May 28, 2002 Mr. John Morrall III Office of Information and Regulatory Affairs NEOB, Room 10235 1725 17th Street, N.W. Washington, D.C. 20503 Re: OMB’s 2002 Draft Report to Congress on the Costs and Benefits of Regulations Dear Mr. Morrall: In past comments, we have focused on the general problems with regulatory accounting, as well as methodological issues. With the change in administrations, however, OIRA has used its new draft report to highlight a shift in approach to regulatory policy. Specifically, this seems to mean a more active OIRA that aggressively asserts cost-benefit analysis and takes a central role in identifying regulatory priorities. Accordingly, our comments focus on these policy changes. In particular, we address transparency issues, which OIRA has made a priority, and offer suggestions for continued improvements. We recommend factors OIRA should consider in preparing its guidance document on cost-benefit analysis. And we offer our views on OIRA’s proper role, focusing on issues such as return letters, risk assessment, and science at OIRA. In addition, we also provide some of our own regulatory priorities for OIRA’s consideration, which are attached, as requested in the draft report. OMB Watch chairs a broad-based coalition called Citizens for Sensible Safeguards, which works to promote strong standards for health, safety, civil rights, and the environment. We are in the process of soliciting regulatory recommendations from our coalition partners. Attached is a preliminary sample of recommendations that have been provided to us for OIRA’s consideration. Once a final list is compiled, we will make it available to the public and share it with OIRA. As discussed further below, we do not believe OIRA should be using this report to rank regulatory priorities, but we feel as if we have no choice but to participate. Finally, we discuss some of the numbers presented in OIRA’s report. We recommend that OIRA drop its continued use of the Hahn-Hird estimates and better discuss uncertainty in monetization, as well as estimates of regulatory burden on small business. Transparency OMB Watch strongly supports recent efforts by OIRA to increase transparency. These efforts are important and should be commended. Yet OIRA should not be complacent. Much still needs to be done. Specifically: OIRA should provide clarity and justification for its rankings of public recommendations for regulatory reforms. We do not believe such priority setting is an appropriate role for OIRA. Yet if OIRA is to continue this practice, it must be transparent. OIRA has given no indication of any sort of process for reviewing recommendations, nor has any justification been given for any of the rankings. OIRA should explain why a rule received a particular ranking, be it “high priority,” “medium priority,” or “low priority,” allowing the public to evaluate OIRA’s reasoning. OIRA should provide greater clarity for its “upfront” involvement in agency rulemakings. OIRA states that “agencies are beginning to invite OIRA staff into earlier phases of regulatory development in order to prevent returns late in the rulemaking process.” This sort of discussion is not covered by E.O. 12866, yet clearly OIRA’s influence is being felt. OIRA should begin to think about how to communicate its role in shaping rules that have yet to be sent to OIRA for review. OMB Watch does not support an “upfront” role for OIRA. Rather, we believe that agencies -- which unlike OIRA, are empowered by Congress to carry out regulation -- should be given more discretion in formulating policy than OIRA seems willing to grant. Yet as OIRA exerts more influence early on in the rulemaking process -- negating the disclosure requirements of the E.O. -- it is important for OIRA to communicate this new role to the public and Congress. In this respect, disclosure of prompt letters (which again, we do not think is an appropriate function for OIRA) is a positive step. Likewise, OIRA should make clear when it becomes involved in a rulemaking, as well as the nature of its recommendations. OIRA should better explain and document “consistent with change.” Based on our experience with a number of regulatory review files at OIRA’s docket library, it is still difficult, if not impossible, to discern what a regulation looked like when OIRA received it, and how it changed during the review process. E.O. 12866 does not require OIRA to provide this documentation. Rather, this is left to the agencies, which are frequently negligent in adequately performing this responsibility. We believe OIRA should provide agencies guidance on how to carry this out. As one possibility, most word-processing software provides editing tools that could clearly indicate how text is changed during OIRA review. Documentation of changes should also be kept at OIRA’s docket. This will make it easier for the public, Congress, and indeed OIRA itself, to assess OIRA’s performance during the regulatory review process. We understand that OIRA does not view changes made during the review process as necessarily its own, but rather the outcome of a dialogue with the agency. This stipulation can be made, but it should not preclude clear and full disclosure of the outcome of the review process. OIRA should also consider adding greater specificity to the ambiguous “consistent with change” label. This label only tells you that change was made, but tells you nothing of the nature of the change. Thus, there is no way to distinguish rules with significant substantive changes from rules with minor changes for clarity where the substance of the rule is unaltered. OIRA should provide guidance on how to better improve transparency of the regulatory process across agencies. There is no government-wide regulatory identification system that would facilitate the tracking of a rule over its life cycle. Ideally, the public should be able to enter in a rule’s identification number into a single web-based search function, and retrieve all related information across federal agencies, OIRA, GAO, and the Federal Register. OIRA should begin an effort to make this a reality. NHTSA provides an example of how an ID system can work. NHTSA keeps all rulemaking information in an electronic docket, which can be searched by entering in a rule’s ID number. This allows the public to easily retrieve a wealth of information, including public comments, the agency’s regulatory impact analysis, and supporting studies. In developing guidance, OIRA should also examine other agency transparency efforts. Many agencies have experimented with electronic rulemaking, for instance. These efforts are crucial, and OMB deserves credit for highlighting the benefits of e-rulemaking in the administration’s e-government initiative. In pursuing e-rulemaking, OIRA should be cautious not to get in the way of innovation, but should serve to highlight the more successful enterprises, and push further improvements. We encourage OIRA to study several models before becoming locked in to one agency approach. Cost-Benefit Factors to Consider OIRA has initiated “a process of refinement to its formal analytic guidance documents.” We have concerns about many of the issues OIRA has identified. The 7 percent discount rate currently advised by OIRA is far too high, whatever you think about the practice of discounting in general. We intend to submit views on these issues in the context of the public comment period committed to by OIRA. In response to OIRA’s request for additional analytical issues that should be addressed in this process, we offer the following suggestions: Willingness to pay vs. willingness to avoid. The use of willingness-to-pay methods may be profoundly affected by socioeconomic characteristics. Specifically, lower-income individuals may report lower willingness-to-pay, not because they actually place a lower value on a particular benefit, but because they are implicitly trading off such payments against basic necessities. We believe agencies should avoid basing willingness-to-pay functions on highly constrained respondents. Indeed, it should be acknowledged that all WTP values are based on income constraints, and that utilization of the WTP methodology implicitly involves a major policy choice on the “correct” level of wealth. Before promoting WTP, OIRA should evaluate which, if any, willingness-to-pay estimates are indeed “generally accepted,” and how such acceptance developed. Willingness-to-avoid methodologies should also be discussed. Willingness-to-avoid is assessed by asking respondents how much money they would have to be paid in order to accept a particular risk -- for example, increased risk of harm to their children’s health resulting from an environmental factor. While in theory willingness-to-pay and willingness-to-avoid should be the same, WTA is generally higher than WTP by a factor of two to ten or more. We believe it is significant that, in the research conducted to date, there is no sum of money that parents are willing to accept for an increased risk to their children’s health. The need to account for technological advancements. Industry learns to adapt to regulation and reduce costs over time through technological advancements, “learning by doing,” and other factors. As a result, agency cost estimates, which do not attempt to anticipate adaptive measures, frequently prove overblown in the real world. An article in the American Prospect by Eban Goodstein and Hart Hodges highlights this tendency in dramatic fashion. In examining estimated costs next to actual costs for 13 major rules, Goodstein and Hodges found estimated costs were at least double the actual costs for all but one. For instance, EPA estimated in 1990 that acid rain controls would cost electrical utilities about $750 per ton of sulfur dioxide emissions; yet the actual cost turned out to be less than $100 per ton, billions of dollars less than what was initially anticipated. OIRA should examine how to incorporate well-documented adaptive effects to improve the accuracy of cost estimates. The need to check industry-supplied data. Regulated entities have a clear self-interest in providing data that makes potential regulation look as expensive as possible. Yet agencies are frequently forced to rely on this data for cost-benefit analysis. OIRA should examine ways to check industry-supplied data to ensure it is “objective,” which OIRA has pressed in the context of its data quality guidelines. The need for transparency in monetization. Monetization involves a host of assumptions and analytical leaps of faith. Agencies should be required to explain these assumptions in a clear and consistent way, so they can be understood by the public and Congress. OIRA’s Role Over the last year OIRA has sought to expand its role and influence in a multitude of ways, as discussed in the draft report. OIRA is not statutorily charged with carrying out health, safety, and environmental protections. Yet OIRA’s draft report demonstrates little reluctance to exert influence in these areas. Specifically, OMB Watch has concerns in the following areas: Ranking of regulatory priorities. The statute mandating the regulatory accounting report asks for “recommendations for reform.” OIRA has inappropriately interpreted this as a broad mandate to push for changes in specific rules. Congress considered specifying such a role in the context of regulatory accounting legislation, yet in the end, decided against it -- and for good reason. OIRA’s regulatory accounting report is a very poor vehicle for soliciting comments on such changes, and for ranking regulatory priorities. Because of the wide-open nature of the request, OIRA may receive just one set of comments on a rule expressing one set of views. This is hardly a sound foundation for priority setting. Outside parties have no way of knowing whether a rule they are interested in might be submitted to OIRA, and subjected to examination. Presumably if OIRA put out a request for comment on a specific rule -- not something we suggest, as this should be left to the agencies -- there would be a greater volume of comments, presenting a fuller picture of the issues involved. OIRA’s priority rules are also constrained by who happens to respond. For the last report, OIRA received 71 suggestions for reforming specific regulations, 44 of which were from George Mason’s conservative Mercatus Center. From this limited pool, a disproportionate number of environmental rules received a “high priority” ranking, signaling OIRA’s agreement and intent for further examination. However, according to OIRA, “a closer examination of OIRA’s decision making process reveals no implicit or explicit intent to target environmental rules for scrutiny. The distribution of rules by agency reflects the concerns raised by public comments, not the interests of OIRA.” Of course, OIRA was free to reject those concerns as unrepresentative, which they clearly were. In the future, OIRA may find itself overwhelmed with recommendations for reform covering a wide range of issues for which it has little or no first-hand expertise. Accordingly, regulatory changes should be left to the agencies, which have the statutory authority delegated by Congress, the technical and scientific expertise, and the proximity to affected parties -- including regulated interests and the intended beneficiaries of regulation -- that OIRA lacks. Data quality guidelines. In discussing data quality guidelines, OIRA states that it “has taken a strong interest in improving the quality of information and analysis used and disseminated by agencies.” Yet the Data Quality Act says nothing about analysis or the interpretation of data. Rather, the focus should be on correction of factual errors. OIRA went far beyond the statutory mandate in requiring agencies to “adopt or adapt” principles for risk assessment in the Safe Drinking Water Act (discussed below), and insisting that “influential” information be “reproducible.” Safe Drinking Water Act requirements for risk assessment. OIRA has pushed agencies to adopt principles for risk assessment laid out in the Safe Drinking Water Act (SDWA), both through its data quality guidelines and its memorandum on regulatory analysis. Across-the-board requirements for risk assessment have been thoroughly debated by Congress since the Contract with America in 1994, and have been rejected. Likewise, OIRA should drop its push for government-wide adoption of SDWA principles. Such significant and far-reaching action must come only at the direction of Congress. Agencies have many different functions, histories, and statutory requirements, and by necessity, carry out risk assessment in different ways. Indeed, in draft data quality guidelines, no agency -- including EPA -- chose to “adopt” SDWA’s risk assessment requirements. The SDWA was written for health-based risk assessment, specifically with cancer prevention in mind. For instance, it does not fit neatly with safety assessments or assessments of ecological impacts. As one example, the Food and Drug Administration points out that many of its actions related to non-cancer-causing hazards -- for example, actions related to adverse effects from drugs -- are based on the judgment of scientific experts, and are essentially qualitative. “Although we analyze the economic costs of the regulations and consider alternatives, regulations like these do not lend themselves to the types of quantitative risk assessments contemplated by the Safe Drinking Water Act principles,” FDA states in its draft guidelines. OIRA should consider agency responses to the SDWA principles as a strong signal that adoption won’t work. Use of return letters. OIRA states that “the degree of OIRA’s actual effectiveness can be questioned when it declines to use its authority to return rules.” Yet more important in evaluating OIRA is the reason for the return. Most of OIRA’s return letters and each of its post-review letters claim inadequacies in the agency’s cost-benefit analysis -- making good on the promise to elevate cost-benefit analysis. In no case does OIRA return a rule for being insufficiently protective of public health, safety, or the environment. OIRA might claim NHTSA’s tire-pressure monitoring standard. But OIRA’s clear overriding concern is cost. For instance, in returning a rule from the Department of Transportation on hazardous materials -- specifically, safety requirements for cargo tanks transporting flammable liquid -- OIRA suggested exempting tanks older than 15 years, questioning the “cost-effectiveness” of the rule, and indicating it feels that DOT is overstating benefits. To satisfy OIRA and “prove” cost-effectiveness, monetization of benefits becomes crucial. This emphasis on cost-benefit analysis, and in particular the monetization of benefits, is very troubling. Cost considerations are frequently easier to monetize than benefits. Benefits on the other hand frequently do not involve goods or services traded on the open market; they may involve the saving of human life, prevention of environmental degradation, or the reduction of injury or disease. Monetizing such benefits involves severe limitations and a host of questionable analytical assumptions. Under the monetization methods favored by OIRA, this process is likely to deflate benefits relative to costs, which can be used to justify a decision not to act. Rhetorically at least, OIRA has recognized the difficulty in monetizing certain benefits and the necessity of incorporating qualitative factors in cost-benefit analysis. Yet through return letters to agencies, as well as OIRA’s memo on regulatory analysis, it is clear OIRA believes agencies are not doing enough to monetize benefits. For instance, in a post-review letter on an EPA proposal to control emissions of certain recreational engines, OIRA objects that EPA failed to monetize environmental benefits associated with the rule. Besides affecting the rule in question, such a public letter also serves to send a message across government, affecting the way other agencies operate -- which OIRA has stated is its intent. Critics, including OMB Watch, have argued that this emphasis on monetization does not bode well for health, safety, and environmental protections, and OIRA has yet to prove us wrong. Cost-benefit analysis vs. underlying statutes. According to OIRA, “The public and Congress have an interest in benefit and cost information, regardless of whether it plays a central role in decisionmaking under the agency’s statute.” Yet this statement can be called into question where Congress has not mandated such information in an agency’s underlying statute, particularly where it has explicitly forbidden cost-benefit analysis as a decision-making tool. If Congress were especially interested in such information, it seems likely it would have been mandated in statute. In such cases, OIRA should recognize an agency’s cost-benefit analysis as less important. Certainly, it should not be a reason for holding up regulatory action, which in most cases, was precisely why Congress prohibited its use in the first place. Peer review. OIRA urges peer review of Regulatory Impact Analyses (RIAs), stating that it will “give a measure of deference to agency analysis that has been developed in conjunction with such peer review procedures.” OIRA needs to recognize the difficulty in conducting such peer review, which has been considered by Congress, and rejected. Serving on such panels is time-consuming, and balance may be difficult, if not impossible, to achieve. Indeed, OIRA may implicitly recognize this, as it does not instruct agency peer review panels to be balanced. Rather, OIRA merely states that peer reviewers should be selected “primarily on the basis of necessary technical expertise.” In the case of EPA’s Science Advisory Board, this has frequently meant a stacked deck of self-interested industry representatives, as documented by GAO. OIRA should instruct that agencies avoid such conflicts. Where there are conflicts, OIRA should make clear that they are to be disclosed to the public, not just the agency (which is OIRA’s current position). OIRA states that its interest in transparency extends partly from an interest in answering suspicion about process and shifting debate to matters of substance. This same logic demands that peer reviewers publicly disclose any conflicts of interest. Prompt letters. OIRA has issued a number of prompt letters. By themselves, none of these letters is truly harmful, and in fact, some may be helpful. However, it raises the question of the proper role for OIRA. OIRA has -- for good reason -- never before attempted to prioritize regulatory issues for agencies, which unlike OIRA, are statutorily charged with carrying out regulation. OIRA has little scientific expertise on staff (even as it tries to expand in this area) and lacks the resources and procedural mechanisms to guarantee public involvement in such important decisions. Indeed, the letter to OSHA on defibrillators in the workplace seems to have been sent after a cursory review of several journal articles. This hardly seems a sound foundation on which to base agency priority setting. The case of the FDA rule on trans fatty acid, however, is less clear-cut. In this case, the agency had already initiated a rulemaking, making it a priority on its own; OIRA’s letter simply asked the agency why it had not completed the rulemaking. The public interest community has long complained that agency rulemakings take too long, sometimes more than a decade. It might be an appropriate role for OIRA to provide the occasional whip to make sure that rules are being completed in a timely fashion, and examine the possible sources of ossification (such as budgetary constraints) where it exists. Science at OIRA. OIRA recently committed to hiring five new analysts “aimed at expanding the Office's scientific capabilities.” This includes a risk assessor, a public health scientist, a health economist/decision scientist, and an engineer. OIRA also notes in its draft report that it is in the process of forming a scientific advisory panel, which will meet twice a year, “that will suggest initiatives to OIRA, evaluate OIRA’s ongoing activities, comment on national and international policy developments of interest to OIRA, and act as a resource and recruitment mechanism for OIRA staff.” This is a significant and troubling development, implying a greater centralization of control over regulatory functions at OIRA. By executive order, OIRA has review authority over agency cost-benefit analysis, and the Paperwork Reduction Act gives review authority over information collection requests, both of which involve economic considerations and analysis. But questions of science that go to the need for a regulation should be the chief purview of the regulatory agencies, such as EPA, DOT, and OSHA. After all, unlike OIRA, agencies have a statutory obligation to address these scientific questions; they must answer to Congress and the courts for regulatory performance; they have the necessary experience, along with the specific built-in expertise; they have relationships with affected communities; and they have processes for public involvement. In turning back NHTSA’s tire pressure standard, the OIRA administrator told the Washington Post, “The question of whether anti-lock braking systems should be made standard equipment in all cars -- and through a regulation -- is worthy of a regulatory proceeding. We at OMB have been thinking about that issue for the last several months." What’s so startling about this statement is that it implies OIRA spearheading action in an area that is clearly in NHTSA’s statutory jurisdiction. This is overstepping by OIRA, and the hiring of scientists, as well as the establishment of a scientific advisory panel, does not bode well for the future. Assessing Costs and Benefits of Regulation OIRA should drop all use of the Hahn-Hird figures. OIRA presents aggregated figures for costs and benefits in its Appendix C, “because they are based substantially on figures that the agencies did not produce and OMB did not review.” OIRA is right to give these figures reduced weight, yet OIRA has not gone far enough. These figures rely heavily on a 1991 study by Robert Hahn and John Hird (not mentioned in the draft report, but discussed in OIRA’s 2001 report), which was actually based on much earlier work from the late 70s. This study fails to incorporate information generated over the last 20 years linking pollution to adverse health effects. For instance, it does not capture the enormous health benefits associated with reduction in levels of airborne lead, or the growing body of work that led to regulation of fine particulates. Nonetheless, OIRA presents Hahn-Hird’s lower-bound estimate to generate the implausible finding that environmental regulation may have resulted in a net loss of $83 billion as of 2001. Does OIRA really believe this is possible? The Hahn-Hird estimates are woefully outdated, and should be dropped entirely. Total costs of regulation should not be overblown. OIRA states that the total costs of regulation are comparable to discretionary spending at $640 billion in 2001. Where is this number from? There is no citation. OIRA presents estimates in its Appendix C that peg regulatory costs from $520 billion to $620 billion. Yet even at the high end, this is still a not-insignificant $20 billion less than $640 billion. Moreover, OIRA disavows these numbers, as discussed above. There is simply no good data available on total regulatory costs, and OIRA should acknowledge that. OIRA should be clearer about uncertainty in monetization. OIRA states that benefits are “highly uncertain.” In a monetized sense, this is undoubtedly true. Many benefits are frequently not captured by monetized cost-benefit analysis, such as morbidity, effects on ecosystems, and equity considerations. Yet this does not mean that benefits are necessarily “highly uncertain.” Rather, while science may be able to establish likely benefits, they may not be reflected in the agency’s cost-benefit analysis. Take EPA’s cost-benefit analysis for arsenic in drinking water, as described by Lisa Heinzerling and Frank Ackerman in their booklet from earlier this year, “Pricing the Priceless.” Arsenic causes cancers of the bladder, lungs, skin, kidneys, nasal passages, liver, and prostate, as well as other cardiovascular, pulmonary, neurological, immunological, and endocrine problems. However, EPA’s dollar estimates only include the health benefits of reductions in bladder and lung cancer, for which there was quantified data, missing other likely benefits. In calculating non-fatal benefits for these cancers, EPA amazingly substituted a previously determined willingness-to-pay value for reducing chronic bronchitis, as no such value was available for non-fatal bladder and lung cancer. In other words, EPA treated nonfatal cases of bladder and lung cancers as equivalent to chronic bronchitis. In this sense, benefits are less uncertain than obscured by monetization, which OIRA should acknowledge. Moreover, while OIRA labels benefits “highly uncertain,” this is not applied to costs, as it should be. It’s true that cost considerations are inherently easier to monetize than benefits. For example, they may involve purchases of new equipment or the hiring of additional personnel. Yet ironically, this does not mean cost estimates are any more accurate. In calculating costs, agencies must rely on data supplied by industry, which has a self-interest to make potential regulation look as expensive as possible. Combine this with well-documented adaptive responses to regulation, such as technological advances or “learning by doing” -- which drive down costs over time, yet are not predicted by cost-benefit analysis -- and agency cost estimates frequently prove overblown in the real world. Costs to small business should be placed in proper context. In discussing impacts on small business, OIRA cites a dire new report, conducted by W. Mark Crain and Thomas D Hopkins for SBA’s Office of Advocacy that sounds a familiar theme: Small businesses are being strangled in a sea of regulation. Indeed, the study estimated that federal regulations cost small firms (less than 20 employees) a mind-boggling $7,000 per employee annually; the total cost of regulation was estimated to be an incredible $497 billion. No estimates were given for benefits, as the focus of Advocacy is clearly on costs. In the press release accompanying the report, Hopkins expresses wonder at how small businesses have still managed to thrive despite such oppressive burden. It doesn’t seem to occur to him that his estimates might be vastly overblown. Given the difficulty OIRA has had in producing its own estimates on the cumulative impact of regulation, the Crain-Hopkins study should be treated with skepticism. A 1996 report from the General Accounting Office -- the investigative arm of Congress -- provides even more reason to be skeptical. GAO examined the regulatory costs of 15 businesses and failed to produce any evidence of significant burden -- or even slight burden for that matter -- despite the fact that participating companies, who volunteered for the study, were largely critical of federal regulation and had a vested interest in proving their case. Indeed, not one company was able to pinpoint its annual regulatory costs, and instead GAO’s study turned into an explanation of why honest regulatory assessment might not be possible. Of course, this is not to say there is no burden at all. We should be concerned about burdens on small business. OMB Watch, for example, has long advocated a system of integrated reporting across agencies that would allow regulated entities to submit required information to one place electronically, thereby reducing duplicative efforts. Recently, we have also worked on legislation (S. 1271) in the Senate that would create a comprehensive inventory of compliance assistance programs, making it easier for small business to take advantage of such programs. At the same time, small business should not be used as an excuse for inaction, as some would clearly have it. According to SBA, the legal definition of small business encompasses 99 percent of American businesses employing 51 percent of the private-sector workforce. This includes, for instance, a general contractor with as much as $17 million in annual revenue, a chemical company with as many as 1,000 employees, and a petroleum refinery with as many as 1,500 employees -- not exactly mom and pop. Under such an expansive definition, it’s impossible to have strong health, safety, and environmental protections without asking small business to do their part. OIRA should make this clear. Thank you for consideration of our comments. Enclosed is a beginning list of rules divided in two parts -- pending administrative actions and new initiatives. Sincerely, Reece Rushing Policy Analyst OMB Watch