REGULATING DRINKING WATER QUALITY INCREMENTALLY: BETTER THAN BENEFIT-COST ANALYSIS?

The use of benefit-cost analysis has expanded over the last twenty five years and now plays a key role in the formulation of new regulations, even regulations for the protection of human health. Regulatory decision making is, of necessity, based on pre-regulatory paper studies of costs and benefits. There has been relatively little post-regulatory validation of these benefit-cost estimates, but what little information is available indicates that they are most often inaccurate. Roughly half of such estimates have been found to over-estimate costs by failing to anticipate technological change driven by the regulation. Roughly a quarter of studies underestimated costs by failing to anticipate unintended consequences of the regulation. Estimates of benefits are often impossible to verify even after the regulation is imposed as they involve small changes in cancer rates that are not distinguishable from random fluctuations in epidemiological data.Basing decisions on a tool that is so often in error would appear to be ill advised. Unfortunately the alternatives are not very attractive. Basing decisions on absolute standards of health protection does not allow for the prioritization of limited resources. In practice it can result either in overly strict regulation, in the hopes of completely eliminating risk, or in inaction, as regulators wait for absolute scientific certainty before imposing stringent and costly regulations.An alternate regulatory procedure based on incremental regulatory changes is described and statutory changes proposed to give EPA discretionary authority to adopt such a strategy for setting drinking water standards. This strategy will aim to avoid large, irreversible investments with uncertain benefits. This strategy will shift the emphasis from the largely inaccurate pre-regulatory benefit-cost assessments towards ongoing studies of regulatory impacts that will provide feedback to governmental agencies.