On Measuring the Social Opportunity Cost of Public Funds

Abstract

The literature on project evaluation abounds with competing recommendations as to what rate of interest should be used to discount to a single point in time the estimated costs and benefits of public-sector projects. Official policy in the United States, as established in the Green Book¹ and in Senate Resolutions, is to base public-sector investment decisions on interest rates prevailing or expected to prevail in the market for government bonds. An alternative, proposed by such authors as Hirshleifer, DeHaven and Milliman, Strotz, and Stockfisch, is to discount benefits and costs at the estimated marginal productivity of capital in the private sector of the economy. A third group, to which Marglin and Sen belong, asserts that market interest rates give an exaggerated picture of the rate of ‘social time preference’, and suggests that that rate be chosen which represents the consensus of the policy-makers (or of the society as a whole) concerning what the social time preference rate really is or should be. I shall defer discussion of social time preference rates, thus arbitrarily defined, to a later point, and shall instead assume that the ‘social rate of time preference’ refers to an appropriately weighted average of the different marginal rates of time preference applicable to the individuals who compose the society.