Before issuing major environmental, health, and safety regulations, administrative agencies are required to assess their proposed rules’ costs and benefits in accordance with President Clinton’s Executive Order 12866. This same Executive Order instructs agencies to consider how such costs and benefits would be distributed across individuals with high or low incomes, the young or the old, members of different racial or ethnic groups, or residents of particular geographic areas. More recently, President Obama’s Executive Order 13563 has reaffirmed the requirement that regulatory agencies incorporate distributional analysis into their cost-benefit analyses.
Have agencies followed through on this requirement? In a recently published paper, Lisa Robinson and Richard Zeckhauser of Harvard University’s John F. Kennedy School of Government and James Hammitt of the Harvard School of Public Health find that agencies typically provide little to no information on distributional impacts in their regulatory analyses.
The authors arrive at their conclusion after examining the regulatory analyses of a dozen major environmental, health, and safety regulations reviewed by the Office of Management and Budget (OMB) during the 2010 and 2011 fiscal years. Focusing specifically on regulations that impose compliance costs on industry in order to achieve health risk reductions, the authors study the extent to which agencies incorporated the distribution of health benefits, industry compliance costs, or net benefits in their analyses.
The process of quantifying health-related benefits in regulatory analyses involves estimating the number of statistical cases averted—that is, the change in the individual risk of illness, injury, and premature mortality multiplied by the number of individuals affected. Once regulators have estimated the likely number of averted cases, they then approximate the monetary value of these risk reductions. According to the authors, the OMB’s Circular A-4, which provides implementing guidance to federal agencies on the development of regulatory analyses as required by Executive Order 12866, suggests that decision makers should be particularly concerned about the extent to which those affected are members of a disadvantaged group or are particularly vulnerable to the regulated hazards due to old age and other health impairments. Moreover, Executive Order 13045 and Executive Order 12898 further require agencies to identify and address any risks that may disproportionately affect children, minorities, and low-income populations.
Yet, only three of the dozen regulations examined by the authors—all U.S. Environmental Protection Agency rules—go beyond certifying that the regulation in question does not violate the applicable executive orders because it has no adverse effects on the health of children, minority groups, and low-income populations. When the risk distribution is discussed, the authors find, agencies focus on incidence—for example, mapping the geographic areas likely to experience air quality improvements—without considering how the value of the risk reductions might vary across affected populations.
Likewise, distributional analyses are absent in agencies’ monetary valuations of health risk reductions, say the authors. For instance, while empirical evidence suggests that individual willingness to pay varies depending on the population affected and the risk involved, the authors find that agencies do not take such variation into account in their current analyses.
In their examination of compliance costs, the authors similarly discover that only some of the regulatory analyses in question estimate how industry costs translate into price, wage, and profit changes, and none look at how these changes might be distributed across different firms or individuals. According to the authors, the analyses for all 12 rules do report the capital and operating expenses necessary to meet regulatory requirements, assessing the distribution of these costs across industries and firms, for-profit and non-profit entities, and small businesses. In fact, some analyses even use a market modeling approach that incorporates behavioral responses, enabling policymakers to calculate the aggregate share of the burden borne by producers and consumers. However, none of the analyses, say the authors, directly address the impact of the regulation in question on the wages of those employed in the targeted industries. In the same vein, the analyses provide no estimate of possible effects on firm profits.
Why have agencies largely eschewed distributional impacts in regulatory impact analyses, ask the authors? They consider three possible explanations. First, policymakers may believe that cost-benefit analysis should solely take into account economic efficiency, leaving equity goals to the tax system or social programs. Second, perhaps decision makers want to integrate distributional effects into traditional cost-benefit analysis but technical difficulties constrain them. Finally, the authors put forward a third explanation as the one they find most promising: policymakers rely on cost-benefit analysis to select the most economically efficient option so long as the regulations do not harm children and other vulnerable populations. This normative motivation—in tandem with pragmatic concerns such as the possibility that distributional analyses may invite additional political resistance and hinder agencies’ ability to issue timely regulations—helps explain why distribution has received scant treatment, say the authors.
Whatever the reason, Robinson, Zeckhauser, and Hammitt conclude that agencies’ inattention to the distribution of regulatory costs and benefits is “disturbing.” They recommend that agencies conduct case studies of rules to identify potential subgroups, provide information on the types and magnitudes of impacts, and test alternative analytic approaches. This information, argue the authors, can help regulators engage in a more meaningful discussion with interest groups, provide better information to the public, and ultimately ensure that the regulatory decision-making process is strong and sound.