Jonathan H. Choi
University of Minnesota Law School
Wednesday, September 30, 2020
8:30 to 9:45 AM Pacific Time
A huge literature contemplates the theoretical relationship between judicial deference and agency rulemaking. But surprisingly little empirical work has studied the actual effect of deference on how agencies draft regulations. As a result, some of the most important questions surrounding deference—whether it encourages agencies to focus on policy analysis instead of legal analysis, its relationship to procedures like notice and comment—have so far been dominated by conjecture and anecdote.
Because Chevron applied simultaneously across agencies, it has been difficult to separate its specific causal effect from other contemporaneous events, like the rise of cost-benefit analysis and the new textualism. This Article contends with this problem by exploiting a unique event in administrative law: the Supreme Court’s 2011 decision in Mayo, which required that courts apply Chevron deference to interpretative tax regulations. By altering the deference regime applicable to one specific category of regulation, Mayo created a natural experiment with a treatment group (interpretative tax regulations) and a control group (all other regulations).
This Article uses natural language processing and various statistical methods to evaluate the causal effect of Chevron deference. These techniques allow the Article to analyze a dataset of 66,691 regulations in a transparent and replicable manner. The Article finds that, after Mayo, the IRS shifted its explanations for new rules to focus more on normative policy concerns and less on statutory interpretation. These results are statistically significant and large in magnitude: a 74.2% increase in language discussing normative goals (like efficiency and fairness) and a 56.9% decrease in language discussing the underlying statute.
In addition, the Article introduces a new theoretical model in which greater judicial deference encourages agencies to exert more effort in following rulemaking procedures. It hypothesizes that agencies counterintuitively view greater procedural effort as the “price” of judicial deference, a price that is more worth paying when courts are more deferential. Empirical analysis supports this hypothesis, finding that the move to Chevron deference caused a 22.8% increase in the length of regulatory preambles and a 36.2% increase in preambles’ discussion of public comments.
These results cast new light on the debate over Chevron, suggesting that Chevron makes agency rulemaking more detailed and policy-focused. This raises the stakes of a potential Chevron reversal and clarifies the arguments of its supporters and critics. Ultimately, this Article underscores the importance of judicial deference regimes in shaping agency behavior.