ABSTRACT
The power of compound interest is enormous over long periods. Discounting over centuries at today's return on capital implicitly makes a commitment that is not credible: that society will keep reinvesting at this rate to compensate distant future generations for damages imposed. A much better alternative is social benefit-cost analysis applying a social rate of time preference (SRTP) to a stream of consumption equivalents that apply a greater than unity shadow price to capital investment effects. This approach tends to ensure that issues with time scales of centuries are not decided with severe bias against future generations. The SRTP equals the rate of pure time preference (impatience), which should be zero for social benefit-cost analysis, plus the elasticity of marginal utility multiplied by the growth rate of per capita income. This yields a discount rate of about 1.5%, except for poor economies where it may be higher. For projects with typical life spans of fifteen years, the penalty from shadow-pricing capital (for example, at 1.5 times as valuable as consumption) about neutralizes the generosity of a lower discount rate (compared with conventional rates of, say, 6% to 8%), so the method does not bias judgments toward low-return projects. For centuries-scale problems, the SRTP approach avoids the outright dismissal of future generations that occurs when conventional discount rates are used. Although I prefer systematic application of the SRTP approach, this paper proposes a compromise approach to resolve the split in the profession on this issue. The more conventional rates (without capital shadow pricing) could be applied for the first thirty years (the present generation), and the SRTP approach would then apply for all subsequent time periods.
