ARTICLE CONTENTS INTRODUCTION 1176 I. BUNDLING DYNAMICS 1182 A. Concept and Measurement 1183 B. Alternate Measures 1190 C. Trends 1192 II. BUNDLERS AND SPLITTERS 1198 A. Internal Regulatory Drafters 1198 1. Intra-agency Bargaining 1198 2. Rule-Production Costs 1202 B. Monitors 1203 1. Facing Oversight 1204 2. Political and Judicial Review 1205 a. Office of Information and Regulatory Affairs 1206 b. Congress 1209 c. Courts 1213 III. IMPLICATIONS 1218 A. Regulatory Process 1219 B. Innovative Controls 1223 1. Line-Item Vetoes 1224 2. Single-Subject Rules 1226 3. Judicial Review 1228 a. Soft Looks 1228 b. Notice and Logical Outgrowth 1232 CONCLUSION 1234 APPENDIX 1235 INTRODUCTION
Administrative agencies can make policy in myriad ways. They can, for example, choose their form of action: whether a rule, an adjudication, an enforcement, or a guidance document. (1) For decades, commenters have analyzed the positive and normative trade-offs of each. (2) More recently, scholars have become attentive to the ways in which these forms can be aggregated. Some have debated adjudicatory aggregation through administrative class actions. (3) Others have remarked upon "crackdowns," or the ways in which enforcement actions can be pursued all at once. (4) This lens has shed new light on agency discretion by drawing upon rich analogies to other areas of the law, such as civil procedure, where aggregation and its discontents are common themes. (5)
Puzzling, then, that more attention has not been paid to the ways in which agencies can aggregate perhaps the most consequential tool at their disposal: legislative rules. (6) Like statutes, legislative rules bind entire classes of individuais. (7) And they too can be combined in multiple ways--a phenomenon that we refer to as regulatory bundling. Regulatory bundling refers to the ability of an agency to choose the scope of a single rulemaking--the number of discrete issues to resolve at a given point in time. Bundling decisions can occur at all stages of the rulemaking process, from drafting to implementation to litigation. For example, an agency can bundle some decisions at the proposed rule stage, only to split them into distinct final rules. It can then reaggregate those issues and revise them through a subsequent regulation.
Consider some examples:
* The Environmental Protection Agency (EPA) recently issued a proposed rule which, in the Agency's own words, combined "three distinct actions." (8) Some of these had previously been issued separately, but were now combined into one proposed rule. The rule's most high-profile action revised the Agency's Clean Power Plan, which was an effort by the Obama Administration to reduce carbon dioxide emissions from coal-burning power plants. (9) In addition, the rule proposed new requirements regarding state implementation of the new emissions requirements. Finally, the rule revised the Agency's New Source Review program, a pre-construction air permitting requirement. Some commented that the Clean Power Plan revisions were actually "distractions" from the more consequential New Source Review provisions. (10)
* Years earlier, EPA had initially set out to regulate greenhouse gases under the Clean Air Act through one rulemaking. (11) After a change in presidential administration, however, the Agency instead decided to issue four separate rules addressing different subjects. (12) The first determined that carbon dioxide "endangered" the public. (13) The second regulated auto emissions in light of that finding. (14) The third dealt with "triggered" permitting requirements for stationary sources, (15) while the fourth "tailor [ed]" the permitting requirements to the largest carbon emissions sources. (16) All of these regulatory decisions could have been packaged into one rulemaking, but EPA chose to split them into four. (17)
* The Occupational Safety and Health Administration (OSHA) traditionally regulated a single substance at a time. (18) Its Air Contaminants Standard rule, however, addressed 428 substances--from sulfur dioxide, to styrene, to wood and grain dust--all at once. (19) After a court struck down the bundled regulation, OSELA was unable to revise any of the individual standards, despite efforts to rebundle them in new ways. (20)
* In response to patient deaths in fialuridine clinical trials, the Food and Drug Administration (FDA) proposed a rule establishing new reporting requirements for human drug and biological products as well as investigational new drug applications. (21) It delayed the issuance of the final rule in response to heavy criticism from commenters, however, and considered breaking up the proposed rules into three separate final rules. (22) The FDA ultimately decided to issue guidance documents addressing some of the issues. (23)
Each of these scenarios illustrates a form of bureaucratic discretion that we call regulatory bundling--a practice with underexplored implications for the administrative state.
Indeed, the phenomenon has newfound significance after a recent Trump Administration executive order aimed at reducing regulatory costs. (24) The order contains, among other things, a "two-for-one" requirement, which directs agencies to repeal two regulations for every significant new one promulgated. (25) Executive agencies now have a greater incentive to pack more regulatory provisions into one rule because doing so delivers more bang for the buck. Assuming a fixed amount of offsetting rules at a given time, an agency can regulate more by aggregating more provisions into a single rulemaking. (26) If, by contrast, the agency splits the provisions, it would be forced to find two additional rules to repeal. One would thus expect to see more regulatory bundling after the executive order went into effect.
As further motivation, consider parallel scholarly observations about the legislative process. Social scientists have long studied the phenomenon of omnibus bills in Congress, through which legislators bundle numerous, often unrelated, provisions. (27) Positive theories propose that omnibus vehicles allow legislators to advance partisan agendas, engage in distributive logrolling, or pass otherwise unpopular measures. (28) Normative reformers, in turn, have often called for statutory unbundling. Single-subject rules found in many state constitutions, for example, limit bills and referenda to one subject. (29) The line-item veto similarly facilitates statutory unbundling by the executive branch. (30)
This Article explores analogous insights in the regulatory context in the hopes of spurring a broader research agenda akin to the decades of studies pursued in the legislative arena. The effort here is primarily one of theory building, and in that spirit we include preliminary empirical analyses to explore intuitions and generate hypotheses for more rigorous testing in future work. This study also attempts to complicate existing debates about agency behavior. Take rule counts. Popular media and academic studies often rely on them to convey the magnitude of agency regulatory activity. (31) But such counting exercises cannot meaningfully capture the scope of an agency's rulemaking. One rule can set standards for an entire industry, while another addresses narrow compliance issues. A third might deal with technical or routine matters and have only a temporary effect, (32) while a fourth is wholly deregulatory. (33) Rule counts are thus often misleading indicators of regulatory activity.
Appreciating agency bundling, however, helps to refine thinking about the relevant units of analysis. If compliance burdens are a concern, for example, it may be more sensible to measure costs than to tally rules. Understanding bundling behavior also enriches and complements work on the strategic timing of agency decisions and the more dynamic aspects of rulemaking behavior. (34) Agencies can simply delay controversial provisions, for example, by splitting them from a particular rule to save for future rulemakings.
Part I analyzes the concept and operationalization of regulatory bundling. It uses a unique dataset obtained from nearly twenty years of rulemaking across a wide range of agencies to provide an initial descriptive picture of the dynamic. (35) The results suggest that bundling is increasingly common. Agencies seem more likely to bundle issues in their rulemaking efforts today than they were even a short time ago. Moreover, agencies appear to adopt a wide variety of practices with respect to bundling: some bundle a great deal, and others do so rarely.
Part II explores the various actors internal and external to the agency that likely influence the agency's bundling decisions. It considers the regulatory drafting process within agencies as well as the ways political and judicial monitors themselves can package and split rules. Basic empirical analysis suggests that independent and executive agencies bundle differently, perhaps reflecting the meaningful influence of presidential review coordinated by the Office of Information and Regulatory Affairs (OIRA). In addition, executive agencies appear to bundle slightly more under divided government; that is, when at least one house of Congress is of a different political party. Elections also seem to have an effect in favor of less bundling.
Finally Part III acknowledges concerns that regulatory bundling raises about political accountability, public participation, and legislative fidelity. Regulatory bundling may allow agencies to overwhelm political and judicial overseers, as well as to short-circuit the notice-and-comment process. Our analysis indeed suggests that agencies bundle more subjects into final, as opposed to proposed, rules. At the same time, this Part recognizes that bundling yields benefits as well and recommends further empirical work to assess the trade-offs. In doing so, it highlights the possibilities and pitfalls of regulatory single-subject rules and the functional...