Cost-benefit analysis: an uncertain guide to public policy.

Abstract

Recognition of the uncertainty inherent in social plans has generated a variety of formal approaches to coping with uncertainty. Perhaps the most heavily touted and widely adopted are variants of cost-benefit analysis. Such analyses ask whether the expected benefits from a proposed activity outweigh its expected costs. Although based on an appealing premise and supported by a sophisticated methodology, these procedures have a number of characteristic limits. One set of limits is imposed by the unavailability of necessary inputs to the analysis. Since neither the values nor the likelihood of many costs and benefits can be assessed by any formal computations, they must be derived by human judgment. Research has shown, however, that probability judgments are often quite unreliable and prone to systematic biases, while judgments of value are highly labile, changing with subtle (and formally irrelevant) shifts in the elicitation procedure. The second set of limits arises when one comes to assess the quality of analyses. There have been few systematic evaluations of formal analyses or attempts to develop a methodology for assessment. Again, one is forced to rely upon judgments which research has shown to be untrustworthy. A third set of limits is the inability of the procedures to address critical issues in the management process they are designed to abet. These issues include the acceptability of the political philosophy underlying such procedures, the feasibility of implementing their recommendations, and the generation of managerial options. The contribution of cost-benefit analysis may be enhanced by reducing whichever of these problems are tractable, acknowledging those that are not, and clarifying the responsible role of cost-benefit analysis in the management process.