ARNOLD C. HARBERGER
UNIVERSITY OF CHICAGO
In this paper, I attempt to bring into focus what I believe to be some of the important practical issues that face development planners in the field of project appraisal. I shall try, insofar as possible, to recognize the handicaps under which planners operate, most importantly the handicaps imposed by imperfect foresight and by the virtual necessity of decentralized decision-making. To elaborate briefly on these handicaps, I think we must take it for granted that our estimates of future costs and benefits (particularly the latter) are inevitably subject to a fairly wide margin of error, in the face of which it makes little sense to focus on subtleties aimed at discriminating accurately between investments that might have an expected yield of 10',4 per cent and those that would yield only 10 per cent per annum. As the first order of business we want to be able to distinguish the 10 per cent investments from those yielding 5 or 15 per cent, while looking forward hopefully to the day when we have so well solved the many problems of project evaluation that we can seriously face up to trying to distinguish 10 per cent yields from those of 9 or 11 per cent.
Moreover, in what follows, I shall try to bear in mind the virtual necessity of decentralized decision-making. Rules and procedures can be imposed which assure a certain rough harmony among the decisions taken in such vastly different areas as roads, irrigation projects, and educational investments, but one cannot realistically expect all investment decisions to be funneled through a single office or authority that exercises more than a general supervisory power. Most of the real work connected with project appraisal must, I believe, be done "close to the ground"; this fact alone limits the range of workable procedures to those
in which a substantial amount of power can in fact be delegated to decentralized bodies.
Within this general framework the focus of the paper is mainly on
the fact that the relevant prices may change through time. The first section discusses the problem of real wage changes. The second section dis-
cusses the problem of future changes in the discount rate; the third section, the choice of a time path for the discount rate; the fourth section, the choice of the level of the discount rate. The fifth section discusses shadow prices for labor and capital, again coming to rest on the problem of selecting time paths. Finally, the sixth section discusses time paths of other prices and of demand functions.
The Problem of Real Wage Changes
Most discussions of project evaluation note that expected price changes should be taken into account, but little more than lip service is paid to this idea when working procedures are outlined. Insofar as the relative prices of commodities are concerned, this neglect of expected changes is understandable. "On the average," our best guess is likely to be that relative prices will remain as they are; cases where we have good reason to believe they will change can probably be regarded as somewhat exceptional, and project analysts can perhaps be presumed to deal with these exceptional cases as they arise.
When, however, we come to the price of labor, the story is very dif-
ferent. A rise in the real wage rate is one of the essential features of economic development, and this means a rise in the price of labor relative to the general price level of the economy. If we normalize on the general price level, we can therefore say that the typical investment is likely to be one in which the price of the product to be produced is ex-
pected to remain constant while the wages paid to labor rise. If a private entrepreneur leaves out of account the expected rise in wages (relative to the general price level), he does so at his peril, for this fact can readily turn a potentially profitable project into an...