| 摘要 | Last December, AEI’s Center for the Study of Government Regulation convened a seminar of experts to assess President Reagan’s year-old program of regulatory reform. These are their edited and updated remarks.
I am going to be talking about what might be called structural reform—attempts to improve regulation across the board by altering administrative procedures. Now that economists have discovered the theory of government failure, making economics a lot of fun, administrative procedure has probably fallen heir to the title “the dismal science.” I will try to make it as little dismal as possible.
About a month after his inauguration, President Reagan issued Executive Order 12291, which might be considered the support beam of his administration’s framework for regulatory reform. It imposes two major new procedures on all executive branch agencies. First, it requires them to conduct regulatory impact analyses (RIAs), both before publishing their proposed rules and before adopting their final rules. (These are intended to ensure compliance with two requirements expressed in the prologue of the order: that the prospective benefits of each rule shall outweigh its prospective costs, and that each rule shall represent the regulatory alternative least burdensome to society.) The second major procedural mechanism established by the order is the requirement that these RIAs and the rules to which they pertain be forwarded to the Office of Management and Budget (OMB) for review and “suggestions” as to how they might be improved.
It is not my purpose to discuss the results that have been achieved since these new procedures have been in place. That task will be undertaken, I presume, by other speakers on this panel who will address various individual fields of regulation. What I want to pursue is the extent to which sound regulation is the predictable effect of these new procedures, so that they may be expected to be useful even when applied by personnel less imbued than the appointees of the current administration with the spirit of regulatory reform.
The first of the two features I have mentioned—the RIA requirement and the underlying prescription of cost-effectiveness and least-onerous alternatives—has received the bulk of public and congressional approbation. I frankly think it is a mirage. The requirement that agencies make sure that benefits exceed costs and choose the least burdensome alternative is really no more than a description of current statutory law. The Administrative Procedure Act invalidates agency rules that are “arbitrary or capricious.” Is it conceivable that a rule would not be arbitrary or capricious if it concluded with a statement to the effect that “we are taking the foregoing action despite the fact that it probably does more harm than good, and even though there are other less onerous means of achieving precisely the same desirable results”? Or a statement that “we really don’t know whether this action does more harm than good; and we really haven’t thought about less costly ways of achieving what we are after”? Surely such a rule would be invalid, unless it were issued under one of the few statutes (such as the Delaney amendment) which in effect say to the agencies, “Do it regardless of cost.” Those statutes, however, are expressly excluded from operation of the executive order.
These substantive prescriptions of the order, therefore, are really nothing new. And the RIA requirement merely spells out in some detail what the agencies are supposed to be doing anyway. One may say, of course, that the agencies in fact haven’t been doing it—or haven’t been doing it right—and that the executive order recalls them to their duty. That is unquestionably true, and to that extent the order is useful. It does not, however, represent any substantive or structural innovation that is likely to affect the nature of regulation even when the government is in the hands of a less reform-minded administration.
The other major aspect of the order—the requirement that RIAs and rules be cleared with OMB—is a different matter. This may represent a basic change in the structure of our regulatory decision making that has enduring importance for the future. It addresses what I regard as the more intractable of the two basic problems in the present system. The first problem—addressed by the RIA requirement—is the slovenliness of individual agency analysis. The second, which remains no matter how careful and painstaking the individual agency analysis may be, is the lack of overall regulatory coordination. After careful and painstaking analysis, the Environmental Protection Agency may conclude, for example, that a particular auto emission limitation is “worth it,” even though it imposes $1 billion in costs on the automobile industry. At the same time, after careful and painstaking analysis, the Department of Transportation may conclude that a particular auto safety feature is “worth it,” even though it imposes $500 million in costs on the industry. And the Occupational Safety and Health Administration that certain auto plant safety requirements are worthwhile, even though they impose $2 billion in costs. But what if the industry can only sustain a total of $3 billion in costs without disastrous impairment of its competitive position? Which one of the agencies is wrong? The answer, of course, is that none of them is—but all of them are. It is the lack of composite evaluation that is fatal.
Even if the RIA requirement of the executive order works ideally, our government is still in the position of a housewife who sends her three boys to the butcher, the baker, and the candlestick maker, telling each of them, “Don’t buy anything unless the price is reasonable, and get the best available bargain”—but who neglects to tell each of them how much he can spend. The predictable result is a houseful of useful bargains, and insolvency.
This is the problem which the OMB clearance process can solve. OMB is (if I may carry forward the analogy without irreverence) the big brother with knowledge of the family’s finances, whom the housewife requires to approve all the separate purchases before they are made. There is only one difficulty with the analogy: The OMB clearance system may be working that way in fact, but it is not supposed to work that way in law. Technically, the executive order only permits OMB to “advise” the agencies regarding RIAs, and not to forbid particular “purchases.” Moreover, it is unlikely that the President has authority to confer such power on OMB even if he wanted to; it would amount to reorganization of the government in defiance of the specified statutory scheme.
As a practical matter, however, the power of OMB to advise comes close to the power to control. In the event the advice is not heeded, the executive order gives OMB the ability to elevate the dispute to the presidential level—where the agencies must expect that OMB will usually be vindicated; and even when OMB loses in such a confrontation, the decision will still be made above the single-agency level (by the housewife herself, instead of the big brother), which is all that is important.
One may say that this coordination by mailed-fist advice is a poor way to run a railroad. But we inhabit, after all, not the best of all possible worlds. One can hardly expect Congress formally to designate David Stockman as Big Brother; and the less precise arrangement created by the executive order gives the agencies a somewhat higher degree of bargaining power, which is perhaps desirable.
I would now like to turn briefly to procedural reform initiatives on the part of the Congress. That body, which has ordinarily been sublimely uninterested in (not to say uninformed about) the utterly sexless subject of administrative procedure, has in the past few years acquired a burning passion for the field. Partly, no doubt, because it feels that regulatory reform legislation must be enacted; but mostly, I think, because it feels that it must enact regulatory reform legislation (if you understand the difference). The Senate has taken the lead, and the major features of the bill it will pass (S. 1080) now seem clear.
First, S. 1080 will place into statutory form the regulatory impact analysis requirements of the executive order I just described. That has the obvious advantage of making those requirements more permanent, not subject to change by a later administration. But since, as I indicated earlier, I do not believe those requirements in and of themselves accomplish a great deal, I do not consider this a major legislative triumph.
Second, it will confer statutory authority for (although it will not mandate) the other major feature of the executive order, the OMB clearance process. This is perhaps the one area where the administration hoped to get something of real value from Congress—namely, the clear ability to extend the clearance process beyond the executive branch to the independent agencies. It is probable (though many would dispute it) that the order could be extended to independents without further legislation, insofar as the RIA requirement and the requirement of receiving advice from OMB are concerned. But the critical element which renders that advice potent—the ability of OMB to take the agency to the mat in the Oval Office if it refuses to heed the advice, with final decision to be made by the President—could not be expanded to the independents under the law as currently understood. It is highly improbable that the final Senate bill will permit this expansion. (Indeed, it will probably not even acknowledge the President’s power to mandate changes in the proposed rules of executive branch agencies.) And OMB review of the independents may be expressly limited to “nonbinding advisory recommendations.” Even though this may be no more than what the President now has technical power to impose without new legislation, it is more than he would dare to impose—so it is worth something.
A third feature of the bill—and one which receives little attention in most of the commentary, since it surpasses even the remainder in intrinsic dullness—revises the administrative procedures required for the adoption of major rules. Essentially, it replaces the simple notice-and-written-comment procedures that are normally required with a more judicialized process, including oral hearings and cross-examination. This resembles the formalized rulemaking procedure imposed on the FTC by the Magnuson-Moss Act—the only clear effect of which has been a considerable increase in the expense and duration (and hence a decrease in the number) of rulemaking proceedings. One suspects that this is precisely what the private-sector proponents of these provisions are after. It seems to me there must be a cheaper way of frustrating agency action.
Fourth, the bill will contain the latest version—or a latest version—of the so-called Bumpers Amendment. I will not describe it other than to say it is meant in some degree, in some manner, to reduce (or eliminate) the deference which courts accord to agencies’ interpretations of their governing statutes, and to require clear statutory authority for agency action. Why a conservative Republican Senate would want thus to increase the power of the overwhelmingly liberal life-tenured judiciary appointed in recent years is beyond me. It is impossible to say how much harm will be done, because it is impossible to predict the precise version of Bumpers that will finally be enacted; and because whatever that may ultimately be, it is sure to be so vague in content as to enable the courts to make of it what they will.
It will not have escaped your attention that there is little in this bill which should have much appeal to the executive branch. Which is why the administration’s support began as lukewarm and has progressively cooled. Of course it could be worse. And that is precisely what makes the administration so edgy: it may well be worse by the time the House gets done.
For example, the Senate bill does not contain a provision for legislative veto of all agency rules. The House has passed such a provision in the past, and may well insist upon its inclusion. A slight and superficially reasonable tinkering with the definition of “major rule” could also have disastrous effect. In the current Senate bill, that definition (which identifies the rules to which the RIA, OMB-clearance, and formal-procedure requirements apply) is carefully crafted to include only rules that increase burdens on the economy. As a discretionary matter, however, the agency (or OMB) is given authority to designate as a major rule any rule that is “likely to result in a substantial increase in costs or prices for wage earners … or geographic regions” or “significant adverse effects on competition, employment, … the environment, public health or safety,” and so on. How tempting it is (to a Democratic House) to change this from a discretionary authority to part of the definition of “major rule.” The result would be that the retarding effect of all these new procedures would extend not merely to new regulation, but to deregulation as well. Similarly tempting is elimination of the provision in the Senate bill that exempts from judicial review the adequacy of the agencies’ compliance with the new procedural requirements. If that is changed, the Regulatory Reform Act of 1982 will provide wonderful opportunity for regulatory hide-and-seek of the sort that developed under the National Environmental Policy Act’s requirement for environmental impact statements. Lawyers will be able to delay rules for years by asserting the inadequacy of RIAs, or wrongful refusals to permit cross-examination of the experts who prepared particular studies.
It may well be that the cause of deregulation—and, coincidentally, the cause of efficient regulation where regulation is retained—will be best served by a stalemate between the Senate and the House. I am put in mind of one of my favorite passages from The Federalist, in which, to justify the proposal for replacing the Continental Congress with a new legislature composed of two separate houses, Madison observes:
It must be acknowledged that this complicated check on legislation may in some instances be injurious as well as beneficial. … But … as the facility and excess of law-making seem to be the diseases to which our governments are most liable, it is not impossible that this part of the constitution may be more convenient in practice than it appears to many in contemplation.
Discussion
Thomas Hopkins: Mr. Scalia argues that putting the benefit-cost principle in the executive order probably didn’t accomplish much because, as he expressed it, to require that a rule meet a benefit-cost test is really no more than saying that a rule should not be “arbitrary or capricious.” I would be delighted if that were the case. But based on my work with regulators over the years, I’m persuaded that they don’t think it is. Indeed, shortly after the executive order was issued, a mid-level rule writer came to us and asked, “Does this mean that a regulation whose costs are $10 billion and quantified benefits are only $7 billion should not be adopted?” I think it’s important that the benefit-cost principle be restated repeatedly as official guidance in regulatory policy. ■
ANTONIN SCALIA, editor of Regulation, is professor of law at the University of Chicago.
|
