• About CPR
   
• Perspectives
  • Corporate Accountability
  • Cost-benefit Analysis
  • Data Quality
  • Devolution
  • Emissions Trading
  • Environmental Enforcement
  • Environmental Indicators
  • Environmental Justice
  • Estimating Regulatory Costs
  • Executive Oversight
  • International Environmental Justice
  • Judicial Review
  • Precautionary Principle
  • Regulatory Toolbox
  • Right to Know
  • Secrecy
  • Statutory Design
     
• Media Room
   
• Contact Us
   
• Mailing List

Estimating Regulatory Costs

Background

Under nearly every law enacted by Congress to protect health, safety and the environment, the cost of compliance is an important consideration in establishing regulatory requirements. If the estimates of regulatory costs that the agencies employ systematically overestimate those costs, the regulators may promulgate less stringent health, safety and environmental requirements out of an unfounded fear that protective standards will be too expensive and possibly result in the loss of jobs. If broad assessments of the costs imposed by all health, safety and environmental regulation are similarly biased upward, the debate over the desirability of protective rules and legislation will become distorted in favor of industry profits and at the expense of public health.

What People are Fighting About

Regulated industries constantly complain about the cost of complying with regulations designed to address the risks that their activities pose to the public from pollution, toxic substances, defective products, unsafe workplaces, and the like. The individual companies that will be subject to particular regulations nearly always raise the specter of debilitating costs when agencies consider how stringent the regulations should be. Industry-sponsored think tanks periodically publish eye-popping estimates of the overall costs of federal regulations with the tacit (and sometimes explicit) suggestion that society is paying far too high a price for health, safety and environmental protections. It’s not just industry that argues that case; it has become a mainstay of anti-regulation conservatives as well.

Supporters of existing regulatory programs respond that the broad cost estimates produced by think tanks with ideological axes to grind have little or no empirical basis and are, in fact, the predictable product of crude economic models designed specifically to produce inflated estimates of regulatory costs. In the context of individual rulemaking proceedings, cost projections invariably depend upon the speculative predictions of the regulated companies themselves, hardly an unbiased source of information on the matter, since they have a clear incentive to inflate prospective regulatory costs. In addition, agency cost assessments typically include costs that companies would have incurred whether or not the regulations were promulgated, thus leading to further inflation of prospective costs. All too often, global assessments of the overall costs that regulatory programs have imposed on regulated industries are based not on retrospective evaluations of the amounts that companies actually expended, but upon unvalidated estimates made at the time that the regulations were being promulgated.

Moreover, when estimating prospectively the costs of individual rules, agencies tend to adopt conservative assumptions that lead to overestimates, because they know that regulatees are likely to challenge agency cost estimates in court. The agencies also tend not to offer a range of cost estimates, but instead offer only a single, inflated number. (For regulatory benefits, by contrast, agencies typically offer a range of estimates.) The agencies also seldom discuss the uncertainty surrounding their estimates of costs. One reason prospective cost estimates are uncertain is that regulation frequently inspires technological and operational innovations that lead to lower compliance costs.

Perhaps the most disturbing aspect of the public debate over the cost of federal regulation is the very modest degree to which it is fueled by empirical data on actual expenditures incurred in complying with past regulations. For example, a recent estimate of the costs imposed by workplace health and safety regulations prepared by the Mercatus Center at George Mason University simply added up the cost projections that the relevant agency, the Occupational Safety and Health Administration (OSHA), had made when it was promulgating 31 major regulations, without undertaking any effort to determine whether the regulations had actually cost as much as the agency had predicted, and multiplied that number by 5.55. The “multiplier” was based upon a determination that companies spend $5.55 on non-major OSHA regulations and OSHA fines for every dollar they spend complying with major regulations, an estimate that was in turn based upon an unpublished and otherwise unavailable 1974 report prepared by the National Association of Manufacturers.

There are, in sum, strong reasons for concluding that prospective cost studies are biased upward:.

• Agency analysts rely heavily upon the regulated companies for the empirical data and assumptions underlying those estimates.
  
• Because they know that regulatees are likely to challenge cost estimates in court, agencies tend to adopt conservative assumptions that lead to overestimates, and they do not offer a range of cost estimates or an assessment of the uncertainty surrounding their estimates of costs.
  
• Agency cost assessments typically include costs that companies would have incurred whether or not the regulations were promulgated.
  
• Most cost assessments fail to anticipate the dynamic and innovative ways in which regulatees react to regulations to minimize regulatory costs in the real world.

The problem with inflated cost estimates is that they distort the regulatory decision-making process. Protective regulatory options may fall by the wayside because they appear to be too expensive, or “economically infeasible” to the regulatory agencies. This can, in turn, lead the agencies to promulgate less stringent protections than Congress intended or give them an excuse to decline to promulgate any protections at all. For example, grossly inflated estimates of the cost of installing airbags in automobiles caused the National Traffic Safety Administration to delay that easily implemented requirement for more than a decade. Similarly, inflated estimates of the overall impact of health, safety and environmental regulations on the economy may distort national debates over the desirability of governmental protections.

CPR's Perspective

Both global and individual cost assessments fail to account for the realities of company compliance with regulatory requirements. When retrospective cost assessments are undertaken, they nearly always conclude that the regulated companies expended far fewer resources in complying with health, safety and environmental regulations than the agencies predicted at the time they promulgated the regulations. Agencies should devote more resources to retrospective analyses aimed at correcting bad estimates and preventing inflated cost estimates in the future. They should also modify their day-to-day prospective cost estimates in the following ways:

• Agency-prepared cost estimates should include a “best case” analysis, in addition to the “worst case” analysis that currently dominates cost estimates, and they should include an “uncertainty analysis” that relates in an understandable way the confidence the agency has in the estimates.
  
• Agencies should include a “technology improvement” scenario in cost assessments in which companies are assumed to come up with innovative ways to comply with proposed standards.
  
• In the absence of solid empirical data on actual compliance costs, agencies should discount regulatory cost projections by a factor representing the probability that technological improvements will result in reduced costs.

In the meantime, policymakers who rely on industry- and agency-prepared cost assessments must take such assessments with several grains of salt. Courts should be less reluctant to allow agencies to “force technology” by promulgating stringent standards that may appear quite expensive when promulgated but usually prove much less costly once the engineers for the companies put pen to paper. Finally, the poor empirical basis for agency cost assessments and the reality of upward bias in those predictions should put to rest suggestions that regulatory agencies should be subject to a “regulatory budget” setting an overall cap on the costs that they may impose on regulatees during any given year.

In Sum

When regulatory agencies promulgate standards and regulations to protect health, safety and the environment, the companies that will ultimately have to comply with those regulations frequently complain that they will cost too much. Surprisingly, the source of the cost estimates that the agencies typically rely upon is the regulated industry itself. Not surprisingly, these estimates tend to be seriously inflated. Agencies should take steps to control cost estimates, and they should not be deterred from promulgating stringent protections by industries crying “wolf.”