POSSIBILITIES FOR INSTITUTIONAL. IMPLEMENTATION
If the case for hierarchically variable deference is appealing in principle, could our judicial system implement it in practice? This Part presents some structural and doctrinal possibilities for realizing a hierarchically variable regime. First, however, it describes some ways in which our judicial system manifests hierarchical variation even now. The analysis of this Article helps to explain and justify these existing patterns of variation.
Ways in Which Our System Already Displays Hierarchically Variable Deference
Unofficial Doctrinal Divergences?
As stated at the outset, official doctrines of deference do not openly embrace hierarchical heterogeneity. (103) Nonetheless, standards of review may be hierarchically variable in practice. Specifically, the Supreme Court may already give agencies less deference than the lower courts typically do. Admittedly, it is hard to be sure: despite the significant and still growing body of empirical literature on deference, (104) the existing research does not allow firm conclusions about differences across courts. Simply comparing agency win rates in different courts will not suffice. For one thing, the Supreme Court's docket is small and highly unrepresentative, reflecting the strategic choices of litigants to seek certiorari and the Justices to grant it. Moreover, it is hard to calculate true levels of deference in any court, for courts might cite a deference regime (or mention facts that would trigger deference) because they plan to defer, rather than the other way around. (Thus, a 100% agency win rate in cases citing Chevron would not necessarily reveal great deference if Chevron went ignored in similar cases that the agency lost.) Nonetheless, despite these complications, there is at least some reason to believe that the Supreme Court is not as deferential as a faithful application of current doctrine would direct. Eskridge and Baer present evidence that the Court does not invoke any deference regime in the majority of cases that involve agency interpretations; this frequently happens even in cases that are, according to prevailing doctrine, Chevron-eligible. (105) The failure to invoke a deference regime is, in turn, associated with lower agency win rates. (106) Further empirical analysis by Raso and Eskridge leads them to the conclusion that the Justices invoke deference regimes episodically and inconsistently, which is not what one would expect if the Court regarded deference regimes as having true stare decisis effect. (107) By way of contrast, there is at least some evidence that the Supreme Court's deference doctrines do have a substantial effect in the lower courts, though to be sure the findings are hardly definitive. (108) Eskridge and Baer float the possibility that the Court regards deference regimes as guides for lower courts but does not regard them as binding or necessary in the Court itself. (109)
Even in the absence of definitive statistical evidence, there are some circumstantial and structural reasons to suspect that the Supreme Court is less deferential than lower courts. First, because the Supreme Court has no reviewing court above it, it need not fear reversal for ignoring or misapplying deference regimes in order to reach particular favored outcomes. Second, the Justices might feel (with some justification) that their relatively favorable decision-making environment-their advantages in resources, time, perceived expertise, and so forth (110)--makes deference less necessary for them than for their more "limited" colleagues in the lower courts. Third, it is easier to write an opinion affirming an agency than an opinion reversing it, and so one imagines that deference is an especially appealing path of least resistance for a busy lower court that lacks the luxury of a discretionary docket. (111)
If it is true that the Court flouts its own deference doctrines, one response is to lament the Court's disobedience. Yet the analysis of this Article suggests another possibility: that the Court's vice is actually a virtue, because deference is not as appropriate for the Supreme Court. That is, this Article can help reconcile us to a feature of existing practice that might otherwise trouble us. And were the Court candidly to announce that it follows different rules, that might reduce the extent to which the Court's non-deferential decisions in administrative cases can mislead the lower courts about how the lower courts should behave.
Similar comments could be made about the United States Court of Appeals for the District of Columbia Circuit. There is evidence that the D.C. Circuit tends to be less deferential than other courts of appeals. (112) Perhaps that court's posture reflects overconfidence or undue activism, (113) but there may be something to say for a relatively assertive stance when one considers the justifications for judicial deference. The D.C. Circuit's semi-specialized docket and lower caseload (114) arguably confer expertise and technical competence in comparison to the regional circuits. The politicized and nationalized nature of the selection process for the D.C. Circuit approximates the process for the Supreme Court, to which the D.C. Circuit is often regarded as a stepping stone. (115) When the D.C. Circuit is the exclusive venue for a certain kind of case, as it sometimes is, (116) there are no problems of uniformity. To the extent the D.C. Circuit is a 'junior varsity" Supreme Court in terms of its institutional context and competencies, (117) reduced deference vis-a-vis other lower courts is defensible. (One could generalize these points beyond the context of administrative review, of course, by considering other (semi-) specialized or expert courts like bankruptcy courts and the Federal Circuit.) (118)
Current Patterns of Jurisdictional Allocation
A second way in which our system already displays hierarchical variation comes in the form of congressional decisions about allocating jurisdiction. The law governing jurisdiction to review agency action is complex. (119) It could hardly be otherwise, given that administrative action is itself ubiquitous and multifarious. The pertinent feature of the jurisdictional pattern for present purposes is that Congress has arranged the jurisdictional statutes so that many cases reviewing agency action bypass the district courts and begin in the courts of appeals. (120) The usual explanation for this arrangement is that proceedings to review agency action have an appellate character. (121) A court typically reviews agency action based on the record compiled by the agency, much as an appellate court reviews the findings and conclusions of a trial court. (122) The distinctive capacity of trial courts to take evidence and adjudicate facts is unneeded in most administrative cases, and so beginning there merely causes duplication and delay. (123)
This Article illuminates an additional set of justifications for Congress's frequent (though not universal) decision to bypass the district courts. To begin, it is not just that review of agency action often has an appellate character but, as explained above, it also often has a political character; to that extent, district judges have less business setting aside the choices of national administrators. (124) In addition, cases reviewing agency interpretations are legally and technically complex in ways that may tax the resources of busy trial judges. (125) Further, the sheer numerosity and heterogeneity of district courts threatens a particularly problematic form of geographic disuniformity. (126) These factors weigh in favor of strong deference, and so the value of having judicial review begin in those courts decreases. Thus, Congress is right to bypass the district courts in many administrative cases. (127)
Potential Changes to Current Law
One could stop there, treating hierarchical heterogeneity as a theory that helps to justify and to explain certain aspects of current practice. But we can also ask whether it would be desirable to change the law in a (more) hierarchically variable direction, and if so, how?
Translating the theoretical case for variable deference into a concrete implementation plan involves several complexities. For one, some variables are continuous (e.g., how much weight to give an agency interpretation), while others are lumpy or dichotomous (e.g., which deference regime applies or whether a court has jurisdiction). In the latter situation, one cannot make adjustments with perfectly calibrated smoothness; instead, one has to decide whether the case for variable deference is strong enough to take a certain step or not. And because each court is part of a larger system, one cannot necessarily assume that changes at the court level will have corresponding, easily predictable effects at the system level.
In addition, the decision to embrace hierarchical variation does not by itself specify the ideal end state of the system and how it should differ from the status quo. If one begins with a baseline in which all courts display basically the same level of deference (which, as just discussed, might not actually describe the current reality), one could implement hierarchical variation from either end: by getting lower courts to defer more or instead by having higher courts defer less or indeed not at all. The choice between those two approaches--ratcheting deference up in some places versus ratcheting it down in others--implicates both normative and empirical questions. On the normative side, we would need to decide how much deference is ideal. Does judicial oversight usefully contribute to balanced government and the formulation of rational policy? Or does judicial review hamper the regulatory system by inviting meddling from inexpert generalists, or worse, agenda-driven partisans? Both views have forceful advocates. (128) On the empirical side, we would need to know how...