In Loper Bright Enterprises v. Raimondo, the Supreme Court will address whether it should replace the Chevron standard of review. I want to pick apart a particular passage of the petitioners’ merits brief:
More troublingly, Chevron has seriously distorted how the political branches operate. Thanks to Chevron, Congress does far less than the Framers envisioned and the executive branch does far more, as roughly half of Congress can count on friends in the executive branch to tackle controversial issues via executive action without the need for compromise, bicameralism, or presentment.
Whether Chevron has decreased congressional productivity and increased delegation to the executive branch is an empirical question. In fact, petitioners’ brief frames it as a causal relationship: An observed decline in congressional productivity is “[t]hanks to Chevron” rather than some preceding or contemporaneous trend. Several of the amici make similar claims.
For these claims to be true, several trends must hold. First, congressional productivity must have decreased post-Chevron. Second, Congress must have increased delegation to the executive branch post-Chevron. Third and finally, Chevron must have caused both trends. Although these trends are empirically observable, petitioners do not cite any studies, facts, or figures to support their claims. Meanwhile, existing empirical scholarship suggests otherwise.
First, Chevron did not cause a decline in congressional productivity. To be clear, Congress has passed fewer statutes since 1963—a trend preceding Chevron by twenty years. But the rate of decline slowed after 1984. Moreover, the number of statutes passed is a poor measure of productivity. It says little about whether Congress is engaging in meaningful lawmaking. Indeed, Congress has passed longer and more complex regulatory statutes since the Court’s decision in Chevron. Statutes also have become more specific rather than more ambiguous (Farhang 2021). According to these metrics, Chevron does not appear to have caused a decline in congressional productivity (see also Baumann 2023).
Even if we believe that congressional productivity has declined (perhaps by some other metric), petitioners still need to show that this decline is attributable to Chevron and not another contemporaneous trend. American politics has undergone rapid transformation during the last forty years. Congress struggles to legislate during periods of increased polarization (Binder 2014, Krehbiel 1996), and polarization has intensified since the mid-1970s (Lee 2015). The number of salient issues on Congress’s agenda has steadily increased since 1987 (Binder 2015). Rising fragmentation within parties may stymie lawmaking even during periods of unified government (Lee 2018). Polarization and partisan fragmentation existed before Chevron; these trends will continue with or without Chevron. Any perceived decline in congressional productivity is likely attributable to broader political trends rather than Chevron.
Second, Chevron did not cause Congress to delegate more often. Obviously, Congress delegated policymaking authority to agencies prior to Chevron. In his 1929 book, Legislative Functions of National Administrative Authorities, John Preston Comer writes, “It is an erroneous supposition that the great mass of administrative legislation now playing such an important part in the government of this nation sprang full-grown out of the complex conditions of modern life. If, for purposes of convenience, one divides the history of national law making roughly into four periods, . . . he will find that Congress has departed all along the way from its general rule of issuing laws complete in themselves and effectual without the interposition of some other will than its own.” The only cases to have invalidated a statute on nondelegation grounds were decided in 1935. Congress has consistently delegated policymaking authority to administrative agencies since the 1940s (Clouser McCann and Shipan 2021, Farhang 2021). The rate of delegation did not significantly increase post-Chevron (Farhang 2021). In fact, according to one study, the rate of delegation has actually declined since the 1990s (Clouser McCann and Shipan 2021). In theory, Chevron may have made delegation more attractive to Congress (Gluck and Bressman 2014). Empirically, the rate of delegation has not increased as a consequence of Chevron, and replacing the standard will not reverse Congress’s pre-Chevron trend of delegation.
Nor does Congress appear to award agencies unfettered discretion. Since 1984, the United States has only experienced ~12 years (30.8%) of unified government. A lengthy literature in political science shows that legislatures respond to divided government by increasing specificity within statutes and lowering levels of discretion. Indeed, Congress has increasingly constrained agency authority since the 1990s (Farhang 2021). Congress provides greater opportunities for judicial review in statutes with higher levels of agency discretion (Clouser McCann, Shipan, and Wang 2023). At times, Congress itself struggles to oversee executive action (Bolton and Thrower 2021). But Congress pairs its delegations with external mechanisms intended to protect the boundaries of agency authority, such as increased provisions for judicial review.
I do not mean to suggest that Chevron has had no effect on Congress, agencies, or the courts. That would be an equally unfounded claim. Chevron results in higher agency win rates than Skidmore (Barnett and Walker 2017, Eskridge and Baer 2008). Congressional drafters recognize Chevron and intend for agencies—not courts—to interpret statutes (Gluck and Bressman 2014). Agency rule drafters are familiar with Chevron, and doctrinal shifts influence the rulemaking process (Choi 2021, O’Connell 2011, Walker 2015). Within the judicial branch, Chevron may constrain partisan decision-making (Barnett, Boyd, and Walker 2018, but see Cross and Tiller 1998). Other briefs in support of petitioners have used some of these studies to support their claims. A faithful interpretation of the empirical record produces mixed evidence for both Chevron’s supporters and opponents.
To understand the potential consequences of replacing Chevron, we must be honest about what Chevron has and has not done. Chevron is just a standard of review (Hickman and Hahn 2020). Chevron does not prevent Congress from delegating policymaking authority to federal agencies. Chevron will not eliminate the incentives that members of Congress have to delegate that authority (regardless of whether they delegate to avoid accountability or harness agency expertise) (Fiornia 1982, Gailmard 2002). As long as Congress delegates, agencies will interpret statutes. Sometimes courts will find clear answers to the meaning of these statutes. Other times, however, courts will exhaust the interpretive canons and find that they must make a policy decision. And, in those cases, courts may adopt the agency’s reasoning as its own—whether or not Chevron remains good law (Bednar and Hickman 2017).
We can only speculate about what will happen if the Supreme Court jettisons Chevron. “[A]bsent empirical evidence, we cannot know whether we are doing more harm than good in fiddling with Chevron’s structure” (Bednar 2019). Petitioners do not need claims about congressional productivity to win at the Supreme Court. Plenty of legal, empirical, and normative arguments exist for why Chevron does not comport with the Constitution or existing law (e.g., Bamzai 2017). But we should be concerned by the ease with which these dubious claims were asserted by the petitioners and some of the amici. We could dismiss these statements as rhetorical flourish but, on some level, the questions asked by the Supreme Court are serious ones, deserving of care and candor rather than hyperbole.
More broadly, we should be worried that unsupported empirical claims will unmoor administrative law from the realities of governance. As political scientist Gary King warns,
In presidency research, we have the luxury (and drudgery) of knowing that many of our recommendations will not be implemented. Nonetheless, prescriptions without adequate judgments of uncertainty are just as irresponsible. If we are listened to at some point, as we occasionally are, improper uncertainty estimates might cause policy makers to act too early. . . . Prior to making prescriptions, we should be asking ourselves whether we are willing to risk the unintended or unknown consequences of proposed institutional reforms.
Advocates before the Supreme Court do not have the luxury of deaf ears. The Court will consider their arguments in issuing its decision. It is one thing to lack evidence and speculate about the potential consequences of a proposed doctrinal change; it is quite another to recite unsupported empirical claims as certain in furtherance of that change.
With this blog post, my audience is not practitioners per se. I also do not believe that the Supreme Court bar is particularly concerned with overly embellished statements about congressional productivity. However, I think legal scholars in particular should think carefully about how they present empirical claims and what obligations they have when they make certain assertions. As Daniel Walters recently tweeted, “It’s not too late to demand that administrative law meet the epistemological standards of every other field of knowledge.” I do not purport to set forth what that standard is or should be. Yet I believe we would benefit from more frequent conversations about the methodology of administrative law. At the most rudimentary epistemological level: How do we know what we know?
Our field derives its credibility from the strength of our arguments. I do not believe that legal scholars have an obligation to pursue archival, qualitative, or quantitative sciences in lieu of traditional legal methods. And I believe scholars are perfectly capable of challenging the conclusions of empirical studies without themselves engaging in empirical research. When we find ourselves unsatisfied with the evidence presented in the literature, we should feel comfortable discarding existing work and presenting an alternative narrative. Doing so requires us to acknowledge the empirical research that conflicts with our own position. Admittedly, we cannot read everything. But, at the very least, we ought to raise an eyebrow when someone makes, without citations, an empirical claim that is the subject of dozens of articles and books. If we allow ourselves to accept loose empirical claims as fact, we weaken our arguments and our credibility as scholars.
Nicholas R. Bednar is an Associate Professor at the University of Minnesota Law School. His research interests include administrative law and bureaucratic politics. The Author thanks Kristin Hickman for her comments.
 I focus here on mechanisms of oversight that Congress imbues within the organic statute (McCubbins, Noll, and Weingast 1987). In general, Congress selects a system of controls to ensure that the agency protects the enacting coalition’s preferences (McCubbins, Noll, and Weingast 1989). Alternatively, Congress could periodically review and amend delegations of authority. Congress rarely engages in this practice. As Chris Walker recently explained on Twitter: “[T]here’s a pretty big empirical assumption build into the assumption—i.e., in a world without Chevron deference, Congress would change its behavior and regularly revisit statutes that govern federal agencies. In other words, you can express concern about Chevron, as Jonathan [Adler] and I have argued, because it allows agencies to use older statutes to do new things the enacting Congress and perhaps a majority of the current Congress might not approve. But at the same time concede that getting rid of Chevron really wouldn’t move the congressional needle in terms of incentives—at least in any meaningful way.”
 To be clear, I am not endorsing these arguments. I would argue that the Supreme Court should reaffirm Chevron’s two-step framework as the appropriate standard of review in Loper Bright. But I acknowledge that many other scholars have made compelling arguments to the contrary.