Inspired by Lisa Bressman and Abbe Gluck’s pioneering empirical study on how congressional staffers approach drafting statutes, I spent months in 2013 surveying federal agency rule drafters on how they interpret statutes and draft regulations. Among its many methodological limitations, my study focused exclusively on agency rulemaking. Agencies, of course, interpret statutes in a variety of other regulatory contexts, including adjudication, enforcement, guidance, permitting, and monitoring—just to name a few. Indeed, it’s fair to say that most agency statutory interpretation takes place outside of the rulemaking context.
Thanks to an exciting new voice in administrative law, we now have some more light shining into this black box of agency statutory interpretation. In An Empirical Examination of Agency Statutory Interpretation, Amy Semet turns her attention to agency adjudication and, in particular, how the multi-member National Labor Relations Board (NLRB) approaches statutory interpretation.1 To conduct this empirical assessment, Semet reviews more than 7,000 cases that the NLRB heard over two dozen years (1993-2016) and three presidential administrations (Clinton, Bush 43, and Obama). This study provides a treasure trove of insights into agency statutory interpretation. Here are a few of the highlights:
First, as a preliminary matter, whereas Semet read more than 7,000 NLRB decisions, very few actually dealt with statutory interpretation. In fact, the number is 121 majority decisions, or 2% of all decisions in the dataset. Semet explains (Pp. 26-28) why this figure might be underinclusive. Even if somewhat underinclusive, however, it may come as a surprise to many administrative law scholars—though probably not many regulatory lawyers—that only a small part of the agency’s adjudicative docket actually implicates novel questions of statutory interpretation. The bulk of the NLRB’s docket merely applies (or at least purports to apply) existing precedent to new facts and circumstances.
Second, the NLRB is a textualist interpreter, but perhaps in name only. Of the 121 majority decisions implicating statutory interpretation, the NLRB engages in some sort of textualist analysis roughly two thirds of the time. Yet, Semet codes (P.32) the statutory text as the “most determinative factor” in less than one in ten cases (8%). Indeed, the NLRB uses non-textualist methods about 95% of the time, including reference to legislative history in roughly two in five majority decisions. Similarly, the NLRB cites policy or practical considerations in 90% of the majority decisions. These findings are consistent with the growing scholarly call for federal agencies to engage in more purposivist statutory interpretation than their judicial peers. One may expect, moreover, that Republican and Democrat NLRB members rely more on textualism and purposivism, respectively, than their political counterparts. Semet finds (Pp. 33-34) some descriptive evidence of this, but none of it is statistically significant.
Third, precedent matters. When I teach 1L legislation and regulation, I often focus somewhat myopically on the distinction between textualism and purposivism and the various tools of statutory interpretation utilized by either/both interpretive schools. Within that interpretation toolbox, I seldom focus on precedent. Yet precedent drives so much of statutory interpretation. That is certainly true at the NLRB. Semet codes (Pp. 32-33) nine in ten NLRB majority decisions as viewing case law as “determinative or influential in the outcome.” Most of the time the NLRB cites its prior decisions, but three in four NLRB majority decisions also cite at least one Supreme Court decision that influences the agency’s interpretation. In one in ten cases, the NLRB “mirrors” a circuit court’s statutory interpretation.
Fourth, legislative history is a double-edged sword. A regular textualist criticism of the use of legislative history is that it encourages the interpreter to advance a broader interpretation of the statute than the plain text would allow. That observation is not without force. But, as Semet finds (Pp. 47-48), legislative history can also narrow the scope of the statutory text. Indeed, in most of the cases where the NLRB relied on legislative history, it did so to more narrowly interpret the statute.
Finally, Republicans and Democrats do not usually agree. Although statutory-interpretation cases make up less than 2% of Semet’s database, when a case does involve statutory interpretation, it usually sparks a dissent—oftentimes a very long dissent. In about three in four NLRB statutory-interpretation cases (77%), the majority decision is accompanied by a dissent. Indeed, Semet finds (P.56) the rate of dissenting has increased from 76% in the Clinton Administration to nearly 90% in the Bush 43 and Obama Administrations. Semet exhaustively details (Pp. 56-65) how the interpretive approaches differ in these various “dueling” opinions, and that section of the article (Part II.D) is definitely worth a close read.
What should we make of these findings?
In Part III of the article, Semet joins the scholarly call, discussed above, for agencies to engage in more-purposivist interpretation to leverage their comparative policy expertise. Having read two dozen years of NLRB decisions, she also seems pretty dissatisfied with statutory interpretation being advanced through adjudication, arguing that rulemaking may be the better vehicle. (As I explore elsewhere, similar arguments could be made for affording Chevron deference only to interpretations promulgated via rulemaking, and not to those advanced in adjudication.)
Semet is rightly cautious about drawing broader conclusions about agency statutory interpretation from her study. After all, while this dataset is impressive, it sheds light on only one agency adjudicative system, and a multi-member commission at that. Much more empirical work needs to be done to understand how agencies approach statutory interpretation in their regulatory activities.
That said, Semet’s study is an important contribution to this literature and has broad implications for administrative law and regulatory practice. Understanding how federal agencies interpret statutes is critical for, among other things, assessing the principal-agent relationship between Congress and agencies and informing how courts should subsequently review such agency statutory interpretations. I look forward to reading more from this promising new voice in our field.
- Fun Fact: Semet was one of two research assistants who worked on the Bressman-Gluck study, assisting with data organization and statistical analysis of the survey responses from congressional staffers.