The Biden Administration has proposed revising the government-wide guidance on how agencies estimate and weigh the costs and benefits of their regulations. These proposed revisions are a further step in a decades-long, bipartisan process of improving the methodology of cost-benefit analysis. They are also a much-needed reminder that government can still work at a time when it can be easier to focus on partisan rancor and dysfunction than the possibilities of thoughtful public administration.
Counting the costs and benefits of regulation might, in principle, sound straightforward. But complex rulemakings have a host of consequences, and evaluating the positive and negative effects of regulation raises difficult empirical, economic, and ethical questions. In the four decades since President Ronald Reagan placed cost-benefit analysis at the heart of the U.S. regulatory system, agencies have struggled with these questions, gradually accumulating a store of practices and methods that help them tackle this difficult but essential task.
Released during the George W. Bush Administration, Circular A-4 aggregated best practices from twenty years of cost-benefit analysis practice in the federal government. It built on President Bill Clinton’s Executive Order 12,866 on regulatory review and offers agencies guidance on how to structure an inquiry into the social consequences of regulatory decisions. Drawing on the intellectual tradition of welfare economics, Circular A-4 provides valuable advice that highlights several fundamental methodological essentials for a sound cost-benefit analysis. These include establishing an appropriate baseline for comparison, conducting marginal analysis, comprehensively examining regulatory effects, avoiding double counting, and recognizing the potential information value of delay.
This document has played an influential role in structuring how agencies conduct cost-benefit analysis for the past twenty years. But it is not, nor should it be, the last word in rational regulatory decisionmaking. Developments in welfare economics, changes in our economy and society, insights learned from regulatory practice, and a new generation of policy challenges all require that methods be updated to reflect the current world. Were cost-benefit analysis methodology to freeze in time, its usefulness for informing and improving regulatory decisions would gradually fade away.
The Biden Administration’s proposed revisions to Circular A-4 keep much of the previous structure in place, but with updates and improvements. Several of the most important of these focus on accounting for the distribution of regulatory costs and benefits.
It is the rare government action that makes every single person better off. More typically, some people are burdened in order to reduce harm to others. Air quality regulation might avoid serious health risks but require regulated firms to install expensive pollution control technologies—costs that may then be passed to consumers. Workplace safety requirements might reduce injuries but increase the cost of hiring new workers—leading to fewer jobs and depressed wages in the regulated sector. The goal of cost-benefit analysis is to determine whether the value created by a regulation for some is sufficient to justify the burdens that it places on others.
Cost-benefit analysis practice has long attended to the temporal distribution of costs and benefits through discounting. When costs or benefits occur in the future, they are typically discounted to the present value. Circular A-4 provides three rationales for this practice: first, investments provide positive returns, implying that “current consumption is more expensive than future consumption”; second, “people generally prefer present to future consumption”; and third, future generations will have more wealth, implying that “consumption will be less valuable in the future than it would be today” due to the diminishing marginal utility of consumption.
The general logic of discounting is still widely accepted in welfare economics, but the specific discounting practices recommended in Circular A-4 are not. The original guidance document recommended that agencies use constant discount rates of 3% and 7% to reflect “the rate that the average saver uses to discount future consumption” and “the average rate of return to capital,” respectively, depending on whether the costs of regulation fall on consumers or investors. These numbers were derived empirically, based on the rate of return on government debt and investment returns as of 2003. They are now badly out of date because the cost of government borrowing has continued to fall over the past twenty years, as have returns on private capital, indicating that the empirical foundation for the 3% and 7% rates have eroded away.
Circular A-4 also recommends a fixed discount rate, meaning that the same discount rate is applied to a future benefit or cost whether it occurs five, ten, fifty, or one hundred years in the future. Over short time horizons, a constant discount rate is a fine approximation, but it has serious problems over long time horizons when discount rates are uncertain. This is due to a simple mathematical fact: the expected value of a discounted amount is greater than the value of the amount using the expected discount rate. What does this mean? If we are uncertain whether the discount rate is 1% or 3%, one might think that the proper rate to use is 2%, but that is wrong. A $1,000 value in 50 years discounted to present day using a 2% rate is $372; if we take the average of the discounted amounts using 1% and 3%, the value is $418. The longer the time horizon, the stronger this effect: over 150 years, the same calculations give us $51 (for the 2% rate) and $118 (for the average using 1% and 3%). Using a fixed rate has the effect of seriously undervaluing consequences that occur over the long term and cannot be justified when there is uncertainty about the correct discount rate.
The proposed revisions to Circular A-4 account for new empirical realities and provide agencies with richer and more nuanced options for discounting, including the use of declining marginal rates for decisions with long-term consequences. These updates reflect broad professional consensus and are long overdue.
A second way that costs and benefits can be unevenly distributed is along geographical lines. Circular A-4 states that “[y]our analysis should focus on benefits and costs that accrue to citizens and residents of the United States.” For many types of regulations, costs and benefits will, as a descriptive matter, accrue to U.S. citizens and residents, so this emphasis will often simply reflect the geographic scope of potential consequences. But in an increasingly connected world, decisions made by regulators in one country can have global effects. The most obvious example is climate change. Greenhouse gases are a global pollutant, so emissions in the United States have as much effect on Paris, France as they do on Paris, Idaho. But climate change is only one area where global effects can be important; others include pandemic responses, anti-terrorism policy, and regulation of cyber risks.
The proposed revisions are an improvement, but they should go further. Importantly, the revisions identify several contexts where “it may be particularly appropriate to include effects experienced by noncitizens residing abroad in your primary analysis,” including where “regulating an externality on the basis of its global effects supports a cooperative international approach to the regulation of the externality by potentially inducing other countries to follow suit or maintain existing efforts.” Agencies are also urged to conduct supplementary analyses, either with global effects (if the primary analysis focuses on U.S. citizens and residents) or domestic-only effects (if the primary analysis is global in scope).
The expanded discussion of global costs and benefits in the proposed revisions appropriately updates guidance to agencies to reflect contemporary reality. Regulatory decisionmaking, though often fully domestic, increasingly has global implications. But the proposed guidance still retains an orientation toward effects “that are experienced by citizens and residents of the United States,” departing from that orientation only under special circumstances. Even these special circumstances have a domestic-effects character to them—for example, focusing on global effects when it might influence the behavior of international actors in ways that generate benefits in the United States.
A better approach would be to urge agencies to be comprehensive in their analysis of aggregate consequences but then provide guidance on how geography or nationality might factor into decisionmaking as a distributional consideration. Regulations sometimes have global effects that are as real as any that occur within the United States. Failing to acknowledge and analyze these effects is simply misleading. Rational decisionmaking requires that agencies consider the consequences of their actions, no matter where those consequences fall. Agencies need not select the regulatory action that maximizes net benefits—there may be distributional, legal, moral, and political factors that agencies may appropriately consider. But there is no justification for failing to understand and appreciate the full range of consequences, whether they are felt in the United States, by a U.S. citizen living abroad, or by that citizen’s neighbor who is a national of another country.
With an aggregate analysis of costs and benefit in hand, agencies can decide whether and how the distribution of regulatory consequences should be considered. Fortunately, the proposed revisions include a much-expanded discussion of distributional analysis. Circular A-4 urges agencies to “provide a separate description of distributional effects,” but it provides little guidance on how to do so or how distributional factors should weigh in the analysis. The proposed revisions have much more to say, offering recommendations on when to perform distributional analysis, how to identify the relevant groups, and how to produce the analysis. One particularly important change is discussing the potential for agencies to use distributional weights to account for the diminishing marginal utility of consumption.
The Biden Administration should be commended for the proposed revisions. They are an extremely promising beginning for a deliberative process that will include public comment and peer review. If adopted, they would bring the guidance more in line with twenty-first century realities and current economic methods. The sections dealing with distributional matters are thoughtful and well-informed. These proposed revisions also recognize that there will always be tough questions that arise when considering how agencies should account for the distribution of regulatory costs and benefits, repeatedly highlighting the need for the reasonable exercise of discretion on the part of the agency.Cost-benefit analysis is a living practice, meant to provide useful information for the public and government decisionmakers in the real world. In a democratic system, public policy will always partially be about politics and partisanship, but it can also be about applying expertise and analysis to pressing social problems. This technocratic side of government is sometimes disparaged, but it is vital to carry out the will of the people as reflected in statutes, budget allocations, and presidential decisions. The proposed revisions are firmly in the technocratic tradition, drawing on a deep well of expertise, but they are also grounded in the democratic need for effective and well-reasoned government action.
