INCOME DISTRIBUTION AND PROJECT SELECTION

In a previous article the author argued that considerations cremental net benefit). The team's results of income distribution should be included in the analysis of were published and comments were sought development projects. Here he illustrates the from various experts inside and outside application of the method to an appraisal of the site for the . The fourth stage was primarily to clear the proposed Third Airport. up all technical issues raised by the re- V. C. Nwaneri search team and other experts. The fifth and final stage was a series of public The economic theory of project analysis the research team at this stage were noise, hearings concentrating on the research is only a branch of the general theory of defense, and surface access costs. The team's findings. In fact, however, the U. K. resource allocation which is the core of medium list of 29 was further reduced to Government eventually rejected the Com- welfare economics. This theory prescribes 15 after considering more factors such as mission's choice of Cublington and selected a simple test for project selection: under land and site preparation and air traffic Foulness instead. the existing pattern of income distribution control costs. Finally, by means of sensi- only those projects should be selected tivity tests on these cost factors, the 15 Deciding the stage of application which—assuming the beneficiaries could proposals were reduced to 9 and subse- Project selection is a continuous process compensate the sufferers—make no one quently to a short list of 4 that excluded of sifting from a whole spectrum of possi- worse off, and at least one person better the original controversial choice of bilities. At each stage in this process the off. However, even if this "efficiency" test Stansted. converitional criteria themselves are often can be satisfied the fact that it ignores the The 4 short-listed sites were the little watered down to simplify the task. The equity issues of existing income distribu- Villages of Cublington (), World Bank's selection process, for ex- tion strips such a criterion of its policy Foulness (Essex Coast), ample, embraces project identification prescriptive value. Of course, the modi- (Hertfordshire), and (Bedford- work (in which other international or local fication of the criterion to meet equity shire). These were carried forward to-the government agencies contribute), project considerations is more easily said than second stage of the investigation whose preparation (which is a long tedious task done. For one thing such modification purpose was to give the public, especially usually undertaken by international con- raises a whole complex of political and local residents, the occasion, through a sulting firms), project preappfaisal (to social questions to which few, if any, series of local hearings, to voice its views settle the preliminary issues), the appraisal political leaders, even in the advanced and criticisms of the Commission's work itself, and the concluding negotiations countries, can provide a ready answer. The thus far. During these local hearings, the (during which a whole complex of under- economist will simply have to fill in the Commission debated such topics as site standings are translated into legal obliga- gaps because the equity issues are too location and alignment, noise, urbanization, tions). The first practical problem is decid- important to be ignored. regional planning, existing infrastructure, ing at which stage in this chain the income agriculture, fishing and ornithology, distribution criteria should be applied. The Selecting a site for the airport amenity, employment, and water and answer seems to be as early as possible— The selection of a third London Airport drainage. i.e., at the identification stage—when one's did not specifically consider income dis- In the-third stage of the selection process, choice of objective function can exert tribution as one of the major issues influ- the Commission's own research team con- strong influence. But quantification of the encing the decision. But it illustrates how ducted its analytical investigation into the distributional effects may become possible noneconomic issues and income distri- four short-listed sites using the conven- only at the appraisal stage when the nec- bution can be introduced along with the tional technique of cost benefit analysis essary numbers become available. Diagram efficiency criterion at a very early stage of modified to take into account such non- A sets out the stages involved in incor- project selection. economic issues as noise, meteorology, air- porating income distribution criteria in the The search for the third London Airport space movements, air safety, defense, edu- final selection of a project. site officially began in 1961. An original cation, health, and recreation in addition to decision for a site at Stansted (about 30 the purely economic costs and benefits re- Incorporating income distribution miles northeast of London), made without • lated to construction, air passenger and criteria consideration of noneconomic issues, was freight, land transport, housing, agricul- Having decided at which stage of project later revoked when public outcry showed ture, commerce, and industry. The in- investigation to apply distribution criteria, that such issues could not be ignored. In cremental costs and benefits (on the above the next problem is to decide the extent of May 1968, the Roskill Commission was economic and noneconomic issues) for such application. A development project appointed to re-examine the choice ob- each of the four sites were computed and affects a large Variety of people within and jectively. While its first public hearings used for ranking the sites. The income outside the country of location; it has pri- were proceeding, a long list of 78 possible distribution factor was not included. The mary, secondary, and tertiary effects and sites was reduced to a medium list of 29 little village of Cublington, about 30 miles each of these creates its own multiplier sites by the Commission's own research northeast of London, was chosen by the effects. It is Utopian, therefore, to expect team using the conventional least cost research team as the site involving the an analyst even using only conventional criterion. The main factors considered by lowest incremental cost (or highest in- analysis to fully investigate all the effects 27

©International Monetary Fund. Not for Redistribution and the affected people. Only the main Subgrouping avoids the difficulty of measuring the so- groups of sufferers and beneficiaries, and Next, the beneficiaries and the sufferers called consumers' surplus (the difference the principal costs and benefits can be must be subgrouped under income classes. between what some consumers would be analyzed by the conventional and the modi- It is usually quite difficult to obtain reliable willing to pay and the lower price they fied methods. In the London Airport study, data on incomes not only because there is may actually pay for a given good or for example, every taxpayer will be affected a lack of unpublished data but also be- service). On this assumption the following unfavorably and every potential air travel- cause of the different types of wage and elements of net benefits were allocated to ler within or outside the United Kingdom nonwage incomes. Some other index such the beneficiaries (1-6 above): Airport will be affected favorably by the project. as value of property occupied by benefi- construction costs, airport services, meteor- However, only those groups bearing the ciaries or sufferers could achieve the same ology, airspace movements, passenger user heaviest and most direct impact of the purpose, assuming that the level of incomes costs, road and rail capital costs, and project can be considered. (wage and nonwage) determine the type finally, air safety costs. As to the sufferers, My proposed method of incorporating of property occupied by most people. In the net costs to them (i.e., costs not income distributional considerations in the the airport study, data on property values readily recoverable from air travelers) appraisal of development projects (see were more reliable than those on incomes. include such items as additional expendi- Finance and Development, March 1973) Therefore, the local communities affected ture on defense and public scientific estab- assumes that the unweighted costs and were subgrouped into low, medium, and lishments as a result of the airport deci- benefits have been quantified and considers high income households including rented sion. In addition, the local residents bear only the problems of classification and and owner-occupied accommodations. The the net costs resulting from the effect of weighting. Let us consider the application evidence collected by the research team the airport on private airfields, residential of these two steps in turn. showed that the average air traveler has a property destroyed or subjected to noise, higher income than the average person in on schools, hospitals, local authority build- Classification—identification of the United Kingdom. It can therefore be ings, on agriculture, local commerce and project beneficiaries and sufferers assumed that the majority of taxpayers industry, and on recreation and urbaniza- In investigating the income distribution belong to the low income group and that tion. issues arising from the third London Air- most of the air travelers belong to the high This appears to yield a stream of costs port study the beneficiaries can be identi- income households. Having classified the larger than of benefits since costs include fied under the following subgroups: households under income groups, these the enumerated benefits plus some addi- (1) the British Airport Authority (and must be further classified under regional tional costs. But, because the benefits con- its employees) which will construct groups. This was simple in the airport tinue over many more years than the costs, and operate the airport; study because the four proposed sites are there is a net benefit during the entire life (2) the British and foreign airline op- located in four different areas. of the project. erators and their employees; Next we must allocate the net benefits (3) British and foreign leisure air Weighting: allocation of unweighted to the individual groups of beneficiaries costs and benefits travelers; (i.e. to a given income group in a given (4) British and foreign employers of In my proposed method, the initial step region) and similarly the net cost to the business air travelers; in the weighting process is to allocate the individual groups of sufferers. To do this, (5) British and foreign air freight total unweighted benefits and costs to the we have to determine the proportion of operators; respective groups of beneficiaries or suf- one subgroup of beneficiaries (or sufferers) (6) industries and urban residents at- ferers. In the airport study the unweighted over the total and then allocate the total tracted to the airport site and de- benefits were measured by the costs in- unweighted benefits (or costs) to these pendent upon it. curred by beneficiaries on the assumption subgroups on the basis of these propor- Similarly the sufferers—those who incur that these costs are minimum measures of tions. In my reappraisal of the airport the cost—can be identified under two such benefits. This common assumption study, this involved determining (for major groups: (1) The communities living in the areas affected by the airport. (2) The people who will bear the bur- den of public financing (by taxa- tion, inflationary financing, reduced expenditure on other services etc.) and any additional private financing. V. C. Nwaneri Most development projects have similar a Nigerian, is a World Bank economist in the groups of beneficiaries and sufferers. In all Public Utilities Division, Europe, Middle East, such projects, the two groups overlap— and North Africa Region. He studied at the the taxpayers are themselves air travelers, London School of Economics and was Economic the airport authority will itself contribute Advisor in the Department of Trade and Industry towards construction costs, the local com- of the British Government. Before joining the munities subjected to noise will themselves ) staff of the Bank in 1971, he worked with the reap the benefits of urbanization arising ; Roskill research team. For his independent : reappraisal of the team's study see the Journal from the airport. To eliminate this over- ! of Transport Economics and Policy, Sept. 1970, lapping we shall consider only the net : and a rejoinder in Sept. 1972. benefits (B) derived by the beneficiaries and only the net costs (C) borne by the sufferers. 28

©International Monetary Fund. Not for Redistribution Diagram A PROJECT SELECTION USING INCOME DISTRIBUTIONAL CRITERIA

PRE-APPRAISAL STAGE PROJECT APPRAISAL POST-APPRAISAL

each of the four proposed sites) the pro- the country and the project concerned. For additional net cost of a dollar from the portion of low income households over example, the effective marginal rates of project is increased by X per cent. the total number of households in that personal income taxes can be accepted as area and similarly for medium and high an already-established basis for income Measuring the distributional effects income households. The total net costs distributional weighting (other possible The distributional weights (X per cent) borne by the sufferers is then allocated to measures that can be applied to sectors applicable to each income regional group the respective groups on the basis of these not taxed were discussed in the previous is easily obtained from a table of marginal proportions. The results are then to be article). These rates have the effect of income tax rates. These weights can then combined with the income distributional equalizing tax incidence. For example, a be multiplied by the net benefits (B) or weights according to some specified for- marginal rate of X per cent means that the net costs (C) in order to obtain the distri- mulae. net reward from an additional dollar is butional effects to respective income/ reduced by X per cent. We can equate regional groups. It must be emphasized Deriving the income distributional these tax rates as our distributional weights that the resulting BX or CX (but not both) weights to equalize project incidence between should be used to measure the distribu- My proposed method clearly left the beneficiaries and sufferers. By so doing tional effect of the project and that all derivation of the weights an open question the additional net benefit of a dollar from projects to be compared must be measured to be determined by the circumstances of the project is reduced by X per cent or the continued on page 33 29

©International Monetary Fund. Not for Redistribution