because of uncertainties at the time of the analysis. For example, if there is an unanticipated increase in energy prices after a decision, the actual cost of a decision could exceed the anticipated cost, and what was thought to be a project of positive net benefit may turn out to have a substantial negative net benefit. That uncertainty, however, is rarely detailed in a BCA or CEA, let alone in the rationale for the decisions that use the BCA or CEA (NRC, 2002a).
We discuss below the uncertainties associated with the estimates of costs—categorized broadly as compliance costs and the costs imposed across the entire economy—and benefits as well as the uncertainties related to discounting those costs and benefits.
Compliance costs are the costs incurred in complying with a proposed regulatory rule, and they include those costs incurred by parties complying with the regulations (for example, the costs to install emission-control technologies in an industrial facility). These compliance costs are borne by regulated entities, of which many, if not most, are in the private sector (EPA, 2010). Additional compliance costs include the government’s costs to monitor and enforce the rule and, more generally, activities to ensure compliance (Harrington et al., 1999). The increased costs borne by regulated entities and public agencies charged with enforcement may also be associated with increased costs borne by other private-sector entities and public agencies at different levels of government. An increase in enforcement effort will generally raise compliance rates and, subsequently, the costs borne by the regulated parties (Shimshack and Ward, 2005, 2008). Some research indicates, however, that the costs of “government administration of environmental statues and regulations” are “rarely considered in regulatory cost estimates” (Harrington et al., 1999, p. 9).
Measuring the capital costs and operating costs incurred by private parties in order to estimate compliance costs can be a difficult task for agencies with the responsibility of implementing a regulation. For example, under a cost-of-service regulation, an agency sets a price per unit of output and must gauge whether the regulated price it sets is sufficient to yield an adequate return to the regulated entity. At the same time, the agency must ensure that the price is not set so high that returns are excessive (Breyer, 1995).
Compliance costs are often estimated using engineering models, with the models often being based on expert opinions of the relationships between input use and outputs for a particular industry or application within