason is that the consumer will learn about the quality and the price/quality ratio of the ra ilway service. When the quality and/or price/quality ratio is repeatedly not in line with the expectations of the consumers, the consumers will search for another transport mode, the operator will exit the market and another operator can enter the market. Consequently, the railway operator will have the incentive to provide a comfortable and safe railway service, and a clear and reliable timetable to build a reputation to attract consumers.

Though, the operator can exploit its information advantage. The operator can exploit the actual travel timetable by arguing that the delay takes more minutes than needed because of own interests (an operator may put more effort in solving a delay on a network with a high density compared to a network with a low density, because the network with a high density generates more revenues). However, the consumer cannot check whether the specified delay is correct. Moreover, the operator may not announce a short delay to circumvent the observations of the delay by other passengers to keep its reputation. However, such unreported delay can be very detrimental for the passengers who will use that railway service. Furthermore, the effort to which extent safety is ensured is difficult to check; to arrive on time, the train driver would not take the safety procedures too serious ly while the passenger will not be aware of that. Moreover, different external factors (e.g. copper theft or weather) can jeopardise safety, which can hide the fa ilure of a train driver. Bes ides that, the consumer has in general not the expe rtise to assess whether the price of a train ticket between particular areas is optimally set as it is hard to determine the costs per kilometre and the initia l fee. Moreover, comparison of a particular train journey with another train journey is also (almost) impossible because of differences in kilometres and number of passengers. At last,

(29)

3. Public interest

the comfort of the trains can be exploited s ince the operator would postpone cleaning as long as this is unnoticed by the passengers.

Separation of the infrastructure manager from the operations can also result in asymmetric information problems. The operator(s) depend(s) on the railway tracks managed and constructed by the infrastructure manager. Although contracts can be s igned, incomplete contracts are not hard to motivate and the infrastructure manager can demonstrate opportunistic behaviour by exploiting the dependency of the operator. The monopolis tic character of the infrastructure market strengthens this. The infrastructure manager can behave in a way that is non-observable for the operator. An example is that the operator cannot control whether the infrastructure manager mainta ins the infrastructure ensuring the safety of the passengers. Furthermore, the infrastructure manager can obtain information about some characteristics of the environment that the operator cannot acquire. An example is that construction or renovation of a ra ilway track will take a longer period than needed because of the infrastructure manager’s interest.

Furthermore, the incumbent(s) can have superior knowledge about the functioning of the railway network relative to the new entrants. In addition, the incumbents have advanced expertise about the coordination between the infrastructure and operations. This increases the entry barriers, which discourages other railway companies to enter the market, which strengthens the market pos ition of the incumbents.

In conclus ion, the asymmetric information is not a failure in itself; the combination of information asymmetry and market power, however, can have socially undesired effects. Hence, the public interests derived from the asymmetric information problem are a reliable actual travel timetable, implementation of safety procedures, a clear tariff structure and a comfortable railway service, transparent coordination between the infrastructure and operator(s), and non-discriminative access to the railway network. The following paragraph discusses how the government can intervene and how the government may fa il.

The operator can exploit its information advantage at the expense of a reliable actual timetable, implementation of safety procedures, clear tariff structures and comfort. To overcome these undes ired

(30)

3. Public interest

effects, the government can require a reliable actual timetable by law. Moreover, the government can set a maximum of minutes of delays. This will incentivize the operator to prevent delays and to optimally resolve a distortion. However, the government may fail when difficulties exist in determining the target value of a maximum delay. Safety can be regulated by liability and safety standards. However, the government may fail when difficulties exist in controlling implementation of the safety procedures. Moreover, if the government set high standards the barriers of entry can increase. Furthermore, the government can regulate the tariff s tructure by requiring transparent cost information. However, the requirement of transparency of the tariff structure can have distorted effects. The railway operator can introduce (high/low) price zones depending on the dens ity of the particular rail track. This can negative ly impact the reachability of (rural) areas. Comfort can also be guaranteed by regulation and standards.

The government can control the cooperation between the infrastructure manager and operator(s) by regulating the user fee and user access to the infrastructure. However, the government can fail in setting the optimal user fee when difficulties exist in determining the user costs. In addition, difficulties can exist in controlling the user access s ince the infrastructure manager can exploit its pos ition. The infrastructure manager can argue that it is not viable to maintain railway tracks during the nights or within a particular time period, while this is possible. This can negative ly impact the functioning of the operator(s). Moreover, even though the government can set the quality of the infrastructure, it is (almost) impossible to control whether all railway tracks are properly maintained. At last, regulation of the cooperation can have distorted effects since it could increase entry ba rriers because the incumbents have superior expertise about the coordination.

The government can regulate the information transition to new entrants from the incumbents to overcome the high entry barriers. However, the government can fail when difficulties exist in determining which information should be transmitted. Furthermore, the monopolistic character of the market can give the incumbents the incentive to hold back relevant information to protect their pos itions. The government can fa il when it is difficult to judge whether all information needed for the new entrants is exchanged, which can hinder competition. Moreover, the government can fail in case she protects the positions of the incumbents against new entrants.

(31)

3. Public interest

3.4 Public goods

Economic theory defines public goods as non-rivalrous and non-excludable goods (Kaul & Mendoza, 2005). Non-riva lrous means that there is an unlimited supply of the good, and therefore if one consumes it, another can still use the same good. Non-excludable means that the provider of the good cannot exclude people from us ing the good. Therefore, free rider behaviour can exist and consumers can benefit from us ing the good without paying for it. However both the infrastructure and operation cannot be classified as public goods. The infrastructure manager can exclude operators from the infrastructure and a limited amount of operators can use the infrastructure. Furthermore, people have to buy a ticket to consume the operating part and the tra ins cons ist of a limited amount of sea ts. However, until all seats are used, the ra ilway service seems non-riva lrous. In case of a probability of overcrowded tra ins, the train service is riva lrous resulting in negative effects of the passenger on another passenger. Moreover, exclus ion seems to be more effective when the departure point is a large station in comparison with a small rural station. Consequently, s ince the railway service is in general excludable and riva l, no public interests are derived from it. However, s ince the railway service has some public good characteristics, the next paragraph discusses how the government may intervene.

The public good characteristics may justify government production. Whether preference should be given to private or government production depends on the mechanisms in place, which could discipline the organizations in a way that the public interests at issue are guaranteed. Public interests do not necessarily require government production; a private firm can provide the service if law and regulation can guarantee the public interest (Boot, 2006). Whether a certain regulatory regime is appropriate depends on the

contractibility of the public interest(s). If a regulatory regime cannot capture all aspects of the public

interest, a private firm could search for other ways to maximize profits by reducing costs at the expense of quality to the detriment of public interests. Consequently, non-contractibility of the public interests can justify government production (Hart et al., 1997; Shleifer, 1998).

(32)

3. Public interest

The government may fail whether government production is less efficient than private production. To overcome (potentia l) inefficiency of publicly owned firms, the railway organizations are often pos itioned at distance from the government. The distance creates a government’s commitment not to

intervene, which gives the railway organizations more independency to operate. However, research suggests that the distance can increase principal-agent problems because the transfer of information is more difficult and transparency in decis ion-making decreases (Boersma, 2015). The government may fa il when difficulties exist in controlling the railway organizations.

3.5 Transaction costs

A market as allocation mechanism requires a well-defined property right system and the feasibility to contract, meaning that transaction costs should be not too high. Transaction costs are defined as all forms of costs of transactions and contracts. Transaction costs occur to find the supplier (search and information costs), to purchase the product or service (bargaining and decis ion costs) and to monitor the quality (enforcement costs). Whether transaction costs are very high, a market can be impeded from emerging (Baarsma & Theeuwes, 2009).

The negative externalities on non-passengers could result in high transaction costs, which can impede welfare-enhancing transactions. These transaction costs are especially high because of the fact that there are a lot of different interest groups of which the preferences are costly to coordinate (high bargaining and decis ion-making costs). When for example a railway track will be constructed, the surrounding towns are benefited by the increase in reachability, while the noise harms not only the residents, but also the animals, which are represented by groups who defend the natural habitat of animals

Transaction costs do not only relate to negative externalities. The enforcement of government intervention often goes hand in hand with high transaction costs. As ymmetric information between the government (principal) and the railway organization(s) (agent(s)) increase transaction costs . For example, the government will have to put a lot of effort in getting the right information. These transaction costs following from principal-agent problems can be reduced by alignment of interests. This way the agent will

(33)

3. Public interest

have less incentive to disrupt the principal from getting the right information. Furthermore, the transaction costs could be high between the infrastructure manager and the operator in case the information exchange is non-transparent.

3.6 Non-economic reasons: paternalism and redistribution of income

Paternalism and redistribution of income are political reasons for government intervention (Baarsma & Theeuwes, 2009). Paternalism exists when the government cons iders that the citizens will not make the best decis ions for themselves. A good example is that people may not be aware or take not into account the environmentally friendly character of the railway service in their transport mode decis ion. They will therefore choose another transport mode, which is more detrimental for the environment, which harms the citizen’s health. Therefore, the government will stimulate the use of the train and tries to influence the citizen’s behaviour. Other examples for paternalism are difficult to come up with. An example could be that the passenger may not take into account the benefit of avoiding travelling during peak hours. This depends on the extent to which the passenger is harmed by other passengers and the benefit of the passenger to travel within another time slot. Though, this seems to be well estimated by the passengers themselves.

To stimulate the use of the railway service instead of the use of other transport modes, the government can response to the main target group, commuter traffic. The government can require the employers to provide railway subscriptions (against a discount) or to forgive parking permits only to employees who live in areas, which are hard to reach by public transport. Furthermore, the government can direct the railway organizations to better respond to the citizen’s preferences to attract more consumers (e.g. improvement of door-to-door travelling). The government can also require that the railway operator will cooperate with large attractions (e.g. amusement parks) to provide a combined ticket for public transport and entrance against an advantageous price. At last, the government can demotivate the use of the car by increasing the road tax, fuel tax or the price of a parking ticket; however, the Dutch taxes and prices

(34)

3. Public interest

are already high in comparison with other countries.16_{Consequently, provis ion of subs idies to projects,}

which will make the railway service more attractive, seems to be welfare enhancing. The government may fa il when it is difficult to determine accurately the amount of subs idy, which may be influenced by lobby groups (e.g. interest groups for passengers can exaggerate the preferences of the passenger). In addition, provis ion of subs idy can have distorting effects, as the number of passengers will increase during peak hours and s ince the taxpayer has to pay.

The redistribution argument concerns s ituations where the government considers that the income distribution is unequal and should be rea llocated, even in s ituations where the market will not fa il (Baarsma & Theeuwes, 2009). Although the argument is based on political preferences, one could argue that this argument concerns the access ibility of public transport for all citizens against a reasonable price because the socia l and economic interest for mobility is high.

Implementation, however, can have distorting effects because of execution costs and inefficient decis ions (Baarsma & Theeuwes, 2009). From economic perspective, the government should implement a redistribution policy that as least as possible distorts (social) optimal decis ion-making. The government can subsidize the railway infrastructure until the railway service is affordable for the citizens with an average income (without subs idies this is impossible because of the high fixed costs). A free card (for everyone) is not welfare enhancing s ince this will incentivize inefficient decis ion-making. Not only because the cardholders will be inclined to use the railway service more often than needed, but also because the taxpayer who occasionally use the railway service is harmed. Next to the subsidy to the infrastructure manager, the government can provide discount cards to specific target groups (e.g. a discount card for elderly, students or the minimum wage group). In addition, the railway service should be made (practically) accessible for the disabled. The government may fail when difficulties exist in determining an “affordable” ticket price and the corresponding cost to conclude the amount of subs idy for the railway organizations. Moreover, the government may fail when difficulties exist in enforcement of the use of the discount card (next to the entitled person others could use the card). In addition, stimulating the demand of the target groups will increase the number of passengers during peak hours; a time -s lot card will be a solution.