Why and How Independent Agencies Should Conduct Regulatory Impact Analysis

Independent regulatory agencies face increasing pressure to conduct high-quality economic analysis of regulations, similar to the regulatory impact analysis conducted by executive branch agencies. Such analysis could be required by evolving judicial doctrines, regulatory reform statutes, or executive order. This article explains how regulatory impact analysis can contribute to smarter regulation, documents the current low quality of such analysis at many independent regulatory agencies, and offers a blueprint that independent agencies can use to build their capacity to conduct objective, high-quality analysis.


I. Introduction
Regulations should solve real problems at a reasonable cost. 1 To know whether a proposed regulation is indeed likely to do this, a regulator needs to know whether a real problem exists, whether a proposed regulation addresses the cause of the problem, and how much of other good things society must forgo to enjoy the expected benefits of the regulation. Regulatory impact analysis is the tool that provides decision makers with this information. A complete regulatory impact analysis assesses the nature and significance of the problem the regulation seeks to solve, identifies alternatives solutions, and estimates the benefits and costs of these alternatives. 2 Since 1981, a series of executive orders has required executive branch agencies to conduct regulatory impact analysis for significant regulations. 3 Independent agencies have not been subject to these executive orders. Independent agencies may face growing pressure, however, to conduct such analysis in the future. That pressure could come from all three branches of government-the courts, Congress, and perhaps even the president. 1 I intentionally avoid the more restrictive normative claim that a regulation's benefits should exceed its costs, and I leave the reader to decide what counts as a "problem" and what amount of cost is "reasonable." However these are defined, a thorough regulatory impact analysis provides useful information. 2 These components can all be found in Exec. Order No. 12866, 58 Fed. Reg. 51,735 (Oct. 4, 1993), which outlines the primary requirements for regulatory impact analysis currently in force for executive branch agencies. For all regulations, agencies are expected to assess the nature and significance of the problem the regulation seeks to solve (Exec. Order No. 12866, § § 1(b)(1) and 6(a)(3)(B)(i)). An assessment of benefits and costs must accompany all "significant" regulations-generally, regulations that have an effect on the economy exceeding $100 million annually; have other material adverse effects; conflict with other agencies' actions; affect federal spending or loan programs materially; or raise novel legal or policy issues (Exec. Order No. 12866, § 6(a)(3)(B)(ii)). Regulations with economic effects exceeding $100 million annually or certain other material adverse effects listed in the executive order must be accompanied by an analysis of the benefits and costs of the regulation and alternatives, with benefits and costs quantified where feasible (Exec. Order No. 12866, § 6(a)(3)(C)). 3 President Reagan's Executive Order 12291 first used the term "regulatory impact analysis." See Exec. Order No. 12291, 46 Fed. Reg. 13,193 (Feb. 19, 1981). would require virtually all regulatory agencies, including the independent ones, to consider the nature and significance of the problem they seek to solve, alternative solutions, and the benefits and costs of alternatives. 9 Agencies would also be required to rely on the best available scientific, technical, and economic information-a provision that would effectively require reasonably thorough regulatory impact analysis. 10 To date, no president has attempted to compel independent agencies to conduct regulatory impact analysis. Key figures in both Republican and Democratic administrationssuch as C. Boyden Gray, coauthor of President Reagan's Executive Order 12291, and Sally Katzen, a principal author of President Clinton's Executive Order 12866-contend that a president has the legal authority to do so, but administrations have sought to avoid a confrontation with Congress over the issue. 11 Still, an executive order requiring independent agencies to conduct regulatory impact analysis remains a definite possibility. Independent agencies may thus be required to conduct regulatory impact analysis, but many of them have yet to develop the capacity to do it. Assessments by agency inspectors general, the US Government Accountability Office (GAO), and independent scholars find that many independent agencies' regulatory analyses lack basic information such as monetized estimates of benefits, monetized estimates of costs (other than paperwork costs), or discussion of benefits and costs of alternatives to the regulation. 12 Some agencies have recognized that producing quality analysis requires significant changes in organizational structure, practices, and culture. The Securities and Exchange Commission (SEC), for example, lost several high-profile court cases because of insufficient economic analysis after courts interpreted language in the SEC's authorizing statute to require benefit-cost analysis of regulations. 13 In response, the SEC in 2012 launched an initiative to improve the quality of economic analysis and the influence of economists in regulatory decisions. The chief economist became a direct report to the chairman, the general counsel and chief economist issued joint guidance on economic analysis based on the principles executive branch agencies must follow, and the commission more than doubled the number of PhD financial economists on staff. 14 The quality of SEC economic analysis has improved measurably since then. 15 The Federal Communications Commission (FCC) provides another example. 12 See Section III infra. 13 The SEC must consider the effects of proposed regulations on competition, efficiency, and capital formation when determining whether the regulation is in the public interest. This language appears in Section 2(b) of the Securities Act of 1933, 15 U.S.C. § 77b; Section 3(f) of the Securities Exchange Act of 1934, 15 U.S.C. § 78c(f); and Section 2(c) of the Investment Company Act of 1940, 15 U.S.C. § § 80a-2(c). This requirement was added to these statutes by the National Securities Markets Improvement Act of 1996. An agency seeking to improve its economic analysis of regulations faces significant challenges in constructing the capacity to do so and ensuring that economists and other analysts have the appropriate incentives and opportunity to conduct objective analysis. This article addresses those challenges. Section II clarifies the role that regulatory impact analysis can play in promoting smart regulation. Section III documents the low quality of such analysis at many independent agencies and presents some brief examples that demonstrate how low-quality analysis can lead to poor decisions. Section IV explains key implementation steps an independent agency can take to improve its analysis and ensure that the analysis is considered when making regulatory decisions. Section V concludes with some suggestions on how an agency can make a credible commitment to produce and use high-quality regulatory impact analysis in the future.

II. The Purpose of Regulatory Impact Analysis: Better Regulation
Citizens expect federal regulation to accomplish many important things, such as protecting us from financial fraudsters, preventing workplace injuries, preserving clean air, and deterring terrorist attacks. Regulation also requires sacrifices; there is no free lunch. Depending on the 16  regulation, consumers may pay more, workers may receive less, our retirement savings may grow more slowly because of reduced corporate profits, and we may have less privacy or less personal freedom. Given the important values at stake, regulatory agencies should craft regulations with knowledge of their likely results. A decision maker's failure or refusal to acquire this knowledge before making a decision is a willful choice to act based on ignorance.
Reasonable people can disagree about the tradeoffs they are willing to make to get the good things regulation provides. 17 However, reasonable people surely can agree that regulators should not adopt a regulation unless they are reasonably certain that it will solve a real problem at some reasonable cost.
Regulatory impact analysis is the tool that helps agencies identify whether alternative regulatory proposals are likely to solve a real problem, and at what cost. The principal elements of regulatory analysis outlined in executive orders and in OMB guidance reflect standard economic principles of policy analysis and government performance management. 18 A thorough regulatory impact analysis should provide four types of critical information: 1) Assess the nature and significance of the problem the agency is trying to solve, so the agency knows whether there is a problem that could be solved through regulation and, if so, whether the agency can tailor a solution that will effectively solve the problem; 2) Identify a wide variety of alternative solutions; 20 3) Define the benefits the agency seeks to achieve in ultimate outcomes that affect citizens' quality of life, and assess each alternative's ability to achieve those outcomes; 21 4) Identify the good things that regulated entities, consumers, and other stakeholders must sacrifice to achieve the desired outcomes under each alternative. 22 In economics jargon, these sacrifices are known as "costs," but just like benefits, costs may involve far more than monetary expenditures. 23 Without this information, agencies are more likely to base regulatory decisions on hopes, intentions, and wishful thinking than reality.
The executive branch has had almost four decades of experience with regulatory impact analysis. 24 Numerous studies document instances in which regulatory analysis helped improve regulatory decisions by providing additional options regulators could consider or by unearthing new information about benefits or costs of particular modifications to the regulation. 25 For example, Scott Farrow, who studied a 2004 Environmental Protection Agency (EPA) regulation requiring power plants to design cooling water intake structures that minimize harm to marine organisms, concluded that the "EPA clearly chose an approach that imposed a significantly lighter burden on society. . . . The record provides substantial evidence that the agency considered a lower-cost alternative to meeting a standard with the potential to save approximately $3 billion in annualized dollars or approximately $40 billion in present value." 26 The primary documented effect of regulatory impact analysis appears to be on the margins of regulations, identifying opportunities to increase benefits or achieve the same outcomes at lower cost. At the conclusion of a generally pessimistic assessment of the impact of economic analysis, Robert Hahn and Paul Tetlock acknowledge that when regulations create billions of dollars' worth of benefit or costs, even marginal changes can be significant for society. 27 They also note that the most important contribution of regulatory impact analysis may be its deterrent value in preventing regulators from advancing some economically unsound proposals. 28 Although it has led to improvements in regulation, regulatory impact analysis is no panacea. GAO studies and scholarly research reveal that in many cases, regulatory impact analyses are not sufficiently complete to serve as a guide to agency decisions. 29 The quality of analysis varies widely, and even the most elaborate analyses still have problems. 30  Despite these shortcomings, regulatory impact analysis can generate significant improvements when agencies perform a thorough analysis and consider the results carefully when making decisions. Unfortunately, most independent agencies lag far behind executive branch agencies in the quality and use of regulatory impact analysis, as Section III documents.

III. The Need for Thorough Regulatory Impact Analysis at Independent Agencies
We do not know whether many of the regulations adopted by independent agencies solve real problems at a reasonable cost, because independent agencies often neglect to conduct thorough regulatory impact analysis of alternatives when developing new regulations.

A. Regulatory Ferment
For the past two decades, regulation by independent agencies has risen steadily. Figure 1 shows the increase in total number of regulatory restrictions from independent agencies since 1980. 31   A "regulatory restriction" is a binding requirement in a regulation that contains the words "shall," "must," "may not," "prohibited," or "required." 33 Independent agencies accounted for 140,915 regulatory restrictions in 2017-about 13 percent of the US government's total. 34 Among independent agencies, the FCC had the largest number of restrictions (28,529), followed by the Nuclear Regulatory Commission (16,603) and the SEC (15,124). The consequences of regulatory accumulation are significant; a recent study estimated that the additional federal regulatory restrictions adopted between 1980 and 2012 could have slowed GDP growth by as much as 1 percent annually. 35

B. Poor Impact Analysis
Unfortunately, some of the ingredients in this ever-expanding pie of regulations leave a lot to be desired. A study prepared for the Administrative Conference of the United States assesses economic analyses of regulations by independent regulatory agencies. 36 It covers evaluations by the GAO, agency inspectors general, and outside researchers. The author also performs some of his own evaluations of agency economic analysis. Key findings of this report include the following: • Independent agencies often perform some type of analysis that considers benefits and costs qualitatively. 37 • Some agencies fail to analyze benefits or costs of parts of the regulation that are required by law. As a result, their analysis does not provide a complete assessment of the benefits and costs of the entire regulation. 38 • Quantification of benefits is uncommon. 39 • Quantification of costs is more common, but it is often confined to paperwork costs. 40 • Costs to agencies are often ignored. 41 • Benefits and costs of alternatives are less likely to be considered or quantified. 42 These findings are consistent with those in other studies by independent scholars. Art Some quantitative estimates of benefits were provided for about 60 percent of executive branch regulations, and quantitative estimates of costs were provided for more than 75 percent. 43 Even some independent financial regulatory agencies that are required by law to consider benefits and costs do not prepare very thorough benefit-cost analyses. 44  before the 2012 economic analysis guidance found that the SEC's assessments of the problem, baseline, alternatives, benefits, and costs were far less complete than those conducted by executive branch agencies. 45 A related but distinct problem is "ready-fire-aim" rulemaking, in which the regulatory agency makes key decisions first and then expects analysts to produce a document that supports those decisions. 46 Before SEC benefit-cost analysis became subject to judicial review, for example, SEC releases typically included a benefit-cost analysis section that merely repeated arguments in favor of the regulation that had already been made elsewhere in the document. 47 A former SEC attorney noted, "Historically, the agency's lawyers have been primarily responsible for drafting these analyses with varying degrees of assistance from the agency's economists." 48 Even when economists are responsible for the economic analysis, they are less likely to conduct high-quality analysis (particularly of alternatives) if they know that the major decisions have already been made.

C. Regulation without Analysis: Cautionary Tales
Consider a few examples of independent agency regulatory proposals or decisions made in the absence of some of the information that a thorough regulatory impact analysis would have provided. 45

Problem analysis: Securities and Exchange Commission's net worth standard for accredited
investors. Companies that issue securities can avoid costly SEC registration requirements if they restrict the sale of those securities to "accredited investors," who are believed to have sufficient sophistication and financial wherewithal that they do not need the protections provided by SEC registration. One way an investor meets the accredited investor test is by demonstrating that he or she has a net worth of $1 million or more. In 2011, the SEC adopted a regulation that excludes the value of an investor's primary residence when determining whether the individual meets the $1 million net worth requirement. The change was required by the Dodd-Frank Wall Street Reform and Consumer Protection Act. The Act also gave the SEC authority to conduct an analysis of the statutory definition of "accredited investor" and modify it as the SEC "may deem appropriate for the protection of investors, in the public interest, and in light of the economy." 49 The SEC explicitly declined to exercise this authority to adjust the definition. 50 If the SEC had considered modifying the definition, it could have evaluated whether the regulation solves a real problem. The SEC could have examined whether individuals whose home values had recently put them above the $1 million threshold actually invested in unregistered securities and suffered any harm from doing so. Perhaps this was a significant problem, or perhaps these individuals made minimal investments in unregistered securities, or perhaps issuers of unregistered securities declined to market them to these potential investors. 51 The answers to these questions could have affected whether or how the SEC opted to change the definition, but the SEC declined to consider this issue. 49  If there is a significant investor protection problem, the adequacy of the net worth requirement to solve the problem is not obvious. The SEC could have considered whether a net worth test would sufficiently protect investors from making bad investment decisions or whether a financial sophistication test or diversification requirement could achieve that objective more effectively. Existing regulations already require broker-dealers to make only "suitable" investment recommendations to their customers. The suitability requirement, perhaps with a heightened level of care when the bulk of the investor's net worth consists of home equity, may have been sufficient to address the problem. 52 The SEC conducted no such analyses, so it is not clear if the regulation solves an actual problem or does so in the most effective way.
Such deficiencies in SEC analysis may become less common as a result of court decisions that remanded several important regulations because of insufficient economic analysis.
Unlike most independent agencies, the SEC's authorizing statute contains language that courts have interpreted to require benefit-cost analysis. 53 23, 2013) (explaining that reading the requirement to consider competition, capital formation, and efficiency to require "that the SEC conduct some sort of broader, wide-ranging benefit analysis simply reads too much into this statutory language"). 54 SEC Division of Risk, Strategy, and Financial Innovation and the Office of General Counsel, supra note 14. 55 Ellig, supra note 15.

Alternatives: Surface Transportation Board's competitive switching proposal for freight
railroads. The Surface Transportation Board (STB) inherited the residual economic regulatory responsibilities of the Interstate Commerce Commission when the latter was abolished in 1996.
In July 2016, the STB proposed new regulations defining when a shipper can require a railroad serving its facilities to switch cars carrying the shipper's freight to a competing railroad. 56 Under current policy, regulators require competitive switching only if the shipper can show that switching is necessary to prevent or remedy some anticompetitive abuse committed by the railroad serving its facilities.
Instead of demonstrating with evidence that anticompetitive abuse is widespread, the STB proposal simply claimed that proving anticompetitive abuse is too difficult. The sole evidence cited in support of this claim is that very few competitive switching cases have been brought before regulators since the current policy was adopted in 1985, and shippers have never present to demonstrate anticompetitive abuse in its particular situation. Another solution was proposed in 2015 by a Transportation Research Board committee on which I served. We suggested that the STB should develop a screening model that uses rate data to identify whether a shipper appears to be paying unusually high rates, and then allow a shipper paying unusually high rates to take its case to an arbitrator. The shipper could ask for competitive switching as a remedy. 57 A thorough regulatory impact analysis would have evaluated the pros and cons of these reasonable alternatives.

Benefits: High-powered magnets as desktop toys. A 2016 court decision remanding the
Consumer Product Safety Commission's (CPSC's) safety standard for magnet sets illustrates some significant flaws in the CPSC's analysis of prospective benefits. 58 Around 2009, several companies began marketing sets of small, high-powered rare earth magnets as desktop toys and stress relievers for adults. In response to reports of injuries to children who ingested magnets, the CPSC in 2011 sent notices of noncompliance to companies that appeared to be marketing or labeling these magnets to appeal to children younger than age 14, and it warned other companies that they should not market or label the magnets to appeal to this age group. In 2012, the CPSC negotiated agreements with 10 of the 13 distributors to cease importation. In 2014, the CPSC adopted a final rule requiring all magnet sets to meet the strength and size standards that previously applied only to magnet sets marketed as children's toys. 59  importer, which required its retailers to restrict sales of the magnets to customers 18 years of age or older, sued the CPSC.
The court found two problems that inflated the CPSC's estimate of prospective benefits. In this case, the errors were caught because the CPSC is required to conduct benefit-cost analysis for product safety standards and courts can review that analysis as part of the record. As CPSC Commissioner Joseph Mohorovic noted, "Although having a rule thrown out is not pleasant for the agency, if we take to heart this reminder of the importance of the . . . analysis, our future rules will be better and sounder for the effort." 62 The CPSC could have prevented these mistakes, and perhaps could have avoided or won the lawsuit, if it had simply followed OMB guidance for preparing regulatory impact analysis.
Additionally, if CPSC regulations were subject to review by the Office of Information and 60 Zen Magnets, LLC, supra note 58, at 11-16. 61 Id. at 16-19. 62  When benefit and cost estimates are uncertain . . . you should report benefit and cost estimates (including benefits of risk reductions) that reflect the full probability distribution of potential consequences. Where possible, present probability distributions of benefits and costs and include the upper and lower bound estimates as complements to central tendency and other estimates.
If fundamental scientific disagreement or lack of knowledge prevents construction of a scientifically defensible probability distribution, you should describe benefits or costs under plausible scenarios and characterize the evidence and assumptions underlying each alternative scenario. 65 The CPSC estimated that the new magnet standard would create $28.6 million in benefits annually by preventing injuries, at a cost of at least $6 million annually. 66

Costs: Federal Communications Commission's digital television receiver mandate. In 2002, as
part of the transition from analog to digital television broadcasts, the FCC phased in a requirement that new television sets must have the capability to receive broadcast digital TV signals. The FCC cited some cost figures submitted by various interested parties, but it did not perform its own independent cost analysis. A consultant's study submitted by broadcasters estimated that a digital tuner would increase the cost of a TV set by $16 in 2006. The Consumer Electronics Association claimed that a tuner would cost $200. The FCC appeared to place most credibility in estimates from two individual manufacturers that ranged between $50 and $75. 67 Without further analysis or elaboration, the commission asserted simply that "the potential price increases under our phase in plan are within an acceptable range." 68 There was no analysis of benefits or other results attributable to this mandate that could be compared with costs to determine whether the costs were acceptable. This omission was especially glaring because, as one commissioner pointed out, about 85 percent of consumers at the time received television signals from cable or satellite companies. These consumers were not receiving overthe-air broadcast signals, but the regulation required them to pay for an over-the-air digital tuner that they would not need. 69 The D.C. Circuit Court of Appeals ruled that the FCC acted within its authority when it made this decision, because "such a shifting of the benefits and burdens of a regulation is well within the authority of the responsible agency." 70 But just because the FCC had the legal authority to make this decision does not mean that the decision was either a fair one or 67  the lowest-cost way to achieve the regulatory objective of ensuring that broadcast households could continue to receive TV signals when analog broadcasts would be phased out.
A thorough cost analysis would have included the following features: • A projected baseline future trend for purchase of TV sets with digital tuners by consumers who actually needed them-the consumers who received only broadcast TV.
• An assessment of how various types of mandates would have affected the per-unit and total costs of producing digital tuners, as well as the rate of adoption.
• A distributional analysis showing how much of the cost would be paid by cable and satellite households who did not need digital broadcast converters.
• A comparison of the cost of mandating digital tuners to the costs of alternatives, such as the subsidies for set-top converters that Congress ultimately adopted in 2005. 71 If there had been no FCC mandate for digital tuners in new TVs, the additional cost of providing subsidized set-top boxes to consumers who bought TVs without digital tuners during the transition period could have been far lower than the cost of mandating digital tuners in all new TVs. 72 The FCC undertook no rigorous comparison of alternatives that would have answered this question before it imposed the mandate.

IV. Key Steps for Better Regulatory Impact Analysis
Five key steps are necessary to ensure that an agency conducts sound regulatory impact analysis and considers it carefully when making regulatory decisions. Some of these steps are best 71 Department of Commerce, National Telecommunications and Information Administration, Rules to Implement and Administer a Coupon Program for Digital-to-Analog Converter Boxes, 72 Fed. Reg. 12,097-12,121 (March 15, 2007). 72 As a participant in the subsidy program, the author received two $40 coupons that allowed him to acquire two simple set-top boxes for a pair of old analog TVs at no additional cost. practices that already have been implemented at some independent agencies. Others are based on lessons drawn from the experience of executive branch agencies that have been conducting this type of analysis for more than three decades.
First, the agency needs to organize and manage economists in a way that promotes highquality, objective analysis. Second, the agency should establish standards for regulatory impact analysis. Third, the analysis should be conducted before regulatory decisions are made. Fourth, the agency should clearly explain how the analysis affected regulatory decisions. Fifth, the agency should invite OIRA to review its regulations and the accompanying analysis, just as OIRA does for executive branch regulations.

A. Organize and Manage Economists to Promote High-Quality, Objective Analysis
The purpose of regulatory impact analysis is to provide decision makers with objective and reliable information about the consequences of alternative courses of action. For this reason, analysts should be organized and managed in a fashion that best protects their ability to produce high-quality, objective analysis.

Functional organization of economists.
Organization theory and the actual experience of federal agencies both suggest that the structure that best accomplishes those goals is the placement of economists in a separate office or bureau that is managed by economists. 73 This structure is sometimes referred to as "functional" organization, because the agency's professionals are organized on the basis of their functions. This organizational structure facilitates better quality control of the economists' work, makes identifying and rewarding 73 "Another way to promote objective analysis is to separate agency economists from the program offices that propose regulations." PRESIDENT'S COUNCIL ON  economic analysis to support decision-making-a function that is facilitated by its existence as a separate organization." 79 The influence of economics at the FTC is widely acknowledged to be both pervasive and difficult to measure. 80 In contrast to many regulatory agencies, a great deal of the FTC's workload-and hence a great deal of its economic analysis-focuses on enforcement cases under the antitrust and consumer protection laws rather than actual writing of regulations. An empirical study found that Bureau of Economics recommendations have a statistically significant effect on FTC decisions on merger cases, but not as large an effect as the Bureau of Competition's recommendations. 81 Jonathan Baker, who served as director of the FTC's Bureau of Economics during the Clinton administration, argues that institutionalizing the bureau's role in commission decisions has created "continuous regulatory reform" in the form of routine application of benefitcost analysis in decision-making. 82 He contends that the FTC's Bureau of Consumer Protection often reconsiders or revises its proposals if it appears they will fail a benefit-cost test. 83 The FTC's "unfairness" standard illustrates the influence of economic thinking at the commission. The FTC Act prohibits "unfair" acts and business practices. 84 The commission commenced numerous consumer protection rulemakings in the 1970s, when the commission's authority to issue rules became clear, but these rulemakings were often based on vague and wide-ranging definitions of what counted as unfair or deceptive. 85 In 1978, the Bureau of Economics established a consumer protection division, and economists became significantly involved in consumer protection rulemakings for the first time. 86 "Economists brought a different set of questions to their analysis. The core questions economists ask revolve around the costs and benefits of regulatory proposals, whether they are pursued through rules or individual cases." 87 After a series of highly controversial rulemakings created significant public backlash, 88 in December 1980 the commission adopted a policy statement to guide future unfairness enforcement actions. To be considered unfair, an action or practice must create substantial injury to consumers, must not be outweighed by any benefits to consumers, and must be an injury that consumers could not reasonably have avoided. 89 Political furor over some of the FTC's regulatory initiatives created a strong incentive for the commission to limit its own discretion in some way, but economic logic provided the solution.  86 Id. at 1062. 87 Id. at 1062-63. 88 Id. at 1064-65. 89  2. Basis for performance evaluation. A separate but related issue is the criteria for evaluation of economists' performance for purposes of pay and promotion.
Unfortunately, regulatory agencies often act as if their job is to produce regulations rather than to produce outcomes. 92 As one agency economist noted, "Success is putting out 10 regulations a year and bigger regulations are bigger successes. They don't say, 'We examined 10 [situations] and we decided that 8 did not warrant regulation. . . .'" Pay, bonuses, career advancement, plaques, and other forms of recognition go to staff members who successfully complete regulatory proceedings. 93 The purpose of a regulatory impact analysis is to provide high-quality, objective information to inform decisions. Therefore, the performance of individual analysts should be evaluated in part on the basis of whether they produce high-quality and relevant analysis. 94 Economics bureaus in government agencies also often perform an R&D function, developing new data, new methods, and original empirical findings to address critical policy questions.
Analysts should also be rewarded on the basis of the quality and usefulness of such work.

B. Establish Agency-Wide Standards for Regulatory Impact Analysis
OMB Circular A-4 provides a great deal of useful guidance on how to conduct regulatory impact analysis. 95 An agency can demonstrate its commitment to conducting high-quality analysis by issuing its own standards that incorporate the concepts in Circular A-4 and explain how to apply them to the particular types of regulations written by the agency. Agency-specific standards explicitly commit the agency to regulatory impact analysis as a matter of policy and help communicate how to carry out the analysis in practice.
1. Basic elements. At a bare minimum, an agency's standards for regulatory impact analysis should identify the four major items any good regulatory impact analysis should cover: analysis of the problem, alternatives, and estimation of the benefits and the costs of each alternative. The SEC's guidance is one example of a document that addresses these items, and it explicitly refers to OMB's much more detailed guidance in Circular A-4. 96 2. Agency-specific factors. Other agencies have gone much further than the basic elements, authoring guidance that helps explain how to conduct various aspects of the analysis for the specific types of regulations written by the agency. The Nuclear Regulatory Commission's guidance includes examples relevant to nuclear power plant safety, lists specific categories of benefits and costs that should be included, and contains a special section on regulatory analysis of "backfits" applicable to existing nuclear power plants. 97  maintains a list of "rulemaking requirements" that refer the reader to relevant executive orders and OMB guidance documents on regulatory analysis. 98 The department also periodically revises and posts on its website the default values for consumers' travel time and the value of a statistical life to be used in regulatory impact analysis. 99 3. The role of benefit-cost analysis and cost-effectiveness analysis. The purpose of benefit-cost analysis is to determine whether a government action can improve economic efficiency and to compare the effects of alternative government actions on economic efficiency. Regulation can improve economic efficiency if it remedies a "market failure." Commonly discussed forms of market failure include externalities, monopoly, public goods, and asymmetric information. 100 A market failure occurs when the private marginal benefits or costs faced by decision makers deviate from the social marginal benefits or costs. This deviation of private and social benefits or costs means that private decisions will not produce the economically efficient result. 101 Governments can also fail to produce the economically efficient result, because the private benefits and costs faced by government decision makers may deviate from social benefits and costs. 102 In cases of both market and government failure, benefit-cost analysis is necessary to determine whether a change in policy will improve economic efficiency. A benefit-cost analysis of alternatives can identify the alternative with the greatest "net benefits" (benefits minus costs). Not all statutory mandates or regulations are intended to improve economic efficiency by remedying market or government failures. Many regulatory policies are intended primarily to ensure fairness in some way that involves redistribution of wealth or income. 103 Others seek to reduce risks to some level that policymakers have decided is desirable, even if that level is below the economically efficient level.
In such cases, a cost-effectiveness analysis can inform decision makers about the lowestcost way to achieve a desired policy outcome. 104 For example, "universal service" programs that subsidize telephone service for low-income and rural subscribers are intended to reduce the number of households that lack telephones. A cost-effectiveness analysis conducted to inform the reform of universal service programs by the Public Utility Commission of Texas found that 4. Distributional analysis. A conventional benefit-cost analysis identifies the net benefits of each alternative. The people who bear the costs, however, may not always be the same as the people 103 Regulation of rates that freight railroads can charge shippers who lack other good transportation options, for example, primarily affects the division of profits between the railroad and the shipper; it reflects a congressional preference that a shipper who lacks other transportation options should not pay dramatically different rates than a similar shipper who has other transportation options. Similarly, the FCC's universal service programs that subsidize broadband and phones for low-income and rural residents do not have much of an economic efficiency rationale; they reflect a congressional desire that all citizens have affordable access to a minimal level of communications services. 104  who receive the benefits. When these groups are significantly different, a separate distributional analysis that identifies disparate impacts may be helpful to decision makers. Distributional analysis should identify who bears costs, who receives benefits, and who has net gains and net losses from the regulation. This seemingly simple type of distributional analysis is rare even in regulatory impact analyses from executive branch agencies. 106 In recent years, the effect of regulations on employment has become a contentious issue.
The question of who gains or loses jobs as the result of an individual regulation is primarily a distributional issue. 107 If employment effects are included in the regulatory impact analysis of an individual regulation, they should be addressed in the distributional analysis.

Standards of evidence.
An honest, objective analysis should meet the standard of evidence articulated in Executive Order 12866: "Each agency shall base its decisions on the best reasonably obtainable scientific, technical, economic, and other information concerning the need for, and consequences of, the intended regulation." 108 This means that analysts should not selectively choose data or studies to support predetermined conclusions. When different studies or data lead to different conclusions, the analyst should use them to identify the range of possible outcomes, identify the most likely outcome, and support this determination with evidence.
6. Full disclosure. The agency's guidance should commit it to full disclosure of all reports, analysis, and data it relied upon when developing the regulation. As recommended by the 106  American Bar Association, this practice would give affected parties a fuller opportunity to evaluate and comment on the factual basis for the regulation. 109

C. Conduct Analysis before Making Decisions
A regulatory impact analysis should inform decisions, not simply justify decisions that have already been made for other reasons. Four steps can help mitigate the tendency for "ready-fireaim" rulemaking. regulations and for enforcement decisions. 113 Over time, attorneys have become adept at using the economic framework to assess both antitrust and consumer protection issues. 114 Within an economics bureau, economists can be organized into subunits that match the agency's different regulation-writing divisions, so that individual analysts and their managers can more easily coordinate their workflow with the regulation writers and the enforcement units.
Second, create an incentive for program staff to consider the economists' advice by allowing the economics office to make its own, independent recommendations to the decision makers. 115 Functional organization of economists should give them greater freedom to offer objective advice and provide greater odds that their advice will reach the ears of higher-level decision makers in the organization. 116 In addition to giving economists greater independence to reach their own conclusions, the FTC also gives the Bureau of Economics independent opportunities to make recommendations to the commission. On the majority of matters before the FTC, the bureau can offer its views both in writing and orally at commission meetings. 117 Economic and legal staffs write separate memoranda to the commission both when the commission is deciding whether to open an investigation and when the matter is ready for final decision. 118 This approach gives the economists' views greater clout in two ways. First, it ensures that advice incorporating an economic perspective is heard directly by the commissioners. Second, it creates an incentive for the attorneys working on a case or other matter to take the economists' advice seriously and reach consensus with the economists before the matter goes to the commission. 119 The inspector general's report states that less than 10 percent of the Bureau of Economics' recommendations to the commission disagree with those from the other bureaus. 120 Third, consult with stakeholders about the need for a regulation and alternative approaches before writing a regulation. A recent study found that agencies tend to produce more thorough analysis when they consult with stakeholders such as state, local, or tribal governments. 121 Fourth, publish a preliminary analysis of the problem the agency seeks to solve and the benefits and costs of alternatives before publishing a regulation. 122 There is evidence that agencies produce more thorough analysis when they first seek public comment on a prior proposal, publish a preliminary analysis, or ask the public for data before they propose a new regulation. 123

D. Explain How the Analysis Affected Decisions
Some commentators present regulatory impact analysis or benefit-cost analysis as a decisionmaking procedure that substitutes the economist's calculations for the decision maker's judgment. The decision maker need merely choose the alternative that produces the greatest difference between benefits and costs-the maximum net benefit. 124 This approach presumes that the decision maker's goal is maximization of economic welfare. For regulations that are intended solely to remedy market failures, this is the appropriate goal. If the decision maker is reasonably certain that all significant benefits and costs have been measured and converted to monetary values accurately, the analyst's calculations can greatly simplify decision-making.
However, some regulations address distributional concerns, unmonetized values, or statutory considerations that are neither benefits nor costs. For these types of regulations, decision makers surely should be aware of the benefit and cost consequences, but benefits and costs may not be the only factors driving the decisions. This is the "soft" benefit-cost test implied by the language in Executive Order 12866 specifying that agencies should regulate only when the benefits "justify" the costs. 125 It is precisely what many federal regulatory economists recommend when surveyed. 126 For this reason, the agency should explain any aspect of the analysis that affected its regulatory decisions-not just present a calculation of net benefits of alternatives. Perhaps the agency did not choose the alternative that maximized net benefits, but the assessment of the problem helped identify the most effective or cost-effective alternative. If unquantified benefits 124 Former OIRA Administrator John Graham refers to this as the "hard" benefit-cost test. See Graham, supra note 17, at 432. 125 Id. 126 Williams, supra note 32, at 6 ("No economist I interviewed thought that the results of a well-done economic analysis, specifically identifying the option that maximizes net benefits, should dictate decisions to a decision maker. But none thought decision makers should be free to ignore the results of benefit-cost analysis, particularly when, for some aspects of regulatory decisions, there were large costs and very small benefits."). and costs, or values that are neither benefits nor costs, affected regulatory decisions, the agency should explain these factors and present evidence that they are significant to citizens. 127

E. Invite Review by the Office of Information and Regulatory Affairs
Independent agencies' regulations and analysis are not currently subject to OIRA review. An independent agency could take a more controversial, but potentially productive, step by inviting OIRA to review its regulations and accompanying regulatory analysis.
Evidence shows that the requirements in the executive orders, coupled with review by OIRA, have induced agencies to engage in more thorough analysis than they would undertake otherwise. For example, "prescriptive" regulations that contain mandates or prohibitions receive more intensive OIRA review than regulations that implement budget programs; prescriptive regulations also tend to have more thorough regulatory impact analysis. 128 Agencies produce higher-quality analysis when OIRA reviews the regulation for a longer time. 129 Agencies also produce higher-quality analysis when the OIRA review is concluded under a presidentially appointed OIRA administrator rather than an acting administrator; OIRA has more clout in the administration when the administrator is a presidential appointee. 130 The concept of submitting independent agencies' regulations and analysis to OIRA is controversial, because it appears to limit their independence from the executive branch. Some argue that the primary reason for creating independent agencies in the first place is to insulate them from political pressure and capture by special interests. 131 If one believes that this type of capture is more likely when the agency is responsible to the president, then the argument against having OIRA review independent agencies' regulations is straightforward. 132 Another traditional argument for independent agencies, however, is that they are more likely to make decisions based on facts and expertise. 133 OIRA review can facilitate this goal by coordinating input from other expert agencies and providing a fresh perspective on the agency's economic analysis. 134 A voluntary arrangement for OIRA review could take one of several forms. OIRA could simply offer comments privately on the agency's proposed regulations and analysis, which could be considered technical assistance rather than formal OIRA review or oversight. 135 Alternatively, the agency and OIRA could agree that OIRA could publicize any concerns about the regulation or the analysis; if the agency disagrees with OIRA, it would have to go on record acting contrary to OIRA's advice. Finally, the agency could agree that it will not move forward with a regulation unless it addresses OIRA's significant concerns with the regulation or the analysis. This approach would effectively mean that the agency has voluntarily agreed to allow OIRA to return regulations to the agency for further work, just as OIRA does with executive branch agencies.

V. Conclusion
This article outlines the role regulatory impact analysis can play at independent regulatory agencies, documents deficiencies in current practice, and suggests five steps that an agency can take to produce and use high-quality, objective regulatory impact analysis. Any agency so inclined faces one additional challenge: how to credibly commit to this change for the long term.
Fortunately, credible commitment mechanisms are available.
One commitment mechanism is bureaucratic inertia. The organizational, incentive, and cultural changes outlined in Section IV will require significant initial effort to accomplish. For this reason, they will also be difficult to reverse once established.
An agency seeking a stronger commitment mechanism can institutionalize many of the policies and procedures described in Section IV by formally adopting them in the Code of Federal Regulations (CFR). For example, when the US Department of Energy considers whether to issue energy efficiency or water efficiency standards for appliances, it follows procedures that are extensively documented in an appendix to the relevant section of the CFR. 136 The department commits to issuing an advance notice of proposed rulemaking that identifies potential standards and discloses all analytical work produced to date, in order to gather stakeholder input before it completes the analysis and selects a standard to propose as a regulation. 137 The appendix outlines the major factors to be considered by the department's analysis, such as the engineering analysis, effects on manufacturers and consumers, and the effectiveness of nonregulatory alternatives. 138 It explains how the analysis of these factors will be conducted and establishes timetables for stakeholder feedback. 139 The appendix also explains how the department will use the results of the analysis to make decisions. 140 136 Energy and Water Conservation Standards, 10 C.F.R. pt. 430C, app'x A. 137 Id. at § 4(c). 138 Id. at § § 9-12. 139 Id. at § § 4, 9-12. 140 Id. at § 5.

40
The Department of Energy states that these commitments do not create new grounds for judicial review of its regulations, but it commits to (1) providing notice and explanation of any deviations in specific instances and (2) publishing a notice in the Federal Register if it permanently alters any of the policies or procedures. 141 An independent agency that wanted to offer an even more credible commitment could specify that noncompliance with its policies and procedures published in the CFR could be grounds for judicial review.
From railroads to broadband and from table saws to financial derivatives, regulation by independent agencies is now a pervasive feature of the US economy. Yet many independent agencies are not required to systematically assess the economic effects of regulations before making regulatory decisions. This article demonstrates why regulatory impact analysis is necessary and explains how independent agencies can build the capacity to conduct objective analysis to inform decisions.