What to Do about Chevron - Nothing

Nicholas R. Bednar*

For thirty-five years, doctrinalists have tormented themselves trying to dissect the Supreme Court's most infamous administrativelaw doctrine: Chevron deference.' We have asked when and how it applies.2 At the same time, we have asked whether Chevron should exist at all.3 In other words, does Chevron have any normative advantages that warrant its continued existence and prolific use? Despite thirty-five years to work out our differences, the academy-and the courts-remain torn on the answers to all of these questions.

Across campus, a different line of inquiry has emerged. Political scientists propose attitudinal and strategic models that assume the ideology and facts of a case institutional structures to produce a result consistent with the judge's preferences.4 Over these models have become more sophisticated. James Gibson writes, "judges' decisions are a function of what they prefer to do, tempered by what they think they ought to do,

interact with the judge's political

the years,

but constrained by what they perceive is feasible to do."5 Scholars writing in the field of judicial decisionmaking have argued that judges must reconcile their preferences with precedent,6 the risk of reversal by higher courts,7 and the risk of interference by other actors (i.e., Congress and the president).8 Yet judges may act strategically within these institutional constraints to maximize ideological preferences.9

Attitudinal models respond dismissively to the concerns of doctrinalists: Chevron applies when judges benefit from its application. Judges defer to interpretations they agree with and reject those that they disagree with. Chevron will persist so long as it produces results that at least five of the Supreme Court justices agree with. To the extent that we observe judicial restraint when lower courts apply Chevron, it stems from a fear of reversal rather than a rule-of-law belief in methodological consistency.10 For the doctrinalist, these answers are overly cynical.

Does

reality reflect

is complicated. In a series of articles, Kent Barnett, Christina Boyd, and the most comprehensive Christopher Walker" have conducted

these attitudinal predictions?

It

empirical study of Chevron to date.12 In Administrative Law's Political Dynamics, the authors challenge the assumptions of the attitudinal models by testing whether Chevron restrains political ideology in the review of agency statutory interpretations. They conclude that Chevron effectively and powerfully restrains political ideology relative to other standards of review, such as the Skidmore standard13 or de novo review. Yet, some attitudinal observations linger. First, when Chevron does not apply, the panel's political ideology appears to have a strong influence on whether the court agrees with an agency statutory interpretation.14 Second, political ideology influences whether panels apply Chevron, suggesting that some panels strategically avoid its application.15 Third and finally, both liberal and conservative panels are more likely to conclude that a statute is unambiguous when the agency statutory interpretation does not align with the panel's political ideology.16

On the one hand, Chevron restrains political ideology more effectively than any other doctrine-including de novo review. On the other hand, political ideology still plays some role in the way courts apply Chevron. In their conclusion, the authors pose a number of questions about Chevron's future, namely how do we reform it in order to better achieve its goal of restraining political judging. In this Response, I offer one possible answer: stop messing with it until we have greater evidence about how these reforms will affect its performance.

is premised on normative assumptions about administrative law. First, Congress expects agencies to interpret some

Chevron

sameness of applications."23 In his Administrative Law Treatise, Kenneth Culp Davis observed:

Davis is undeniably correct that Chevron's verbiage is "vague, abstract, uncertain, and conflicting." The opinion in Chevron itself is opaque,25 in part because Justice Stevens did not view it as a departure from traditional doctrines of administrative law.26 Throughout the decades, the Supreme Court has provided minimal guidance on the application of Chevron's two steps. As a result, Chevron has organically evolved as judges applied it to new situations, using different tools of statutory interpretations and understandings of "reasonableness."27 The restatement of Chevron's canonical two steps in the "standard of review" section of every circuit court opinion does little to inform the reader about how the panel will apply Chevron in that case.

Beyond the verbiage, however, legal realists-and attitudinal scholars-argue that judges do not take standards of review seriously in application. Rather, they opine that courts use deference to justify preferred outcomes.28 If so, Chevron should exhibit little restraint of judges' political ideology.

Before continuing,

the definition of "ideology" presents a quandary in the legal context.29 To legal academics, "conservative" ideology may refer to judicial restraint, textualism, and originalism;30

traditionally

"conservative" meaning values

"liberal" ideology may refer to judicial activism, purposivism, and living constitutionalism.31 Here, the authors and I refer to political ideology embraced by as Republicans and "liberal" meaning values traditionally embraced by Democrats. The authors use Judicial Common Space ("JCS") scores to measure political ideology. JCS scores impute the ideological scores of the president and the judge's home senators on the judicial nominees to estimate the judge's political ideology.32 A judge has a liberal political ideology if she has a score between -1 and 0; a judge has a conservative political ideology if she has a score between 0 and 1. JCS scores are an imperfect measure, but finding alternative exogenous measures for every federal judge is nearly impossible.

judicial decisionmaking,

Returning to the issue of whether standards of review seriously the authors' empirical evidence inform provides a resounding answer of "yes." In fact, the authors' data shows that Chevron is remarkably successful at curtailing the influence of political ideology. Table 1 reports the predicted probabilities that the most extreme liberal or conservative panel agrees with an agency statutory interpretation. The data supports the authors' hypothesis that Chevron acts as a structural force that constrains the ability of panels ideological preferences on judicial decisionmaking.

impart their own

this distribution.36 The average panel leans slightly conservative (median JCS = 0.067; mean JCS = 0.069). Nevertheless, approximately two-thirds of the panels have an average JCS score between -0.154 and 0.291.37 Accordingly, the average panel is rather moderate. In fact, the probability that a liberalpanel has an average JCS score of greater than -0.4 is 0.018 (1.8%) and the probability that a conservative panel has an average JCS score of greater than .4 is 0.068 (6.8%). The chance that we observe a panel as extreme as the ones used to illustrate the model is low.

8-

8-

How do these more moderate panels differ from their more extreme peers? Using the authors' model and predicted probabilities calculations,38 I estimate predicted probabilities for a moderate liberal panel (JCS = -0.154) and a moderate conservative panel (JCS = 0.291) to capture the likely outcomes in two-thirds of cases. For purposes of comparison, I also estimate predicted probabilities of a neutral panel (JCS = 0).39 Table 2 reports the results.

As expected, moderate panels agree with agency interpretations of the opposing ideology more readily than extreme panels. Chevron performs well at washing out disparities between moderate liberal and moderate conservative panels. When Chevron applies, the model predicts that liberal panels agree with liberal interpretations in 3% more cases than a neutral panel, while conservative panels agree with liberal interpretations in 8% fewer cases. Likewise, the model predicts that conservative panels agree with conservative interpretations in 4% more cases than a neutral panel, while liberal panels agree with conservative interpretations in 3% fewer cases. The Chevron we most often observe has the effect of reducing the influence of political ideology to mere percentage -point differences.40

Consistent with

the Supreme Court's hope, Chevron successfully curtails the influence of judges' own political preferences in review of agency statutory interpretations. Chevron performs far better at reducing ideological influence than either Skidmore or de novo review. Still, the rule of law necessitates that judges decide cases using neutral principles of law rather than personal preferences.41 We observe a minor influence of political ideology among moderate panels applying Chevron but that influence grows as the panels grow more extreme.

How can the Supreme Court and the circuit courts reform Chevron to reduce any traces of political ideology in judicial decisionmaking?

The remainder of this Response argues that the appropriate answer to this question is that the Supreme Court should stop fiddling with Cheron-at least until we have evidence about how reforms affect the delicate balance Chevron strikes. Chevron accomplishes its goal of restraining the influence of judges' own personal preferences. Skidmore and de novo review do not accomplish this goal. The constant tweaking of Chevron risks creating so many exceptions that judges may always reliably find a way to avoid applying Chevron. Until the Supreme Court knows how a particular change will affect Chevron's ability to restrain political ideology, Chevron is best left alone.

I want to avoid overstating my thesis. I shiver at the thought of leaving the impression that judges behave in blatantly political ways. For the most part, I believe that judges have a strong sense of rule of law that constrains them from acting politically.42 Moreover, to the extent that political ideology may subconsciously inform a judge's decision, those cases are few and far between. Law and facts come together often enough to present clear-ish solutions. 43 In a number of "hard" and "very hard" cases where doctrine and interpretive methodology support both sides equally, political ideology may tip the scales in favor of the judge's personal preferences.44 To the extent political ideology creeps into the Chevron analysis, it comes in unconsciously, without malice, and infrequently.

If the Supreme Court is to improve Chevron's ability to constrain political ideology, the Court must know what about Chevron requires reform. There are two possibilities. First, courts apply the substance of Chevron's two steps in a way that promotes some interpretations while

penalizing others. Second, courts strategically apply Chevron in order to avoid deferring to the interpretations that run contrary to their political ideology. This Part discusses the former; Part III discusses the latter.

If political ideology enters Chevron through its two steps, which step presents the most concern? At step one, the court asks whether Congress has "directly spoken to the precise question at issue."45 In other words, to interpret unambiguous? At step two, the court asks whether the agency's interpretation is "reasonable."46

the agency purports

the statute

is

The authors' results tell us that step one presents the greatest concerns. Ideological differences play, at best, a marginal role in the application of step two.47 At step one, however, political ideology appears to influence the likelihood that a panel finds the statute unambiguous. Conservative panels are more likely to conclude that the statute is clear when presented with a liberal interpretation, whereas liberal panels are more likely to conclude that the statute is clear when presented with a conservative interpretation.48

That step one presents the greatest concern should come as no surprise to administrative law scholars.49 First, step one is highly determinative of whether the court defers to an agency statutory interpretation. In an earlier study, Barnett and Walker find that courts end the Chevron analysis at step one in 30.0% of cases and agree with

interpretation in only 39.0% of those cases.50 By the agency's comparison, when courts reach step two, they refuse deference in only 6.2% of cases despite previously concluding that the statute is ambiguous.51 Second, step one has the most potential to ensure that a judge's position sticks. If the court concludes that the statute is unambiguous, the court's interpretation.52 Accordingly, judges may conclude that the statute unambiguously requires the judge's preferred interpretation, thereby preventing the agency from ever adopting a contrary one.

forever bound by

agency

the

is

How do we reform step one to prevent political biases from influencing courts' assessments of whether a statute is ambiguous? Step one has two doctrinal features that determine its application: (1) the meaning of "clarity," and (2) the tools of interpretation that courts use to determine whether a statute is unambiguous. Neither has suitable solutions to this problem.

A. How Clear is Clear?

When is a statute sufficiently clear to foreclose interpretation of the statute?53 Ambiguity is an amorphous concept.54 Some judges will "find ambiguity in a stop sign";5 others argue that they have never found a statute ambiguous.56 If we are concerned about political ideology influencing the application of Chevron, the definition of "clarity" seems like a reasonable suspect. Judges have significant discretion in setting the bounds of the "zone of ambiguity."57 Indeed, the authors find that conservative panels discriminate against liberal interpretations when deciding whether a statute is unambiguous. Liberal panels do the same to conservative interpretations.

In order to set the requisite level of clarity, the Supreme Court would have to identify the boundaries of the "zone of ambiguity" that to governs at step one. However, "ambiguity"

is not susceptible

The difficulty of setting uniform standards for "clarity" has not prevented Justice Gorsuch from trying to push the Supreme Court toward a "clear enough" standard.61 Other textualists, including Justice Kavanaugh, have also expressed desires for a more searching step one inquiry.62 Of course, a more searching step-one inquiry provides greater room for a judge to find that the statute unambiguously requires her preferred statutory interpretation.

interpretation

contrary

agency

over

a

Moreover, changing the level of clarity does not rectify the ideological imbalance. If the Supreme Court decreases the rigor with which courts assess clarity, liberal panels will more often find that the statute does not prohibit a conservative interpretation but will also more often find that the statute does not prohibit a liberal to all agency statutory interpretations interpretation. Deference increases but the ideological imbalance remains. A similar result holds for conservative panels' review of agency statutory interpretations. Finally, if anything, step one requires a more rigorous search for clarity.63 Increasing the rigor of step one would not resolve the imbalance but would further decrease the likelihood that panels defer to ideologically-opposed interpretations.

If Chevron has a problem with political ideology, the clarity question of step is partially to blame. However, adjusting the standards of "clarity" and "ambiguity" is not a feasible solution to preventing this political influence.

B. Interpretive Tools at Step One

Alternatively, the Supreme Court could prevent judges from using certain tools of statutory interpretation at step one. The political ideology of judges correlates strongly-although not perfectly-with the interpretation. they use interpretive methodologies Conservative judges tend to favor textualism and liberal judges tend to favor purposivism. Each methodology offers a different toolkit. Textualists search for the statutory text's ordinary meaning using textual canons and dictionaries.64 Purposivists search for the statute's purpose using legislative history and substantive canons.65 The strong correlation between political ideology and interpretive methodology likely explains some of the appearance of ideological behavior in the authors' findings.66

in statutory

1. The Agency Interpretation Hypothesis

First, like their judicial counterparts, liberal agencies may use purposivism to form and defend their interpretations. Conservative agencies may use textualism for the same purposes. If so, judges may naturally gravitate toward those interpretations that comport with their preferred interpretive methodology. As a result, liberal judges agree more often with liberal interpretations and conservative judges agree more often with conservative interpretations.

that agencies consistently apply a single

The trouble with this first hypothesis is that we have little interpretive evidence methodology. The study of internal agency statutory interpretation is new but burgeoning.67 In a groundbreaking survey of agency rule drafters, Christopher Walker observes both consensus and variation in drafters' use of interpretive tools.68 He finds that the vast majority of drafters use legislative history (76%) but there is far more variation in (25%-79%), and the use of dictionaries substantive canons (13%-47%).69 The high use of legislative history that many agencies use purposivism.70 In qualitative suggests responses, sympathized with conservative interpretive methodologies suggested that legislative history is a more valuable tool for agencies than courts. A forthcoming study by Amy Semet demonstrates that the NLRB does not consistently employ one interpretive methodology its adjudications; Republican and Democratic appointees both use purposivist and textualist tools.71 Accordingly, the empirical literature does not suggest that agencies sort neatly into purposivist and textualist camps based on political leanings. Nor should we expect that agencies and courts interpret statutes in the same way. In announcing its decision in Chevron, the Supreme Court envisioned that the standard would apply to interpretations that "really center H on the wisdom of the agency's policy."72 By this logic, an

some drafters who

agency should guide its interpretation of a statute by policy and expertise rather than neutral principles of statutory interpretation.73 Moreover, each agency has its own mission, culture, and structure that informs how it interprets statutes and shapes policy.74 Even assuming that political ideology affects an agency's choice of interpretive methodology, it is just one of a number of factors that influences that decision.

Moreover,

this hypothesis assumes

that courts do not meaningfully examine contextual factors when reviewing an agency statutory interpretation. In Barnhart v. Walton, Justice Breyer stated that courts should examine "the interstitial nature of the legal question, the related expertise of the Agency, the importance of the question to the administration of the statute, the complexity of that administration, and the careful consideration the Agency has given the question over a long period of time" when deciding whether to review the agency's interpretation under Chevron.75 Barnett and Walker's empirical study demonstrates that circuit courts reference agency expertise (18.4% of cases), the longstanding nature of the interpretation (10.7% of cases), contemporaneity (1.9% of cases), and public reliance (0.7% of cases) in their Chevron analyses.

Beyond the factors referenced in Barnhart, other agency traits influence the court's willingness to defer to the agency's interpretation. The subject matter of the agency statutory interpretation also has a significant effect on the willingness of courts to accord deference. Circuit courts generally accord high levels of deference in cases involving telecommunications, Indian affairs, and pensions, but low levels of deference in cases involving civil rights, housing, and prisons.76 But the agency's political ideology and subject-matter jurisdiction alone cannot explain some disparities. Both the NLRB and the EEOC handle labor and employment policy, and both are among the most liberal agencies.77 The NLRB receives some of the highest rates of deference but the EEOC receives the lowest.78 The skill of the agencies' workforces may offer one explanation for this disparity: the NLRB is perceived as

significantly more skilled than the EEOC.79 All of these contextual factors suggest that courts consider more than just the agency statutory interpretation when deciding whether to accord Chevron deference to the interpretation.

interpretive methodology and whether

Absent more studies on the interpretive methodologies of agencies, it is difficult to assess whether agencies consistently use a the use of these single methodologies fall along liberal/conservative lines. Early evidence suggests that agencies interpret statutes differently than courts. Even if agency same interpretive methodologies do fall along liberal/conservative lines as the courts, other evidence suggests that the agency's interpretive methodology alone does not drive deference rates. Given the current literature, the Agency Interpretation Hypothesis seems too tenuous to explain the ideological imbalances present in Chevron step one.

the

2. The Political Leanings of Interpretive Tools Hypothesis

The second hypothesis is that purposivist tools produce liberal results and textualist tools produce conservative results. If true, the Supreme Court could forbid courts from using tools with an overtly political bend at Chevron step one.80 Current empirical evidence is insufficient to support this hypothesis.81

If interpretive tools produce ideological results, purposivist tools should produce liberal results. In a study of the Supreme Court justices' use of legislative history from 1953 to 2006, David S. Law and David Zaring find "no statistically significant relationship between whether an opinion cited legislative history and whether the opinion arrived at a liberal or conservative result."82 Other scholars conclude that legislative history corresponds to ideological results. Frank Cross finds that the use of legislative history is associated with more liberal outcomes even when conservative justices employed legislative history, but the effects are "quite small."83 Examining the Supreme Court's use of legislative history in employment cases from 1969 to 2006, James

Brudney and Corey Ditslear find that the justices reach conservative results more often in opinions using legislative history relative to opinions using textualist tools.84 Accordingly, the literature presents inconclusive findings about whether legislative history produces liberal, conservative, or neutral results.

Empirical evidence also fails to support that substantive canons produce liberal results. Brudney and Ditslear find that the Supreme Court's application of substantive canons did not produce liberal decisions more often than conservative decisions.85 Rather, justices use substantive canons at higher rates when the outcome aligns with their similar own political conclusions, but further finds that the Supreme Court rarely invokes these canons anyway.86 Although judges may use substantive canons to obtain ideological results, substantive canons do not categorically produce liberal results.

ideology. Anita Krishnakumar

reaches

Likewise, studies of textualist tools demonstrate that textualism does not consistently lead to conservative outcomes. In a separate study, Brudney and Ditslear find that both liberal and conservative justices apply textual canons at similar rates in employment law cases.87 Moreover, the application of textual canons does not produce more conservative outcomes than liberal outcomes.88 Cross observes that conservative justices use textualism more frequently than liberal justices. However, for six of nine justices, the use of textualism is greater in opinions with liberal outcomes, suggesting that textualism does not necessarily have a conservative or liberal bend.89

These studies tell us two things. First, liberal judges tend to use more purposivist tools and conservative judges tend to use more textualist tools. However, they also frequently use tools of the other interpretive methodology. Second, the use of purposivism or textualism does not cleanly correspond to either liberal or conservative outcomes. Whether a specific tool produces a liberal or conservative outcome depends on how the judge uses it.90 Purposivism serves liberal ends

when judges use it to produce liberal results. Textualism serves conservative ends when judges use it to produce conservative results.91 What does this tell us about reforming Chevron step one to judicial further decisionmaking? Simply forbidding the use of certain interpretive tools will not produce more balanced results. If judges are truly motivated to obtain political results, they can use whatever tools the Supreme Court deems admissible to achieve those results.

the exercise of political

ideology

restrain

III. CHANGES TO CHEVRON'S SCOPE

The authors' most troubling finding is that some judges choose to avoid applying Chevron when ideologically advantageous to do so.92 Courts agree with fewer agency statutory interpretations under lesser standards of review, such as Skidmore or de novo review. Indeed, Barnett and Walker observe that the court agreed with agency statutory interpretations in 77.4% of cases applying Chevron, 56.0% of cases applying Skidmore, and 38.5% of cases applying de novo review.93 These lower standards of review provide judges with more discretion to overrule the agency's interpretation in favor of the judge's own preferred interpretation. Recall from Part I that the authors' data shows substantial gaps in the willingness of panels to agree with liberal or conservative interpretations depending on the panel's own political ideology. If courts discriminate on the basis of political ideology when deciding whether to apply Chevron, Chevron's constraints of political ideology have little effect.

Fortunately, curing the influence of political ideology on judges' decisions to apply Chevron is easier than fixing its two steps. Two solutions present themselves. First, the Supreme Court should continue to clarify when Chevron applies, preferably expanding its reach in more cases to take advantage of its constraining powers. If Chevron applies in more cases, political the uncertainty left by the Supreme Court's decision in Mead v. United States leaves open the possibility that a court will apply Skidmore in place of Chevron when presented with an agency statutory interpretation it disagrees with. The authors' findings suggest that judges' political ideology influences whether they apply Chevron or Skidmore in reviewing a particular interpretation. Part III.A. examines how Mead's uncertainty contributes to political judging.

ideology drives fewer results. Yet

Second,

the Court should cease carving out amorphous exceptions to the Chevron doctrine. In the last decade, some Supreme court justices have called for Chevron's demise.94 This anti-Chevron rhetoric has resulted in further limitations on Chevron's scope, even if the Court did not intend to impose these limitations.95 Part III.B. discusses how one exception-the major-questions doctrine-threatens to provide more opportunities for political ideology to govern when judges apply Chevron.

A. Mead v. United States

its

issued

interpretation

In Mead Corp. v. United States, the Court held that courts should review an agency statutory interpretation under Chevron when the agency has the exercise of congressionally delegated authority to act with the force of law.96 Dissenting, Justice Scalia warned that the Court "will be sorting out ..for years to come."97 Scholars the consequences of the Mead doctrine . have expressed confusion about what agency interpretations are promulgated with the "force of law."98 Although the Supreme Court has provided some post-Mead clarity on Chevron's scope,99 indeterminacy about Mead's force-of-law standard remains.

The confusion surrounding Mead results in circuit splits that may appear political in nature but stem from legitimate disagreements about the meaning of the force of law. Following Mead, circuit courts

both parties envisioned that the standard was important to the case.108 The Supreme Court's silence has left some wondering whether the Court intends for Chevron to apply in the immigration context.109

The difficulty of applying Mead to agency adjudications is not limited to the immigration context.110 The Federal Circuit denies Chevron deference and Trademark Office's the Patent interpretations of substantive patent law created during the course of adjudication.' Some have proposed abandoning Chevron's application to agency statutory interpretations adopted in adjudications all together.112 The uncertainty surrounding the application of Chevron to agency statutory interpretations adopted in adjudications, and other forms of agency action, creates more potential for political ideology to sneak into the Mead framework.

The failure of the Supreme Court to provide clarity on the definition of "force of law" leaves open the possibility that those seeking to shirk Chevron for political reasons may succeed at doing so. Even when a lower court's decision to apply Skidmore or de novo review stems from a legitimate reading of Mead, the Court risks that circuit courts will split in a way that creates the appearance of impropriety. The Supreme Court ought to provide greater guidance to avoid situations where liberal and conservative panels reach different conclusions about Mead's meaning.

B. The Major-Questions Doctrine

In several cases, the Supreme Court has applied the "majorquestions doctrine" to prevent agencies from deciding interpretive questions of "political and economic significance" without express statutory approval from Congress.113 Most recently, the Supreme Court revived the doctrine in King v. Burwell. Writing for the majority, Chief Justice Roberts held that the IRS's interpretation did not warrant

consideration under Chevron because its interpretation of the Affordable Care Act involved billions of dollars and affected the healthcare plans of millions of people.114

The major-questions doctrine suffers from a similar definitional problem as Mead. When does an issue present a question of "political and economic significance"? Consider the Court's holding in King v. Burwell. Many tax provisions involve billions of dollars and affect millions of people. Yet the Supreme Court has instructed lower courts to accord Chevron deference to Treasury Department regulations.115 Kristin Hickman has argued that King's vague directive may lead lower courts to apply the major-questions doctrine in unintended ways in the tax context.116 Like Mead's directive to apply Chevron when an agency acts with the "force of law," King's directive to avoid Chevron for questions of "political and economic significance" lacks precision and risks providing judges an out from the constraints of Chevron.

Perhaps there is some benefit, as Barnett and Walker argue, to allowing both Mead and the major-questions doctrine to percolate in the lower courts.117 Maybe. The circuits managed to work out their differences regarding how Chevron should apply to BIA decisions. And for any question that becomes dire, the Supreme Court can always grant cert or Congress can always amend a statute directing courts to apply or disregard Chevron when interpreting a specific statute.

Still, the Supreme Court's anti-Chevron streak runs the risk of upsetting constraints Chevron currently offers. If the Supreme Court carves out too many exceptions, we will observe a greater amount of ideological avoidance of Chevron and more political influence in the review of agency statutory interpretations.

CONCLUSION

In 1993, political scientist and presidency scholar Gary King

made the following remark about his field:

The Chevron literature is full of prescriptions based on perceived problems and normative predictions about what should occur as a result of the changes. Courts sometimes adopt these recommendations. Yet absent empirical evidence, we cannot know whether we are doing more harm than good in fiddling with Chevron's structure.

Barnett, Boyd, and Walker's study serves as a template for empirical legal scholarship. They take a doctrine and test how it handles a normative concern in law: the influence of political ideology on judicial decisionmaking. Their results present an answer that some Chevron scholars may gasp at: Chevron actually works at restraining political ideology.

In the end, the authors leave us with a series of questions about how their study should inform future efforts to reform Chevron. My preference: leave well enough alone-for now. Chevron works and does what the Supreme Court intended it to do. Administrative law has too many salient policy questions to allow the political ideologies of judges to subconsciously sneak into judicial review. I do not believe judges intend to impart their political beliefs on cases, but the authors' evidence suggests that political influence sneaks in. All said, Chevron is an effective structural tool for restraining the possibility of political influence-whether intentional or not. Absent future research, we cannot know how reforms may affect the delicate balance Chevron strikes.

Of course, this assumes that Chevron primarily serves to constrain the effect of political ideology on judicial decisionmaking. to benefit from its Increasing the rate of Chevron's application constraining effects would come at the expense of allowing agencies to engage in more binding statutory interpretation. This may cut against the delegation theory of Chevron, which demands that courts engage in a searching inquiry to ascertain whether Congress actually intended to delegate a policymaking decision to the agency.119 I am not arguing that we ought to increase Chevron's application at the expense of its theoretical underpinnings. The delegation theory should serve an important role in future research on how Chevron can better allow

courts disrupting its ability to constrain judges' political ideologies.

legitimate congressional delegations without

identify

Chevron could do a better job. But we do not have enough empirical evidence to suggest how structural changes to Chevron may jeopardize its ability to restrain political ideology. I do not mean to suggest that Chevron ought to never be reformed. Rather, let us slow the effort to overhaul Chevron and allow the Supreme Court to provide clarity to existing doctrines. Meanwhile, we should follow Barnett, Boyd, and Walker's lead in conducting more empirical studies on (1) how administrative law does (or does not) protect normative concerns, and (2) how changes to doctrine may upset these normative concerns.

--- FOOTNOTES ---

J.D., University of Minnesota Law School; B.A., University of Minnesota; Ph.D. student, Department of Political Science, Vanderbilt University. I would like to thank Christina Boyd, Kent Barnett, Kristin Hickman, David Lewis, Christopher Piper, and Christopher walker for their indispensable comments.

1.

Chevron U.S.A. Inc. v. Nat'l Res. Def. Council, 467 U.S. 837, 842-43 (1984). In brief, Chevron deference is a two-step standard of review for determining whether a court should defer to an agency interpretation of a statute. At step one, the court asks whether the statute is clear or unambiguous. Id. If so, "the intent of Congress is clear, [and] that is the end of the matter; for the court, as well as the agency, must give effect to the unambiguously expressed intent of Congress." If not, the court asks at step two whether the agency's interpretation is "reasonable" or "based on a permissible construction of the statute." Id. at 843-44.

2.

For a literature review of the Chevron standard's many proposals, see Nicholas R. Bednar and Kristin E. Hickman, Chevron's Inevitability, 85 GEO. WASH. L. REV. 1392, 1399-1441 (2017).

3.

For a literature review of the many calls to end the Chevron standard, see Christopher J. Walker, Attacking Auer and Chevron Deference: A Literature Review, 16 GEO. J.L. & PUB. POL'Y 103, 110-18 (2018).

4.

See Jeffrey A. Segal, Judicial Behavior, in THE OXFORD HANDBOOK OF POLITICAL SCIENCE 280 (2011) (providing an overview of the judicial behavior literature); Lee Epstein & Jeffrey A. Segal, Trumping the First Amendment?, 21 WASH. U. J.L. & POL'Y 81 (2006) (describing the interactions of facts and ideology in the First Amendment context).

In fairness, some formal models of judicial review of executive action focus on whether the president has exceeded his discretion. See, e.g., WILLIAM G. HOWELL, POWER WITHOUT PERSUASION: THE POLITICS OF DIRECT PRESIDENTIAL ACTION 26-30 (2003).

5.

James L. Gibson, From Simplicity to Complexity: The Development of Theory in the Study

of Judicial Behavior, 5 POL. BEHAV. 7, 9 (1983).

6.

See Frank Cross, Appellate Court Adherence to Precedent, 2 J. EMPIRICAL LEGAL STUD. 369, 389-98 (2005); Donald R. Songer & Stefanie A. Lindquist, Not the Whole Story: The Impact of Justices' Values on Supreme Court Decision Making, 40 AM. J. POL. SCI. 1049, 1051-54 (1996).

This argument alone seems to concede the importance of doctrine to judicial decisionmaking. As Judge Harry T. Edwards and Michael A. Livermore state, "[t]he attitudinal model has been a consistent target for attack, and for good reasons: it does not adequately account for the role of law and precedence in judicial decisionmaking, it indulges fanciful assumptions about the nature of judicial preferences, it fails to account for judicial deliberations, and it has an impoverished account of ideology and law." Harry T. Edwards & Michael A. Livermore, Pitfalls of Empirical Studies that Attempt to Understand the Factors Affecting Decisionmaking, 58 DUKE L.J. 1895, 1914-16 (2009). Attitudinal scholars respond that if the adherence to doctrine or a particular interpretive ideology correlates with political preferences, "it is politics all the way down." Margaret H. Lemos, The Politics of Statutory Interpretation, 89 NOTRE DAME L. REV. 849 (2013) (book review).

7.

See Christina L. Boyd, The Hierarchical Influence of Court of Appeals on District Courts, 44 J. LEGAL STUD. 113 (2015); Donald R. Songer et al., The Hierarchy of Justice: Testing a Principal-Agent Model of Supreme Court-Circuit Court Interactions, 38 AM. J. POL. SCI. 673, 693 (1994). 8.

See HOWELL, supra note 4, at 136-74; Rafael Gely & Pablo T. Spiller, A Rational Choice Theory of Supreme Court Statutory Decisions with Applications to the State Farm and Grove City Cases, 6 J.L. ECON. & ORG. 263, 264 (1990); McNollgast, Politics and the Courts: A Positive Theory of Judicial Doctrine and the Rule of Law, 68 S. CAL. L. REV. 1631 (1995).

9. McNollgast, supra note 8, at 1649. 10. Cf id. at 1635 ("Stare decisis, respect for precedent and the rule of law, is the by-product of the strategic and political use of doctrine. Stare decisis reflects a self-enforcing equilibrium of doctrinal preferences among the courts. The properties of stare decisis do not in fact depend on whether judges actually respect precedent and the rule of law.").

11. For brevity, I refer to Barnett, Boyd, and Walker as "the authors" throughout.

12. For the first article in this series, see Kent Barnett & Christopher J. Walker, Chevron in

the Circuit Courts, 116 MICH. L. REV. 1 (2017). See also Kent Barnett, Christina L. Boyd & Christopher J. Walker, Administrative Law's Political Dynamics, 71 VAND. L. REV. 1463 (2018); Kent Barnett, Christina L. Boyd & Christopher J. Walker, The Politics of Selecting Chevron Deference, 15 J. EMPIRICAL LEGAL STUD. 597 (2018); Kent Barnett & Christopher J. Walker, Chevron Step Two's Domain, 93 NOTRE DAME L. REV. 1441 (2018); Kent Barnett & Christopher J. Walker, Short-Circuiting the New Major Questions Doctrine, 70 VAND. L. REV. EN BANC 147 (2017). The Barnett and Walker dataset is inordinately impressive. Barnett and Walker coded every published circuit court case from 2003 to 2013 that cited Chevron or Skidmore. Barnett & Walker, Chevron in the Circuit Courts, supra note 12, at 27. Their dataset is comprised of 1,558 instances in which a court reviewed an agency statutory interpretation. Id.

13.

See Skidmore v. Swift & Co., 323 U.S. 134, 140 (1944) (instructing courts to consider factors such as "the thoroughness evident in [the agency's] consideration, the validity of its reasoning, its consistency with earlier and later pronouncements, and all those factors which give it power to persuade, if lacking power to control" when deciding whether to defer to an agency statutory interpretation).

14. Barnett, Boyd & Walker, Administrative Law's Political Dynamics, supra note 12, at

1494-1502. 15.

Id. at 1506-07; see also Barnett, Boyd & Walker, The Politics of Selecting Chevron

Deference, supra note 12.

16. Barnett, Boyd & Walker, Administrative Law's Political Dynamics, supra note 12, at

1512-18.

into

judges

intruding

improperly

should avoid

statutory provisions and for courts to defer to those interpretations.17 the Second, policymaking realm.18 Administrative law requires a greater exercise of judicial restraint than any other area of law, except perhaps constitutional law. It brings many politically salient issues to federal court, including travel bans,19 protecting polar bears,20 and the Affordable Care Act.21 Cheuron sprang from the Supreme Court's recognition that "[j]udges are not experts in the field, and are not part of either political branch of the Government. Courts must, in some cases, reconcile competing political interests, but not on the basis of the judges' personal policy preferences."22 Focusing exclusively on this standpoint, Chevron "works" so long as judges resist imposing their personal policy preferences in the cases they adjudicate.

to conform

This Response concerns how reforms may jeopardize the delicate balance Chevron has struck. Note, I do not address whether Chevron requires reforms the Constitution or the Administrative Procedure Act. Part I reviews the authors' claim that Chevron effectively restrains political ideology. Parts II and III turn to possible reforms of Chevron step one and its scope, concluding that we have too little evidence about how these reforms would affect Chevron to prescribe any reforms at this time.

to either

I. CHEVRON IS EFFECTIVE

Standards of review enforce norms of judicial restraint by ensuring that reviewing courts are adequately deferential to the decisions of other actors. Justice Frankfurter described the standard of review as a mood that "must be respected, even though it can only serve as a standard for judgment and not as a body of rigid rules assuring

17.

See Abbe R. Gluck & Lisa Schultz Bressman, Statutory Interpretation from the Inside-

An Empirical Study of Congressional Drafting, Delegation, and the Canons: Part 1, 65 STAN. L. REV. 901, 994 (2013).

18.

See Gregory v. Ashcroft, 501 U.S. 452, 487 n.1 (1991) (Blackmun, J., dissenting) ("Ajudge is first and foremost one who resolves disputes, and not one charged with the duty to fashion broad policies establishing the rights and duties of citizens. That task is reserved primarily for legislators.").

19.

See Trump v. Hawaii, 138 S. Ct. 2392, 2407-15 (2018) (holding that the Immigration &

Nationality Act delegates significant authority to the President to bar entry of covered aliens into the United States).

20. See In re Polar Bear Endangered Species Act, 709 F.3d 1, 9-12 (D.C. Cir. 2013) (upholding the sufficiency of the Fish and Wildlife Service's determination that polar bears are endangered species).

21. See King v. Burwell, 135 S. Ct. 2480, 2488-49 (2015) (denying Chevron deference to an IRS interpretation of the ACA because the provision concerned questions of "economic and political significance").

22. Chevron U.S.A. Inc. v. Nat'l Res. Def. Council, , 865-66 (1984).

Probably more than 500 pages a year are devoted to detailed statements about scope of review of administrative action; most of that verbiage is harmless, for neitherjudges nor the readers of opinions take it seriously. Whether the verbiage about scope of review is helpful is doubtful, for it is typically vague, abstract, uncertain, and

conflicting. 24

23. Universal Camera Corp. v. NLRB, 340 U.S. 474, 487 (1951).

5 KENNETH CULP DAVIS, ADMINISTRATIVE LAW TREATISE § 29:2 (2d ed. 1984).

24. 25. Bednar & Hickman, supra note 2, at 1419-23 ("Chevron's own rhetoric supports different expositions of its two steps depending on which snippets of language one chooses to emphasize.").

26. See Thomas W. Merrill, The Story of Chevron: The Making of an Accidental Landmark, in ADMINISTRATIVE LAW STORIES 398, 412-20 (Peter L. Strauss ed., 2006) (describing how the Supreme Court did not view Chevron as a significant administrative law case); Cass R. Sunstein, Beyond Marbury: The Executive's Power to Say What the Law Is, 115 YALE L.J. 2580, 2595-96 (2006) (" [T]he Court itself may have had limited ambitions for its decision in Chevron[,] [but] the decision was soon viewed as a kind of revolution.").

27. Bednar & Hickman, supra note 2, at 1418 (describing the Chevron standard's many

versions).

28. See Colin S. Diver, Statutory Interpretation in the Administrative State, 133 U. PA. L. REV.

549, 645 (1985) (refuting these concerns).

29. For a lengthier discussion of the definitional problems of "ideology" in the legal context,

see generally Lemos, supra note 6.

30. For discussions of "conservative" legal principles, see generally STEVEN M. TELES, THE RISE OF THE CONSERVATIVE LEGAL MOVEMENT (2008); ILAN WURMAN, A DEBT AGAINST THE LIVING: AN INTRODUCTION TO ORIGINALISM (2017); William N. Eskridge, Jr., The New Textualism,

37 UCLA L. REV. 621 (1990); Keith E. Whittington, Is Originalism Too Conservative?, 34 HARV. J.L. & PUB. POL'Y 29 (2011).

31. For discussions of "liberal" legal principles, see JACK M. BALKIN, LIVING ORIGINALISM (2011); ROBERT A. KATZMANN, JUDGING STATUTES (2014); William J. Brennan, Jr., The Constitution of the United States: Contemporary Ratification, Lecture at Georgetown University (Oct. 12, 1985).

32. See Lee Epstein et al., The Judicial Common Space, 23 J.L. ECON. & ORG. 303, 306-09

(2007) (describing the methodology behind the calculation of JCS scores).

TABLE 1: PREDICTED PROBABILITIES OF CIRCUIT-PANEL AGREEMENT WITH AGENCY INTERPRETATION33

Agency Interpretation

Standard of Review

Liberal Interpretation Conservative Interpretation

/ oi

rol

Most Liberal Panel (JCS=-.502) .81 .91 .24 1

Most Conservative Panel (JCS=.538) .18 .6 .60 .74

Yet, an immediate concern emerges from this data. Yes, Chevron increases the probability that a panel agrees with the agency's interpretation. But a large margin of disagreement still exists between liberal and conservative panels. When Chevron applies, the model predicts that the most liberal panels agree with liberal interpretations in 25% more cases than the most conservative panels, while the most conservative panels agree with conservative interpretations in 23% more cases than the most liberal panels. More troubling, that gap is significantly larger when a lesser standard of review, such as the Skidmore standard or de novo review, applies. When the Chevron standard does not apply, the most conservative panels agree with liberal interpretations in 63% fewer cases than the most liberal panels. Likewise, agree with conservative interpretations in 36% fewer cases than the most conservative panels.

liberal panels

the most

But Table 1 provides predictions for the most extreme panels. These examples demonstrate the full strength of Chevron's power. However, these predictions do not account for the distribution of political ideology among panels.34

In reality, the vast majority of panels are far more moderate than these extreme panels. I use the authors' data to calculate the distribution of the average JCS scores of the panels.35 Figure 1 shows

33. See Barnett, Boyd & Walker, Administrative Law's Political Dynamics, supra note 12, at

1499-1502.

35.

34. Figures 3-8 show the range of predicted probabilities for JCS scores ranging from -0.5 to 0.54. Id. From these figures, one can derive the probability that a more moderate panel agrees with the agency's interpretation. However, these figures do not describe the distribution of panels. Individual judges are similarly distributed but exhibit more polarization. On average, circuit court judges tend to lean conservative (mean JCS score = 0.114). Among liberal judges, ideologies range from -0.06 (most moderate) to -0.52 (most extreme) and the mean JCS score of liberal judges is 0.29. Among conservative judges, ideologies range from 0.01 (most moderate) to 0.60 (most extreme) and the mean JCS score of conservative judges is 0.35.

36. Each observation is reported as a single instance of agency interpretation. Some cases involve multiple instances of agency interpretation. Therefore, some panels appear multiple times within the dataset and the distribution.

37. The data has a mean of 0.069 and a standard deviation of 0.222, which places the bounds of the first standard deviation at -1.54 (-1 s.d.) and .291 (+1 s.d.). In a normal distribution, approximately two-thirds of the data falls within one standard deviation of the mean. See KOSUKE IMAI, QUANTITATIVE SOCIAL SCIENCE: AN INTRODUCTION 288 (2017).

FIGURE 2: DISTRIBUTION OF AVERAGE JCS SCORES OF PANELS

I---------------08

-0.502

-0.089

0---07-- 0 0.067

0.218

0.538

IU-

CD

-0.6

i -0.4

I -0.2

I 0.0

I 0.2

I 0.4

i 0.6

38.

I used the same fixed values for the variables as the authors in the original estimation of

their predicted probabilities.

39. Note, the liberal panel and the conservative panel are set according to standard deviations. However, the neutral panel does not equal the mean average JCS score. Therefore, the liberal panel and the conservative panel are not equidistant from the neutral panel.

TABLE 2: PREDICTED PROBABILITIES OF CIRCUIT-PANEL AGREEMENT

WITH AGENCY INTERPRETATIONS

Agency Interpretation

Standard of Review

Liberal Interpretation

'1churoll

Conservative Interpretation

Most Liberal Panel (JCS=-.502) .81' (+.30) .91

.24 (-.16)

(-.12)

Liberal Panel (-1 s.d.) (JCS=-. 154) .61 (+.10)

(±09 (±3)

()

(08)

.34 (-.06)

(-0)(0)

Neutral Panel

-

(JCS=0)

.51 (0)

.40 (0)

Conservative Panel (+1 s.d.) (JCS=.291) .31 (-.20)

.50 (+.10)

(-,.) 1

Most Conservative Panel (JCS=.538) .18 (-.33)

.60 (+ 20)

The top number reports the probability that the panel agrees with the agency's interpretation. The bottom numlb the difference from the probability of the neutral panel.

40. Absent the application of Chevron, we still observe concerning disparities between liberal and conservative panels. Perhaps we ought to consider whether Skidmore and de novo review sufficiently constrain political ideology in the administrative law context. Aside from this observation, I offer no insight into this issue here.

41. Robert A. Stein, The Rule of Law, in THE RULE OF LAW IN THE 21ST CENTURY 13 (Robert A. Stein & Richard J. Goldstone eds., 2015) ("The judicial power must be exercised independently of either the executive or legislative powers, and individual judges must base their decisions solely on the laws and the facts of individual cases.").

II. CHANGES TO STEP ONE

42. Political science has long faced a similar problem in trying to ascertain why people vote. Rational citizens should not vote because they are unlikely to influence the election and incur costs in making the decision to vote. Yet citizens vote. See William H. Riker & Peter C. Ordershook, A Theory of the Calculus of Voting, 62 AM. POL. SCI. REV. 25, 25 (1968). Similarly, judges should not adhere to the rule of law because doing so comes at the cost of losing the opportunity to decide cases according to their own beliefs. Yet judges decide cases against their own ideological interests. Political scientists have concluded that many citizens vote because of a strong sense of civic duty; I suspect a similar psychological force operates in judicial decisionmaking. Id. at 28.

43. RICHARD A. POSNER, THE PROBLEMS OF JURISPRUDENCE 32 (1990) (describing law as 'strongly objective in easy cases, weakly objective in difficult ones, but rarely either highly determinant or merely political").

44. See Edwards & Livermore, supra note 6, at 1898 ("I have estimated in only 5 to 15 percent of the disputes that come before me in any given term do I conclude, after reviewing the record and all pertinent legal material, that the competing arguments drawn from those sources are equally strong. Put differently, only in those few cases do I feel ...that to dispose of the appeal I must rely on some significant measure of discretion.").

45. Chevron U.S.A. Inc. v. Nat. Res. Def. Council, , 842-43 (1984). 46. Because this section concerns Chevron step one, I do not discuss the variations in step twd's application. Broadly speaking, there are two conceptions of step two. Step two may ask whether the agency has arrived at a "permissible construction of the statute" through the use of traditional tools of statutory interpretation. Id. at 843; see Frederick Liu, Chevron as a Doctrine of Hard Cases, 66 AIDMIN. L. REV. 285, 315-17 (2014) (defining step two in positivist terms). Alternatively, step two may ask whether the agency's interpretation is "reasonable" meaning neither arbitrary nor capricious. Id. at 844; see Ronald M. Levin, The Anatomy of Chevron: Step Two Reconsidered, 72 CHI.-KENT L. REV. 1253, 1267-69 (1997) (arguing for a hard-look interpretation of step two); Mark Seidenfeld, A Syncopated Chevron: Emphasizing Reasoned Decision-Making in Reviewing Agency Interpretations of Statutes, 73 TEx. L. REV. 83, 129 (1994) (requiring "the agency to identify the concerns that the statute addresses and explain how the agencys interpretation took those concerns into account"). Both conceptions are permissible interpretations of the Chevron opinion.

47. Barnett, Boyd & Walker, Administrative Law's Political Dynamics, supra note 12, at 1518 to liberal (finding that, among cases that reach step two, extreme liberal panels defer interpretations in 90.48% of cases and conservative interpretations in 78.05%, and extreme conservative panels defer to liberal interpretations in 96.55% of cases and conservative interpretations in 95.74% of cases). There is a moderate indication that political influence creeps into liberal panel's review of conservative interpretations at step two. However, this effect would substantially decline as the panel grows more moderate and approaches the mean liberal panel.

Id. at 1513-15.

48. 49. Antonin Scalia, Judicial Deference to Administrative Interpretations of Law, 1989 DUKE L.J. 511, 520-21 (predicting "that the future battles over acceptance of agency interpretations of law will be fought" at step one).

50. Barnett & Walker, Chevron in the Circuit Courts, supra note 12, at 34. 51. Id. 52. Nat'1 Cable & Telecomms. Ass'n v. Brand X Internet Servs., 545 U.S. 967, 980-83 (2005). 53. Scalia, supra note 49, at 520 ("How clear is clear?"). 54. Brett M. Kavanaugh, Fixing Statutory Interpretation, 129 HARV. L. REV. 2118, 2136-37 (2016) (book review) ("If the statute is 60-40 in one direction, is that enough to class it clear? How about 80-20? Who knows?").

55. Bednar & Hickman, supra note 2, at 1453 (quoting an anonymous judge). 56. Raymond M. Kethledge, Ambiguities and Agency Cases: Reflections After (Almost) Ten

Years on the Bench, 70 VAND. L. REV. EN BANC 315, 320 (2017).

57. See Matthew C. Stephenson & Adrian Vermeule, Chevron Has Only One Step, 95 VA. L.

REV. 597, 601 (2009) (using this phrase).

definition as a bright-line rule.58 Justice Kavanaugh argues that every judge has a different standard for deciding whether a statute is clear or ambiguous:59

First, judges must decide how much clarity is needed to call a statute clear. If the statute is 60-40 in one direction, is that enough to call it clear? How about 80-20? Who knows?

Second,

let's imagine that we could agree on an 80-20 clarity threshold.... Even if we say that 80-20 is the necessary level of clear, how do we then apply that 80-20 formula to particular statutory text? Again, who knows? Determining the level of ambiguity in a given piece of statutory language is often not possible in any rational way. One judge's clarity is another judge's ambiguity. It is difficult for judges (or anyone else) to perform that kind of task in a neutral, impartial, and predictable fashion. 6o

58. See Nicholas R. Bednar, The Clear-Statement Chevron Canon, 66 DEPAUL L. REV. 819,

831-32 (2017); Kavanaugh, supra note 54.

Id. at 2137.

59. Kavanaugh, supra note 54, at 2134-54. 60. 61. Wis. Cent. Ltd. v. United States, 138 S. Ct. 2067, 2074(2018) ("In light of all the textual and structural clues before us, we think it's clear enough that the term 'money' excludes 'stock,' leaving no ambiguity for the agency to fill.").

62. Chris Walker, Gorsuch's "Clear Enough" & Kennedy's Anti-"Reflexive Deference": Two Potential Limits on Chevron Deference, YALE J. ON REG.: NOTICE & COMMENT (June 22, 2018), http://yalejreg.com/nc/gorsuchs-clear-enough-kennedys-anti-reflexive-deference-two-potentiallimits-on-chevron-deference/.

63. See Kavanaugh, supra note 54; Kethledge, supra note 56; Scalia, supra note 49.

However, interpretive methodology alone cannot explain all of the political ideology that seeps into Chevron step one. If interpretive methodology completely determined outcomes, we would expect that liberal panels and conservative panels defer to agency statutory interpretations at different rates but do so consistently across both conservative and liberal interpretations. In other words, a liberal panel employing purposivist ideology should agree with conservative and liberal interpretations at equal rates. However, we observe an imbalance. Liberal panels agree with liberal interpretations more and conservative interpretations less. The same is true of conservative panels.

Two non-political hypotheses may explain why we observe this imbalance. Given the state of the literature, neither hypothesis has sufficient support to provide a basis from which to offer new reforms.

64. See ANTONIN SCALIA & BRYAN A. GARNER, READING LAW: THE INTERPRETATION OF LEGAL TEXTS (2012) (laying out a complete method of textualist interpretation); John F. Manning, SecondGeneration Textualism, 98 CALIF. L. REV. 1287, 1303-17 (2010) (describing modern textualism's interpretive tools); Jennifer L. Mascott, The Dictionary as a Specialized Corpus, 2017 BYU L. REV. 1557, 1564-72 (advocating the use of dictionaries in statutory interpretation).

65. See KATZMANN, supra note 31, at 35-39 (discussing the benefits of legislative history). 66. See Richard H. Fallon, Jr., Three Symmetries Between Textualist and Purposicist Theories of Statutory Interpretation And the Irreducible Roles of Values and Judgment Within Both, 99 CORNELL L. REV. 685, 688-97 (2014) (arguing that both textualists and purposivists use value judgments in interpreting statutes).

(39%),

textual canons

67. See Jerry Mashaw, Norms, Practices, and the Paradox ofDeference: A Preliminary Inquiry into Agency Statutory Interpretation, 57 ADMIN. L. REV. 501, 536 (calling on administrative law scholars to mine the "rich veins of interpretive ore" to gain empirical insights into agency interpretation).

68. Christopher J. Walker, Inside Agency Statutory Interpretation, 67 STAN. L. REV. 999,1020

Id. at 1020-34 (textual and substantive canons), 1034-48 (legislative history).

(2015). 69. 70. Cf Kevin M. Stack, Purposicism in the Executive Branch: How Agencies Interpret Statutes, 109 Nw. U. L. REV. 871 (2015) (arguing that agencies should use a purposivist methodology when interpreting statutes).

71. Amy Semet, An Empirical Examination of Agency Statutory Interpretation, 103 MINN. L.

REV. (forthcoming Spring 2019).

72. Chevron U.S.A. Inc. v. Nat. Res. Defense Council, (1984) ("When a challenge to an agency construction of a statutory provision, fairly conceptualized, really centers on the

wisdom of the agency's policy, rather than whether it is a reasonable choice within a gap left open by Congress, the challenge must fail.").

73.

See ADRIAN VERMEULE, JUDGING UNDER UNCERTAINTY: AN INSTITUTIONAL THEORY OF LEGAL INTERPRETATION 213-14 (2006); Stack, supra note 70, at 887-904; William N. Eskridge, Jr., Expanding Chevron's Domain: A Comparative Institutional Analysis of the Relative Competence of Courts and Agencies to Interpret Statutes, 2013 WIS. L. REV. 411, 420-23.

74.

See generally JAMES Q. WILSON, BUREAUCRACY: WHAT GOVERNMENT AGENCIES DO AND

WHY THEY Do IT (1989).

75. Barnhart v. Walton, 535 U.S. 212, 222 (2002). 76. Barnett & Walker, Chevron in the Circuit Courts, supra note 12, at 50. 77. Mark D. Richardson et al., Elite Perceptions of Agency Ideology and Workforce Skill, 80

J. OF POL. 303 (2018) (collecting data on agency ideology and workforce skill).

78.

Id. at 54.

79. Richardson et al., supra note 77. 80. There may be other reasons that these interpretive tools should not apply in the Chevron analysis. For example, John Manning argues that Chevron precludes the use of legislative history because Chevron is founded on the assumption that Congress intends for the agency to resolve the ambiguity. Therefore, the court's use of legislative history to clarify the statute contradicts the assumption that Congress delegated authority to the agency to interpret the statute. John F. Manning, Chevron and Legislative History, 82 GEO. WASH. L. REV. 1517, 1520-21 (2014).

81. See Lemos, supra note 6, at 874-77 (reviewing this empirical literature). 82. David S. Law & David Zaring, Law Versus Ideology: The Supreme Court and the Use of

Legislative History, 51 WM. & MARY L. REV. 1653, 1726 (2010).

83. FRANK B. CROSS, THE THEORY AND PRACTICE OF STATUTORY INTERPRETATION 172 (2009).

84. James J. Brudney & Corey Ditslear, Liberal Justices' Reliance on Legislative History:

Principle, Strategy, and the Scalia Effect, 29 BERKELEY J. EMP. & LABOR L. 117, 120-21 (2008).

85.

Id. at 156-62 (finding that, among cases employing substantive canons, 42.5% had a

liberal outcome and 50.7% had a conservative outcome).

86. Anita S. Krishnakumar, Reconsidering Substantive Canons, 84 U. CHI. L. REV. 825, 859-

62 (2017).

87. James J. Brudney & Corey Ditslear, Canons of Construction and the Elusive Quest for Neutral Reasoning, 58 VAND. L. REv. 1, 58 (2005) (finding that, among cases employing textual canons, 49.1% had a liberal outcome and 43.5% had a conservative outcome).

88.

Id. at 56 (finding that, among cases employing textual canons, 49.1% had a liberal

outcome and 43.5% had a conservative outcome).

89. CROSS, supra note 83, at 169. The three justices for whom this was not true were Justices

Ginsburg, Rehnquist, and Scalia.

90. Lemos, supra note 6, at 884-91.

Id. at 891-903.

91. 92. See Barnett, Boyd & Walker, The Politics of Selecting Chevron Deference, supra note 12, at 599 (concluding that liberal judges are less likely to invoke Chevron when reviewing conservative interpretations).

93. Barnett & Walker, Chevron in the Circuit Courts, supra note 12, at 30.

94. See, e.g., Pereira v. Sessions, 138 S. Ct. 2105, 2120-21 (2018) (Kennedy, J., concurring); Perez v. Mortg. Bankers Ass'n, 135 S. Ct. 1199, 1211 (2015) (Scalia, J., concurring); Michigan v. EPA, 135 S. Ct. 2699 (2015) (Thomas, J., concurring); Christensen v. Harris Cty., 529 U.S. 576, 596-97 (2000) (Breyer, J., dissenting) (arguing that Chevron made no relevant change to Skidmore); Gutierrez-Brizuela v. Lynch, 834 F.3d 1142, 1149 (10th Cir. 2016).

95. For some articles discussing these limitations, see Kristin E. Hickman, The (Perhaps) Unintended Consequences of King v. Burwell, 2015 PEPP. L. REV. 58; Nicholas R. Bednar, Coping with Chevron: Justice Gorsuch's Majority and Justice Breyer's Dissent in SAS Institute, YALE J. ON REG: NOTICE & COMMENT (Apr. 25, 2018), http://yalejreg.com/nc/coping-with-chevron-justicegorsuchs-majority-and-justice-breyers-dissent-in-sas-institute-by-nicholas-r-bednar/; Chris Walker, Should Courts Defer to Administrative Interpretations of Criminal Law?, YALE J. ON REG: NOTICE & COMMENT (Nov. 11, 2014), http://yalejreg.com/nc/should-courts-defer-to-administrativeinterpretations-of-criminal-law/.

Id. at 239-40 (Scalia, J., dissenting).

96. United States v. Mead Corp., 533 U.S. 218, 226-27 (2001). 97. 98. See Lisa Schultz Bressman, How Mead Has Muddled Judicial Review of Agency Action, 58 VAND L. REV. 1443, 1444 (2005); Kristin E. Hickman, The Three Phases of Mead, 83 FORDHAM L. REV. 527, 528 (2014); Adrian Vermeule, Introduction: Mead in the Trenches, 71 GEO. WASH. L. REV. 347, 347 (2003).

99. See City of Arlington, Tex. v. FCC, 569 U.S. 290, 304-05 (2013) (holding that Chevron applies to an agency's "jurisdictional" interpretation); Mayo Found. For Med. Edue. & Research v. United States, 562 U.S. 44, 54-58 (2011) (holding that Chevron applies to Treasury Department regulations); Nat'l Cable & Telecomms. Ass'n v. Brand X Internet Sers., 545 U.S. 967, 980-83 (2005) (holding that an agency's new interpretation of an ambiguous statute trumps a judicial interpretation of the same statute).

interpretations experienced significant confusion about whether contained within Board of Immigration Appeals ("BIA") decisions were subject to Chevron. The BIA issues decisions in one of three ways: (1) published three-member decisions, (2) unpublished three-member decisions, and (3) single-member decisions. Its interpretations are overwhelmingly conservative-of the 377 immigration interpretations arrived at in the course of an adjudicatory proceeding in the authors' dataset, only 8 (2.1%) are liberal interpretations.100

In the mid-2000s, circuits split on the issue of whether Chevron applied to the BIA's unpublished and single-member decisions. In Garcia-Quintero v. Gonzales, a Ninth Circuit panel held that singlemember decisions of the BIA did not meet the force-of-law requirements of Mead because they were non-precedential decisions.101 The Seventh Circuit, however, applied Chevron to single-member decisions.102 Other circuits followed the Ninth Circuit's reasoning.103 In 2011, the Seventh Circuit overruled its decision.104

Although this circuit split appears to have worked itself out, it is illustrative of how indeterminacies in Mead appear political to researchers and their models. With one exception, all of the cases holding that Chevron did not apply to single-member BIA decisions were conservative interpretations.105 The Seventh Circuit case holding that Chevron applied was decided by a conservative panel reviewing a conservative interpretation.106

reviewing

decided

panels

liberal

by

Yet the risk that circuits will again split over review of BIA ignores interpretations remains. The Supreme Court frequently Chevron in immigration cases.107 Most recently, it failed to even mention Chevron or "deference" in Nieslen v. Preap, despite the fact that

100. The 377 decisions would also include decisions by the Executive Office of Immigration Review and by the Attorney General. The vast majority of the cases in the dataset review Board of Immigration Appeals' interpretations. The inability to sort out the precise number of interpretations made by each agency does not change the fact that the interpretations are predominantly conservative, regardless of their source.

101. Garcia-Quintero v. Gonzales, 455 F.3d 1006, 1011-14 (9th Cir. 2006). 102. See Gutnik v. Gonzales, 469 F.3d 683, 689-91 (7th Cir. 2006). 103. See, e.g., Carpiov. Holder, 592 F.3d 1091, 1097 (10th Cir. 2010); Quinchia v. U.S. Attorney Gen., 552 F.3d 1255, 1258 (11th Cir. 2008); Rotimi v. Gonzales, 473 F.3d 55, 57-58 (2d Cir. 2007).

104. Arobelidze v. Holder, 653 F.3d 513 (7th Cir. 2011). 105. I obtained this data from the author's dataset: Garcia-Quintero, 455 F.3d 1006 (average JCS score = -.253); Carpio, 592 F.3d 1091 (average JCS score = -.052); and Rotimi, 473 F.3d 55 (average JCS score = -.2245). The average JCS score of the panel in Quinchia, 552 F.3d 1255, was .019 but two judges were moderately liberal (-0.06 and -.205) whereas one was conservative (0.324).

106. Gutnik, 469 F.3d 683 (average JCS score = .218). 107. See Michael Kagan, Chevron's Liberty Exception, 104 IOWA L. REV. 491, 494 (2019)

(highlighting the Supreme Court's inconsistent application of Chevron in immigration cases).

108. Nielsen v. Preap, No. 16-1363, slip op. (Mar. 19, 2019); Michael Kagan, Chevron Goes Missing in an Immigration Case. Again., YALE J. ON REG.: NOTICE & COMMENT (Mar. 19, 2019), http://yalejreg.com/nc/chevron-goes-missing-in-an-immigration-case-again/.

109. Kagan, supra note 107. 110. See Kent Barnett, Against Administrative Judges, 49 UC DAVIS L. REV. 1643, 1686-89 (2016) (suggesting that agencies should pursue formal adjudication to increases odds they receive Chevron deference).

111. See Aqua Prods., Inc. v. Matal, 872 F.3d 1290, 1328-34 (Fed. Cir. 2017) (en banc) (Moore, J., concurring) (describing at length why the PTO must promulgate interpretations through rulemaking); Brand v. Miller, 487 F.3d 862, 869 n.3 (Fed. Cir. 2007); Merck & Co., Inc. v. Kessler, 80 F.3d 1543, 1549-50 (1996).

112. See Christopher J. Walker, Constitutional Tensions in Agency Adjudication, 104 IOWA L.

REV. (forthcoming 2019) (citing some of these arguments).

113. See King v. Burwell, 135 S. Ct. 2480, 2488 (2015); FDA v. Brown & Williamson Tobacco Corp., 529 U.S. 120, 159-60 (2000); MCI Telecomms. Corp. v. AT&T Co., 512 U.S. 218, 231 (1994).

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

114. King, 135 S. Ct. at 2488. 115. Mayo Found. For Med. Edue. & Research v. United States, 562 U.S. 44, 54-58 (2011). 116. See Kristin E. Hickman, The (Perhaps) Unintended Consequences of King v. Burwell, 2015

PEPP. L. REV. 56, 58.

117. Barnett & Walker, Short-Circuiting the New Major Questions Doctrine, supra note 12, at

159-62.

we occasionally are, improper uncertainty estimates might cause policy makers to act too early, perhaps even doing significant damage by creating political instability or even civil war. Prior to making prescriptions, we should be asking ourselves whether we are willing to risk the unintended or unknown consequences of proposed institutional reforms. 118

118. Gary King, et. al., The Methodology of Presidential Research, in Researching the

Presidency: Vital Questions, New Approaches 387, 391 (George C. Edwards III, et. al eds., 1993).

119. See Kent Barnett, Codifying Cheumore, 90 N.Y.U. L. REV. 1, 11-17; Lisa Schultz Bressman, Reclaiming the Legal Fiction of Congressional Delegation, 97 VA. L. REV. 2009, 201415 (2011); Thomas W. Merrill & Kristin E. Hickman, Chevron's Domain, 89 GEO. L.J. 833, 870-72 (2001).