Behind the Scenes of Regulation: Measuring Business Preferences Across Two Lobbying Venues in U.S. Federal Rulemaking

Federal administrative rulemaking has become a major venue for regulatory policymaking in the United States, and firms invest substantially in influencing it. Empirical research on business influence over federal rulemaking has focused primarily on the public commenting procedure. There is also an emerging literature on a second venue: the Office of Information and Regulatory Affairs (OIRA), part of the Executive Office of the President, which reviews significant federal rules and is required to grant meetings to outside parties on request during review. This lets interest groups lobby the same rules in a much less transparent manner and to address directly the office that sits above agencies in the institutional hierarchy. Given that each venue offers distinct advantages, interest groups would be expected to use both. However, the literature has examined these venues separately, or compared formal and informal channels within a single institution, but not lobbying across the agency and OIRA institutional divide. In addition, the literature in this field often assumes that regulated business industries are inherently anti-regulatory and analyzes them as a single bloc, opposed to other types of organizations. While it is widely recognized that business interests are not homogeneous, the empirical rulemaking literature analyses participation in a way that cannot observe conflict within this group. This paper addresses both limitations. I assemble a dataset linking OIRA review and meeting records with proposed rules, final rules, and public comments, covering 318 proposed-final rule pairs and 95258 comments. Using generative AI, I extract authorship and organization type from comment text, match commenters to OIRA meeting attendees, and code the regulatory stringency positions of business interests participating in both venues. Three findings follow. Dual-venue lobbying is widespread: in 69.8% of rules, at least one organization both commented and attended an OIRA meeting, and this is not driven by a small set of the most active lobbyists - 74.1% of dual-venue business organizations appear in only one rule. The composition of participants differs significantly between venues: individuals are the largest group of commenters but appear in no OIRA meetings, while business interests account for more than two thirds of dual-venue organizations. Finally, business preferences are neither uniform nor uniformly opposed to regulation: 62.6% of coded comments request looser requirements against 3.1% requesting stricter ones, yet 38.5% of rules show divergence among commercial interests and 7.6% show directly opposing preferences. These results establish unequal representation across the two venues and document within-business conflict. Two case studies illustrate the divergence in the underlying comment text. The paper is work in progress: extending position coding beyond dual-venue participants and measuring effects on regulatory outcomes remain for future work.