A reality check on an empirical study: comments on "inside the administrative state".

• Author: Katzen, Sally
• Position: Response to Lisa Schultz Bressman and Michael P. Vandenbergh, Michigan Law Review, vol. 105, p. 47

TABLE OF CONTENTS INTRODUCTION I. IS THERE A BIAS IN THE DATA? II. WHAT DO THE DATA SHOW? A. Political Accountability B. Regulatory Efficacy III. WHERE DO WE GO FROM HERE? INTRODUCTION

Presidential control is the term used for the process (or some would say, the model) by which agency decision-making (more particularly, rule-making) is brought under the direction of the president to "render such decision-making accountable and effective." (1) Until now scholars, who have generally endorsed both the theory and the practice of the process, have written from the perspective of those who exercise presidential control (2)--those at the White House or the Office of Information and Regulatory Affairs ("OIRA"). (3)

In a recent article in the Michigan Law Review, (4) Lisa Schultz Bressman and Michael Vandenbergh ("the authors") decided to study presidential control from the perspective of those who experience it--those inside the agencies. They undertook a significant empirical study, interviewing top political appointees who served at the Environmental Protection Agency ("EPA") during the George H.W. Bush and William J. Clinton Administrations. The authors find, based on their data, that presidential control "is more complex and less positive than previous accounts acknowledge" (5) and that presidential involvement in rule-making "may not sufficiently enhance political accountability" (6) and "may not achieve regulatory efficacy." (7) Because they acknowledge that the president "has a role to play in controlling agency decision-making," they conclude that the model "requires reworking" and identify "next steps in that direction." (8)

I served as the Administrator of OIRA during the Clinton Administration, I am unabashedly a proponent of centralized review of rule-making, and I have a very different take on some of the data presented in Bressman & Vandenbergh's article. Nonetheless, I value continued productive dialogue, rather than an adversarial contest, on the issues. My comments in this Article, therefore, are presented as just that--comments, as in the thoughts, observations, or reactions of someone who has been involved in the process--with the hope that they may be useful to those doing follow-on work in this field.

First, however, let me give praise where praise is due. The authors quite rightly decided to look at the "other side" of the process. Not only is it wholly appropriate and highly desirable to have the perspective of those who are subject to presidential control, but the information serves as a valuable counterpoint to the material presented by those who have studied the process from the perspective of the White House or OIRA. More broadly, more data are always helpful in gaining an understanding of a process and testing preconceived notions about it. The Bressman & Vandenbergh article indisputably makes an important positive contribution to the on-going discussion and debate.

That said, it is important to recognize that just as rule-making is only one piece of an administration's activities, (9) the centralized review of rule-makings is only one piece of presidential control. The Office of Management and Budget ("OMB") presides over the whole mosaic, where review of rule-making occurs along side review of legislative proposals, review of Executive Orders, and, very importantly, review of resource (budget) decisions. Future studies of presidential control should, therefore, consider how these other pieces fit together with the review of rule-makings, because they dramatically affect agencies' responsibilities, priorities, and capabilities.

Turning now to Bressman & Vandenbergh's article, I look first at the data generated by the study, then some of the findings and conclusions drawn from those data, and lastly the authors' recommendations for next steps.

1. IS THERE A BIAS IN THE DATA?

   The authors wanted to get "inside the administrative state," (10) and they chose EPA for the period 1989 to 2001, covering the terms of the first President Bush ("Bush I") and President Clinton. (11) The authors explain that EPA had "more high-cost rules than any other" agency and that the time period selected would encompass both a Republican and a Democratic Administration. (12)

   EPA is certainly worth studying, but it is an atypical agency in almost every relevant respect. Although some degree of self-selection occurs in almost all agencies, EPA has a singular history of attracting highly motivated and intelligent political and career staff dedicated to its institutional mission. During both the Bush I and the Clinton Administrations there were a significant number of senior officials (within the study group identified by the authors) who had previously been environmental activists or state or local environmental officials and who had forcefully advocated the importance of protecting--indeed, improving--the environment. (13) This does not constitute "agency capture" in the traditional sense of the word, (14) nor does it necessarily produce "environmental overreaching." (15) But it is an answer to the authors' concern that the Clinton White House did not issue any directives to EPA "to energize [it] and induce [it] to take action." (16) In any event, in my experience senior political appointees at EPA clearly stand out from their colleagues at other agencies for both the intensity of their enthusiasm for their agency's mission and their faith in regulatory solutions.

   EPA is also unusual in that its authorizing statutes provide multiple mandates, allowing ample opportunities (if not affirmative obligations) for developing new regulations. And, more importantly for present purposes, EPA is atypical in terms of its relations with the White House because EPA is the only agency with an office in the Executive Office of the President dedicated to its subject-matter jurisdiction--the Council on Environmental Quality ("CEQ"). (17)

   To be sure, EPA produces many high-cost regulations, but there are many other agencies which produce appreciably more and varied regulations, and which would therefore provide a more representative sample. For example, for Bush I the agency with the most economically significant rules (18) was the Department of Agriculture ("USDA") (111), followed by Health and Human Services ("HHS") (92), EPA (86), and the Department of Transportation ("DOT") (41); for the Clinton years, EPA had the most economically significant rules (158) reviewed by OIRA, followed by USDA (144), HHS (137), and DOT (57). (19) When all the rules reviewed by ORIA in those administrations are considered, EPA's relative position drops to third in the Clinton Administration and fifth in the Bush I Administration. (20)

   With respect to the time period chosen, I agree with the authors that it is useful to look at administrations of both political parties. But there was no Senate-confirmed OIRA Administrator for most of the Bush I period; instead, the "political" oversight was exercised by the Competitiveness Council--which was placed in the White House and thus logically viewed by the Bush respondents as a White House office. (21) Also, the Bush I OIRA was operating under Executive Order 12,291, (22) while the Clinton OIRA operated under Executive Order 12,866. (23) There are significant differences between the two Executive Orders, (24) although except for the provisions relating to openness, the differences probably would not have a significant effect on the conclusions of the study. Nonetheless, I would hope that any follow-up study would compare the eight years under Clinton with the eight years under Bush II, which operated under Executive Order 12,866 and a Senate-confirmed Administrator for the majority of the time.

   In addition, there are aspects of the survey itself that lead me to question whether it is appropriate to accept its data uncritically and to draw conclusions (and recommendations) from it without qualification. First, the authors are clear that just because OMB says something, that is not necessarily what the agency hears. (25) So too, what the agency hears is not necessarily what OMB is saying. Second, just as White House officials might introduce their biases for centralized review into their comments about the effect of their interventions in rule-makings, so too agency officials might incorporate their biases for greater agency autonomy into their comments about relations between the White House (and/or OMB) and their agency. (26) This would not be the first time that strategic responses were supplied for an academic survey. Third, there is not always a good fit between the respondents and the questions asked of them. The survey's subjects were all political appointees, who are often the recipients of messages from the White House and can therefore speak to those messages. But political appointees are simply not positioned to be reliable sources of information about other topics they were asked about. (27) They are personally involved in only a fraction of the interactions between an agency and OIRA (or the White House), and they are bound to resent any suggestions that they are told have come from the civil servants at OMB.

   There is also a potential problem from the way some of the questions were framed. (28) At a minimum, the survey was not completely successful in capturing the dynamics of the review process, which the authors ultimately recognize is "more complex than scholars generally have acknowledged." (29) I agree that the process is more complex than has been depicted in...