Journal Issue

Evaluating Transport Projects

International Monetary Fund. External Relations Dept.
Published Date:
March 1966
  • ShareShare
Show Summary Details

Hans A. Adler

UNTIL A FEW YEARS AGO transport evaluation was virtually the exclusive domain of engineers. Economists were not interested—even though in this field close cooperation between economists and engineers is vitally important. As a result, the evaluation of transport projects is handicapped by the frequent failure to apply economic criteria.

In less developed countries, basic statistics are often lacking. Many countries, for example, have only very recently begun to collect highway traffic data. Where statistics are available, they are usually limited to simple traffic counts, whereas effective highway planning calls for much more sophisticated data—for example, the origin and destination of traffic or the types of commodities carried. This is hardly ever available. Nor is much known about the costs of operating vehicles on different types of highways or the costs of maintaining different surfaces. As a result, new investment and maintenance expenditures have usually been undertaken without detailed study. And the absence of statistics is only a segment of a vicious circle; because until recently economists have not focused on the right questions, there has been little incentive to collect the right statistics.

Preliminary Steps

Before a specific transport project can be evaluated, two steps are needed in order to discard less rewarding alternatives. The first step is a general economic survey of the country; this has two major functions. One is to establish the country’s over-all transportation needs by investigating, for example, the rate of economic growth and the resultant expansion in traffic. The other is to provide a basis for comparing these needs with the country’s other economic needs. This cannot be done very precisely, and it depends heavily on qualitative judgments. Even so, the results have sometimes been surprising, suggesting that too much was being invested in transport. A recent survey of Colombia, for example, found that investments in education, housing, and health would probably be more rewarding than the marginal investments in transport. Such surveys are also needed to help decide whether by changes in the location of industries the total demand for transport can be reduced, and at what cost.

When a country’s transport needs have been related to its total economic needs, then we can examine the transport needs more closely. There should be a detailed transportation survey of the country. Such transportation surveys have been made recently, under World Bank auspicies, in Argentina, Colombia, Ecuador, and the Republic of China. If they are to be really useful, the surveys should not only establish the priorities for each method of transport—such as the listing of highways in order of their importance; they should also indicate the proper role for each method and the priorities among them.

Problems of Project Evaluation

Economic evaluation of a project means measuring its economic costs and benefits in order to decide whether the benefits are at least as great as those obtainable from other marginal investment opportunities.

It is sometimes stated that the value of a project should be measured by its contribution to the growth of national income as conventionally measured. This is unfortunately not a very practical measure; it excludes, for example, the time saved and used instead as leisure time, which would not be reflected in national income. In any case, the national income approach is impossible in less developed countries because of the lack of statistical information. However, this is useful in focusing on costs and benefits from the point of view of the economy as a whole and not merely from that of the parties directly involved.

In evaluating a project which consists of a number of separable and independent sub-projects, separate analyses should be made of each subproject. Otherwise, it is quite possible that the extra large benefits of one subproject may hide the insufficient benefits of another. In a port expansion project in Central America, the engineers recommended the construction of two new wharves. Surveys indicated an economic rate of return on the investment of about 12 per cent, which was a satisfactory rate in the particular country. However, when separate analyses were made for each wharf, it turned out that the rate of return on one was nearly 20 per cent, while that on the other was only about 4 per cent, even after deducting the extra costs of building it separately. The second wharf was clearly not justified. The same principle applies to degrees of highway improvement and frequently also to different highway sections.

Economic benefits and costs must be measured and compared in monetary terms, but these do present a problem. In most countries there is only limited competition among the modes of transport, and market prices may not reflect real costs. Moreover, in most developing countries gasoline taxes and other charges on the beneficiaries do not cover the costs of highways (including maintenance, depreciation, interest, and administration); even where they cover total costs, they are not allocated to the costs of the individual transport services, such as those of trucks, buses, and passenger cars. Yet monetary terms provide the only practical common denominator.


Measuring the benefits of transport projects is usually much more difficult than measuring their costs. Some benefits, even though they may be very decided—such as the increased comfort and convenience from an improved road—are difficult to express in monetary terms since they have no market price. Some benefits that can be measured, such as reduced transport costs, are spread over a great number of people and a long period of time, requiring difficult long-range forecasts. Third, many benefits are indirect, such as the stimulation to the economy from improved transport; and for these benefits to materialize other investments are frequently also needed.

Still, it is our job to provide the best measure of benefits that we can. In the following list of possible benefits, those that come first are easiest to measure: (1) reduced operating expenses both for users of the new facility and also usually for those who continue to use the existing facilities; (2) lower maintenance costs; (3) fewer accidents; (4) savings in time for both passengers and freight; (5) increased comfort and convenience; and (6) stimulation of economic development.

It is worth considering the more important of these benefits individually, but first there is one question which, considering its importance, is not asked often enough: just who is the beneficiary? For example, if the improvement of a port reduces the turn-round time of ships, much of the benefit might go initially to foreign shipowners; the degree to which they pass it on to the country paying for the investment depends largely on the degree of competition in shipping. A government can, of course, recoup some or most of these benefits by charging the users. Such user charges are very important; in the absence of a well-conceived charging policy the benefits of a transport project may sometimes go to the wrong people.

Reduced Operating Expenses

The most direct benefit from a new or improved transport facility is the reduction of transport costs. Not only is this the most direct; often it is both the most important and the one most easily measurable in monetary terms. While it is bestowed initially upon the users of the facility, competition or the desire to maximize profits leads them to share it in various degrees with other groups, such as producers, shippers, and consumers. Before long, therefore, cost reduction benefits not merely the users of the facility but the nation as a whole.

By what standard should the saving in operating costs be measured? Often, the standard is the “before and after” test: what were the costs before the new facility was constructed and what will they be afterward? This is a very inadequate test which usually leads to seriously underestimated economic benefits.

For example, in evaluating a new expressway in Japan, the authorities measured the operating costs of a truck on the existing highway in 1958; they were equivalent to 15 U.S. cents a kilometer, excluding taxes. The costs on the new expressway, which is scheduled to be opened in 1969, were estimated at 11 cents, or a saving of 4 cents per truck/kilometer. This saving was then applied to the estimated truck traffic for the years 1969 to 1979; no increase in traffic was assumed after 1979, because the capacity of the expressway would then be reached and vehicle operating costs would thereafter begin to increase.

This example of a “before and after” test illustrates a number of fallacies. The comparison of costs on the existing highway in 1958 with those on the new expressway in 1969 fails to take into account the important fact that the increasing congestion on the existing highway would have increased operating expenses considerably by 1969 over those prevailing in 1958. Secondly, the operating costs on the existing highway would have continued to increase after 1969, while those on the new expressway are likely to remain relatively stable for ten years and the increase thereafter is likely to be less sharp than on the existing highway.

The “before and after” test can lead to even more striking errors. Applied in the Syrian Arab Republic, it showed vehicle operating costs on the existing highway to be quite reasonable; the highway had a fair surface and a satisfactory width. Unfortunately, the highway was not constructed to carry the prevailing heavy loads, and engineers advised that it would break up in about two years and that (even with heavy maintenance expenditures) a complete reconstruction would be necessary. However, vehicle operating costs would not be significantly lower thereafter. The “before and after” test indicated that the reconstruction would not be justified.

The proper standard to apply is the “with and without” test: what (we must ask) are the costs going to be with the new facility, and what would they be without it? In the Syrian case just cited, for example, the “with and without” test indicated that without the new investment vehicle operating costs would go up very sharply, to say nothing of maintenance costs; the avoidance of this increase would in this case have been the proper basis for the economic evaluation of the benefits.

The first step in measuring the benefit from reduced costs is to estimate the traffic that the new highway will carry during its useful life—that is, until it in turn becomes economically out of date. This traffic can be broken down into three main types: “normal,” “diverted,” and “generated” traffic. Normal traffic growth is the growth that would have taken place on the existing facilities in any case, even without the new investment. Since this type of traffic would have operated with unchanged or even increased costs, it clearly gains from reduced costs.

The second type of traffic is that diverted to a new facility either from other modes of transport or from other routes. The benefit for diverted traffic is measured by the difference in transport costs on the old route and on the new. The relevant costs are not the average costs of transport on both facilities but the amounts that would be saved. Thus, if the traffic is diverted from a railway to a new highway, the benefits cannot be measured by comparing the transport costs on the new road with either railway charges or even average railway costs, but by comparing them with the marginal costs of carrying the diverted traffic by railway. If, as often happens, the diverted traffic is only a small part of the total traffic of a railway that has excess capacity, the marginal savings would be substantially less than indicated by a comparison of average costs.

Comparing costs of different transport modes presents a further practical problem in that the transport services provided by each mode usually differ substantially and must therefore be reduced to a common denominator. We should not look at shipping costs alone; total distribution costs are what really matter. For example, a comparison of the costs of coastal shipping traffic diverted to a highway must take into account not merely shipping costs, but also such additional costs as loading and unloading, storage, insurance, breakage, delays, etc. These additional costs are far from negligible; they may readily add as much as 50 per cent to the basic shipping costs. Similarly, trucks transport goods from door to door, while railway service will generally require two loadings and unloadings, which, in addition to the direct costs, frequently involve delays and breakage.

The third type of traffic is that which previously did not exist at all and is created by lower transport costs. This includes traffic both from increases in industrial or agricultural production and transport for commodities previously sold locally but now sent to markets where better prices can be obtained.

Accident Reduction

Accident reduction is clearly an economic benefit, but not every transport improvement reduces accidents. An improved highway may initially increase not only the number of accidents, but, even more unfortunately, the accident rate per vehicle/kilometer and the severity of each accident. This could happen where the increased speed is not offset by additional safety factors, especially in a less developed country where automobile driving skill is itself underdeveloped. Accident reduction is apparently best achieved by expressways with divided lanes and carefully planned access.

Time Savings

Even though most transport improvements reduce travel time, the value of time for passengers and freight is frequently omitted from project evaluations. This may lead to a serious underestimate of benefits as time savings can be substantial.

As far as people are concerned, time can be money but it need not be. Whether it is depends primarily on how the opportunities made possible by the increased availability of time are used—whether for increased production or voluntary leisure, on the one hand, or for involuntary idleness, on the other. Unfortunately, in many developing countries there is extensive underemployment, so that time savings may merely make the situation worse. But even here, time savings for entrepreneurs, for example, may be very valuable.

For purposes of measuring the value of time in the Japanese expressway project, all travelers were divided into two classes: the relatively few who can afford to travel in private cars, and the many who travel in buses. The average value of time was related to the per capita income of the two classes. This showed that in one hour travelers in cars could earn at least the equivalent of 1 U.S. dollar, and those in buses at least 20 U.S. cents. Since there are ample employment opportunities in Japan, this calculation was not unreasonable; it was in fact supported by a separate study of what passengers are willing to pay for time savings.

Time saved on the shipment of freight may well be more valuable in the less developed countries than in countries already more advanced. Freight tied up during transit is in fact capital and is therefore of particular importance when capital is short. This saving can be measured by the price of capital, i.e., the rate of interest. In addition, faster delivery, which usually goes along with more reliable delivery, reduces spoilage and makes possible lower inventories, which in turn are an additional form of capital savings. Beyond this, a delay may immobilize other resources, as when delay in delivering a spare part may keep expensive equipment idle.

Economic Development

It is frequently assumed that all transport improvements stimulate economic development. The truth is that some do but unfortunately some do not, and even some of those that do may not be economically justified in the sense that there may be better investment opportunities.

Before any transport improvement can be said to have stimulated economic development at all, a number of conditions must be met. First, it must be clear that the economic development would not have taken place in any event, even without the transport improvement. A second condition is that the resources used in the new development would otherwise have remained unused or would have been used less productively. Finally, it is essential that the economic activity stimulated does not replace activity that otherwise would have taken place.

When a transport facility does lead to increased output, the net value of this additional output is the proper measure of the economic benefit. Often, however, the transport facility is not the only new investment needed to achieve the increased production. This raises the problem of allocating the benefit, i.e., the increased production, among the transport and the other investments. There is no correct theoretical answer to this problem and the wisest course is not to make an allocation at all but to relate the total benefits to the total investments.

Where the transport facility enlarges the market for commodities previously produced, the economic benefit consists of the difference in value of the commodity in the old and the new market minus the new costs of transport. For example, the price of a commodity in the old market may be 10 cents; in a second market it is 20 cents, but because transport costs are 12 cents, shipment to this market is uneconomic. Assuming a transport improvement that cuts transport costs in half, to 6 cents, the commodity can be delivered to the second market for 16 cents and there sold for 20 cents. The benefits from the new investment would be 4 cents per unit. The increased supply may affect prices in both markets; if so, the benefit is usually valued at prices prevailing after the transport improvement is completed.

Little research has been done on the benefit conferred on national economies by transport projects, and if the main purpose of a transport facility is to stimulate economic development, greater efforts must clearly be made to measure this benefit.


Measuring the economic costs of a project is a much simpler process than measuring its economic benefits. It usually comes down to adjusting the actual expenses to reflect real economic costs. There are three classes of costs for which such adjustments are usually necessary, i.e., where it is appropriate to provide “shadow prices.”

The Use of Shadow Prices

The first example is sales taxes and indirect taxes. The tax on gasoline, for example, is a cost to those who pay the tax, but it does not necessarily reflect economic costs to the country as a whole. An increase in the tax per gallon does not mean that more economic resources are required to produce that gallon. Similarly, license fees and import duties should be excluded, and adjustments should be made for the costs of imports subject to some form of subsidization.

A second example is wages. In most countries minimum wage laws and other inflexibilities have the result that some wages actually paid do not correctly measure the real costs of labor. Where an economy is marked by extensive unemployment or underemployment, the real costs of the type of labor involved are much less than actual wage rates. Where this condition prevails widely and is likely to remain prevalent for some time, as in many less developed countries, the cost of labor, especially unskilled labor, should be calculated at substantially less than actual wage payments. On the other hand, the real costs of skilled labor may be greater than the wages paid. The same considerations are also applicable on the benefit side. In considering the benefits of laborsaving equipment, it must be remembered that the real benefit is substantially less if the replaced labor remains unemployed for a significant period during the economic life of the equipment.

A final example is interest. Interest actually paid is the financial cost of capital, which frequently has no relation to its economic cost. For example, investment funds provided by governments for transportation are often made available at rates below the cost to the government; and even if they cover the government’s costs, the latter do not reflect economic costs if the funds were obtained by the government under direct or indirect compulsion, such as by forced subscription to government bond issues.

In calculating the costs of a project, engineers usually include a contingency for unforeseen expenses. These are of two types. First, costs may be greater than anticipated because the work turns out to be more difficult or more extensive. Then, costs may be greater because inflation increases wages and prices. For the purpose of economic analysis, this second element of the contingency allowance should not be included under costs, nor should a general inflation in the prices of benefits be taken into account.


Once costs and benefits have been measured in monetary terms, the results can be put into at least three different forms: the rate of return on the investment, the benefit-cost ratio, or the pay-back period.

There is unfortunately no uniformity in applying these different forms. In some benefit-cost ratios, for example, gross costs are compared with gross benefits, while in others, some costs are first deducted from the benefits; such differences can affect the ratio very substantially. In rate-of-return calculations, the benefits are sometimes measured against the investment costs (with or without allowance for depreciation) or sometimes by the internal rate of return. It is essential to know exactly what formula was used if the final result is to be correctly interpreted.

While the basic ingredients—the value of the costs and benefits—are the same regardless of the final form in which they are compared, the usefulness of the various forms is different, depending on the purpose. A short pay-back period is important where the future is particularly uncertain, where better investment opportunities are likely to arise soon, or where long-term funds are not available. These considerations are much more important for private businesses than for governments. Also, the fact that the benefits of an investment are large in the beginning may give no indication of what they may be over the life of the investment, so that this method is a particularly poor one for comparing investments of which the benefits vary differently over time. Furthermore, there are superior techniques for incorporating uncertainty into investment analysis.

Discounting benefits and costs by the economic cost of capital (i.e., its opportunity cost) is theoretically the best way of comparing different projects. The main disadvantage of this approach is that a particular interest rate must be chosen for discounting. In practice, the interest rate mistakenly selected is frequently the one being paid, but this may or may not have any relation to the economic cost of capital in the country. Unfortunately, the economic cost of capital is frequently not known or can be estimated only with a considerable margin of error. This is crucial, since the discount rate chosen is one of the major determinants of the benefit-cost comparison.

This disadvantage can be reduced by expressing benefits and costs in terms of the internal rate of return on the investment, i.e., discovering the rate of discount at which annual costs and benefits are equal. In this case, the economic cost of capital becomes important only in the marginal cases where the internal rate of return is not clearly above or below the estimated opportunity cost of capital. For example, it would be virtually certain that an investment in Japan with a rate of return of 12 per cent would be justified, since the economic cost of capital there is less, probably between 6 and 10 per cent. But even where the two rates may be relatively close, the internal rate-of-return formula has the advantage of focusing directly on the crucial question: how the particular investment compares with other investment opportunities. The benefit-cost ratio tends to hide this crucial point by assuming a certain interest rate.

On the other hand, the internal rate-of-return formula also has its disadvantages. While, as a practical matter, it usually leads to a correct choice of projects, it may sometimes be misleading in comparing projects having different lives and different time streams of costs and benefits. It does, however, have the practical advantage that economists, financial experts, and many businessmen have some concept of what an interest rate is, so that a rate of return is probably more meaningful to many audiences than a benefit-cost ratio. On balance, therefore, and unless the economic cost of capital is known within a reasonably narrow range, the internal rate of return on the investment, when used carefully, is usually, but not invariably, a satisfactory form in which to express benefits and costs of transportation projects in the less developed countries.

This article is based upon a paper entitled “Economic Evaluation of Transport Projects,” published in Transport Investment and Economic Development, The Brookings Institution, Washington, D.C.

Other Resources Citing This Publication