Journal Issue
Finance & Development, September 1973

Income Distribution and Project Selection: The third London Airport serves as an example of how noneconomic issues and income distribution might have been introduced along with the efficiency criterion at a very early stage of project selection. If it had, the site chosen might be different….

International Monetary Fund. External Relations Dept.
Published Date:
September 1973
  • ShareShare
Show Summary Details

V.C. Nwaneri

The economic theory of project analysis is only a branch of the general theory of resource allocation which is the core of welfare economics. This theory prescribes a simple test for project selection: under the existing pattern of income distribution only those projects should be selected which—assuming the beneficiaries could compensate the sufferers—make no one worse off, and at least one person better off. However, even if this “efficiency” test can be satisfied the fact that it ignores the equity issues of existing income distribution strips such a criterion of its policy prescriptive value. Of course, the modification of the criterion to meet equity considerations is more easily said than done. For one thing such modification raises a whole complex of political and social questions to which few, if any, political leaders, even in the advanced countries, can provide a ready answer. The economist will simply have to fill in the gaps because the equity issues are too important to be ignored.

Selecting a site for the airport

The selection of a third London Airport did not specifically consider income distribution as one of the major issues influencing the decision. But it illustrates how noneconomic issues and income distribution can be introduced along with the efficiency criterion at a very early stage of project selection.

The search for the third London Airport site officially began in 1961. An original decision for a site at Stansted (about 30 miles northeast of London), made without consideration of noneconomic issues, was later revoked when public outcry showed that such issues could not be ignored. In May 1968, the Roskill Commission was appointed to re-examine the choice objectively. While its first public hearings were proceeding, a long list of 78 possible sites was reduced to a medium list of 29 sites by the Commission’s own research team using the conventional least cost criterion. The main factors considered by the research team at this stage were noise, defense, and surface access costs. The medium list of 29 was further reduced to 15 after considering more factors such as land and site preparation and air traffic control costs. Finally, by means of sensitivity tests on these cost factors, the 15 proposals were reduced to 9 and subsequently to a short list of 4 that excluded the original controversial choice of Stansted.

The 4 short-listed sites were the little villages of Cublington (Buckinghamshire), Foulness (Essex Coast), Nuthampstead (Hertfordshire), and Thurleigh (Bedfordshire). These were carried forward to-the second stage of the investigation whose purpose was to give the public, especially local residents, the occasion, through a series of local hearings, to voice its views and criticisms of the Commission’s work thus far. During these local hearings, the Commission debated such topics as site location and alignment, noise, urbanization, regional planning, existing infrastructure, agriculture, fishing and ornithology, amenity, employment, and water and drainage.

In the-third stage of the selection process, the Commission’s own research team conducted its analytical investigation into the four short-listed sites using the conventional technique of cost benefit analysis modified to take into account such non-economic issues as noise, meteorology, airspace movements, air safety, defense, education, health, and recreation in addition to the purely economic costs and benefits related to construction, air passenger and freight, land transport, housing, agriculture, commerce, and industry. The incremental costs and benefits (on the above economic and noneconomic issues) for each of the four sites were computed and used for ranking the sites. The income distribution factor was not included. The little village of Cublington, about 30 miles northeast of London, was chosen by the research team as the site involving the lowest incremental cost (or highest incremental net benefit). The team’s results were published and comments were sought from various experts inside and outside the United Kingdom.

The fourth stage was primarily to clear up all technical issues raised by the research team and other experts. The fifth and final stage was a series of public hearings concentrating on the research team’s findings. In fact, however, the U. K. Government eventually rejected the Commission’s choice of Cublington and selected Foulness instead.

Deciding the stage of application

Project selection is a continuous process of sifting from a whole spectrum of possibilities. At each stage in this process the conventional criteria themselves are often watered down to simplify the task. The World Bank’s selection process, for example, embraces project identification work (in which other international or local government agencies contribute), project preparation (which is a long tedious task usually undertaken by international consulting firms), project preappraisal (to settle the preliminary issues), the appraisal itself, and the concluding negotiations (during which a whole complex of understandings are translated into legal obligations). The first practical problem is deciding at which stage in this chain the income distribution criteria should be applied. The answer seems to be as early as possible—i.e., at the identification stage—when one’s choice of objective function can exert strong influence. But quantification of the distributional effects may become possible only at the appraisal stage when the necessary numbers become available. Diagram A sets out the stages involved in incorporating income distribution criteria in the final selection of a project.


Incorporating income distribution criteria

Having decided at which stage of project investigation to apply distribution criteria, the next problem is to decide the extent of such application. A development project affects a large variety of people within and outside the country of location; it has primary, secondary, and tertiary effects and each of these creates its own multiplier effects. It is Utopian, therefore, to expect an analyst even using only conventional analysis to fully investigate all the effects and the affected people. Only the main groups of sufferers and beneficiaries, and the principal costs and benefits can be analyzed by the conventional and the modified methods. In the London Airport study, for example, every taxpayer will be affected unfavorably and every potential air traveller within or outside the United Kingdom will be affected favorably by the project. However, only those groups bearing the heaviest and most direct impact of the project can be considered.

My proposed method of incorporating income distributional considerations in the appraisal of development projects (see Finance and Development, March 1973) assumes that the unweighted costs and benefits have been quantified and considers only the problems of classification and weighting. Let us consider the application of these two steps in turn.

Classification—identification of project beneficiaries and sufferers

In investigating the income distribution issues arising from the third London Airport study the beneficiaries can be identified under the following subgroups:

  • (1) the British Airport Authority (and its employees) which will construct and operate the airport;

  • (2) the British and foreign airline operators and their employees;

  • (3) British and foreign leisure air travelers;

  • (4) British and foreign employers of business air travelers;

  • (5) British and foreign air freight operators;

  • (6) industries and urban residents attracted to the airport site and dependent upon it.

Similarly the sufferers—those who incur the cost—can be identified under two major groups:

  • (1) The communities living in the areas affected by the airport.

  • (2) The people who will bear the burden of public financing (by taxation, inflationary financing, reduced expenditure on other services etc.) and any additional private financing.

Most development projects have similar groups of beneficiaries and sufferers. In all such projects, the two groups overlap—the taxpayers are themselves air travelers, the airport authority will itself contribute towards construction costs, the local communities subjected to noise will themselves reap the benefits of urbanization arising from the airport. To eliminate this overlapping we shall consider only the net benefits (B) derived by the beneficiaries and only the net costs (C) borne by the sufferers.


Next, the beneficiaries and the sufferers must be subgrouped under income classes. It is usually quite difficult to obtain reliable data on incomes not only because there is a lack of unpublished data but also because of the different types of wage and nonwage incomes. Some other index such as value of property occupied by beneficiaries or sufferers could achieve the same purpose, assuming that the level of incomes (wage and nonwage) determine the type of property occupied by most people. In the airport study, data on property values were more reliable than those on incomes. Therefore, the local communities affected were subgrouped into low, medium, and high income households including rented and owner-occupied accommodations. The evidence collected by the research team showed that the average air traveler has a higher income than the average person in the United Kingdom. It can therefore be assumed that the majority of taxpayers belong to the low income group and that most of the air travelers belong to the high income households. Having classified the households under income groups, these must be further classified under regional groups. This was simple in the airport study because the four proposed sites are located in four different areas.

Weighting: allocation of unweighted costs and benefits

In my proposed method, the initial step in the weighting process is to allocate the total unweighted benefits and costs to the respective groups of beneficiaries or sufferers. In the airport study the unweighted benefits were measured by the costs incurred by beneficiaries on the assumption that these costs are minimum measures of such benefits. This common assumption avoids the difficulty of measuring the so-called consumers’ surplus (the difference between what some consumers would be willing to pay and the lower price they may actually pay for a given good or service). On this assumption the following elements of net benefits were allocated to the beneficiaries (1-6 above): Airport construction costs, airport services, meteorology, airspace movements, passenger user costs, road and rail capital costs, and finally, air safety costs. As to the sufferers, the net costs to them (i.e., costs not readily recoverable from air travelers) include such items as additional expenditure on defense and public scientific establishments as a result of the airport decision. In addition, the local residents bear the net costs resulting from the effect of the airport on private airfields, residential property destroyed or subjected to noise, on schools, hospitals, local authority buildings, on agriculture, local commerce and industry, and on recreation and urbanization.

This appears to yield a stream of costs larger than of benefits since costs include the enumerated benefits plus some additional costs. But, because the benefits continue over many more years than the costs, there is a net benefit during the entire life of the project.

Next we must allocate the net benefits to the individual groups of beneficiaries (i.e. to a given income group in a given region) and similarly the net cost to the individual groups of sufferers. To do this, we have to determine the proportion of one subgroup of beneficiaries (or sufferers) over the total and then allocate the total unweighted benefits (or costs) to these subgroups on the basis of these proportions. In my reappraisal of the airport study, this involved determining (for each of the four proposed sites) the proportion of low income households over the total number of households in that area and similarly for medium and high income households. The total net costs borne by the sufferers is then allocated to the respective groups on the basis of these proportions. The results are then to be combined with the income distributional weights according to some specified formulae.

Deriving the income distributional weights

My proposed method clearly left the derivation of the weights an open question to be determined by the circumstances of the country and the project concerned. For example, the effective marginal rates of personal income taxes can be accepted as an already-established basis for income distributional weighting (other possible measures that can be applied to sectors not taxed were discussed in the previous article). These rates have the effect of equalizing tax incidence. For example, a marginal rate of X per cent means that the net reward from an additional dollar is reduced by X per cent. We can equate these tax rates as our distributional weights to equalize project incidence between beneficiaries and sufferers. By so doing the additional net benefit of a dollar from the project is reduced by X per cent or the additional net cost of a dollar from the project is increased by X per cent.

Measuring the distributional effects

The distributional weights (X per cent) applicable to each income regional group is easily obtained from a table of marginal income tax rates. These weights can then be multiplied by the net benefits (B) or net costs (C) in order to obtain the distributional effects to respective income/ regional groups. It must be emphasized that the resulting BX or CX (but not both) should be used to measure the distributional effect of the project and that all projects to be compared must be measured in the same way. If BX and CX are used together then the result will be double counted. I prefer to measure the distributional effects by BX (weighted benefits) instead of CX (weighted costs). First, BX was larger than CX. Second, assuming the initial egalitarian objective of choosing projects that benefit the poor more than the rich, we chose BX to weigh down the benefits accruing to the rich, who would be the principle beneficiaries.

Sensitivity analysis

The results (BX) can be tested for sensitivity to some of the important assumptions, as is usually done in conventional cost-benefit analysis. In my reappraisal of the third London Airport study, the sensitivity of the BX-results to various levels of distributional weights was tested. For instance, the weights (equated to marginal tax rates) were tested for sensitivity to such factors as the varying numbers of people affected (the size of total income or total households does not necessarily reflect the total individuals involved), and the varying degrees of depreciation in house prices within the same noise zone. It may also be felt that the weights are different from those based on factors other than marginal income tax rates. For example, in the airport study, distributional weights, derived on the basis of diminishing marginal utility of income (defined for specified income elasticities), were used in the tests.

Project ranking

If project selection is to be undertaken strictly according to the principles of resource allocation, every project must be compared (ranked) against all possible alternatives ensuring that the one with the highest social and economic returns be taken. On this principle, the four proposed sites ought to be ranked on the basis of the economic and distributional results. This means that the results of the economic analysis (i.e., the unweighted cost-benefit results for each site as derived in the Roskill Commission’s Stage III analysis) should be added to the results of the distributional analysis (i.e., the distributional effects measured by BX for each site). The four sites can then be ranked on the basis of the sum of these two results. It could be argued that the intersite ranking should be based entirely on BX and not the sum of BX and the cost-benefit results but such approach clearly ignores economic issues and bases the project judgment entirely on distributional criteria. On the other hand, the additive approach bases the choice of a site on both economic and distributional criteria. On this basis, the results of my reappraisal of the airport study show Thurleigh as the preferred site.

quoting or reprinting Finance and Development articles

No permission is required to quote or reproduce material appearing in Finance and Development.

A credit line or other acknowledgment is requested.

The Editor would be glad to receive two copies of publications containing reprints or quotations.

Other Resources Citing This Publication