THE WORLD BANK GROUP A World Free of Poverty

Toll Roads and Concessions

Overview Tolling Options
International Experience Mechanisms for Involving the Private Sector
Objectives and Potential Consequences of Tolling Terms of Reference
Network and Planning Issues Selected References
Tariff Setting and Policy Issues Public-Private Options for Developing, Operating, and Maintaining Highways: A Toolkit for Policymakers *New*


This knowledge base deals with the general issue of toll roads. It also covers contractual options for private sector involvement (including concessions). The knowledge base covers the extent of toll road provision internationally, the objectives, benefits, and costs of a toll road program, tariff setting and development issues, and involvement of the private sector. This key issues document is based on extensive experience in the sector worldwide and follows on from the work being carried out on behalf of the World Bank and Japanese Ministry of Construction on the development of toll roads in Asia. Click here to access the seminar proceedings for "Asian Toll Road Development in an Era of Financial Crisis," March 1999.

Key Issues
The following list focuses on the key issues for toll roads, whether implemented through the public or private sector. It covers data on international experience; objectives and potential consequences of tolling, network and planning issues, tariff setting and policy issues, tolling options, and mechanisms for involving the private sector. It also includes examples of Terms of Reference and Selected References.

  1. International Experience

    Most countries have no toll roads. Where there are toll roads the tolled network typically comprises less than 5 percent of the road network. Click here to display a document presenting information on tolled roads in selected countries.

    Tolls for passenger cars average around $0.03 to $0.08 per kilometer. There are variations between and within countries, but these variations are not dependent on national income levels. Generally goods vehicles pay at least twice the passenger car toll depending on the number of axles and sometimes the number of tires. Tolls for bridges and tunnels are considerably higher per kilometer reflecting both the higher cost of construction and the lack of alternative routes.

    In most countries with toll roads the private sector has been heavily involved in development of the roads and often thereafter in their operation. In Latin America in particular there has also been extensive involvement of the private sector in maintenance and on-going operation of roads that were built by the public sector. The USA, Japan, and France are the key exceptions to this rule, where most toll roads are owned and operated by public corporations. (Though in France this follows an initial period under which most toll road development and management was in the private sector and the subsequent bankruptcy of the private companies.) Click here to display a document presenting the extent of private sector involvement in toll roads in selected countries.

    Even where toll roads are operated in the private sector, Government support has been considerable, in almost all cases. Click here to display a document presenting the level of Government support for private toll roads in Asia. This picture has been similar across the world.

    Return to top of document

  2. Objectives and Potential Consequences of Tolling

    One or more of the following objectives are typically the reason for introducing tolls on the road network:

    • New Source of Finance. The toll revenue is a "new" source of revenue, where previously roads were supported out of general Government revenues. This has been a major objective in many countries, given pressure to reduce taxes. For example, in Norway toll revenue represents 32 percent of the budget for the national road system. In Spain the figure is around 46 percent. Norway is rare in using some of these funds to support public transport in urban areas of Oslo and Trondheim, not only the road network.

    • "Stable" Source of Finance. Tolls provide an ongoing revenue source, which is not tied to the annual Government budgetary process. This can be particularly important for raising debt finance outside the national accounts and requires a Government Corporation or private sector operator.

    • Dedicated Source of Finance. The funds from toll revenues can be dedicated to the support of construction and maintenance for a particular road thereby ensuring that maintenance funds in particular do not compete with the requirements of other roads in the network.

    • User Pays and Internalizing of Externalities. Some Governments have introduced tolls in pursuit of a general policy to increase the extent of "use related payment" or with the goal of reducing road use and internalizing the negative effects of road usage (for example, congestion related prices). This is central to a 'sustainable' transport policy. In the Netherlands tolls are levied with the express intention of directing road users to other means of transport, both to ease road traffic conditions and to encourage use of the railways and inland waterways.

    • Regional Equity Issues. Some countries have introduced tolls on one road in order to support the development of infrastructure networks in less developed regions. Such schemes can help to transfer wealth from one region of a country to another.

    • Private Sector Development. Some Governments have sought private sector participation in roads where they wanted to develop the road network, and to develop the private sector within their economy at the same time. In addition the involvement of the private sector can allow the government to finance at least part of the road development off balance sheet. Some governments also use the private sector concessions as a mechanism for promoting/explaining the introduction and increase of tolls to a reluctant public using phrases such as "it is not what the Government would want but our hands are tied by the contract."

    It is vital that the government understands its objectives in pursuing a road tolling policy since these objectives will direct all activities undertaken, both in the early years and during the operation of the road as regulatory questions arise. Most importantly perhaps the objectives should affect the toll rate decisions.

    A summary table below presents the options, which face Governments once they have decided to toll the road.

Summary of Actions Available to Meet Government Objectives for Toll Roads


Some Actions available to Government to Achieve Objectives


Involve Private Sector

Provide on/off ramps and access and feeder roads

Specify land clearance and resettlement arrangements

Specify lower tolls for local residents/buses etc.

Provide Financial support

Specify lower tolls by time of day/level of congestion etc.

User Pays/Internalize Externalities







Regional Equity







New, stable & dedicated funds







Private Sector Participation







    The potential consequences of road tolling include:

    • The Raising of Revenue. Traffic and toll levels may not be sufficient to cover all costs, including construction, operation and maintenance. In developing countries where traffic levels are low or where construction costs are high (particularly likely in urban areas because of land acquisition and clearance costs), it is unlikely that the tolls will ever cover more than operation and maintenance and perhaps a part of the construction cost. For a rule of thumb in financing road provision in France, click here to display a document. Note however that these figures are based on simple assumptions about financial structures etc.

    • Development of the Private Sector. New areas for private sector investment can be developed. However there is a danger that this may crowd out other investments (which makes Government strategy development particularly important) or create long term liabilities for Government. For information on the effects of the Asian economic crisis on government liabilities in toll roads, click here to read more.

    • Diversion of Traffic away from the New Road. Price elasticity of demand and the provision of free alternatives to the tolled road, will affect the level of traffic on the facility. In turn this may mean that some potential economic benefits of the new road are lost since the objective of new road provision is to move people and goods more reliably and quickly. However when well designed the cost of tolling for revenue should be lower than those of any other system of revenue collection.

    • Social Impacts. Just as with any road, toll roads can have significant social impacts in the manner and location of their construction and in their operation. These can be both positive (providing improved access for some regions of a country) and negative (degrading the environment around the road, for example underneath an elevated urban expressway). However there are additional consequences which result from the tolls. For example, tolls

      1. can discourage unnecessary trips and therefore provide environmental benefits,

      2. may be too high for the poor to benefit from the new facilities, or

      3. may be so high that traffic diverts off the new road onto parallel roads which pass through residential neighborhoods therefore reducing the environmental benefits that the new road could have provided.

    • Political Opposition. Political opposition to road tolling has been significant in many countries. The opposition has meant that toll rates have not been increased as planned or untolled facilities have been developed to provide an alternative. Both actions have negatively affected the financial outcome of the toll project and have affected the economic return from the road.

    Return to top of document

  1. Network and Planning Issues

    Where a strategic transport and road network plan is in place Governments are able to:

    • prioritize investments (both public and private),

    • provide a transparent environment for infrastructure provision and thereby make it easier to attract the private sector,

    • assess the potential effects of tolling new facilities, and

    • identify capacity requirements at different stages in network development¾there is scope for roads to be widened/extended only once traffic levels reach a certain threshold for example

    Governments should therefore seek to institute a periodic network planning process which identifies critical road segments, undertakes initial feasibility and economic and financial analysis and determines priorities for investment and implementation. In some countries where the private sector has played an important developmental role in road provision, the Government has relinquished this responsibility, taking the view that private sector planning and project identification will suffice. However the ramifications of this¾that networks developed by different private sector providers do not enhance one another or cannot be connected and roads which are required from an economic point of view, but do not have a good financial return, are not provided¾are significant.

    Interface with other parts of the network/other modes¾the new tolled road should be accessible from different parts of the road system. To ensure that the road plays the full role in the network expected of it, the connections to other parts of the network (and to other modes where necessary) are vitally important. Where the Government is not providing the road itself and a private sector partner is undertaking construction the delineation of responsibilities and phasing of development by the different parties is particularly important. This has been a significant issue in the development of the tolled road network in Bangkok for example. For more information about the Bangkok experience, click here to read more.

    Dense Road Networks/Multiple Road Providers. Where road networks are dense the following issues may arise.

    • There may be constraints on the level of the toll (since it may not be politically possible to levy different tolls/kilometer on different parts of the network), or

    • it may be difficult to forecast traffic and therefore revenue, because of the extent of the competing network, or

    • the tolling systems chosen by each provider may not be compatible, or

    • users may need to stop at several toll plazas within a short distance and therefore slow their travel considerably.

    All of this typically means that revenues are lower and more difficult to predict than would be the case with less dense road networks and a single provider. This implies higher commercial risks and correspondingly higher Government costs if the private sector is involved. With higher costs to Government and reduced economic returns from the road, careful consideration needs to be given to these elements. The Malaysian and French Governments in particular, with extensive toll road networks, are beginning to find that this commercial risk issue is particularly significant. In both cases the Government has had to provide considerable support for the toll road industry.

    Is it necessary to have a parallel free road? This will depend on the reasons for provision of the new road. If the road is provided simply to add capacity in a constrained area, for example, then leaving the existing network untolled should be appropriate. The argument in favor of free parallel roads is one of social equity, to ensure that the poor can still have access to the road network. However it also detracts from the effectiveness of the new tolled road in alleviating congestion and may well cause problems for financial cost recovery if even the revenue-maximizing toll cannot produce enough revenue. With a parallel road however it is possible to provide for slower moving traffic (animal drawn vehicles for example) where a toll road does not. Click here to display a document and read about the negative revenue effects of parallel free roads in Mexico and Hungary.

    Cross subsidy between different parts of the network. Some existing roads have been tolled in order to provide revenue for construction of new segments in the network. The French pioneered this technique whereby new roads with higher construction costs are supported from operating surpluses on existing roads built at lower cost, and with existing revenue streams from tolls. The Japanese too have been very committed to this concept. Click here to read more on the Japanese experience. More generally it is a very common approach to 'create' profitability of the new road or the concession arrangement. In Malaysia the Government transferred several completed sections of the North-South Highway to the concessionaire at the start of the concession. Click here to read more on the Malaysian example.

    While these examples illustrate that tolling can assist in releasing funds for new construction, it is also possible that resources can be misallocated with cross subsidy because there is insufficient consideration of the effects. This reinforces the need for significant Government involvement in the design of the road network and setting road and transport policy generally. Careful project evaluation is required whether or not there is an existing revenue stream from which the new road can be built. Without this there is a danger that better investments are missed or even that private sector resources being channeled into road development will crowd out other more important investments in the country.

    Other concerns about cross subsidy related to the question of transferring resources from one group of consumers to another. Those who are paying tolls on the existing road are thereby paying for the construction of a new road, which would otherwise have been funded by taxpayers and will provide benefits for other future users. This may be part of a Government program of regional development but needs to be explicitly recognized.

    Toll roads are often developed in congested corridors of a capital city, because good revenue streams are easier to predict. Where this is the case, the investment in the road is benefiting more affluent areas of the country. If this crowds out other investments in the less affluent areas of the country, then other regional equity issues are raised.

    Public or Private? Whether tolled roads are provided by the public or private sector is a planning issue for early consideration by Government since the private sector involvement will change the project economics¾with new costs and benefits associated with the private sector contract.

    Both France and the USA have public sector toll road authorities. These authorities are arms length bodies and can raise bonds in the financial markets, on the strength of their future toll revenues. They are not considered true private sector entities in this case. Click here to go to a web site which provides connections to available USA toll road authorities' web sites. These typically give information about the prevailing tolls, the use of the revenue and the functions/obligations of the authority.

    The Japanese too have developed their toll road network through the public sector with a "toll revenue pooling" policy¾allowing them to build and operate 9,000 km of non-urban toll roads by the end of 1998. Click here to read more about the Japanese revenue pooling system.

    The earliest toll roads were concessioned in the 19th century. More recently the standard format has been Build-Operate-Transfer (BOT) concessions in most of the world. In South-East Asia the BOT mechanism has been paramount and few other forms of private sector participation have been used.

    A BOT concession typically lasts 30 years, and grants the concessionaire the right to construct, maintain, and toll the road over that period after which it is transferred to Government. However there are other options for private sector involvement. Some of these may be more appropriate than BOT where traffic levels are low. Even where BOT is required because the Government desires new capacity, it may be that a "negative concession" should be considered, under which the Government provides a level of support to the road. Click here to see a spectrum of alternatives for involvement of the private sector in provision of toll roads.

    Alternative approaches have been used in for example Argentina (where the nature of private sector involvement has evolved over time and where road concessions often allow for staged construction as traffic grows) and the United Kingdom where maintenance and operation concessions have been developed with a shadow toll mechanism.

    Legal Issues. When a country embarks on a toll road program it is likely that new legislation will be required to allow the collection of tolls (typically this is banned on the majority of public sector roads). Similarly if private sector involvement is considered, this often needs specific enabling legislation.

    Other legislation is required to cover issues such as:

    • signing to the road from the rest of the network,

    • advertising controls on the road,

    • operational procedures (such as arrangements for emergency vehicles and information disclosure rules, which are particularly important where tolls are levied electronically),

    • defining the enforcement regulations for non payment,

    • provisions for land acquisition and clearance, and

    • structure for involvement of the private sector in the provision of roads.

    Return to top of document

  2. Tariff Setting and Policy Issues

    Typically 95 percent of revenues for a toll road come from toll receipts themselves. The other 5 percent come from advertising or small concessions such as service areas. Hence the toll level and collection are of critical importance.

    Sensitivity of Demand. In economic evaluation of the road scheme it is vital that an accurate estimate of travel on the road is produced. Even when a new road is built and is not tolled, the estimation of traffic is not a simple matter because of the complexities of interaction between different parts of the network etc. However with the additional complexity of tolls, forecasting is more complex still. It has not been successfully achieved in many cases. Click here to read a World Bank Private Sector Viewpoint Note detailing the experience of Mexico.

    It is important to understand however that the forecasts are required in order to:

    • estimate an economic benefits from the road, and

    • estimate the net revenue maximizing toll (i.e., net of costs of revenue collection)

    Governments should therefore seek the most comprehensive forecasts they can and consider a range of sensitivity tests to determine the robustness of the proposed financial arrangements.

    Maximizing net toll revenue and maximizing economic return from the road are contradictory objectives. Hence the Government needs to understand where the appropriate balance lies. This balance will be set where economic return and revenue raised meet Government objectives, and the revenue raised is sufficient to finance the project as structured. This will usually mean that the Government needs external financial advice¾Investment Bankers may well take a different view on "revenues sufficient to finance the project" from bureaucrats. All of this means that given the sensitivity and multiple functions of the toll rate, the Government is likely to want to specify this to private sector bidders in a tendering process. Click here to jump to the section in this key issues document which gives more detail on different approaches to bidding.

    Duration of toll. Potential objectives in raising tolls are manifold. They include:

    • covering construction or rehabilitation costs¾which requires a long period of tolling if it possible at all (depending on financial structure, the construction costs, and the amount of traffic),

    • covering maintenance and operations costs¾which requires continued tolling,

    • creating a stream of revenues which can be used to develop the road network and which are not under control of the annual Government funding process. This requires tolling to continue until the required network is complete and requires the earlier roads to generate surplus revenues for investment in new ones. Few Governments admit to having completed a road network however, and

    • ensuring that road users bear the full cost of their travel directly, by pricing the externalities¾which also requires continued tolling.

    The objective for each road will affect both the level of toll and the period of tolling. Where tolls are simply required to cover construction costs, and the traffic levels are high, then tolls will only be required for a relatively short period, for example. The Japanese Government has made 61 roads toll free following recovery of the total project cost.

    Where Governments seek to cover maintenance and operations costs, tolls required are low and need to be levied continuously. Where tolls are levied to allow extensions to the road network, the tolls would be required only as long as further developments of the network are required. The Chinese government has used this approach most extensively, though the Malaysian Government has used the funds from tolling of the North-South Expressway to support other ailing infrastructure investments.

    Toll variations. Typically tolls vary with distance traveled along the road and according to the number of axles on the vehicle. This approach was first adopted because it can act as a proxy for the road space used by the vehicle and the damage that the vehicle inflicts on the road pavement. However other options (which can be combined) include:

    1. Time of Day or Day of Week. Variation by time of day is typically used where congestion causes considerable delays to travelers. This has been tested in France, and seems to have encouraged 10 percent of travelers to retime their journey when tolls were increased 50 percent in the peak periods. These figures will vary depending on the value of time and the decision making criteria of travelers and therefore is very country specific.

    2. Cost of road construction. Tolls vary across a country or region, because of the different costs of road construction through different areas, for example roads cut through mountains are likely to cost more to construct than those across flat open landscapes. Political incentives for unified toll structures across a country may however prevent differentials in toll across the country.

    3. Social considerations. Some countries make political decisions to encourage use of new facilities by higher occupancy vehicles. Malaysia for example has set bus tolls considerably lower than for other vehicles.

    4. Tolling by area. Singapore has long had an area pricing scheme (the Singapore scheme is now also congestion charging - the next point), tolling those entering the central area of Singapore. This has now been fully automated and the tolls are variable with the time of day and day of week, to reflect congestion (for more on congestion charging see point v). There is similar toll ring system in operation in Trondheim in Norway. Other toll rings operate for example, in Oslo, but do not vary tolls with the time of day. The problems associated with cordon pricing are concentrated on the edge of the area and related to irregular users. The appropriateness of the system depends on the nature of the area in question. This is likely to change as technology develops.

    5. Congestion Related Tolling. Recently, in California, congested related tolling has been introduced on major roads. The tolls are levied on new facilities (whether full new roads or specific additional lanes), and may be waived for high occupancy vehicles. For more information on the California experience, click here. For information on other international congestion pricing schemes click here.

    6. Loyalty programs and other discounts. With the introduction of electronic tolling systems this sort of commercial action has been more feasible. Where the private sector are involved however the concession arrangements must allow for toll variations. On the M5 in Hungary a whole program of discounts has been introduced with:

      • 40 percent reductions for regular users,

      • 20 percent reductions for fleet owners,

      • 20 percent reductions for local residents,

      • vouchers for users of the Southern Food Market in Budapest paid for by the Food Market at 30 to 40 percent discounts,

      • 20 percent reductions for agricultural producers in the four counties around the road, and

      • 20 percent discounts on monthly tickets for car-pools (4-passenger).

    7. Other systems of area tolling rely less directly on road use. For example, the Vignette system in use in some European countries. Under this system the vignette holder has access to all roads within a particular geographic area, as necessary during a specified period. This is typically only applied in the case of expressways and passes can vary from one week to one year in duration. They give the pass holder the right to travel on the expressways at any time during that period. These tolls are less directly related to use of the road therefore, since frequent users are paying less per trip than infrequent users.

    Increasing tolls over time. When the private sector is involved in the operation of the road, almost all agreements include a provision for toll increases. Most concession agreements include a toll escalation formula that is related to the consumer price index (often foreign exchange devaluation too) and allows for toll increases either every year or every few years. Usually there is also a clause providing for governmental approval of any increases and compensation arrangements if that approval is withheld. However Government approval is not always given and operators do not always receive the specified compensation immediately. This is an area of considerable political interest, and increases can meet violent opposition from the public, particularly if the economy is weak, as was witnessed recently in Malaysia and has also been seen elsewhere for example, India. In Texas, the toll rates on the Dallas North Tollway have not been increased since 1982. This is not atypical where toll roads are managed in the public sector.

    This does not imply that the practice of introducing toll increases should not be promoted¾small annual increases can be easier for consumers to incorporate in their spending patterns than larger less frequent rises. Economic forecasts and financial analysis should however consider, as one sensitivity test, the failure to increase tolls for several years and a failure to agree/pay specified compensation.

    Early gaps in the funding of the road from toll revenues. Often, particularly in BOT or ROT (rehabilitate - operate - transfer) concessions, the early revenues from the tolls are insufficient to cover the expenditures for the road. Hence there is a requirement for additional loans. There are at least two revenue-maximizing strategies¾a high level, which maximizes short run revenue, and a lower level which encourages use of the road and maximizes long run revenues. These trade offs need to be considered explicitly by the Government in determining the appropriate toll rates. Some Governments have faced this issue by providing soft loans for the early years of the concession.

    Return to top of document

  3. Tolling Options

    Getting the approach to tolling right, given the situation of the particular road and the options available, is important, since toll collection typically accounts for around 10 percent of total revenue collected in Europe. There is a wider variation around the world with figures varying from 5 to 20 percent. The cost of installing the toll collection equipment may be between 2 to 8 percent of initial construction cost. Where schemes are poorly designed leakage through toll avoidance, pilferage or underpayment can be significant

    Different technologies for tolling¾There are three basic options:

    1. Manual tolling.

    2. Mixed tolling¾some manual tolling, some electronic.

    3. Electronic¾where there is no manual toll collection.

    The most common method is still manual. The drawbacks are that it is a slow system and therefore requires more toll booths/lanes than any other to achieve the same traffic flow. Set up costs may also be high if land acquisition is costly. In Germany, for example, the estimated cost of installing toll booths on the motorway network (where the roads are often passing through dense development and therefore land acquisition would be problematic) have been used to argue against tolling.

    There are few all-electronic systems in operation¾Singapore's new area pricing system being the most comprehensive application of electronic tolling. More such schemes are being developed however, for example, in Melbourne. Electronic systems require all users to carry tags in their vehicles and to pass the toll gates at a slow speed, but without stopping.

    There has been some opposition to electronic tolling because of the level of information which it allows road operators to collect about individual users movements. There is likely to be more enthusiasm where a manual system is already in operation and vehicles are subject to severe delays at toll plazas, or where a discount for frequent use is available with the electronic system.

    Enforcement can be difficult with an electronic system since it requires accurate records of car owners' addresses that are not always available. Another drawback can be compatibility between different systems where there are several toll road operators each with different electronic toll collection equipment. This is something, which Government can however prevent by careful structuring of agreements or perhaps with legislative control.

    Click here to display a document containing more information on electronic tolling options.

    Open/Closed. A system is typically described as "open" when there are no toll booths on entry or exit ramps to the road. The tolls are collected at points along the road, and are therefore unrelated to distance traveled. Open tolling is generally used in urban or semi-urban areas where the traffic flows are high and the cost of toll plazas at all access and egress points would be too great. The revenue collected will be less than in a closed system, but careful design can minimize this. In a "Closed" system, all exit and entry points are monitored and tolls collected on exit, so that all travelers make payments and the payments are directly related to distance traveled. The closed system tends to be more expensive¾requiring more infrastructure and more operating staff but, if most trips on the road are short, can be necessary to achieve anything nearing full revenue collection.

    Shadow Tolls. Shadow tolls have been used in the UK, Finland and the Netherlands, and the terminology has confused many people. No tolls are levied from road users under this approach. Instead the shadow tolls are paid by Government to the operator, based on traffic counts on the road and an agreed rate per vehicle/vehicle type. The benefits of this system do not therefore stem from the development of a new source of funds, or from making users internalize the external costs of their travel, but rather from:

    • the Government commitment to continued financial support over several years

    • the involvement of the private sector and their responsibility for efficient delivery of service.

    The shadow toll approach does not require traffic to slow for toll collection (and does not require additional land take for widening the road around toll booths). However because it requires the Government and private sector to agree the vehicle counts and because of the difficulties surrounding legal arrangements, the transaction costs can be very high. This has led to significant criticism of the approach in the Netherlands. Click here for more on the rationale used for the British shadow toll experience.

    Toll evasion. There are many ways to avoid tolls, and the ease of avoiding toll booths is one reason why interurban roads have often been considered as difficult to toll. However with good policing, sensible design and rule of law the problems can be overcome.

    Pilferage of the toll revenues. Often Governments have considered privatization of toll collection because of the difficulty of ensuring that toll revenues are not removed illegally by members of the collection or counting team. It is assumed that private sector profit incentives will prevent this pilferage.

    Informal Tolls. In some countries such as China and Pakistan "informal tolls" or fines/fees levied by police, bandits, or others, can be considerable, and make a significant difference in people's behavior. It is important that traffic and revenue forecasts acknowledge these. Forecasting of toll revenue requires an understanding of the costs of travel along the road and its alternatives. Those costs include time taken to travel the distance, the tolls levied on the different roads, the costs of operating the vehicle along the road, and any informal tolls levied on the user of a road.

    Return to top of document

  4. Mechanisms for Involving the Private Sector

    Toll roads were out of favor for many years, because of the concerns about traffic diversion away from the new facility. However there has been a revival of interest in the last ten years, principally to raise off-balance sheet financing and to involve the private sector in the provision of roads. This coincides with a growing interest in private sector involvement in infrastructure more generally.

    There are additional complexities (risks, and costs) when the private sector is involved in the provision of toll roads. These depend on the mechanism selected, and are offset by benefits (off balance sheet financing, risk transfer, efficiency etc.) in some cases.

    Some of these complexities, some poor forecasting, and the Latin American and Asian financial difficulties, have all come together to make potential project sponsors and lenders more wary. This reinforces the need for Governments to get good advice and to think through their objectives for the road and for private sector involvement in its' provision.

    Depending on Government objectives the method of private sector involvement should vary. The strategy setting process should have identified objectives and potential roads for tolling. There are four different basic options:

    1. Maintain and operate contracts.
    2. Rehabilitate, Operate, Transfer (ROT) contracts.
    3. BOT contracts.
    4. Corridor management/maintenance contracts.
    Each allows new revenues to be raised (if the contract covers an existing road this is only true if the road has not been tolled before). Only the BOT schemes and the UK's corridor management schemes provide new roads. (The UK scheme includes both new construction and maintenance of whole corridors.) All should improve maintenance. This is summarized in the table below.

Summary of Objectives met by different styles of PSP contract


Maintain and Operate



Corridor management/maintenance

New revenues





New construction




Improved maintenance





    BOT concessioning has dominated the Asian approach, while maintenance concessions are most common in Latin America. It is becoming clear that BOT concessions are most likely to be successful under the following conditions :

    • Projects in existing high traffic corridors with missing links such as river crossings. These offer the prospect of high traffic on opening and minimize the land costs.

    • Inter-urban projects because the implementation costs are minimized. Traffic risks are also likely to be lower than in dense urban networks. In urban areas projects at-grade or elevated will keep construction costs low.

    • Projects in middle income countries where there is a willingness to pay tolls.

    • Projects where tolls are set at, or close to, the revenue maximizing tariff and toll escalation formulae are invoked.

    • Projects where there is an existing income stream from which to draw revenues from day one, perhaps even during the construction period.

    • Well prepared projects where feasibility issues, particular for land acquisition, have been addressed and therefore the implementation risks are minimized.

    • Large projects¾ because there are high, relatively fixed, bidding costs for BOT projects.

    • Maintenance concessions on all but very low-traffic roads on the other hand, with the lower initial expenditures, can usually recoup their costs from the toll revenues.

    Some Governments, and even some public sector toll road authorities, have entered into contracts with the private sector simply for collection of tolls and daily operation of the road. This is the approach in Norway, where there are 26 toll companies.

    Government Support. When evaluating a road project in terms of its economic effects, the Government needs to consider too whether an acceptable financial return can be generated. If not, some redesign of the project and/or its financing must be considered. For example the Government can consider a "negative" concession. This may in turn effect the economic returns. Click here to display a presentation containing a flow diagram illustrating these connections.

    All too often however the Government support for projects is increased far above expected levels when the sponsors are well connected politically, have better advisors, or threaten to withdraw at the last minute. To prevent this is not difficult¾it requires Government to be well prepared, with the specification and design of the Government support part of that preparation.

    Considerable government support is typically required for new build projects¾at least in the form of land acquisition and clearance. This can be a large portion of the cost of the road (averaging around 15 percent but up to 50 percent in some areas, particularly cities where both land and relocation costs are high) and can be the most time consuming element in the construction process. Further Government financial assistance may be required too, perhaps through allowing the operator to toll an existing road to give revenue streams on day one (as was the case with the North South Expressway in Malaysia). Alternatively the Government may provide soft loans for the early years of the concession, or consider a "negative concession".

    There is a range of options for the provision of Government assistance to projects including:

    • land purchase and clearance,

    • ancillary facilities,

    • direct revenue support,

    • loan/equity guarantees,

    • subsidies/grants,

    • preferential and subordinated loans,

    • shadow tolls,

    • revenues from tolls on existing roads,

    • tax relief,

    • foreign exchange guarantees, and

    • additional rights for development around the road

    The most advantageous for the concessionaire are those which provide early funding streams (when revenues from the toll road are low or non existent during the construction period) and which give guarantees for unexpected problems (for example, exchange rate guarantees). The least significant are those which themselves are unpredictable i.e., additional rights for development around the road.

    The bidding process. Once the road for tolling has been identified, the Government has decided to involve the private sector, and given a desire to introduce competition into the process , a bidding process needs to be determined. The key aspects for Government consideration are:

    • Detailed development of the project. This must be set out clearly in the bidding documents, specifying the responsibilities of the concessionaire and the relevant Government departments.

    • A draft concession agreement, to ensure that all bidders make similar assumptions. This should be prepared by the Government and included in the bidding documents.

    • The maximum level and nature of Government support. Economic analysis of the project will have determined the appropriate level of Government support, and concurrent financial analyses will have determined whether this is feasible, given likely private sector requirements.

    • The decision criteria. The most transparent approach is to select a single criterion, perhaps the lowest level of Government support required (if any) for a given toll. The decision should be made only following a technical evaluation, and only those bidders who meet the technical standards should be considered for the concession.

    Where Governments have used several different criteria for selecting the preferred bidder, the outcome has been problematic since it makes economic evaluation of the scheme complex, and can be less transparent. In addition, where several variables are used in selection of the concessionaire, these variables are often important in regulation. This makes the process of regulation more complex, and perhaps less transparent.

    The possible criteria include: the total capital cost of the project, the construction period, the toll rate at opening, the NPV of Government subsidy required, and the length of the concession. Click here to display a document illustrating some approaches used in European decision making. Click here to read about an alternative approach where the term of the concession is not fixed.

    Without clear bidding documents and detailed specification of requirements of the project, the Government may find that either the bids are more costly for Government than expected or there are fewer bidders than expected. This is because bidders will view the transaction costs of preparing the bid as too high, or may fear that there will be significant regulatory difficulties during the course of the project that would reduce their expected return.

    Unsolicited proposals have been particularly problematic in Asia. The typical approach is for a proposal to be submitted to one arm of Government, for the proponent to be given exclusive rights to develop a project in a defined corridor, and for a proposal subsequently to be submitted to Government. In some countries this proposal is then submitted to competition, following which the initial proponent may be allowed to match the highest bidders terms. In other countries the Government simply negotiates a deal directly with the proponent. Neither of these approaches is likely to achieve the most cost effective deal for the Government. The benefit of unsolicited proposals is that they allow the private sector to be part of the planning process. The danger is that without proper network consideration the full effects of the project may not be understood and the financial aspects will be paramount.

    Risk sharing/responsibilities of different parties. The standard advice on risk sharing is that "the party best able to shoulder the risk should continue to bear it". A table which defines the different risk categories and attempts to define the party most able to shoulder the risk for BOT roads, is forthcoming (Annex 15)

    However this statement does not acknowledge that the process of private sector participation itself introduces risks which are not present under a direct Government project. The risks of toll revenues not meeting required levels for supporting financial arrangements for example is one of these new risks. There are many reasons why this may be true in practice. Predicting them before signing of the concession, is complex, and suggests that thirty or more years can be understood in advance. However this is not the case as has been witnessed recently with the Asian financial crisis for example. It is therefore more pragmatic to develop a framework which establishes the robustness of project performance against realistic future scenarios. However if the parameters are too many or too wide, the scope for private sector evasion of responsibility and risk bearing can be great. This is a sensitive, complex area, and one which has been tested deeply of late through the Latin American and more recent Asian, financial crises. For details about a new approach adopted recently in Hong Kong, click here to read more.

    Even where regulatory arrangements are well defined, Governments are reluctant to let private sector concessionaires fail. Hence for example, in France during the 1970s three of the four private sector providers of toll roads were nationalized when financial difficulties set in. In East Asia the recent financial crisis has meant that with exchange rate problems, and a reduction in toll revenues, the operators have looked for further Government assistance, which has often been forthcoming. The Asian situation is complicated by the extent to which Governments are guaranteeing loans taken out by the private sector investors, and therefore would be liable were the companies to fail, but the message is clear¾Governments don't tend to let toll road companies fail. The investments are too high profile and the tolls themselves too political, for that to happen. Of course it may also be that the economic crisis as it has developed is far outside the realm of prediction and therefore more akin to a "force majeure" event. If that is so, it is not in anyone's interests to bankrupt the concessionaires and Governments are taking a pragmatic line in continuing to provide support.

    Duration. The most appropriate duration of the concession is affected by the:

    • reason for revenue collection¾is it to cover construction costs? To internalize externalities of travel to the user?

    • the financial instruments available¾in some markets there is no possibility of long term financing from the private sector,

    • political imperatives/security concerns.

    Typical BOT concessions are 20 to 30 years in length, whereas maintenance concessions tender to be shorter¾typically 5 to 15 years. They differ in length because of the different financial requirements.

    The duration of the concession may either be set by the Government in advance or be part of the decision criteria in selecting the concessionaire. One of the problems for the Mexican program was that this was the sole criteria therefore forcing toll rates very high (which given the price elasticities did not increase revenues as expected) and eventually causing many toll road concessionaires to fail and require significant Government support.

    Development of the private sector road industry. In some countries, where the private sector has built roads under standard Government direct procurement contracts, the objective of giving the private sector a greater role in road provision is to develop a new industry in road management. These techniques (such as marketing, revenue control, life cycle construction/maintenance costing) are not typically found in construction companies, and under the first BOT concessions, construction companies did not need to develop them. However under schemes such as the Design, Build, Finance and Operate (DBFO) approach used in Britain these skills are being developed and companies are emerging which can provide road management staff. Typically construction companies, who take their profits from the construction contract, dominate concessionaires. Hence the development of operating companies, who also view the operating period as a profitable business, prevents this focus on the construction period alone. For more information about the British approach, and the companies developing around it see "DBFO: Value in Roads" Case Study. To access a presentation on DBFO contracts given at a World Bank Seminar on Tolls Roads in February 2000, click here.

    Regulation. It is important that the legal environment under which the concessions are granted is clearly structured. This may require general enabling legislation or can be achieved through specific concession arrangements. Whichever approach is selected the arrangements must make responsibilities of the different parties clear. There are instances in which, despite the enabling law being in place, it is still unclear where ultimate responsibilities lie. In the Philippines, for example, projects may be proposed by the Toll Regulatory Board, the Philippines National Construction Company (under the terms of their original concession), the Department of Public Works, or an unsolicited private sector bidder. The concession may be regulated by any of the first three. These competing interests mean that Government expertise is not collected together and that common evaluation of projects can be difficult to achieve. Similarly in Hungary the legal arrangements were not strong enough to prevent a legal challenge to the toll rate. These examples illustrate the importance of clear, well defined and complete legal arrangements. It also suggests that the Government must take external legal advice on the structure of these arrangements before introducing them.

    Where international law is also to be considered, or where there are common legal arrangements within a trading block for example, these also need to be taken into account. (The EC directives stipulate, for example, that toll rates "must be linked to the construction, operating and development costs of the infrastructure network concerned" and that Government should seek to introduce distance related tolls.)

    Without transparent processes, either at the bidding stage, or during the concession, problems in the regulation can mean long delays. Delays in the bidding stage can mean that the economic value of the project is reduced because of delayed implementation and higher transaction costs. Delays during the operating phase, particularly for example if Governments fail to authorize toll increases can simply mean that the required Government support for the project increases (and therefore perhaps the economic return is reduced).

    Issues involved in regulation and contract management. The regulatory and contract management responsibilities and their differences must be clearly understood in Government and by the project sponsors. Typically regulatory bodies consider economic variables such as the toll rates while the contract management agency's considerations include:

    • land acquisition and clearance,

    • other Government contributions

    • traffic/revenue levels,

    • quality of construction,

    • maintenance of the pavement,

    • congestion levels on the road and at toll plazas,

    • lane availability

    • safety management systems,

    • commercial activities, and

    • payments to Government

    The contract management body is typically an arm of the Government Roads Agency. It tends to rely heavily on the private sector for its information (since reporting by the concessionaire is a requirement of the agreement).

    The regulatory body tends to be separate, often part of a general regulatory agency for transport or for infrastructure. It is responsible to reviewing the tariff changes under the contract, particularly where a maximum rate of return is permitted. It also deals with requests for additional Government support during the project and with requests for extension or widening of the road.

    Either regulator or contract management agency can provide a point of contact for road users¾the information which they provide giving more information on which the agencies can act.

    Both regulation and contract management tend to take place at the National level. China is the exception to this rule where local Governments have been heavily involved in toll road development.

Return to top of document

Selected Terms of Reference

To see examples of Terms of Reference, click below:

Return to top of document

Selected References

Publications available on-line at this Web site

Analysis of Highway Concessions in Europe, French Study for the DERD/WERD, 2001.

Estache, Antonio and Josť Carbajo, Designing Toll Road Concessions-Lessons from Argentina, Public Policy for the Private Sector, The World Bank, Private Sector Development Department, Note No. 99, December 1996

Fishbein, Gregory and Suman Babbar. 1996. Private Financing of Toll Roads. World Bank, Washington, DC.

Halperin, Ricardo and Patrick Malone, Overview of World Experience in Private Financing in the Road Sector, The World Bank, ECSIN Working Paper No.2, 1999. Available at World Bank, Europe & Central Asia, Infrastructure Sector Web site

Heggie, Ian G. and Piers Vickers. 1998. Commercial Management and Financing of Roads. Technical Paper 409. World Bank, Washington, DC (6,058KB PDF).

Nickesen, Alfred and Mitchel Stanfield. 2000. Topic Note RH-3, Toll Road Securitization in China. World Bank, Washington, DC.

Review of Recent Toll Road Experience in Selected Countries and Preliminary Tool Kit for Toll Road Development. 1999. Study prepared by Padeco Ltd. for the World Bank and Ministry of Construction of Japan (MOCJ) as part of the Asian Toll Road Development Program.

Ruster, Jeff. A Retrospective on the Mexican Toll Road Program (1989-1994), The Private Sector in Infrastructure: Strategy, Regulation, and Risk, The World Bank Viewpoint Note No. 125, September 1997.

World Bank, Country Management Unit for Colombia, Ecuador and Venezuela (LCC4C), Latin America and Caribbean Regional Office, Project Appraisal Document on a Proposed Loan in the Amount of US$137.1 million to Colombia for a Toll Road Concession Project, Report No. 17986, June 11, 1998.

Publications available for purchase or downloading from other organizations/Web sites

Wright, Charles L. and Daniel J. Freire Coloma, Toll-Road Partnerships: What Works, What Doesn't, and Why?," Transportation Quarterly, Vol. 51, No. 4, Fall 1997, pp. 85-101. For more information on this publication, visit the Eno Foundation for Transportation Web site.

European Bank for Reconstruction and Development, Project Evaluation Department, Operation Performance Evaluation Review of Budapest Orbital Motorway, Hungary (A State Sector Investment Operation), December 1996. For more information on this publication contact EBRD at Tel: +44 20 7338 7553 ; Fax: +44 20 7338 6102 ; E-mail:

Publications available from the World Bank InfoShop
To order by phone or fax: Phone 1-800-645-7247 or 703-661-1580; Fax 703-661-1501; E-mail

Developing Best Practices for Promoting Private Sector Investment in Infrastructure: Roads. 2000. Asian Development Bank (available from the World Bank InfoShop Development Bookstore).

Publications available through regular library services

Financing Private Infrastructure, IFC Lessons of Experience 1996. IFC, Washington, DC. ISBN 0-8213-3822-6.

DPWH Philippines, General Policies for Financial Analysis of Negotiated Toll Road Proposals, 1998.

DPWH Philippines, General Features of Financial Aspects of Negotiated Toll Road Agreements, 1998.

Return to top of document

Ports & Logistics | Railways | Roads & Highways
Rural Transport | Transport Economics & Sector Policy | Urban Transport

Help in Using This Site