A new approach to the economic The calculation of an economic rate of return has become a intensity of farm operations, or the ex- standard feature of project evaluation. Yet, as traditionally defined, pected prices for the project's output. This the rate of return does not focus adequately on-the scarcity of identification helps to improve the project investable resources faced by developing countries, nor on their design, or at least to indicate the chances need to alleviate extreme poverty. A new approach in project of the project having its expected benefits. analysis has therefore been developed which provides a framework Trade-offs between different policy objec- tives are analyzed by testing how a proj- designed to assist systematic decision making in this area. ect's net benefits increase or decrease as, say, the project design is changed to give Anandarup Ray and Herman G. van der Tak more benefits to poorer income groups. The framework for cost-benefit analysis along these lines has been extensively dis- The World Bank lends for projects that cussed in recent years within and outside contribtite to the development objectives of the World Bank, resulting basically in two the borrowing countries-primarily faster types of improvements. First, some of the economic growth and the alleviation of ex- old concepts of analysis, such as the treme poverty. The economic analysis of a shadow exchange rate, have been rede- project assesses its likely impact on the rel- fined and in the process made more pre- evant development objectives by compar- cise. Second, an attempt has been made to ing the various ways in which the scarce make the framework more relevant to pol- resources required by the project might be icy objectives in developing countries, used instead. These resources may include stressing the flexibility needed to adapt the different types of labor and skills, land, im- analysis to the great diversity of situations ported and domestic equipment and ma- to which it is to be applied. This article is terials, and so on. The costs of the project concerned especially with this second as- are the forgone benefits which these re- pect. sources would have produced elsewhere, Cost-benefit analysis which must of course be less than the proj- ect benefits if the project is to be a sensible The costs and benefits of a proposed one. project are always measured against an al- Cost-benefit calculations also help to ternative situation-generally that of not identify the critical parameters of a project. proceeding with the project at all. Thus, In an agricultural project, for example, the the benefits and costs are those expected key measures that determine the outcome, from the project over and above those eX- and therefore need to be closely examined, pected without it. Net benefits to be real- might be the yield per hectare, the labor- ized over future years are given a present 28 Finance & Development I March, 1979 Onalysis of projects value and are expressed in constant prices project will create more net benefits to the (adjusted for purely nominal changes due economy than any other option. To be sure, to inflation) to demonstrate whether the the search for a better option may be lim- total net benefits over the life of the project ited by practical considerations, including will be positive or.negative. its cost; but it is always wasteful to proceed Another approach, equivalent to the first, with the project if a better option is known compares the return on the investment in to be feasible. Since it is not sufficient for the project with the return on investment the calculation to show only that undertak- at the margin in the economy, that is, the ing the project is better than doing noth- "opportunity cost of capital." When the ing, it is necessary to define costs and ben- economic rate of return on the project is efits carefully in most cases. above the opportunity cost of capital, the The definitions of costs and benefits used project clearly helps the economy; con- in the economic analysis of a project de- versely, if it is below, the project will in- pend on the national objectives that are to volve an outright waste of resources. be included in that analysis. When the only It is often thought that a project needs to objective is the naximization of the total be analyzed carefully only when it appears income of the economy, then the costs are marginal. But a project with a high rate of the reductions in income suffered else- return, of, say, 100 per cent, is not neces- where due to the project's use of scarce re- sarily an acceptable investment, since there sources, and the benefits are the additions may be better ways of designing the proj- to the total income brought about by the 8 ect. A highway may be designed according project. If a second objective were to be in- to different standards or it may be started cluded, say, the reduction of income in- later; an irrigation project may be designed equality (the "equity" objective), then the to supply water thinly over a large com- project's effects on equality would have to mand area or concentrated in a smaller be taken into account-an increase in in- area; there may be several hydroelectric come disparity in the country due to the sites or different techniques for generating project would be a cost, and a reduction a enough power to meet the growth in de- corresponding benefit. Another objective mand, and so on. Project analysis attempts could be the alleviation of absolute pov- to ensure that the chosen option for a proj- erty, as distinct from merely reducing the ect is the best possible-not only in terms income gap between rich and poor. These of its size, technology, and location, but last two objectives would involve weight- also in terms of the ultimate beneficiaries ing the income gains flowing to the poor and the quality of output. The analysis, in more heavily than the gains flowing to the short'nust demonstrate that the proposed affluent. Finance & Development March, 2979 , 29 E.~~ 29- An attempt to calculate the effects of a surplus" objective). To be precise, the tra- does, however, treat the gains and#losse, project on such broad objectives as growth, ditional approach is defined in terms of to- accruing in different periods differently- poverty, or equity, and to assign Weights tal real consumption of goods and services future gains and losses being discounted tcI to them according to a countty's socioeco- in the economy, rather than of incomes, make them comparable to changes in con- nomic preferences, poses difficult prob- since the economic welfare of individuals sumption during the current period. Once lems for economic analysis, since market is related to their levels of consumption aggregate consumption is defined in this prices do not necessarib provide a satisfac- rather than to their incomes per se. A proj- way, the cost-benefit analysis can proceed tory basis for measurement. Prices which ect investment reduces the total goods and to measure the project's net impact on to- do reflect the proper weights to be given to services available for current consumption tal consumption over time. the various objectives are called accounting but increases the level of consumption This traditional framework has been very or shadow prices. If, for example, a unit of possible in the future. Projects also change helpful in organizing thought and focusing labor is used in a project, the resulting sac- the relative consumption levels of various attention on the economy-wide changes in rifice in the economy's total income would individuals in the economy, both at a point total income and consumption that result be the shadow price of that labor, if maxi- in time and over time. In order to judge the from a project. However, the choice of a mizing total income were to be the only se- worth of a project from the national point discount rate for making changes in future lected objective. If equity were also an ob- of view, it is necessary to aggregate the consumption comparable to changes in jective, then a different shadow price would various gains and losses accruing to differ- current consumption can be a source of be used which would also take into ac- ent individuals over different periods into major inconsistencies. The lower the dis- count the project's effect on equity. To dis- a single gain/loss measure. For this, some count rate, for example, the more weight is tinguish between these different types of rules or conventions need to be chosen to given to future gains in consumption rela- shadow prices, the shadow prices related define how the different gains and losses tive to sacrifices in current consumption, to the income objective only are usually can be compared. and hence the greater the importance given called "efficiency" prices; by contrast, the The traditional practice has been to re- to savings and growth. A low discount ;hadow prices reflecting total income nea- gard all gains and losses at a point in time rate---of, say, 2-6 per cent-may be ap- sured with differential income weighting to be equivalent, regardless of whether they propriate for cost-benefit analysis in a de- are called "social" prices. affect the poor or the rich. The practice veloping country which has a commitment Not all objectives need to be, nor indeed to rapid growth. However, the opportu- can be, reflected in each cost-benefit anal- nity cost of capital in such a country, re- ysis. Suppose that a country is not partic- Anandarup Ray flecting the yield expected on investment, ularly concerned about reducing poverty, may in fact be much higher because the or that it can do so more effectively through level of investment is low in relation to ex- means other than the project. It would isting opportunities and available funds then be proper to exclude poverty-alleviat- are invested efficiently. ing aspects altogether from the design of Economic Adviser, Projects Premium on savings the project, let alone from its economic Advisory Staff, joined the analysis. On the other hand, if the allevia- World Bank's Young If the yield on investments in an econ- tion of poverty were a prominent consid- Professional Program in omy exceeds the yield necessary to com- eration, then it must be included in the an- t 1967. An Indian citizen, pensate peor e for lower current con- alytical framework if systematic decisions he holds degrees in sumption, then the level of investment is are to be made about the relative merits of 2 economics from the Universities of Calcutta, clearly inadequate-a situation which iS projects which have. different effects on Cambridge, and Chicago, and is the author ofa presumed to be a key feature of most de- poverty. However, if the analysis tries to &number of papers on investment analysis, pricing veloping countries. In such a case, simply incorpoiatetepolicies, and related topics. discounting future costs and benefits by than three-it may become too complex for . the opportunity cost of capital, as in the Oractical use. Herman G. van der Tak traditional approach, gives incorrect re- The issues addressed and the precision & sults' since consumption gains and losses it desired in the analysis tend to vary over different periods are not properly co'- the project cycle. The study of an irrigation pared. If, instead, the rate appropriate for project might begin with the choice of the discounting future consumption-the areas to be irrigated, move on to the choices a Dutch citizen, holds "consumption rate of interest"-is used, regarding the operation of the particular - graduate degrees in law this will also lead to errors as it underest- schemes devised within a project, and then ... and economics from the mates the productivity of investments and proceed to alternative methods of cost re- University of Utrecht and thereby causes additional investments re- covery. The economic analysis of alterna- the London School of sulting from the project to be undervalued* tives is likely to be relevant to all such de- Economics. Prior to joining To reflect properly both the relative value cisions, Even though the analysis of project the Bank in 1961 lie served as economist with the of current and future consumtion and the UN Regional Economic Commissions in Europe and c a t designs is bound to be rather crude in the Asia. His positions at the Bank have included Head unsatisfactory level of investment, I early stages, it should still incorporate the of Sector and Project Studies, Economic Adviser, necessary to use the "consumption rate relevant socioeconomic objectives. Projects, and currently, Director, Projects Advisory interest" as the discount rate in combira- The traditional approach Staff. In 1970-71 he was a Visiting Professor at the tion with a special premium for adjusting Institute of Social Studies in The Hague. He is the the value of investmentexpenditures. Thus, Cost-benefit analysis has traditionally author of several articles and monographs, the most if a 5 per cent return is all that is necessa' focused only on maximizing incomes (an recent being Economic Analysis of Projects (Johns to compensate for a sacrifice of $1 in CU' objective variously referred to as the "eco- Hopkins, 1975, jointly with Lyn Squire). rent consumption, but if that $1 when in noric," the "efficiency," or the "social vested yields 10 per cent, then investre 30 Finance & Development I March, 1979 the margin should be regarded as twice levels of investment in the economy, then such a employment, income inequality, a valuablu as current consumption. the scarcity of foreign exchange is reflected and the alleviation of poverty are to enter This introduction of a premium on in- in the premium on investment. It has re- into the economic analvsi,. Gain, and ,etment, and thus on savings, requires centiv become clear that the so-called losses to different income groups will then ,hc project analyst to judge how much of "shadow exchange rate," or "shadow price be weighted differently to reflect these he income created by the project would be of foreign exchange," as used in tradi- concerns, by giving more value to benefits 1ved. Since the public sector and the pri- tional analysis, does not bear on the scar- to the poorer groups. For this purpose it ;ate sector, and the different income groups city of foreign exchange in this sense. This would be necessary to assess which in- ,ithin the private sector, save at different "shadow exchange rate" is only a device come groups are expected to gain or lose ates, one needs to estimate how the incre- for correcting the distortions in the relative from the project. A broad distinction be- rental income derived from the project is prices of internationally traded and non- tween only a few income groups is likely to eoing to be distributed among the various traded goods, and for that purpose it is suffice in practice. An even simpler dis- leneficiaries. Great precision is not neces- also used in the new method. tinction of beneficiaries into only two ,ary in this estimation. A distinction be- There are practical difficulties, ut course, groups, abovc or below a threshold level of tween, say, three income groups in the pri- with the use of a premium on investment. -poverty, would suffice if the reduction of vate sector-the very rich, the very poor, It is often hard to decide the proper size of absolute povei-ty is the desired objective, and a large middle-income group-may be the premium, and estimating the increases It is sometimes thought that even though sufficient. in income and savings of different groups governments may be concerned with in- An investment premium makes invest- from a project may be a demanding task. come distribution and poverty alleviation, irents more attractive in those public or Would it then not be better to rely solely they need not introduce such concerns into private sector enterprises that reinvest a on qualitative judgments in this respect? project decisions, but instead should rely greater share of their profits productively. The World Bank, for example, has always on other instruments of policy. Even though most governments have many pol- In many developing countries. .. the introduction icy instruments available which could di- of equity or poverty objectives into rectly or indirectly affect equity and pov- erty ani- .ig their populations, the majority project selection tends to be an important of developing countries seem to have found complement to other policy measures. poverty redressal or the alteratipn of in- come distribution very difficult. The redis- On the other hand, any gains derived by placed strong emphasis on financial viabil- tribution of land, for example, is generally the poor tend to be penalized, insofar as ity, high levels of cost recovery, replicabil- crucial to redistributing incomes in most of these classes tend to save less of their in- ity, and other policies which directly or in- these countries; but effective land reform come gains than the rich. Investments in directly reflect concern about the scarcity of has often proven infeasible, and land taxes heavy industries, such as steel and petro- investable resources. Unfortunately, qual- are notoriously difficult to administer. The chemicals, and in revenue-generating util- itative adjustments -in project decisions imposition of progressive income taxes also ities, such as power and telecommunica- rarely work satisfactorily. Suppose the has practical limits, especially if serious ad- tions, are likely to become relatively more economic rate of return of a project, mea- verse effects on earning incentives are to attractive. Large-scale mechanized farming sured without an investment premium, is be avoided. Moreover, reliance on indirect and estate plantations will perhaps also be marginally above the cutoff rate; can this taxes, or on inflationary finance, would af- favored. In other words, the premium will project still be rejected if all of its gains are fect the allocation of resources adversely tend to make capital-intensive projects expected to be spent on additional con- and tend to weigh more heavily on the rel- more attractive and reduce the emphasis sumption? Or, can a project with an unsat- afively poorer groups. on employment generation. Higher taxes isfactory rate of return be accepted if all of In many developing countries, there- on consumption goods, on income, and on its gains are expected to be reinvested? It is fore, the introduction of equity or poverty land will appear more desirable, assuming clear that answers to questions such as objectives into project selection tends to be that the government uses the tax revenues these implicitly involve a quantification of an important complement to other policy productively. The allocation of investment the value of savings and investment. Such measures. It is usually easier to locate proj- funds between private and public sectors implicit, ad hoc quantification can, how- ects in backward areas or to design them may also be affected, insofar as these sec- ever, lead to grossly inconsistent project for urban or rural poverty groups than, for tors have different propensities to reinvest decisions. example, to change the tax system or to re- and different levels of efficiency. If such distribute assets directly through a national differences are considered significant, they Inequality, poverty, and basic needs land reform. It is much harder to shift the should be reflected in different investment distribution of existing assets than to direct premiums for the public and private sec- Introducing the investment premium the creation of new assets in favor of the tors. does not require any change in the basic poor-although the allocation of public A primary purpose of the new cost- economic objective of traditional cost- sector investments also has political con- benefit analysis is to take proper account of benefit analysis, which will still treat con- straints. the "scarcity of foreign exchange" faced sump6on gains or losses to different indi- The current practice in the World Bank by many developing countries. It is often viduals equally. The premium focuses on treats the alleviation of absolute poverty as thought that this "scarcity" is also al- the correct assessment of a project's im- a very important aspect of many of the lowed for in traditional analysis. However, pact on total consumption, but does not af- projects it finances. In order to orient proj- the scarcity value of foreign exchange de- fect the concept of the costs and benefits ects toward this goal, several informal rules Pends on the economic objectives it ad- that are being aggregated. of search are used in the identification versely affects. If the benefit of additional The concepts themselves, however, will stage, such as upper limits for the cost per foreign exchane is that it permits higher need to be changed if concerns with issues job created, or'for the acceptable cost per Finance & Development I March, 1979 31 beneficiary. However, search rules are not The cutoff point for subsidies might then treats certain goods and services as basic an adequate substitute for a fully inte- be at an income level which is only, say, 25 needs or "merit wants" that should be grated analysis of the conflicts between ob- per cent of the national average. Subsidies satisfied as a matter of government policy, jectives, such as more employment or in- will thus tend to be restricted to the very rather than being met through charity de. come for the poor versus more rapid growth poor groups in such countries. If, on the pendent on private preferences. The plan- in output. In land settlement projects, for other hand, the country is more concerned ners or policymakers decide therefore example, the question frequently arises with equity or with alleviating poverty, as which needs are basic, and what quantity whether to allocate small units to each set- are, say, Sri Lanka and Tanzania, the ap- and quality of service should be provided. tier and thus spread the benefits widely or propriate critical consumption level would They fix the weights that determine the to allocate fewer, larger units in the inter- be much higher, perhaps as high as 75 per importance to be given to additional con- ests of higher productivity of land use. A cent of average income. The critical con- sumption of the goods or services which rule restricting the cost per beneficiary may sumption level is a relative income meas- meet the basic needs of various (usually in- be counterproductive in such cases unless ure in the country and is usually well above come) groups, and how soon these needs it is derived from a full analysis of the levels representing absolute poverty. It is should be fully satisfied in relation to other trade-offs involved. widely used in practice, especially in the objectives of growth and distribution. These If poverty or equity objectives are intro- context of project-related pricing and cost- specific basic need weights are a straight- duced in the analysis then suitable rules recovery policies. forward variation on, and complement, the must be specified for aggregating the var- The critical consumption level is one of more general distribution weights dis- ious gains and losses accruing to different the benchmarks for assigning distribution cussed above. individuals into overall benefit and cost weights that reflect a country's policy The economic rate of return.of a project figures for each year of the life of the proj- priorities. There are many other tests that T efe noters of re th a rorat ect. The decision maker who rules on a can be devised to determine the most rea- defined in terms of prices that incorporate project, or the advisor who recommends a sonable weighting scheme for the country be called its "social rate of return" It will project, must necessarily use a scheme for concerned. Generally speaking, assigning frequently differ from that calculated on weighting the gains and losses of different equal weights to different income groups, the basis of traditional efficiency prices. income groups. The question for any par- as in the traditional economic analysis, There is no built-in tendency for social ticular country is then which type of would appear to be appropriate only in ex- rates to be higher than the traditional rates weighting scheme is most realistic and rel- ceptional cases. Some degree of differen- of return. The new approach is designed in evant for this purpose? Should one choose tiai'n between income groups, at least to such a way that the social rates will be the equal weights used in traditional prac- take account of extreme wealth and ex- higher only to the extent that any increases tice or should one differentiate according treme poverty, is usually likely to be more in consumption due to the project accrue to income groups? The answer to this realistic. to those below the poverty line, and will be question obviously depends on the specific The introduction of different weights for to the extent that the project in- socioeconomic priorities of the country for different income groups would counteract cerase the consumption of the relatively which the project is planned, and no single some of the effects of giving special weight creaue Se ns is do t makely to te geeraion f aditinal nve tt aslento Social analysis does not make it weighting scheme is universally applica- to the generation of additional investmen e o ble. But these priorities are usually not ex- Projects which lead to additional savings dfern justify projects, but it tends to jus plicitly formulated, and the analyst is faced and reinvestment will still be favored, other tify ifferent projects, that is, projects that with having to deduce their relative impor- things being equal, unless the benefits ac- investments in the economy. tance. crue to those below the critical consump- It is, however, possible to test the plau- tion level. Labor-intensive operations and The rigid adherence to one particular set sibility of relative weights reflecting differ- employment generation will be favored to of weights, as in traditional cost-benefit ent policy objectives by analyzing various the extent that the additional labor income analysis, appears too doctrinaire to be ap- national policies. For example, equity is accrues to the poor. propriate for all deveioping countries, or often an important aspect of taxation pol- The differences between countries can even for the same country at different stages icy, and there is always an exemption limit be easily reflected in the analysis since the of its development. In countries where the for income taxes. Moreover, many govern- emphasis given to employment, equity, or distribution of project benefits is impor- ments run large subsidy programs for the poverty alleviation can be "controlled" by tant, the traditional way of analyzing proj- poor. Such policies suggest that if a person varying the critical consumption level: the ects is only a partial indicator of the eco- is poor enough, then an extra dollar to him lower the level, the less the importance nomic impact of a project and is not is valued more highly than an extra dollar given to such concerns. However, since necessarily a reliable guide to project deci- of government revenue, and therefore there- the same differential income weights and sions. The new approach, on the other is a critical or break-even point-the "crit- the same critical consumption level are to hand, focuses directly on the hard choices ical consumption level"-at which mar- be use, .r all projects within a country, it facing developing countries between ginal private gains are socially worth just is clear that ad hoc judgments are avoided growth and redistribution and is likely to about as much as marginal increases in by this method. The use of poverty or eq- improve the decision-making process. As government revenues. A person should uity considerations on an ad hoc basis tends- the experience with social pricing in cost- not receive subsidies unless the level of to give a "free license" to accept any and benefit analysis accumulates and the meth* consumption he can afford is below this all projects that help the poorer groups. In odology is adapted accordingly, it is likely critical level. If the country concerned is contrast, the new approach demands con- to become a widely employed tool of anal' deeply committed to growth, as are Brazil, sistency and discipline in project choice. ysis, not only in the World Bank but also i" Ivory Coast, and Korea, for instance, then Another important objective for many other international and national institu- heavy weight is given to generating in- developing countries is to meet the "basic tions with responsibility for selecting Projs comes for investment, and hence this crit. needs" of their people. Definitions of basic ects which best meet the policy objectiv ical consumption level should be yery low, needs vary, but the principal interpretation of the country concerned. 32 F Dmme D iopmient i March, 1979