When President Bush stood in front of an old coal-burning power plant last September to defend his decision to relax clean air regulations for these huge polluters, he had his explanation ready. It "makes sense," he said simply. Likewise, it has made sense, to this administration, to reject federal standards to protect workers from workplace injuries; raise exposure limits for toxic chemicals like hexavalent chromium; lower energy efficiency standards for air conditioners; and ignore Clean Water Act violations by major industrial plants. It even made sense to raise the proposed limit on the permissible level of arsenic in drinking water – until public outcry forced the Bush administration to capitulate and adopt the Clinton standard.
The campaign to freeze or roll back health, safety and environmental protections may appeal to the president and his staff in part because it pleases their corporate backers. But it plays in the heartland because millions of Americans have been led to believe that regulators are over-zealous, and that the costs of many federal regulations hugely outweigh their benefits.
Where does this belief come from? Is it valid? The search for answers to these questions led me to an in-depth investigation of the stories and the studies most widely cited as evidence for the thesis that federal health, safety and environmental regulations are pervasively irrational. I found that anti-regulatory sentiment in this country draws on two main sources.
The first is a stream of anecdotes of government zealotry and caprice that have circulated in the media and congressional committee rooms for years – usually with little or no prior investigation of the facts. Some are true. Many turn out, on further examination, to be either hyped or atypical of agency practice. A disturbing number – like Rep. Tom DeLay's (R-TX) story about the time EPA declared "footprints of cows" to be protected wetlands – come close to pure invention. Through sheer repetition, such stories have taken on life of their own as "urban legends."
Besides the anecdotes, critics of regulation have generated at least three "regulatory scorecards." These studies typically tabulate a few summary statistics (costs, benefits, cost-per-life-saved) for a range of major regulations, and conclude that the costs of many government regulations outweigh their benefits. One study by John Morrall, an OMB economist, claims that government regulations cost up to $72 billion per life saved. Another scorecard – compiled by Robert Hahn of the AEI-Brookings Joint Center for Regulatory Studies – claims that over half of all major regulations issued since 1981 fail cost-benefit tests. A third study, co-authored by Bush's regulatory "czar," John Graham, reached the sensational conclusion that over 60,000 people lose their lives each year due to irrational government regulation – a situation Graham has called "statistical murder." These widely-cited studies have fueled scathing critiques of government for years. Only quite recently have the source data and methods of the studies themselves come under scrutiny.
Close examination reveals major problems: all three studies rely on undisclosed data and non-replicable calculations; use biased regulatory samples; misrepresent ex ante guesses about costs and benefits as actual measurements; underestimate the value of lives saved or the number of lives saved, or both; exclude all unquantified costs and benefits; disregard all questions about the fairness of the distribution of cost and risk; and conceal the large uncertainties that are present in virtually every regulatory analysis. In short, these studies are so fundamentally flawed that they prove nothing at all about the rationality of government regulation.
To cite just a few examples: Morrall acknowledges that he "revised" agency cost and benefit estimates (sometimes by orders of magnitude and almost always in the direction of higher costs or lower benefits). But he has yet to name the studies that justify his revisions, much less demonstrate their superiority over agency estimates.
Hahn claims to have improved on Morrall's table by "using the government's numbers." But his hitherto unpublished spreadsheet reveals, amazingly, that 41 of the 136 major rules in his database are assigned a "zero" benefit – including a rule to require contingency plans for cleaning up major oil spills, a rule to protect 3.9 million agricultural workers from acute pesticide poisoning, and a rule requiring the public reporting of releases of toxic chemicals from large manufacturing facilities. Those zeroes are not "government numbers." They are supplied by Hahn.
What is going on here? It turns out that Hahn's accounting system simply doesn't recognize whole categories of benefit – like environmental protection (in most cases), or prevention of acute poisoning, or improved enforcement of other rules, or any benefits that are not quantified. These benefits are summarily excluded from the tally. Rules that provide such benefits exclusively get an automatic zero. Hence the 41 zero-benefit rules. Given such accounting methods, it is no wonder that many rules "fail" cost-benefit analysis. What is remarkable is that so many rules pass.
Graham's "statistical murder" charge gained him nationwide publicity and delivered another black eye to the reputation of federal agencies. But Graham's claim rests entirely on the covert, counter-factual assumption that a dollar spent on low risk A is automatically subtracted from spending on some other risk B where that dollar would save more lives. In reality, no such cost triage now occurs, nor should it, given that ours is a 10 trillion dollar economy which spends only a small fraction of its resources on risk reduction of any kind. If money spent cleaning up hazardous waste sites might save more lives if redirected to combat smoking, then so might a portion of the $36 billion spent each year on lottery sales, the $92 billion spent on alcoholic beverages, or the $54 billion spent on tobacco. Spending money on low risks may be inefficient in some sense. But it isn't killing anyone.
In fact, if you look carefully at Graham's data you will notice that they don't really show even widespread inefficiency in the current system. It turns out that two-thirds of the 60,000 additional lives saved by hypothetical re-allocation of funds are accounted for by just two interventions (out of the 185 in his database), and 95 percent by just nine interventions. What these figures show, then, is not a pervasive pattern of regulatory irrationality, but a handful of life-saving opportunities which Graham's sources considered deserving of more money.
For all these reasons, and others too numerous to mention here, regulatory scorecards do not prove that government regulation is pervasively over-zealous or irrational. They also do not show that all regulations (or failures to regulate) are wise. What they prove is the propensity of quantitative cost-benefit analysis to cause confusion about complex matters of vital interest to the public=
Regulations exist to protect us from harm. For the most part, they do that. Of course policy-makers should weigh the costs and the benefits of all manner of public policies. The real lesson of scorecards is that this balancing will never be accomplished, rationally or fairly, by the pseudo-science of strictly numerical cost-benefit analysis.
To read Richard Parker's report, Grading the Government, click here. For the report's executive summary, click here.
Richard Parker is a professor of law at University of Connecticut School of Law and chair of the ABA Administrative Law Section’s Regulatory Policy Committee. He previously served as special counsel to the Deputy Administrator of Environmental Protection Agency, and Assistant General Counsel at the Office of the United States Trade Representative.