Chevron’s Mistake

TABLE OF CONTENTS Introduction.............................................................................................550 I. The Dominant Theories and Their Administrative

Law Doctrine................................................................................559 A. The Search for Statutory Meaning...........................................559 B. The Possibility of Agency Delegation.....................................566 C. The Possibility of Deliberate Ambiguity.................................571 II. A Delegation-Respecting Theory...................................................575 A. Reasons for Interpretive Delegation.......................................576 B. Conditions of Interpretive Delegation....................................580 C. Application Issues......................................................................583 1. Scope of Agency Interpretive Authority............................583 2. Reasonableness of Agency Interpretations........................585 3. Exercise of Judicial Interpretive Authority........................586 III. Illustrations.......................................................................................589

A. Zuni Public School District No. 89 v.

Department of Education.......................................................589 B. Gonzales v. Oregon....................................................................597 IV. Objections and Responses..............................................................602 A. Institutional Costs......................................................................605 B. Nondelegation Issues.................................................................609 1. The Nondelegation Tools......................................................610 2. The Normative Difficulty......................................................613 3. The Administrative Law Alternative..................................617 Conclusion................................................................................................619

Statutes that delegate authority to administrative agencies are different from the rest. Congress may delegate to an agency not only the authority to implement the statute but, implicitly, the authority to

interpret it as well—that is, to specify its meaning. For other statutes, courts resolve any interpretive questions. Chevron U.S.A. Inc. v. Natural Resources Defense Council, Inc.1 recognizes this essential difference between regulatory statutes and other statutes. Chevron asks courts to approach regulatory statutes by determining whether involved the relevant Congress has delegated to the agency interpretive authority. This Article argues, however, that the decision fails to supply courts with the tools for providing a proper answer.

Chevron directs courts to determine “whether Congress has directly spoken to the precise question at issue,” applying the “traditional tools of statutory construction.”2 Only when courts find no “clear” meaning for the statutory text do they infer a delegation of interpretive authority to the agency. In applying Chevron, courts rely heavily on the dominant theories of statutory interpretation: intentionalism, purposivism, or textualism. Those theories, despite their differences, all invite courts to construct a meaning for statutory text as if Congress intended the text to carry a relatively specific meaning. For example, intentionalism sees Congress as intending a in the chosen text.3 meaning, albeit expressing Purposivism understands Congress as motivated by a general aim in enacting statutes, often imperfectly expressing that aim.4 Both permit courts to consult legislative history in pursuit of legislative intent or

imperfectly

it

from

purpose. Meanwhile, textualism denies that legislative history is a permissible interpretive source as this theory has a strikingly different view of legislative behavior. Textualism tends to conform to public choice theory and the claim that Congress has no intent or purpose text.5 distinct Nevertheless, modern textualism invites courts to discern a meaning for awkward or imprecise language. It sees the statutory meaning as the likely product of strategic legislative compromise, which courts should not unsettle through resort to legislative history. Textualists also have a constitutional objection to the use of legislative history because Congress only enacted the text.6

those explicitly stated

the statutory

The search for meaning, common to the dominant theories of statutory interpretation, does not square with what a good deal of positive political theory and legal scholarship has stated about regulatory statutes. Congress often designs these statutes with the aim of delegating authority, including interpretive authority, to agencies.7

Furthermore, Congress likely delegates authority to agencies under certain circumstances. For example, it likely delegates to avoid complex issues, which conserves legislative resources and capitalizes on agency expertise.8 It also likely delegates to avoid contentious issues and obtain consensus on legislation.9 Whatever the reason, Congress seeks to ensure that subsequent agency action will not depart too far from legislative preferences. It can do so through the structure of regulatory statutes, imposing procedural requirements that enable future legislative coalitions to monitor agency action by placing constituents in the administrative process.10 It can consider positions that the agency has steadfastly maintained in the past or that the agency has offered in the course of legislative drafting.11 Congress does not always delegate expressly but often leaves interpretive questions for agencies to resolve.12

Chevron recognizes such “delegating” factors; its mistake is failing to make those factors central to its doctrinal inquiry. The factors operate only as justifications for agency delegation, not as guides for determining the existence of that delegation. Thus, Chevron acknowledges that Congress may implicitly delegate interpretive authority to agencies and may do so to capitalize on agency expertise or to obtain legislative consensus.13 But rather than

implementing these insights as part of the doctrinal analysis, Chevron reverts to the conventional theories of statutory interpretation in search of a meaning.14 Agency delegation is an afterthought, if a thought at all. The very search for a meaning directs courts away from what many scholars have been saying about regulatory statutes— namely, that Congress can consider others factors.15

law,

force of

Nor does United States v. Mead Corp.16 correct the problem. Mead recognizes that when Congress intends for an agency to issue interpretations with it authorizes certain administrative procedures for that purpose.17 By focusing on procedures rather than the statutory text, Mead makes delegation rather than meaning the key question. Notwithstanding this important insight, the decision does not provide the best way to judge delegation. It requires that Congress authorize (and agencies utilize) sufficiently formal procedures but does not connect those procedures to their legislative function. Rather, Mead connects procedures to the “force of law.” Procedures can be important to Congress because they facilitate legislative monitoring and not necessarily because they promote rule-of-law values. Courts interested in delegation should ensure that procedures are adequate for their strategic legislative function. And courts should regard sufficiently formal procedures as affirmative indications of agency delegation. Under Mead, these procedures merely enable courts to apply Chevron as usual.

This Article imagines what interpretive theory would look like for regulatory statutes if courts were to truly gauge the existence of agency delegation. In short, courts would look for evidence that Congress has delegated the task of specifying the meaning of a statute

to the agency charged with implementing it. Rather than focusing heavily or exclusively on the relative clarity of statutory language, courts would examine the sorts of considerations that positive political theorists and legal scholars have identified as indications of legislative intent to delegate interpretive issues to agencies. Assuming Congress likely delegates for reasons related to the complexity or contentiousness of the issue addressed by a statute, courts should consider the complexity or contentiousness of the issue by examining the statute or the surrounding context. Congress also likely delegates under conditions that minimize principle-agent concerns. It can use administrative procedures to monitor agency interpretations. Or it can rely on a position that the agency has long maintained in the past or offered in the course of legislative drafting. Courts should look for signs of these factors as well, either in the statute or the surrounding context. The stronger the case for interpretive delegation, the more courts should hesitate to read the statutory text as “clear” on the theory that Congress intended such clarity. Ordinarily, courts search for a meaning to avoid undermining statutory purposes or unsettling legislative deals. Under delegating assumptions, courts would show genuine respect for Congress by not allowing the relative clarity of the language to defeat what is most plausibly understood as an interpretive delegation.

In this Article, I use two Supreme Court opinions to illustrate the framework that I envision. In Zuni Public School District No. 89 v. Department of Education,18 Justice Breyer, writing for the majority, consciously set aside Chevron’s question of statutory meaning and examined the political factors relevant to interpretive delegation.19 He upheld the agency interpretation on the basis of these political factors, drawing criticism from virtually every other Justice for his unorthodox analysis, even those who agreed with his ultimate disposition of the case.20 In Gonzales v. Oregon,21 Justice Kennedy, writing for the majority, found that the issue was not the sort that Congress was likely to have delegated to the agency involved.22 He made this determination based on realistic assumptions about

legislative behavior, including the lack of reasons for and conditions of interpretive delegation.23 Neither Justice explained precisely why these unconventional analyses were appropriate. But both implicitly recognized the need to depart from the norm.

I argue that this departure is justified because it conforms to congressional will and for the other well-known reasons that Chevron captures: agencies generally should elaborate regulatory policy because they are more expert and politically accountable than courts.24 At the same time, this approach does not call for abject judicial deference or judicial abdication. For example, courts would not allow agencies to assert interpretive authority when Congress has not authorized proper procedures. In so doing, courts would serve as “faithful agents” by enabling Congress to control when not to delegate. By the same token, they would reinforce rule-of-law values because procedural formality tends to ensure that agency action is rational and fair. In addition to preserving delegation, courts would still police its exercise. Thus, courts would ensure that agencies remain within the scope of their delegated authority and issue reasonable interpretations.25

I acknowledge that widespread adoption of a “delegationrespecting” approach has downsides. Even if the Supreme Court is capable of applying it with a reasonable degree of accuracy, lower courts might face considerable difficulties. Mead has already introduced confusion on the issue of legislative intent to delegate, and that case is restricted to a single procedural consideration.26 A delegation-respecting approach, which essentially articulates and develops Mead’s basic insight, might make matters worse. Although these concerns have merit, I believe that courts and commentators should not condemn the approach based solely on lower courts’ experience with Mead. The Court has made applying that decision

more complicated than it needs to be, and my theory offers significant clarifications. If courts and commentators reject all other suggestions from this Article, they should accept the suggestion to embrace and reform Mead. Mead moves in the right direction by making agency delegation rather than statutory meaning the threshold question. Even restricted to a gatekeeper function, precluding delegations rather than validating them, Mead advances legislative supremacy and rule-of-law values. Refocusing on the legislative use of procedures would provide courts with more guidance and impose fewer institutional costs.

In the end, I acknowledge that no theory based on legislative intent to delegate will yield a simple rule. Ultimately, simplicity is not the sole benchmark for evaluating any interpretive theory. An acceptable theory should reflect a reasonable balance among the various goals of statutory interpretation.27 If courts and commentators nevertheless reject a delegation-respecting theory on institutional grounds, as Justice Scalia advocated in Mead,28 they should also reject the search for meaning on other grounds. The overarching lesson from this Article is that the search for meaning gives courts too much power to frustrate interpretive delegations when Congress is likely to intend them and when they are likely to promote other normative values. Even if sufficiently simple from an institutional standpoint, existing law cannot supply the relevant “rule” for normative reasons. In its place, courts and commentators should be prepared to adopt a presumption of judicial deference to agency interpretations—one stronger than Chevron’s presumption ever has been.

Finally, I discuss the implications for the use of statutory interpretation to constrain broad delegations. Ever since the demise of the constitutional nondelegation doctrine, courts have relied on statutory interpretation as a surrogate for addressing concerns about broad delegations.29 Agencies can use broad delegations to produce

undemocratic policies, impair states’ rights, raise constitutional questions, and impose harsh criminal penalties. On one view, Chevron and Mead are not wrong and are framed to allow courts to address these concerns. My argument suggests this view can no longer be premised on congressional will, as it sometimes is. Congress is likely to intend the very delegations that courts (and commentators) aim to restrict. Still, a delegation-respecting theory does not preclude courts from using statutory interpretation to address delegation worries if they justify that practice on other grounds. There is a better way. Analytically, it makes more sense to address nondelegation worries statutory interpretation. Specifically, I argue that courts should remand agency interpretations rather than narrowing regulatory statutes. This role allows courts to express their views without overstepping their bounds. In addition, it affords agencies an opportunity to craft better policy by applying their specialized expertise and consulting with the political branches.

through administrative

rather

than

law

This Article proceeds as follows. Part I demonstrates that the conventional theories of statutory interpretation share a basic search for meaning, which administrative law doctrine does not sufficiently alter. Yet the search for meaning is inappropriate because it does not match how Congress designs regulatory statutes. Part II sketches an interpretive theory that better matches legislative realities, departing in significant respect from existing doctrine. Part III offers two Supreme Court opinions as examples of how the theory might work in actual practice. Part IV considers institutional and normative objections to the theory and offers some responses.

Although the dominant theories of statutory interpretation differ in many ways, they share an important feature: they invite courts to construe statutory text as if Congress intended that text to have a relatively specific meaning. That presumption does not reflect how Congress designs regulatory statutes, as administrative law doctrine recognizes. In this Part, I show that existing doctrine, in particular Chevron and Mead, nonetheless allows this presumption to dominate and therefore allows courts to deprive agencies of interpretive authority more often than Congress intends. To demonstrate the full extent of the problem, I set forth the core aspects of regulatory statutes that positive political theorists and legal scholars have identified.

A. The Search for Statutory Meaning

The conventional theories of statutory interpretation take varying approaches to awkward or imprecise text in part because they hold different pictures about legislative behavior. Intentionalists and purposivists see legislative inadvertence or, more specifically, the inevitable difficulty of capturing all the aspects or applications of a policy in a relatively few words.30 According to these theorists, Congress has a meaning or a purpose in mind but sometimes chooses words that poorly or incompletely express it. Intentionalists seek to recover the meaning when interpreting words, using legislative history when relevant, and purposivists look for indications of broader statutory purposes.31 Modern proponents of these theories

look for more objective evidence on the assumption that actual legislative intent may be hard to reconstruct given the complexities of the legislative process; but they assume that every enacted law has a reasonable purpose or an “intentionalist stance.”32

By contrast, modern textualists trace awkward or imprecise text to legislative compromise.33 They view the legislative process as chaotic and messy. In this environment, legislators cut deals to obtain consensus, and awkward words reflect those deals.34 Modern textualists adhere to the ordinary meaning of those words to give effect to whatever deal they may manifest.35 Textualists also maintain a rule-of-law or constitutional defense against the use of legislative history because only the text is enacted.36

On the basis of their respective visions, proponents of each theory ask courts to announce the statutory meaning that best reflects what Congress was after—for example, a broad statutory purpose or a specific legislative deal. By construing language in this fashion, those proponents all can claim to position courts as faithful agents of Congress. Undoubtedly, the different camps believe that their theory will produce a statutory meaning that most often reflects the legislative design.

The basic law governing interpretation of regulatory statutes reflects a different picture of legislative behavior. Chevron establishes a two-step test for courts to apply when reviewing agency interpretations of regulatory statutes. The Chevron test tells courts first to determine whether “Congress has directly spoken to the precise question at issue,” applying the “traditional tools of statutory

construction.37 But when courts find no meaning for the statutory text, Chevron instructs them to defer to the agency interpretation as long as that interpretation is reasonable.38 Chevron recognizes that Congress may intend for agencies rather than courts to fill gaps in regulatory statutes.39 It notes that Congress may have a variety of interpretive authority to agencies—for reasons for delegating example, to capitalize on agency expertise or to obtain legislative consensus.40 Regardless of the particular reason, Chevron directs courts to accept the legislative assignment of interpretive authority and defer to reasonable agency interpretations.41 Doing so is consistent not only with congressional delegation but also with administrative expertise and political accountability. Agencies possess more expertise than courts for handling regulatory schemes that are “technical and complex” and for reconciling the “competing interests” that regulatory decisions often involve.42 Agencies are also accountable to the people, not directly but through the president, and “it is entirely appropriate for this political branch of Government to make such policy choices.”43

Mead recognizes that whether Congress intends an agency to issue an interpretation with the force of law depends on whether Congress has authorized and the agency has used certain procedures for that purpose. Mead involved an interpretation that the United States Customs Service had issued through a ruling letter addressed to a party seeking guidance on a tariff classification.44 These letters come from forty-six different offices of the agency at a rate of 10,000 to 15,000 per year without an opportunity for public participation and are generally unaccompanied by a reasoned explanation.45 The Court held that Congress would not intend to delegate interpretive authority to an agency through such a procedure.46 In this way, Mead suggests that Congress would not intend to delegate interpretive

formal procedures,

authority to an agency absent the authorization (and use) of such as notice-and-comment sufficiently rulemaking or formal adjudication. Procedural formality ensures “fairness and deliberation” and “bespeak[s] the legislative type of activity that would naturally bind more than the parties to the ruling.”47

Although Chevron and Mead offer a view of legislative behavior that is appropriately different from the standard view, neither directs courts to adopt a sufficiently different (or better) method of statutory interpretive interpretation. Chevron mentions the reasons for delegation as a justification for judicial deference. But it actually directs courts to approach awkward or imprecise statutory language by asking the same question as the standard theories of statutory interpretation: whether “Congress has directly spoken to the precise question at issue.”48 And it tells courts to answer that question applying the “traditional tools of statutory construction.”49 Thus, courts effectively approach interpretive questions mindful of giving effect to broad statutory purposes or finely tuned legislative deals. Only when that effort reveals no clear meaning do courts infer a delegation of interpretive authority.50 In other words, courts treat a lack of clarity as the exclusive proxy for interpretive delegation. Perhaps, then, it is no wonder that a wide range of legal scholars have characterized the congressional delegation rationale for Chevron as a fiction.51 Because Congress probably does not draft statutes with Chevron in mind, courts can justify judicial deference, if at all, using the other values that Chevron cites—agency expertise and political accountability.52

Mead adds a procedural dimension to the analysis,53 making legislative intent to delegate the threshold inquiry of statutory interpretation. But the Mead opinion does not carry the procedural consideration through to its proper conclusion. Mead deprives agencies of interpretive authority when procedures are deficient because the procedures lack the law-like features that Mead emphasizes. Mead has essentially no effect, however, when the statute authorizes and the agency uses relatively formal procedures—as when an agency interpretation is the product of notice-and-comment rulemaking. Even if the agency passes the Mead analysis, courts proceed to apply Chevron as usual, effectively regarding words as poor expressions or concrete deals unless that approach fails to produce a clear meaning. Because Mead gives way to Chevron in routine cases, it does not go far enough to alter the standard search for statutory meaning.

The result is that, Chevron and Mead notwithstanding, courts approach regulatory statutes with much the same mindset as they approach other statutes. It follows that they are just as likely to find a specific meaning in a case involving an agency as in a case that does not. This is not an empirical claim, although it well might be true. Rather, it is an analytical claim about the framework that courts apply. Because courts approach awkward or imprecise text in regulatory statutes mindful of respecting broad statutory purposes or particular legislative deals, they are unlikely to grant agencies their delegated interpretive authority as often as Congress intends.

Chevron cannot overcome the problem by simply instructing courts to set aside their meaning in favor of an agency’s “reasonable”

meaning. The thought is a good one, but it is easier said than done. Once courts work actively to construct a meaning for statutory text, as they do under all the dominant theories of statutory interpretation, they will have considerable difficulty recognizing the reasonableness of other interpretations. As Professor Thomas Merrill has observed about textualism:

ingenuity

less risk of in judicial Professor Merrill sees intentionalism because the judge remains more cognizant of legislative intent and has more tools available for finding such intent.55 In my view, intentionalism and purposivism are still vulnerable to judicial craft because they also ask courts to construct a meaning for statutory text on the assumption that the text has one right answer. This way of thinking makes it difficult to see other possible meanings and, moreover, to accept the assignment of interpretive authority to the agency. Professors Jacob Gersen and Adrian Vermeule have gone even further to suggest that judges cannot set aside their meaning for the agency’s meaning.56 Once a court has found that the text is susceptible to a best reading, Professors Gersen and Vermeule question what it means as a conceptual matter for the text to have

other “reasonable” readings.57 How can an incorrect interpretation nevertheless be reasonable? They also question whether courts psychologically can abandon the view to which they have become committed.58 The authors quote Justice Breyer, who wrote,

search

for meaning

Somewhat counterintuitive,

is problematic whether a court finds a meaning that precludes an agency interpretation or permits it. Most worrisome is when a court finds a meaning that precludes an agency interpretation because, under such circumstances, the court is able to substitute its idea of wise policy for that of the agency. As Chevron recognizes, courts should restrain themselves in this regard because the agency is the congressional delegate as well as the more technically sophisticated and politically responsive policymaker.60 Yet the search for meaning is also worrisome when a court finds a meaning that is consistent with an agency interpretation. It is worrisome because an agency, though victorious in the short term, may lack authority to alter its interpretation as circumstances change. Chevron anticipates that agency interpretations may evolve over time, whether as a result of emerging technologies or new administrations.61 But the Court has clarified in a subsequent decision that such interpretive room only exists when the statutory text is ambiguous, not when a court has

summarize,

conventional

previously determined that it is clear.62 When a court finds that the statutory text has a specific meaning, it forecloses future changes, even if flexibility is part of why Congress created the agency in the first instance. To

statutory interpretation invite courts to answer a question about statutory meaning, and neither Chevron nor Mead sufficiently displaces that inquiry. As a result, courts are likely to find a specific meaning for statutory text more often than Congress intends. The remaining is Sections explain more fully why the search for meaning inappropriate for regulatory statutes. It does not match what many positive political theorists and legal scholars have been saying about how Congress designs regulatory statutes.

theories of

B. The Possibility of Agency Delegation

Political scientists have sought to explain why Congress delegates authority to agencies.63 Although many strands of such work might be useful in illustrating the set of factors that motivate Congress, I rely here on positive political theory (PPT) because it offers hypotheses about legislative behavior that are comparable to those on which the dominant theories of statutory interpretation have tended to rely. PPT thus provides a helpful platform from which to evaluate those theories.

Positive political theorists start from the premise that Congress is composed of members who, as rational actors, primarily seek to improve their reelection chances.64 They may have other goals—“such as the desire for power, rewarding friends, and good government”— but they cannot achieve any of these goals unless they hold office.65 To improve their reelection prospects, members of Congress seek policies that track their preferences or, more specifically, the

preferences of their constituents.66 One way to ensure that policy tracks legislative preferences is for Congress to specify that policy— that is, to write “detailed, exacting laws.”67 But this approach is not always best. In certain circumstances, Congress may determine that delegation to agencies is actually the better way to serve its interests.

Consider first and foremost that Congress has limits on its time and expertise.68 Every resource devoted to specifying policy is a resource unavailable for other “electorally productive activities,” including campaign activities.69 When Congress enlists the aid of agencies to set regulatory policy, it might do so because it lacks the requisite time and expertise to formulate the details of such policy. As Professors David Epstein and Sharyn O’Halloran explain, the reasons are twofold:

Thus, Congress cannot gain expertise as efficiently as agencies, and it cannot transform its expertise into policy solutions as easily as agencies. The costs of writing specific legislation are high, indeed wastefully so. Congress therefore is likely to delegate more authority to agencies “[t]he more complex [the] policy area” is.71

Consider next that Congress balances a variety of other interests in deciding how much authority to delegate to agencies. For example, Congress might attempt to avoid blame for controversial policy choices by shifting them to agencies, while still claiming credit for broad solutions to public problems.72 In other words, Congress might aim to write just enough policy to receive a positive response for its actions, while deflecting any negative attention for the burdensome details to the agency. Legislators also might consider whether to allocate interpretive authority to courts or agencies by assessing the relative risk that each legislative preferences. For example, a legislator who holds views consistent with the preferences of the median legislator might prefer judicial interpretation, whereas a legislator who holds views different from the median legislator might prefer agency interpretation. The hypothesis is that courts are more likely to issue decisions that track the preferences of the median legislator, and agencies may diverge in other directions.73 Legislators may base this choice on a particular distinction. Judicial decisions are more predictable over time but more variable across issues than agency decisions.74 Congress’s

institution will depart from

allocation of interpretive authority in a particular instance depends on whether the median legislator prefers policy stability (courts) or coherence (agencies).75

Another dimension is also relevant: political alignment between the legislative branch and the executive branch.76 Professors Epstein and O’Halloran have argued, for example, that Congress is likely to delegate more authority to executive branch agencies when the president is of the same political party, that is, during periods of unified government.77 In such periods, Congress may reasonably assume that presidential preferences are less likely to diverge from legislative preferences.78 By the same token, Congress is likely to delegate to independent agencies, which are run by officials who are protected by statute from plenary presidential removal, if it desires delegation but the president is from the opposite political party.79

These observations connect to a broader point: whatever the motivation for delegating to agencies, positive political theorists agree that Congress remains aware that agencies may choose policies that depart from legislative preferences.80 Put simply, delegation creates a principal-agent problem.81 Agencies are influenced from a number of directions—“by the President, by interest groups, by the courts, and by the bureaucrats themselves”—and those pressures may push regulatory policies away from legislative preferences.82 Congress

therefore seeks to monitor agency action, but it must devise mechanisms for acquiring information about that action.

Professors Mathew McCubbins and Thomas Schwartz have identified two sorts of monitoring mechanisms: “police patrols” and “fire alarms.”83 Police patrols are committee hearings and other direct forms of oversight.84 Fire alarms are indirect forms of oversight; they enlist private parties to gather information about agency action.85 Professors McCubbins, Roger Noll, and Barry Weingast (McNollgast) have argued that Congress can use the procedures in the Administrative Procedure Act as fire alarms.86 More specifically, Congress can use administrative procedures to place constituents into the administrative process, where they may acquire the information necessary to evaluate agency action.87 If displeased with that action, they may alert members of Congress to intervene before the agency has altered the status quo.88 McNollgast also contend that Congress can use procedures to “stack the deck” in favor of the same constituents that supported a regulatory statute by affording those constituents access to agency decisions before they are final.89 Administrative procedures are particularly efficient and effective monitoring mechanisms, shifting the monitoring costs from Congress to its constituents.90

For Congress, ex post oversight mechanisms,

including administrative procedures, form a “precondition” of delegation91 and also “make the benefits of statutory control less compelling.”92 Thus, ex post oversight mechanisms and statutory control (through specific

wording) can act as “substitutes” for one another.93 The presence of one can suggest less need for the other.

All of these observations furnish assumptions about the considerations that matter to Congress in allocating regulatory authority. None is without criticisms. The important point is that positive political theorists are in basic agreement that Congress is strategic in how it designs regulatory statutes, and that such strategic behavior is relevant in examining regulatory statutes.

C. The Possibility of Deliberate Ambiguity

Another phenomenon is also relevant. Both PPT scholars and legal scholars have observed that a divided Congress may choose deliberately ambiguous words to obtain consensus, thereby delegating interpretive authority to agencies or courts.94 Thus, Congress can use awkward words to achieve legislative compromise, as textualists assert. But the compromises are more like dodges than deals.95 Congress chooses words that are imprecise enough for legislators with opposing views each to claim victory. Meanwhile, the language also allows those legislators to press for their favored positions at the later administrative or judicial level.96 The ultimate losers may avoid blame by laying it at the feet of the responsible agency or court. By choosing words that “mean all things to all people,”97 Congress can obtain the

requisite support to enact a bill while preserving opportunities to recommence the battle at another time and in another place.

Professors Victoria Nourse and Jane Schacter have identified the phenomenon of deliberate ambiguity empirically by interviewing legislative staffers.98 According to Professors Nourse and Schacter, the staffers recognized that a decision to create ambiguity left future resolution to courts and agencies.99 Nevertheless, the staffers stated that unless legislators can get a bill passed, they can do nothing at all.100 But for some, the delegation was as intentional as the ambiguity. Staffers “realized that statutory ambiguity created an opportunity to let an agency, as opposed to a court, resolve the issue, and sometimes they specifically desired this result as well.”101

Professors Joseph Grundfest and Adam Pritchard have identified an example of this phenomenon in the Private Securities Litigation Reform Act of 1995 (PSLRA).102 I set out this example in some detail because I refer to it later.103 The PSLRA arose under circumstances suggesting that Congress could not have enacted the statute without agreeing to disagree about the applicable pleading standard for liability under Section 10(b) of the Securities Exchange Act of 1934.104 Before the PSLRA, the Supreme Court had held that mere negligence did not satisfy the scienter requirement.105 Lower courts that addressed the issue diverged on whether a “barely reckless” or “highly reckless” standard applied.106 Furthermore, the lower courts divided on the actual pleading requirements: whether plaintiffs must “allege facts that give rise to a strong inference of fraudulent intent,” as the Second Circuit determined,107 or could get away with merely

“saying that scienter existed,” as the Ninth Circuit found.108 The PSLRA “purported to resolve this conflict by adopting a uniform standard for pleading scienter, the ‘strong inference’ standard.”109 But the pleading issue actually implicated the underlying scienter issue: whether recklessness suffices, and if so, what recklessness means.

Initially, the House bill required actual knowledge, precluding recklessness altogether, but the version that passed contained a “strong inference” pleading standard for recklessness.110 The Senate bill also contained the strong inference standard, and the relevant committee report explains this choice as following that of the Second Circuit.111 But the issue was not that simple. The Second Circuit had not specified exactly what the strong inference standard required. Senator Arlen Specter introduced an amendment that included more precise considerations based on Second Circuit case law.112 The Senate adopted the amendment, but the conference committee deleted it.113 The conference committee report contains conflicting signals on whether the committee preferred a weaker or stronger standard.114

Meanwhile, the PSLRA itself “goes to lengths to sidestep the issue by using ‘required state of mind’ in the pleading provision, rather than knowledge or recklessness.”115 As Professors Grundfest and Pritchard explain,

According to Professors Grundfest and Pritchard, this agreement to disagree was essential to the passage of the legislation. As further evidence, the authors note that the Senate barely mustered enough votes to override a presidential veto—President Clinton’s first, and one based expressly on the pleading standard.117 In his veto message, President Clinton stated that he opposed a pleading standard higher than that of the Second Circuit and highlighted language in the conference committee report that might support such a position.118 Preserving the deliberate ambiguity on this issue was critical for members of Congress: “The loss of even two votes in the Senate due to disagreement over the ‘strong inference’ standard would have doomed the legislation.”119 Thus, in the floor debate concerning an override vote, the Senate managers of the PSLRA who shared the president’s view of the pleading standard nevertheless distanced themselves from the president’s remarks as well as the related conference committee report statements to gain the requisite support for the legislation.120

Professors Grundfest and Pritchard join others in contending that the possibility of deliberate ambiguity creates complications for the dominant theories of statutory interpretation.121 Those theories, despite their differences, have a central commonality. They seek clarity when none may exist or at least where Congress may intend that none exists. Professors Nourse and Schacter contend that the dominant theories are therefore either implausible or disingenuous: “[T]he fictions deployed in judicial opinions are a kind of diversion, allowing judges to exercise significant discretion in determining

statutory meaning while attributing their discretionary choices to Congress.”122

In sum, positive political theorists and legal scholars posit that Congress may delegate to agencies the task of giving statutory provisions more precise meaning for certain reasons and under certain conditions. Although they differ on the details, they agree that these reasons and conditions are important to thinking about the legislative design of regulatory statutes. Chevron and Mead are in general agreement, but they do not adequately integrate these reasons and conditions into statutory interpretation. For the most part, they leave in place the conventional theories and those theories’ shared search for statutory meaning. Under these circumstances, it is questionable whether courts can accurately or reliably determine whether Congress has delegated interpretive authority to agencies.

This Part imagines what interpretive theory would look like if it were based on the considerations that are thought to induce Congress to delegate interpretive authority to agencies. A delegationrespecting theory would require courts to consider indications of interpretive delegation like the ones that PPT and legal scholarship have identified. To be clear, I do not claim that either PPT or legal scholarship provides a means for translating their general hypotheses into an or empirical observations about interpretive theory. Rather, I argue that PPT and legal scholarship identify the right sort of signals for courts to track when interpreting regulatory statutes. Courts tracking these signals would be less likely to read statutory language as clear on the assumption that Congress intended that result.

legislative behavior

This Part also addresses how courts, applying a delegationrespecting theory, would handle related issues. If courts generally disregard statutory language, how would they ensure that agencies remain within the scope of their delegated interpretive authority? If courts are generally interested in respecting interpretive delegations, would they continue to review the reasonableness of agency interpretations? In the absence of sufficient evidence that Congress

delegated interpretive authority to an agency, how would courts resolve interpretive questions? This Part saves evaluation of the theory for the final Part.

A. Reasons for Interpretive Delegation

Congress

likely to delegate authority to agencies for predictable reasons—primarily to avoid resolving complex and contentious issues that, if resolved, would defeat the enactment of legislation or would generate other costs to legislators.123 Under a theory directed at this behavior, courts would identify one or both of these reasons as evidence of an interpretive delegation by Congress to the relevant agency. I do not suggest that courts engage in legislative mind reading, trying to recover actual legislative reasons for interpretive delegation. Rather, courts should recognize that Congress is likely to delegate certain issues based on their nature. Moreover, courts should recognize that this delegating impulse does not vary whether the grant of authority to the agency is explicit or implicit. A complex or contentious issue should motivate delegation to an agency whether Congress responds by enacting an express mandate or choosing an unspecific term. There is one difference. Without an express mandate, courts must determine whether an issue is sufficiently complex or contentious to prompt an interpretive delegation. To do so, they would examine the text, purpose, and background of the statute.

Chevron is not contrary to this general focus on the nature of the issue. It observes that Congress may delegate interpretive authority to an agency “thinking that those with great expertise and charged with responsibility for administering the provision would be in a better position to do so.”124 Chevron also acknowledges that Congress may delegate interpretive authority to obtain legislative consensus.125 But Chevron does not ask courts to examine whether an issue is complex or contentious as part of their interpretive delegation analysis, which is the suggestion here.

If courts are to evaluate the nature of the issue, practical questions abound. First, what constitutes a complex issue? From a legislative standpoint, a complex issue is a technical issue, requiring

both time and expertise. If the statutory provision later to be interpreted concerns a technical matter, requiring the acquisition or assessment of specialized information, it is likely the sort of subject that generalist staffers or legislators are unwilling or unable to handle, even with the benefit of outside consultants.126 A provision that depends on science, economics, or statistics falls into this category. If a provision requires applying specialized skill, it is also likely the kind that Congress may delegate to agencies.127 Thus, Congress may possess information but still encounter difficulty translating that information into appropriate rules—for example, rules governing acceptable risk of illness or injury, an acceptable level of drug effectiveness, or an acceptable degree of market concentration.

A complex issue may also be a subset of a more general issue. Chevron included as a reason for interpretive delegation that perhaps Congress “simply did not consider the question at this level.”128 Congress can opt for generality to conserve time.129 By delegating the details, Congress may claim credit for broad statutory responses while avoiding blame for inevitable regulatory burdens.130 Courts need not focus on the precise legislative gains from generality. Rather, they might simply attempt to find such generality in the text or structure of the statute.

Under this approach, when would a court conclude that a regulatory issue is not so complex as to suggest delegation? Truthfully, the answer is infrequently. Congress will predictably enlist agency judgment not only to arrive at wise solutions to public problems but to avoid the obligation to choose those solutions itself. Moreover, if Congress has delegated authority to an agency because of the complexity of a regulatory regime, it likely did not intend to withhold authority from the agency over interpretive questions. Those questions are likely to share the same general character as the

regulatory regime to which they are relevant. Thus, courts might expect more often than not to start their interpretive analysis with a thumb on the scale for interpretive delegation. Under the existing framework, courts do not start from here. The search for meaning more likely predisposes them away from interpretive delegation. In this sense, the approach that I advocate is potentially more deferential than Chevron.

One might still imagine that some issues are the sort that Congress did not intend to commit to an agency either because the agency lacked the requisite expertise or because Congress is capable of resolving the issue itself. For example, Congress can make moral or value judgments, such as whether to impose a criminal sanction for a regulatory violation.131 Even though the regulatory regime is complex, the sanctions question within involve specialized knowledge or skill. In this sort of case, complexity, standing alone, does not indicate an interpretative delegation.

it does not

Another common reason for supposing Congress has delegated interpretive authority to an agency is the need to short-circuit extended legislative battles over contentious issues. As Chevron contemplates, it will sometimes be the case that Congress “was unable to forge a coalition on either side of the question, and those on each side decided to take their chances with the scheme devised by the agency.”132 A contentious issue is one subject to active debate

between legislative coalitions, the resolution of which in the statute for either side might derail the law’s passage. Unlike complex issues, contentious issues are not evident from the text or structure of the statute. Courts must look to the surrounding circumstances, as Professors Grundfest and Pritchard did in examining the PSLRA. As summarized in Part I, they performed an extensive reconstruction of the legislative history and political backdrop of the statute, finding an agreement to disagree on particular issues and consequent delegation of interpretive authority on those issues.133

In their example, the interpretive delegation flowed to the courts not the SEC,134 which illustrates an important point about this factor. Courts must uncover evidence not only of an agreement to disagree but also of a consequent delegation to the agency. Congress likely intended courts to interpret the PSLRA or at least knew that they would. The explanation is straightforward: courts have been the primary interpreters of securities law in the context of private class actions.135 With respect to other statutes, a court might discover an indication, either in the legislative history or a background norm, that Congress intended or expected the agency to exercise interpretive authority.

A court might discover no indication that Congress had any preference other than to avoid resolving the contentious issue. Under such circumstances, the contentiousness of the issue alone would not support a finding of interpretive authority to the agency. Whether the agency can assert interpretive authority would depend on other factors. Note, however, that even if courts can claim interpretive authority, they might exercise that authority differently, as I explain in Section C.3.

The bottom line is that courts would focus on legislative reasons for interpretive delegation by examining the nature of the issue.

Courts would view Congress as having delegated interpretive authority to an agency when confronted with a complex or contentious issue. When courts identified issues as such, they would be building an affirmative case for interpretive delegation. The stronger the affirmative case, the less courts should regard the text as a legislative attempt at a clear meaning that bars the interpretive delegation.

B. Conditions of Interpretive Delegation

that

subsequent agency action

A focus on legislative reasons is only half the case. When Congress delegates for any reason, it has an ongoing interest in ensuring legislative preferences.136 Competing legislative coalitions wish to influence subsequent agency interpretations in their preferred directions.137 Yet Congress cannot influence what it cannot see. Thus, Congress is more likely to delegate under certain conditions—those that facilitate legislative monitoring of agency interpretations. If courts are going to build an affirmative case for delegation, they must also focus on these monitoring conditions.

tracks

Administrative procedures are a mechanism that facilitates legislative monitoring.138 Mead aligns with this view, making procedures relevant on the question of legislative intent. This is an important move the direction of recognizing appropriate assumptions about legislative behavior in the regulatory context.139 Based on that premise, it is no stretch to suggest that courts ought to recognize procedures as a legislative condition of interpretive delegation. But Mead botches the implementation. First, it fails to connect procedures with legislative monitoring. Instead, it links procedures to rule-of-law values, stating that proper procedures “foster fairness and deliberation” and “bespeak the legislative type of activity that would naturally bind more than the parties to the ruling.”140 Meanwhile, Barnhart v. Walton,141 which the Court decided just one term later, highlights other factors, including “the interstitial nature of the legal question, the related expertise of the Agency, the

importance of the question to 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.”142 These cases produce competing frameworks for atypical procedures, which has confused lower courts and produced conflicting results.

Moreover, neither approach adequately produces a result that reflects what Congress needs for monitoring purposes. For example, a court might reject an interpretation rendered through an informal adjudication because the informal adjudication lacked the “fairness and deliberation” in the sense that Mead means.143 Yet that particular procedure may have been sufficient for legislative monitoring purposes because it provided constituents with access to information about the agency interpretation ahead of time. A court might approve an interpretation announced in a brief prepared for court litigation because the brief reflected the expertise and “careful consideration” that Barnhart contemplates.144 At the same time, the brief may have evaded legislative monitoring because it was not available to constituents in advance of litigation. When focusing on administrative procedures, courts should consider their strategic use to Congress in overseeing agency interpretations. In so doing, they would better track the conditions under which Congress delegates interpretive authority to agencies.

Second, Mead fails to recognize that sufficiently formal procedures, if a necessary condition of delegation to an agency, should count in favor of an interpretive delegation to the agency, not just against it. Mead deprives agencies of interpretive authority unless they possess and use sufficiently formal procedures. Thus, it can have a negative effect on agency interpretive authority. But once an agency passes Mead muster, courts apply Chevron as usual. In other words, they search for a meaning behind particular language, as if Congress intended such a meaning. Courts could infer more of an affirmative signal from adequate procedures, understanding these procedures as part of the case for interpretive delegation to the agency. In this way, procedures would help to displace the search for meaning rather than facilitate it.145

Congress can seek other means for ensuring that agency action roughly tracks legislative preferences.146 It can have a proxy, such as the party of the president. Or, to go further, Congress can have more concrete opportunities to assess agency interpretations. For example, an agency may be involved in drafting the legislation, sharing its interpretations with legislative staff. If those understandings are sufficiently specific, Congress may rely on them when enacting legislation. In addition, an agency may maintain a long-standing policy or practice that Congress can observe when enacting legislation. Unlike procedures, agency involvement or long-standing practice is not evident on the face of the statute. Rather, courts must determine these aspects from the surrounding circumstances.

interstitial nature of the

Agency involvement or long-standing practice can function in lieu of procedures from a legislative standpoint, providing sufficient assurance that agency interpretations will roughly track legislative preferences. This is one way to understand Barnhart. In Barnhart, the Court upheld an interpretive delegation not because the procedures met with the criteria that Mead identifies for “force of law” purposes but for a host of seemingly unrelated factors. Viewed through a legislative lens, most of the factors sound like reasons for delegation, legal question,” “the including “the importance of the question to the administration of the statute,” and “the complexity of that administration.”147 The final factor—the “careful consideration that the Agency has given the question over a long period of time” suggests that the agency policy was well worked out and unlikely to change.148 Congress could observe it and reasonably rely on it. Thus, the Court upheld an interpretive delegation because Congress had a reason for the delegation and an assurance that the agency interpretation would mirror its longstanding practice. On this account, Barnhart does not offer an alternative test for measuring the adequacy of procedures, as lower courts have concluded. Rather, it offers an example in which Congress could evaluate the agency interpretation for consistency with legislative preferences through other means.

Congress can include procedures notwithstanding a prior agency interpretation, and it is likely to do so. Procedures are useful for making policy decisions in general. In addition, authorizing procedures may suggest that Congress anticipated the possibility of change and provided for legislative monitoring in the event of change. Under such circumstances, Congress intended a delegation but not exclusively in reliance on the prior agency interpretation. Procedures, more so than the prior agency interpretation, furnish the requisite assurance legislative preferences.

interpretations will

that agency

track

To summarize, Congress is more likely to delegate when it can ensure that subsequent agency interpretations will roughly track legislative preferences. It can use procedures for that purpose or rely on positions that the agency has maintained before or taken during the course of legislative drafting. On the account I defend, courts would treat the presence of these factors as indications that Congress intended an interpretive delegation.

C. Application Issues

The purpose of focusing courts on political indicia of interpretive delegation would be to prevent them from reading statutory language as clear or otherwise depriving an agency of interpretive authority based on inappropriate assumptions about legislative behavior. This Section addresses several residual issues that courts would have to confront if they adopted this framework: (1) ensuring that agencies remain within the scope of their delegated interpretive authority, (2) reviewing the reasonableness of agency interpretations, and (3) resolving interpretive questions in the absence of sufficient indications of interpretive delegation to an agency.

1. Scope of Agency Interpretive Authority. Even if courts found sufficient indications of interpretive delegation to an agency, they may also encounter instances in which an agency has exceeded the scope of its authority. Courts would still police this boundary, as they always have done.149 An agency cannot interpret a statute to reach a subject that it does not address or to include a policy that it rules

out.150 A simple illustration: an agency cannot interpret the word “clothing” to cover “automobiles,” no matter how broad the interpretive delegation of that term.151 Congress must be able to rely on the communicative aspect of words at some very general level, to limit the reach of a statute and preclude certain considerations. In a sense, the constitutional nondelegation doctrine so requires.152 Under the nondelegation doctrine, every regulatory statute must contain an “intelligible principle” constraining and guiding agency action.153 That intelligible principle can be quite minimal or vague, and it usually is.154 Nevertheless, the statute cannot confer a blank check on the agency.155 Congress must employ some language that can be understood to delineate the boundaries of administrative authority.

But courts must be restrained in making scope-of-authority determinations. When courts consult statutory language to ensure that the agency has remained within the scope of its power, they must avoid reintroducing the problematic search for meaning. Thus, courts should not ask what “clothing” means in a relatively specific sense, either as written or in light of broad statutory purposes. Rather, they should seek to determine whether “clothing” as conventionally defined or understood can include “automobile” without depriving the statute of a communicative limit. This approach is appropriate because Congress need not precisely define “clothing” to exclude “automobile.”

As a practical matter, courts should rarely find that an agency has exceeded the scope of its authority. The point about “clothing” and “automobiles” is that agencies may not choose readings that are

so illogical as to constitute virtual category mistakes or polar opposites. Thus, a statute directing an agency to regulate x does not authorize an agency to regulate y.156 But the legislative act of precluding an agency from selecting y does not obligate an agency to regulate x in any particular manner. Indeed, that act says very little about x. When an agency regulates x in an unreasonable manner, a court should not conclude that the agency has exceeded the scope of its authority. Courts should instead address such errors through a different, more forgiving channel, as Section C.2 shows.

2. Reasonableness of Agency Interpretations. Asking courts to use different proxies for determining whether Congress has delegated interpretive authority would not diminish courts’ responsibility to evaluate the reasonableness of an agency interpretation. Courts have always reviewed agency interpretations for reasonableness whether under Chevron157 or the arbitrary and capricious test from the Administrative Procedure Act (APA).158 Although the relationship between the Chevron inquiry and the arbitrary and capricious test has confused courts, the effect of each is much the same.159 Agency interpretations, like all agency policy decisions, must comport with the reasoned decisionmaking requirement.160

The reasoned decisionmaking requirement has a long lineage in law, predating the APA.161 It rests on certain administrative assumptions about the administrative process—that absent judicial supervision, agencies would be insufficiently attentive to certain considerations, whether statutory factors, policy aspects, alternative solutions, or party comments.162 Courts have sought to address these pathologies by asking agencies to issue along with their decisions an explanation for those decisions.163 The reasoned decisionmaking requirement serves important normative values. For example, the requirement ensures that agency decisionmaking comports with ruleof-law values, like rationality and consistency.164 In addition, it ensures that the administrative process is transparent and accessible, enhancing political accountability and public participation.165

It is possible to challenge the validity of these assumptions or the pursuit of these goals. I believe that preserving reasonableness review becomes even more important once courts adopt a different approach to statutory interpretation. Administrative law becomes the main vehicle for disciplining agency interpretations that depart from accepted norms, such as rationality and accountability. I discuss in Part IV the role of administrative law in responding to additional concerns about agency interpretations.

3. Exercise of Judicial Interpretive Authority. If a court fails to find an interpretive delegation to the agency involved, the court possesses the relevant interpretive authority. This is not surprising: if an agency has no interpretive authority, then it follows that the court must. But the issue is actually more complicated. How courts should exercise interpretive authority depends on why they possess such authority. Existing law tracks this insight in some places and not in others.

Assume a court finds that an agency lacks interpretive authority because Congress had no reason to delegate. The issue is not technical. The agency involved possesses no relevant expertise for resolving it. No evidence suggests that the issue generated any controversy that might have obstructed the law’s passage. The reasonable inference under these circumstances is that Congress did not intend to delegate the issue at all—for resolution by any other actor, agency, or court. Congress either misstated its intent or struck a deal. As a result, courts should interpret any awkwardness or imprecision in the statutory text as matching their customary assumptions about legislative behavior. If ever there is a place for Chevron as usual, this is it.

Assume a court finds instead that Congress agreed to disagree but designated the courts as the relevant interpreter—as in the PSLRA context. Courts have special experience interpreting securities laws in class action suits. When courts are recipients of delegated interpretive authority, should they exercise their authority by applying the conventional theories of statutory interpretation? Because Congress had in mind no specific meaning for the relevant statutory language, courts are justified in relying less on the traditional tools of statutory construction, particularly textual analysis. For guidance, they might consult other sources, such as the agency’s practical experience with the regulatory regime. This approach mirrors the one from Skidmore v. Swift & Co.,166 which directs courts to consult the agency’s views of wise policy if they are “persuasive.”167

What if Congress failed to authorize procedures that facilitate legislative monitoring? This deficit suggests that Congress did not intend to delegate interpretive authority to the agency. Otherwise, it would have sought a means for ensuring that the agency did not depart from legislative preferences. Assuming no agency involvement or longstanding practice, Congress likely intended the statutory language to be clear, or it meant to delegate interpretive authority to courts. In either case, courts fill any gaps. Existing law does not attempt to isolate the difference, but the difference is worth knowing if possible. Again, if the statute delegates interpretive authority to less on traditional tools of statutory courts, they could rely

construction and more on the agency’s practical experience with the regulatory regime.

Finally, Congress might have provided sufficiently formal procedures, but the agency failed to employ them. Because the agency failed to abide by an essential condition of its delegation, courts are the default interpreters. The idea is not that Congress would prefer a judicial interpretation. It is that agencies must internalize the costs of foregoing proper procedures. Agencies must face a penalty for fencing Congress out of the process. They run the risk of losing their preferred interpretations, especially as those interpretations are more adventurous than courts would be willing to accept based on their standard approaches.168 The result is that agencies choose procedures for interpretations that make significant moves from the status quo.

But when Congress provided procedures that the agency failed to follow, courts are not the permanent interpreters. They possess authority not because Congress preferred judicial resolution but because the agency involved failed to abide by a condition of its delegation. All else equal, Congress preferred administrative resolution. Consequently, an agency may reverse the judicial interpretation if it uses sufficiently formal procedures in the future. Duly authorized agency interpretations take priority over judicial interpretations, much as statutes do. This result tracks the reasoning in National Cable & Telecommunications Ass’n v. Brand X Internet Services.169

Courts must consider the reasons for and conditions of interpretive delegation to agencies in recognizing the existence of such delegation. Courts identifying these factors would be less likely to find clear meaning when Congress likely intended none to exist. The existing framework views Congress as pursuing particular legislative deals or broad statutory purposes—

if they are

interested

and, in both cases, commits courts to search for a meaning behind statutory language.

Under this new approach, courts would still retain their essential functions. They would ensure that agencies stay within the scope of their delegated authority and issue reasonable interpretations. Courts would also possess interpretive authority when an agency lacks it. Existing administrative law doctrine provides the basic guidance for these inquiries. It is not far off in this context. But even here, the judicial posture would be different.

So far this discussion has been fairly abstract, suggesting particular factors for courts to examine in interpreting regulatory statutes and showing how these factors differ from the existing framework. This Part offers two examples in which members of the Court actually may be understood as applying approaches remarkably like the interpretive theory that this Article envisions. The examples help to flesh out the theory.

The first example is Zuni Public School District No. 89 v. Department of Education. Justice Breyer, writing for the majority, examined the considerations that matter to Congress when delegating authority to agencies and regarded the statutory text as having a less significant role. His colleagues, including those that agreed with his conclusion, took issue with his approach on the very question that this Article raises: which interpretive theory best positions courts to respect congressional will in the regulatory context?

The second example is Gonzales v. Oregon. The majority opinion, written by Justice Kennedy, is best explained in terms of the delegation considerations that this Article presents. The decision is instructive at another level. It shows that the Court, employing better assumptions about legislative behavior, does not always find an interpretive delegation to the agency involved. Meanwhile, the dissenting opinion demonstrates the pitfalls of textualism, even when used to uphold an agency interpretation.

A. Zuni Public School District No. 89 v. Department of Education

Zuni involved what looked like a difficult but not atypical issue of statutory interpretation. The Federal Impact Aid Act contains a method for the secretary of education to use when determining funding program “equalizes whether a state’s public school

to “disregard”

expenditures” among that state’s public school districts.170 Only if a state’s program equalizes expenditures may the state use federal aid to reduce its own local funding.171 According to the statute, a state’s program equalizes expenditures when the disparity in per-pupil expenditures does not exceed 25 percent.172 But, when doing so, the school districts “with per-pupil secretary expenditures . . . above the 95th percentile or below the 5th percentile of such expenditures . . . in the State.”173 The secretary issued a series of regulations, in place for thirty years, interpreting the so-called “‘disregard’ instruction”174 essentially to allow the secretary to consider “the number of the district’s pupils as well as . . . the size of the district’s expenditures per pupil.”175 For New Mexico’s local district aid program, which applied to the petitioner, the secretary calculated expenditures, consistent with its regulations, as follows for fiscal year 2000:

Zuni Public School District sought review of the agency’s findings, conceding that the calculations were correct under the regulations but arguing that the regulations were inconsistent with the

statute.177 Zuni contended that the statute required the secretary to calculate the upper and lower percentile cutoffs “solely on the basis of the number of school districts (ranked by their per-pupil expenditures), without any consideration of the number of pupils in those districts.”178 Under Zuni’s method, “only 10 districts (accounting for less than 2 percent of all students) would have been identified as the outliers” and “[t]he difference, as a result, between the highest and lowest per-pupil expenditures of the remaining districts (26.9 percent) would exceed 25 percent.”179 The statute thus would “forbid New Mexico to take account of federal impact aid as it decides how to equalize school funding across the State.”180

After the Tenth Circuit affirmed the secretary’s decision by a split vote of a twelve-member en banc panel, the Court took the case to consider the interpretive question.181 Justice Breyer wrote the opinion for the majority, which Justices Stevens, Kennedy, Ginsburg, and Alito joined with several separate concurrences.182 Justice Breyer began the customary way, citing Chevron and noting that “Zuni’s strongest argument rests upon the literal language of the statute.”183 But rather than considering that argument at the outset, as Chevron instructs, Justice Breyer made an unusual move. He wrote, “For purposes of exposition, we depart from a normal order of discussion, namely an order that first considers Zuni’s statutory language argument. Instead, because of the technical nature of the language in question, we shall first examine the provision’s background and basic purposes.”184 He found that “[c]onsiderations other than language provide us with unusually strong indications that Congress intended to leave the Secretary free to use the calculation method before us and that the Secretary’s chosen method is a reasonable one.”185

The first consideration was the complexity of the issue. Justice Breyer stated that “the matter at issue—i.e., the calculation method for determining whether a ‘equalizes expenditures’—is the kind of highly technical, specialized interstitial

state aid program

matter that Congress often does not decide itself, but delegates to specialized agencies to decide.”186 The second consideration was the “history” of the statute, suggesting that Congress has ample opportunity to monitor the agency interpretation, which predated the statute.187 The secretary had promulgated the regulation at issue in 1976 under a 1974 version of the federal aid statute.188 Meanwhile, in the ensuing twenty years until Congress enacted the version of the statute at issue in the case, “no Member of Congress has ever criticized the method the 1976 regulation sets forth nor suggested at any time that it be revised or reconsidered.”189 In addition, the secretary himself had sent the relevant statutory language to Congress in 1994, and Congress “adopted that language without comment or clarification.”190

The third consideration was the reasonableness of the secretary’s interpretation in light of the purpose of the statutory disregard instruction.191 Justice Breyer relied on the secretary’s explanation for how the secretary’s interpretation squared with the purpose of the disregard instruction, rather than supplying his own.192 The secretary had provided that explanation along with its regulation in 1976.193

The final consideration was the language of the statute, which Justice Breyer consulted to ensure that the agency had not acted outside the basic scope of its authority.194 What Justice Breyer looked for was a truly clear meaning or an “absolute literalness,” which he illustrated as follows: “A customs statute that imposes a tariff on ‘clothing’ does not impose a tariff on automobiles, no matter how strong the policy arguments for treating the two kinds of goods alike.”195 The words of the Federal Impact Aid Act were not as easily understood as “clothing” or “automobiles.” Justice Breyer therefore sought outside assistance. He found that every mainstream and

technical dictionary—Webster’s Third New International Dictionary, an economics dictionary, a mathematics dictionary, a science dictionary, and a medical dictionary—defined “per-pupil” the way that the secretary did.196 Furthermore, other education-related statutes also supported the interpretation.197 In addition, the “more general circumstance[s]” surrounding the statute confirmed the the phrase.198 Finally, Justice Breyer gathered ambiguity of “reassurance from the fact that no group of statisticians, nor any individual statistician, has told us directly . . . that the language before us cannot be read as we have read it.”199 He found this consideration “significant” because the words are “technical, and we are not statisticians.”200 The “upshot,” he said, is that “the language of the statute is broad enough to permit the Secretary’s reading.”201 For reasons that he had previously stated, he concluded that “the Secretary’s reading is a reasonable reading.”202

Justice Breyer’s opinion drew criticism from all of his colleagues except Justice Ginsburg.203 Four Justices wrote separate statements, each reflecting a commitment to an alternative theory of statutory interpretation—or at least obedience to the Chevron doctrine as traditionally applied. Justice Stevens preferred intentionalism.204 He joined the majority opinion because he agreed that the text was “sufficiently ambiguous to justify the Court’s exegesis.”205 He said that the better approach, however, was to acknowledge that the “legislative history is pellucidly clear and the statutory text is difficult to fathom.”206 In his view, “this is a quintessential example of a case in which the statutory text was obviously enacted to adopt the rule that the Secretary administered both before and after the enactment of the rather confusing language” at issue.207 Thus, Congress had a clear

“intention on the precise question at issue.”208 Citing Church of the Holy Trinity v. United States,209 Justice Stevens stated that he would not upset this interpretation even if he thought that the literal reading were correct.210

Justice Kennedy concurred,

joined by Justice Alito, also expressing concern about the Court’s analysis.211 In his view, the Court was “correct to find that the plain language of the statute is ambiguous.”212 Nevertheless, Justice Kennedy noted that “[t]he opinion of the Court . . . inverts Chevron’s logical progression.”213 He worried that “[w]ere the inversion to become systemic, it would create the impression that agency policy concerns, rather than the traditional tools of statutory construction, are shaping the judicial interpretation of statutes.”214 Although he believed that the Court had not fulfilled its “obligation to set a good example,” he was willing this time to give Justice Breyer the benefit of the doubt “in matters of exposition.”215

Justice Scalia dissented, joined by Chief Justice Roberts and Justice Thomas, and in part by Justice Souter.216 He took issue with the Court’s chosen interpretation as well as its overarching mode of analysis. First, he castigated Justice Breyer for treating the interpretive question in the case as a “scary math problem,” requiring a “hypothetical cadre of number-crunching amici to guide [the] way.”217 He also decried the “sheer applesauce” of Justice Breyer’s interpretation, which required a number of steps to get from local school districts to populations in those school districts to average perpupil expenditures.218 He read the words as they were written, refusing to consult dictionaries or statisticians.219 The contested phrase

was “local educational agencies with per-pupil expenditures.”220 Carefully dismantling the Court’s interpretation, he argued that perpupil expenditure could not refer to the students themselves, but only to the local school districts.221

Turning to the interpretive methodology, Justice Scalia rebuked Justice Breyer for applying intentionalism. Declaring that “today Church of the Holy Trinity arises, Phoenix-like, from the ashes,” he characterized Justice Breyer’s opinion as “nothing other than the elevation of judge-supposed legislative intent over clear statutory text.”222 Against this approach he offered a broad defense of textualism. Justice Scalia stated that “once one departs from ‘strict interpretation of the text’ . . . fidelity to the intent of Congress is a chancy thing” because “[t]he only thing we know for certain [is that] both Houses of Congress (and the President, if he signed the legislation) agreed upon . . . the text.”223 He dissected each piece of evidence on which Justice Breyer relied, showing that Congress could intended to depart from the secretary’s have interpretation.224

just as easily

These concurring and dissenting Justices, though reflecting differences among themselves, all thought that Justice Breyer failed to serve as a faithful agent of Congress. But they misperceived his approach. Justice Scalia assumed that it was intentionalism. But an intentionalist would have found, as Justice Stevens did, that Congress intended to select a specific policy—the methodology that the secretary had adopted—whereas Justice Breyer found only that Congress has intended to delegate interpretive authority to the secretary, who in turn had selected the policy. Justice Kennedy thought that Justice Breyer departed from legislative intent and undermined judicial authority by elevating agency policy over the traditional tools of statutory construction. If he did, it was in pursuit of a theory that reflected better assumptions about legislative

behavior. Indeed, the traditional tools prevented the other Justices from serving as faithful agents of Congress. Nowhere was this more evident than in the dissent. Justice Scalia focused on the words as written, on the assumption that interpreters can know no more about Congress than what it writes. Justice Breyer showed that interpreters might know much more about Congress if only they read the words of the statute in their political context rather than for their narrowest meaning.

But the critics were right in one respect. Justice Breyer did not fully acknowledge the significance of his interpretive approach.225 He did not merely invert Chevron. He changed the unit of analysis from statutory meaning to interpretive delegation. Why did he? As Professors Thomas Miles and Cass Sunstein have determined, Justice Breyer is already “the most deferential justice” on the sitting Court without the benefit of any novel approach.226 Furthermore, he had five votes for deference in this case under the traditional approach. Would a new approach allow Justice Breyer to teach others what his early experience with the legislative process had taught him?227 If so, his lesson is worth serious consideration.

B. Gonzales v. Oregon

Gonzales concerned an interpretation extending the reach of the Controlled Substances Act (CSA) to the regulation of physicianassisted suicide.228 Attorney General John Ashcroft issued this interpretation229 in the wake of an Oregon law that legalized physician-assisted suicide, “without consulting Oregon or apparently anyone outside his Department.”230 According to the interpretation, physician-assisted suicide is not a “legitimate medical purpose” for which physicians might dispense and prescribe controlled substances under the CSA and accompanying regulations.231 Physicians who dispense and prescribe controlled substances to assist suicide violate the CSA and jeopardize their federal registrations to prescribe controlled substances for other purposes, even if “state law authorizes or permits such conduct.”232

The Court held that Congress has not delegated authority to the attorney general to issue the interpretation.233 Justice Kennedy wrote the opinion for the majority, joined by Justices Stevens, O’Connor, Souter, Ginsburg, and Breyer.234 The Court acknowledged that the language of the CSA was ambiguous, which ordinarily supports agency delegation under Chevron.235 Nevertheless, it refused to presume that Congress implicitly would grant the executive branch such “broad and unusual” authority to criminalize an act that is legal under state law.236 Quoting earlier cases applying the “extraordinary” question principle, it noted that Congress is unlikely to “hide elephants in mouseholes.”237 The Court continued, explaining that

“[t]he importance of the issue of physician-assisted suicide, which has been the subject of an ‘earnest and profound debate’ across the country, makes the oblique form of claimed delegation all the more suspect.”238

reasons

the ordinary

If the Court had said only this much, it would have failed to grasp the significance of this case. The problem was not simply one of elephants and mouseholes—for Congress can hide elephants in mouseholes when those elephants raise technical issues or impede legislative consensus. But this was not such a case. The question was extraordinary239 precisely because for interpretive delegation were not present. As the Court observed, the decision whether to permit physician-assisted suicide does not raise a technical issue but a moral, legal, and practical issue.240 Consequently, there was no reason to presume that Congress intended to delegate the issue of its meaning to the attorney general. In fact, Oregon had handled the issue through popular means of a voter ballot measure.241 Furthermore, no evidence suggested that delegating physicianassistant suicide to the attorney general was essential to the passage of the CSA. Indeed, Congress generally anticipated state laws regulating the medical profession, like Oregon’s physician-assisted suicide law.242 The Court stated that “[t]he structure and operation of the CSA presume and rely upon a functioning medical profession regulated under the States’ police powers.”243 If anything, Congress left the issue of regulating physician-assisted suicide to the states.

The Court offered other considerations that reflected realistic assumptions about the legislative process. The Court noted that the attorney general lacks broad rulemaking power to enforce the relevant provisions of the statute.244 The Court did not connect the attorney general’s lack of authority with concerns about legislative monitoring, although no one outside the Department of Justice could observe the interpretation until it was a fait accompli.245 The Court did

through

its chosen procedure, an

not even decide whether the attorney general could ever issue binding interpretive interpretations bulletin.246 Nevertheless, it acknowledged the importance of the procedural feature to the delegation analysis.247 Because Congress had not granted rulemaking power to the attorney general over medical practices such as physician-assisted suicide, it has not delegated interpretive authority to the attorney general over those practices.248

Relatedly, the Court also stated that Attorney General Ashcroft exercised authority over an issue outside his area of expertise. The CSA provides the attorney general with authority to regulate “the registration and control of the manufacture, distribution, and dispensing of controlled substances.”249 Attorney General Ashcroft asserted a different sort of authority—the authority to determine whether physician-assisted suicide constitutes a legitimate medical practice.250 But the definition of medical practice is a technical issue. The Court stated that Congress tends to delegate in accordance with whether a particular agency possesses “historical familiarity and policymaking expertise.”251 The Court further recognized that Congress can expect no benefit from delegating the definition of medical practice to the attorney general, even if it might expect a benefit from delegating that issue to another federal agency.252

Although Justice Kennedy purported to apply a fairly standard interpretive principle—the extraordinary question principle—he actually engaged a more accurate delegation analysis than that principle might suggest in other cases. He claimed that Congress was unlikely to delegate authority over physician-assisted suicide because of the normative significance of such a practice.253 Yet he demonstrated, more realistically, that Congress had little reason to delegate the issue to any agency and had little opportunity to monitor

how the attorney general would resolve the issue.254 He reasoned essentially that when neither the reasons nor the conditions for interpretive delegation are present, courts should conclude that Congress intended no such delegation.255

Finding no interpretive delegation, the Court considered whether the attorney general’s interpretation nevertheless was “persuasive” under Skidmore.256 It concluded that the interpretation was not.257 The Court read the statute as if Congress had not meant to delegate authority over conduct like physician-assisted suicide at all, neither to the agency nor to the courts. Thus, the Court stated that the CSA “manifests no intent to regulate the practice of medicine generally” and, in fact, “presupposes” statutes like Oregon’s.258 Furthermore, what limited authority the CSA does delegate to regulate medical practices goes to the secretary of Health and Human Services, not the attorney general.259 Finally, the Court applied a traditional tool, a federalism canon, to resolve any residual statutory ambiguity: “The background principles of our federal system also belie the notion that Congress would use such an obscure grant of authority to regulate areas traditionally supervised by the States’ police power.”260 Given these reasons, the attorney general’s interpretation.261

to approve

it refused

Justice Scalia dissented, offering a characteristically narrow reading of the language but this time to support the agency’s interpretation. Noting that the attorney general has undisputed authority to regulate prescriptions, Justice Scalia reasoned that the attorney general therefore has authority to determine that “the dispensation of a Schedule II substance for the purpose of assisting a suicide is not a ‘prescription’ within the meaning of [the statute]”262 because the act of assisting suicide is not a “legitimate medical purpose” or otherwise in the public interest.263 The attorney general

may further determine that any physician who writes illegitimate prescriptions “may ‘render his registration . . . inconsistent with the public interest’ and therefore subject to possible suspension or revocation under [the statute].”264 Justice Scalia found that this interpretation was “the most reasonable” reading of the CSA and would control regardless of whether the interpretive ruling was entitled to deference.265

To summarize, in contrast to Justice Scalia’s textual analysis, Justice Kennedy’s analysis considered broader indications of legislative intent to delegate interpretive authority to the attorney general. In this way, his analysis enabled him to serve as a faithful agent of Congress. His approach also furthered other important normative values. By leaving authority to regulate physician-assisted suicide in the states, Justice Kennedy’s analysis advanced the principle of federalism. It allowed the states to continue to serve as laboratories for the experimentation with the regulation of physicianassisted suicide. In addition, it ensured that agencies do not receive delegated interpretive authority unless they possess the requisite expertise and accountability for exercising it.

Zuni and Gonzales demonstrate how courts (or the Court) can focus on factors other than the clarity of statutory language to determine whether Congress likely delegated interpretive authority to the agency involved. Justice Breyer inverted Chevron and examined strategic delegation considerations before determining that the agency interpretation was reasonable and within the scope of the statute. Justice Kennedy purported to apply a standard interpretive principle but did so in appreciation of more realistic assumptions about the legislative process. Together, these Justices demonstrated that the focus on indications of interpretive delegation is not one directional: it can lead to a decision upholding an interpretive delegation or to a decision rejecting an interpretive delegation.

These opinions also begin to reveal the normative significance of updating statutory interpretation to better capture the legislative interpretive delegation to agencies. Neither Justice indicia of

explained the reason for such a change or even fully acknowledged how much their opinions departed from the typical framework. Indeed, it might upset Justice Kennedy to know that his analysis was unconventional, given his objection to Justice Breyer’s analysis in Zuni. Nevertheless, each might have felt that their particular approach would best respect the will of Congress. At the same time, their approaches placed interpretive authority where other normative values suggested that it belonged. Agency expertise—and possibly political accountability—suggested that the agency should have the interpretive authority in Zuni. But the opposite was true in Gonzales. Moreover, federalism strongly suggested that the states should have the authority in Gonzales.

The evaluation of any interpretive theory lies in the balance of the goals that it accomplishes. Before turning to objections, it is worth highlighting the normative advantages of the theory that I have sketched. As I have shown, my delegation-respecting theory would enable courts to ascertain congressional will when interpreting regulatory statutes better than existing theories do. A delegationrespecting theory would therefore better promote the principle of legislative supremacy in the regulatory context. In enacting regulatory statutes, Congress can delegate interpretive authority to agencies for the strategic benefit of the act of delegation itself. A delegationrespecting theory would recognize this meta-aspect of legislative intent by offering courts appropriate proxies for detecting interpretive delegations to agencies. Not insignificantly, it also would place interpretive authority in the hands of those officials best able to discern and further substantive legislative purposes. Unlike courts, agencies have a continuous relationship with Congress and may have a better understanding of the general aims of legislation.266 Put simply,

agencies are likely to be the superior purposivists when it comes to resolving interpretive questions that Congress left open.267

The theory proposed in this Article has other normative advantages. The reasons for which Congress delegates certain interpretive questions often make those questions especially appropriate for agency resolution, as Chevron recognizes.268 Complex questions typically benefit from agency expertise, and contentious legislative issues typically benefit from continuing political debate at the administrative level. Chevron notes that agencies are more accountable than courts—not directly but through the president.269 Beyond these abstractions, both the White House and Congress maintain active interest in agency decisionmaking as it unfolds.270 The White House has a regulatory review apparatus and an informal network that purpose. Congress can use administrative procedures to facilitate legislative monitoring. A theory that allows agencies to possess interpretive authority as often as Congress intends would promote both agency expertise and political accountability. A theory that makes proper procedures a condition of delegation also

for

would reinforce rule-of-law values. As Mead suggests, procedural formality tends to ensure that agency decisionmaking is rational and fair.271

At the same time, a theory that deprives agencies of interpretive authority absent sufficient indications of legislative intent to delegate would serve beneficial purposes. This negative side is essential to maintain a claim of legislative supremacy, allowing Congress to control when not to delegate. In addition, the negative aspect of the theory would preserve rule-of-law values by withholding interpretive authority when an agency fails to use proper procedures. Less intuitively, facilitate agency expertise and political accountability. Without sufficiently formal procedures, there is less assurance that agency decisionmaking reflects specialized knowledge rather than tunnel vision or political pressure. There is also less assurance that agency decisionmaking is transparent and therefore amenable to political oversight as well as judicial review. Signaling to agencies that they possess interpretive authority only when Congress intends would serve legislative supremacy and rule-of-law values, as well as promote agency expertise and political accountability.

it would

Yet this theory has serious downsides. If Zuni and Gonzales help to illustrate the theory in this Article, they also confirm that the theory is complicated—multifactored and context dependent, perhaps more so than a search for the meaning of statutory words. Justice Scalia was alarmed in Zuni to think that Justice Breyer was resorting to Holy Trinity intentionalism, elevating legislative history over statutory text. But the possibility that Justice Breyer was actually expanding Mead to include a variety of considerations would send him through the roof. Justice Scalia has castigated Mead for its uncertainty from the start, and his prediction about its potential to confuse lower courts has come to pass.272 In this Part, I address the institutional costs of the theory outlined in this Article.

In addition, this theory squarely presents the difficulty of using statutory interpretation to constrain broad delegations. Since the demise of the nondelegation doctrine, courts have relied on statutory

interpretation for this important purpose. In this way, one might say, Chevron and Mead function just as they should, providing courts with select opportunities to discipline broad delegations. I show that the general adoption of a delegation-respecting theory does not prevent courts from using statutory interpretation in a delegation-restricting fashion. At the same time, neither courts nor commentators can claim that such a practice is consistent with the aims of Congress, except in cases like Gonzales, when the evidence revealed that Congress likely did not intend the asserted delegation anyway. In light of this fact and for other reasons, I prefer an alternative method for disciplining broad delegations when the circumstances suggest a need.

A. Institutional Costs

The objection based on institutional costs is an enlarged version of the objection that Justice Scalia voiced when he dissented in Mead. He stated that the decision to inquire into legislative intent to delegate interpretive authority “makes an ‘avulsive change’ in judicial review of agency action, the consequences of which ‘will be enormous, and almost uniformly bad.’”273 As he saw matters, “[w]hat was previously a general presumption of authority in agencies to resolve ambiguity in the statutes they have been authorized to enforce has been changed to a presumption of no such authority, which must be overcome by affirmative legislative intent to the contrary.”274 He warned that the decision would confuse the lower courts. He stated, “We will be sorting out the consequences of the Mead doctrine, which has today replaced the Chevron doctrine, for years to come.”275

Justice Scalia was correct about the effect of Mead on lower courts. Those courts have applied inconsistent analyses, vacillating between the factors that Mead provides and those that Barnhart provides.276 Indeed, courts in different circuits have split as to the same procedural format.277 Many have chosen instead to avoid resolving the procedural question altogether by upholding the agency

interpretation under Skidmore rather than Chevron.278 Although this avoidance is a practical solution to Mead’s complications, it may create difficulties should an agency seek to change its interpretation in the future. When a court upholds an agency interpretation under Skidmore, it effectively adopts that interpretation. Under Brand X, an agency may overrule a judicial interpretation only if the underlying statutory language is ambiguous.279 But the court may have avoided resolving this issue along with the procedural one, leaving the agency in a state of uncertainty. Finally, lower courts have used Mead as a basis for the extraordinary question principle, even though the two can reflect very different assumptions about legislative behavior.280 That Congress demands procedures as a condition of delegation does not mean that it withholds issues for normative reasons, such as the national significance or jurisdictional nature of the issues.

If Mead alone generates confusion among lower courts, a more elaborate version could produce disaster. Each factor would permit room for judgment and variation, potentially producing inconsistent analyses of the same interpretive issue. The theory would encourage courts to embrace Skidmore more often than before, reducing agency flexibility. Finally, the theory would create new opportunities and incentives for courts to forge questionable default rules for ascertaining legislative intent.

Although these concerns are legitimate, they should not be overstated on the basis of lower courts’ experience with Mead. The Court has made Mead more complicated than it needs to be, and a theory tied to better assumptions about legislative behavior might help repair the problem. First, as discussed in Part II.B, a delegationrespecting theory would focus on procedures that facilitate legislative monitoring, rather than only on abstract rule-of-law values such as fairness or deliberation. The message to courts would be that proper procedures are those that furnish constituents with information about agency action before such action is a fait accompli, so that those the constituents may alert administrative process as necessary. Mead was correct under this

legislators

intervene

their

to

analysis. U.S. Customs Service ruling letters are virtually impossible for Congress to monitor, through its constituents or otherwise.281 They are issued at a rate of 10,000 to 15,000 per year by forty-six different offices of the agency without any participatory process or reasoned explanation.282 Other procedures are better for monitoring purposes. Notice-and-comment rulemaking is the paradigm from a legislative monitoring perspective because is the most accessible and informative, largely as a result of heavy judicial regulation.283 But other procedures might suffice, such as a public hearing preceded by notice and followed by an explanation of the decision.284

it

Second, a delegation-respecting theory could address the shift to Skidmore by changing the application of that decision. Under the existing framework, courts can use Skidmore to uphold an agency interpretation without resolving the procedural question (and often without declaring whether the statutory text is clear or ambiguous).

Under the theory proposed in this Article, Skidmore would no longer provide an escape route. Courts would undertake the delegation analysis, which includes the procedural question, in all cases. Thus, courts could not avoid the procedural question simply by choosing Skidmore. Courts would still rely on Skidmore-style deference when an agency failed to use sufficiently formal procedures, but they would not thereby hamper agency flexibility. Rather, courts would possess interpretive authority unless the agency overrules the judicial interpretation by reissuing its own interpretation using sufficiently formal procedures.

Finally, a delegation-respecting theory could minimize the misuse of Mead by exposing the assumptions about legislative behavior on which the decision implicitly relies but fails to fully implement. Once the procedural factor is understood as mattering to Congress for strategic reasons, the extraordinary question principle no longer seems categorically justifiable to discern legislative intent. That principle reflects a view of Congress as hanging on to certain issues because of their normative importance—their national significance or their jurisdictional nature.285 But Congress is unlikely to hold on to any issue unless doing so serves its own political interests. Thus, an extraordinary question principle is justifiable as a matter of legislative intent only to the extent that it overlaps with the factors that already form part of the new interpretive framework. Any blanket exemption for matters of great importance does not follow from Mead’s emphasis on delegation. To survive, the extraordinary question principle requires a separate democracyforcing justification, as I address in further detail in the next Section.

If critics were to reject all other suggestions from this Article on institutional grounds, they should still accept the suggestion to embrace and reform Mead. Mead moves in the right direction by making delegation rather than meaning the threshold question. Although it does not go far enough for my purposes, it still serves a valuable gatekeeper function. By insisting on sufficiently formal procedures, Mead blocks certain interpretive delegations—those Congress does not interpretative delegations that transgress rule-of-law values. At the same time, a

intend. Likewise,

it blocks

focus on the legislative use of procedures refines the inquiry and reduces institutional costs.

If Mead is still too much to take despite the clarifications herein, then one must be prepared to draw a broader lesson from this Article. Critics of Mead might seek a simpler test, but it should not be the existing framework. That framework is complicated in its own right, with courts using different interpretative tools and even using the same interpretive tools differently. Even if it is simple enough from an institutional perspective, it is deficient on other grounds. The search for meaning affords courts too much power to defeat interpretive delegations when Congress likely intended them and when they are likely to promote agency expertise and political accountability. It cannot supply the operative rule, whatever the institutional strengths.

The appropriate rule under these circumstances is not one that acts as a judicial gatekeeper. Rather, it is a presumption of deference. By a presumption of deference, I mean a rule that upholds an agency interpretation so long as it is reasonable.286 (Admittedly, even this rule provides room for judicial judgment.) A presumption of deference strikes a better balance between institutional and normative goals, in light of legislative realties, than the existing framework. At one time, Chevron might have furnished a presumption of deference based on what the Court observed in that case about legislative intent to delegate. Well before Mead, the search for meaning got in the way.

B. Nondelegation Issues

Another potential problem with a delegation-respecting theory is how to use statutory interpretation to constrain broad delegations in light of long used statutory construction for this purpose, with a significant degree of scholarly approval. In this Section, I quickly catalogue the main ways courts have used statutory interpretation to constrain broad delegations. I

legislative realities. Courts have

briefly demonstrate that such an approach can no longer be associated, even nominally, with congressional will. Nevertheless, I show that a delegation-respecting theory does not preclude courts or commentators from pursuing it. I prefer an alternative approach for addressing the concerns about broad delegations.

1. The Nondelegation Tools. Ever since the demise of the nondelegation doctrine in 1935, courts have relied on statutory interpretation to address their persistent worries about broad delegations.287 Broad delegations enable to avoid responsibility for making hard policy choices while claiming credit for statutory responses.288 They place policymaking in the hands of unelected bureaucrats. They supply few standards to prevent agencies from rendering arbitrary decisions.289 And they erode the structural safeguards for other values, including federalism and criminal lenity.290 Although the Court has been unwilling to enforce constitutional restrictions on broad delegations, it has used statutory interpretation to respond to these underlying concerns.291

legislators

It is possible to characterize many interpretive principles—and even an interpretive theory—as ways to restrict broad delegations. For example, Professor John Manning has argued that textualism can function as a nondelegation doctrine.292 By compelling courts to adhere to the text, textualism effectively prevents Congress from

delegating interpretive authority to individual legislators through legislative history.293 Legislators have incentives to carefully consider the deals they strike because they cannot effectively undo those deals through legislative history. In addition, selected legislators are not making law for the whole. The people are entitled to policy made by a majority of their elected representatives, not a few.

Textualism can operate as a nondelegation doctrine in another sense. As Justice Scalia has noted, textualism is useful for defeating the sort of ambiguity that affords agencies room to expand their regulatory authority.294 Courts can construe language as clear and as foreclosing the asserted delegation. In this way, courts can prevent statutes from going too far.

So-called “nondelegation canons” and clear statement rules serve a similar purpose, although they do not always deny textual ambiguity.295 Rather, these principles refuse to take that ambiguity as authorizing an expansive agency interpretation. Thus, the Court has held that Congress would not delegate interpretive authority to agencies over extraordinary questions without expressly so stating. In MCI Telecommunications Corp. v. American Telephone & Telegraph Co.,296 the Court refused to permit the Federal Communications Commission (FCC) to read the words “modify any requirement”297 as allowing the agency to essentially eliminate a central requirement for

all long-distance carriers except the dominant one.298 Likewise, in FDA v. Brown & Williamson Tobacco Corp.,299 the Court refused to permit the Food and Drug Administration (FDA) to interpret the words “drug” and “delivery device” as allowing the agency to regulate the manufacture and sale of cigarettes and other tobacco products.300 As Justice Scalia remarked, Congress does not “hide elephants in mouseholes.”301 In other words, Congress does not delegate interpretive authority to agencies over issues of such economic or social significance through mere ambiguity.

The Court has sought to protect federalism interests through an analogous strategy. Federal statutes often conflict with state laws, yet they contain an ambiguous preemption provision or no preemption provision at all.302 Agencies then issue interpretations concerning the preemptive effect of their statutes and regulations. In many of these cases, the Court has invoked a presumption against preemption.303 It has demanded a clear statement from Congress before allowing an agency interpretation to preempt state law.304

The Court has also refused to read ambiguities as granting agencies the authority to raise constitutional questions. In Solid Waste

Agency of Northern Cook County v. U.S. Army Corps of Engineers,305 the Court applied the “avoidance of questions” canon to invalidate an agency interpretation that raised a question under the Commerce Clause.306 In this way, the avoidance canon prevents agencies from taking statutes into constitutional gray zones.

The Court has held that Congress does not intend for agencies to interpret ambiguities in criminal sanctions provisions. Rather, the Court has invoked the rule of lenity to interpret those provisions when necessary to protect potential criminal violators.307 Although the rule of lenity is not a constitutional requirement, it is a wellestablished norm for interpreting ambiguities in federal criminal statutes.308 It affords criminal defendants notice of prohibited conduct, protection against overzealous prosecution, and freedom from unauthorized liberty deprivation.309

2. The Normative Difficulty. The theory proposed in this Article does not prevent courts from vindicating nondelegation or other values through statutory interpretation. But courts pursuing any of these strategies must stick to their normative justifications. Many of these tools have always had at least some nominal association with the legislative design of statutes. Thus, textualism presumes that Congress cuts deals, and legislative history should not frustrate them. The nondelegation canons presume that Congress intends to withhold certain issues from agency control because of the normative values that they implicate—whether matters of constitutional structure or individual liberty.310 The difficulty is that courts and commentators can no longer regard the nondelegation tools in the same light.

To see this point, take two examples: extraordinary questions and preemption questions. Congress can delegate interpretive power to agencies over extraordinary questions because of the complexity of the underlying regulatory issues.311 The FCC must ensure that the dominant carrier does not possess monopoly power while allowing new entrants to flourish, calibrating the statute’s requirements accordingly.312 The FDA must consider the health effects of particular products, evaluating their safety and efficacy by relying on scientific studies and medical judgments. Furthermore, Congress authorized the agencies to use notice-and-comment rulemaking for regulating in their respective domains.313 Congress may not have anticipated the precise interpretations, concerning the tariff-filing requirement and tobacco products. But Congress frequently thinks only in general terms. Courts cannot effectively undermine the delegation by later attributing the extraordinary nature of the question to Congress. In fact, the evidence that Congress did not intend the FDA to regulate tobacco did not have to do with the enacting Congress. Rather, it concerned subsequent Congresses and subsequent legislation.314

Similarly, Congress can delegate preemption questions when they are complex or technical. Banking law is an illustration. The National Bank Act grants national banks the power to engage in mortgage lending, subject to regulation by the Office of the

Comptroller of the Currency (OCC).315 It also grants national banks “all such incidental powers as shall be necessary to carry on the business of banking.”316 The act further grants “visitorial powers” (the power to “inspect, examine, supervise, and regulate the affairs of an entity and to enforce compliance with applicable laws”317) over national banks to the OCC and provides that “[n]o national bank shall be subject to any visitorial powers except as authorized by Federal law.”318 Thus, states are effectively preempted from exercising visitorial powers over national banks under their own laws. The OCC issued a rule interpreting this provision as preempting the states from exercising visitorial powers over operating subsidiaries of national banks.319 National banks have authority to conduct business through operating subsidiaries as part of their “incidental powers,” and the incorporated rule defined operating subsidiaries as separately divisions or departments of the parent bank.320 As a result of the rule, Michigan could not permit its officials to exercise examination and lending by operating enforcement authority over mortgage subsidiaries of a national bank.321

In Watters v. Wachovia Bank, N.A.322 the Court confronted the question whether the OCC rule was entitled to Chevron deference.323 The Court refused to decide, finding the level of deference “academic” because the interpretation matched the Court’s own interpretation of the statute.324 The theory offered in this Article might reach the same result but for a different reason—one more in line with legislative intent. The interaction between federal and state banking systems is so highly complex as to indicate that Congress likely delegated it to the agency. As the Court acknowledged, the OCC had to determine whether “duplicative state examination, that undoubtedly would burden supervision, and regulation” mortgage lending by national banks would similarly burden that

lending “when engaged in by an operating subsidiary.”325 The agency’s determination on these matters should have controlled because Congress so it reflected a reasonable reconciliation of the competing concerns underlying the statute.

intended and because

Congress also has reasons to delegate preemption questions when necessary to obtain consensus. The 2007 fuel economy legislation offers an illustration. That legislation set an average fuel economy standard of thirty-five miles per gallon for new automobiles by 2020.326 During negotiations, Representative John Dingell (D-Mich.) and House Speaker Nancy Pelosi (D-Calif.) waged a heated battle over many issues, including whether to preempt California and other states from setting their own more stringent standards.327 Representative Dingell asked for specific language that would have preempted state standards, and Speaker Pelosi resisted.328 The legislators agreed to leave the issue ambiguous.329 The fight recommenced promptly at the administrative level.330

The point of these examples

is not to show that the nondelegation canons are unjustified. Rather, the point is to show that the nondelegation canons must be justified on normative grounds. The Court, for its part, has not maintained a steadfast commitment to the nondelegation canons in the face of evidence that Congress may have intended the delegation.331 Thus, the Court has shown ambivalence about which values ought to prevail. A delegation-respecting theory does not prevent more rigorous defense

of nondelegation tools. But it underscores the necessity of that defense.

3. The Administrative Law Alternative. My preference is not to embrace the nondelegation canons or textualism to constrain broad delegations, even though I am sympathetic to the problem that they seek to address. I would address this problem another, better way. Although this is not the place to fully elaborate my suggestion, the basic idea is that courts should reframe nondelegation worries in terms of administrative behavior rather than legislative behavior. Broad delegations give agencies considerable latitude in answering interpretive questions. And agencies can fall short of proper performance. Specifically, agencies can fail to adequately consider all the relevant factors or interests involved, particularly if a particular result is important to the agency’s priorities or to the administration’s priorities.332 Under these circumstances, a court could remand the interpretation to the agency for further consideration under the arbitrary and capricious test of administrative law.333

The administrative law strategy would send agencies a strong message about a court’s own view of the question involved but, at the same time, would have certain modesty. First, it would not pit the court against Congress when circumstances suggest that Congress is likely to have intended the interpretive delegation.334 For a court to impose its own views or call for a clear statement when Congress intended the agency to balance the competing interests does not promote good interbranch relations.

Second, an administrative law approach would not enable courts to assume excessive traditional nondelegation strategies can do. Given the complexities of the

interpretive control, as

legislative process, Congress rarely provides express agency authorization in response to a clear statement rule or restrictive judicial interpretation.335 Rather, Congress usually leaves in place the court’s narrowing construction, whether or not that decision reflects legislative preferences.336 Consequently, courts alone can determine that a statute does not reach a particular subject, preempt state law, raise a constitutional question, or impose a harsh criminal sanction. Furthermore, courts can do so more often than they should because statutory interpretation provides a relatively low-cost method of constitutional protecting decisionmaking, statutory interpretation often attracts less public attention and represents less legal change. The Court can use statutory interpretation to protect federalism values in a decision like Gonzales without drawing the reaction that it would for overruling the leading Commerce Clause or Tenth Amendment decisions. Lower courts can use statutory interpretation to protect normative values without incurring the wrath of the Court for usurping the power to alter constitutional law.337

values. Compared

normative

to

Finally and related, an administrative law approach would afford agencies an opportunity to reweigh interpretive questions. Agencies are still in the best position to resolve interpretive questions, even if courts can help clarify the interests at stake.338 Agencies have a better understanding of the regulatory scheme and its complexities. Furthermore, agencies have a connection to both political branches. Congress as well as the White House can participate in the formulation of policy at the administrative level.339 By remanding an issue to the agency, courts could signal Congress to take note. This strategy may not force Congress to take responsibility for policy choices when drafting legislation.340 But it could prompt Congress to

play a more active role in the selection of such choices by the agency, which is an advantage from a democratic perspective.

Administrative law is not a perfect substitute for statutory interpretation precisely because it does not allow courts to lock in a particular interpretation on normative grounds. An agency has a second shot, albeit one that occurs in the shadow of a judicial remand and with the increased scrutiny of interested parties. To my mind, that is a desirable result. The formulation of regulatory policy often involves reconciling conflicting interests, suggesting that judges should not simply elevate one value at the expense of other important considerations. The best way to discipline broad delegations, in light of all the factors at stake, is for courts to ensure that agencies adequately consider those factors and defend their resulting policy choices.

In sum, the objections to the theory I propose in this Article are not without responses. If courts are to respect agency delegation, they must be willing to reorient their approach. They would encounter tradeoffs with other goals. This Article is not meant to discount those other goals but to call for balance among them. The existing interpretive framework does a poor job of enabling courts to respect agency delegation. A theory that aims to improve on that front and secure related values would carry institutional costs, perhaps not much more so than the existing framework though certainly more so than a default deference rule. It would also present starkly the degree to which Congress likely delegates interpretive authority to agencies despite judicial and scholarly preferences for nondelegation. Courts and commentators must examine how best to vindicate nondelegation values once they fully appreciate the distance between those values and legislative realities.

I have argued that existing administrative law inadequately positions courts to recognize the existence of agency delegation because it invites courts to apply the traditional theories of statutory interpretation. For all their differences, the traditional theories operate in a similar fashion, directing courts to construe statutory text as if Congress was aiming for a relatively specific meaning. This basic presumption about legislative behavior cannot be squared with what

many political scientists and legal scholars have been saying for some time about how Congress designs regulatory statutes. And it is this basic presumption that Chevron and Mead fail to sufficiently dislodge. In this Article, I imagine a theory of statutory interpretation that focuses courts on the factors that matter to Congress when delegating interpretive authority to agencies. For example, Congress is likely to delegate when an issue is complex, both to conserve legislative time and harness agency expertise. It is likely to delegate when an issue is contentious and an obstacle to a law’s enactment. Whatever the reason for delegating, Congress is likely to seek means for ensuring that agency interpretations roughly track legislative preferences. Congress can use administrative procedures to place constituents in the administrative process, where those constituents can demand legislative intervention if agency interpretations stray too far. It can also observe agency interpretations directly if an agency is involved in the longstanding interpretation. To create a better theory, courts would focus on these factors rather than the relative clarity of language as a proxy for interpretive delegation. Even one reason for delegating and one condition on the delegation might suffice to build a case for interpretive delegation. But the stronger the case, the less courts should be inclined to read the statutory text as clear and as foreclosing the delegation.

legislative drafting process or has a prior

Although my theory departs from the framework that courts generally apply, I have shown that it is not without precedent. Members of the Court can be understood to have pursued a strikingly similar approach. I offer these cases, not as examples of a burgeoning trend in statutory interpretation, but, more modestly, to show how the Court can incorporate more realistic assumptions of legislative behavior into statutory interpretation if it chooses.

The theory that this Article envisions has many advantages. Taking account of legislative realities is worthwhile in its own right and to promote agency expertise and political accountability, as well as rule-of-law values. Agencies should possess interpretive authority as often as Congress delegates such authority—and no more often technically that than sophisticated, publicly responsive, and consistently applied.

regulatory policy

that—to ensure

I have acknowledged that a delegation-respecting theory raises an institutional objection. Lower courts might struggle to apply it, as they have Mead. I have argued against condemning the new theory on the basis of lower courts’ experience with Mead. It promises to

if a delegation-respecting theory

simplify Mead while integrating more appropriate assumptions about is legislative behavior. Even intolerably complicated, I have contended that Chevron still fails to provide the proper rule. That decision has never been particularly easy for courts to apply. Moreover, I have argued that it cannot for normative reasons. The only supply institutionally simple and normatively justifiable proposal is a true presumption of judicial deference, directing reviewing courts to uphold agency interpretations more so than Chevron’s presumption has ever done.

the relevant

test

Finally, I have noted that a delegation-respecting theory presents a challenge for the use of statutory interpretation to constrain broad delegations. As a general matter, my theory does not preclude courts from pursuing this strategy. Yet it does suggest that courts can no longer base their interpretive practices on how Congress designs statutes. Furthermore, there is a better way to address the underlying concerns about broad delegations. Courts could use administrative law to police how agencies exercise their interpretive authority rather than using statutory interpretation to narrow delegations that Congress likely intended to make. I believe that this approach would offer courts a better role to play. Courts could make known their views on issues that implicate normative values, but agencies would reconcile the competing interests through application of their specialized expertise and in consultation with the political branches.

When courts think about the proper method for interpreting regulatory statutes, they recognize the need to account for the possibility of agency delegation. But existing law confines them to interpretive practices ill suited to this context. As a result, courts construct a meaning for text in regulatory statutes more often than they should. A more deferential approach, with more restrained opportunities intervention, would produce better regulatory policy. It would allow agencies the leeway to set regulatory policy under judicial supervision rather than subject to judicial domination.

judicial

for

--- FOOTNOTES ---

LISA SCHULTZ BRESSMAN†

ABSTRACT

statutory

interpretation—whether

to determine whether Congress has delegated

Chevron U.S.A. Inc. v. Natural Resources Defense Council, Inc. to asks courts administrative agencies the authority to resolve questions about the meaning of statutes that those agencies implement, but the decision does not give courts the tools for providing a proper answer. Chevron directs courts to construe statutory text by applying the traditional intentionalism, theories of purposivism, or textualism—and to infer a delegation of agency interpretive authority only if they fail to find a relatively specific meaning. But the traditional theories, despite their differences, all invite courts to construe statutory text as if Congress intended that text to have a relatively specific meaning. The presumption of a specific meaning does not match the reality of how Congress designs regulatory statutes. Congress is more likely to eschew specificity in favor of agency delegation under certain circumstances—for example, if an issue is complex and if legislators can monitor subsequent agency interpretations through administrative procedures. Although

Copyright © 2009 by Lisa Schultz Bressman.

† Professor of Law, Vanderbilt Law School. For helpful comments and discussions, I thank Rachel Barkow, Michael Bressman, Bill Eskridge, John Goldberg, Maggie Lemos, Jerry Mashaw, Cathy Sharkey, Suzanna Sherry, Kevin Stack, Matthew Stephenson, Cristina Rodriguez, Adrian Vermeule, Ernie Young, as well as participants at the Duke University School of Law faculty workshop, the Georgetown Law Center legislation seminar, the Harvard Law School faculty workshop, and the New York University Law School Colloquium on Public Law. For excellent research assistance, I thank Leah Bressack and John Greer.

Chevron recognizes such “delegating” factors, it fails to sufficiently credit them. Even United States v. Mead Corp., which makes delegation the key question, falls short. This Article imagines what interpretive theory would look like for regulatory statutes if it actually incorporated realistic assumptions about legislative behavior. The theory would engage factors such as the complexity of the issue and the existence of administrative procedures as indications of interpretive delegation more satisfactorily than existing law does. In the process, it would produce a better role for courts in overseeing the delegation of authority to agencies.

INTRODUCTION

1. Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837 (1984). 2. See id. at 842, 843 n.9. 3. E.g., T. Alexander Aleinikoff, Updating Statutory Interpretation, 87 MICH. L. REV. 20, 24 (1988) (“[Intentionalists] would scrutinize the legislative materials to see if the legislature actually considered and expressed an opinion on the question under review.”); Cass R. Sunstein, Interpreting Statutes in the Regulatory State, 103 HARV. L. REV. 405, 429 (1989) (“[For intentbased views,] the goal is not to look at a general legislative aim or purpose, but instead to see more particularly how the enacting legislature would have resolved the question, or how it intended that question to be resolved, if it had been presented.”). The exception is a more “objective” version of this theory. See, e.g., Cheryl Boudreau, Mathew D. McCubbins & Daniel B. Rodriguez, Statutory Interpretation and the Intentional(ist) Stance, 38 LOY. L.A. L. REV. 2131, 2137–38, 2143–46 (2005) (focusing interpretation on the objective intentionality that words reflect).

4. E.g., Martin H. Redish & Theodore T. Chung, Democratic Theory and the Legislative Process: Mourning the Death of Originalism in Statutory Interpretation, 68 TUL. L. REV. 803, 815 (1994) (“[P]urposivism calls on judges to identify the statute’s broader purposes and to resolve the interpretive question in light of those purposes.”); see also, e.g., HENRY M. HART, JR. & ALBERT SACKS, THE LEGAL PROCESS: BASIC PROBLEMS IN THE MAKING AND APPLICATION OF LAW 166–67, 1378 (William N. Eskridge, Jr. & Philip P. Frickey eds., 1958) (focusing interpretation on the broader purposes embodied in statutes and asking courts to assume, “unless the contrary unmistakably appears,” that “the legislature was made up of reasonable persons pursuing reasonable purposes reasonably”).

5. E.g., Frank H. Easterbrook, Text, History, and Structure in Statutory Interpretation, 17 HARV. J.L. & PUB. POL’Y 61, 68 (1994) (“Intent is empty. Peer inside the heads of legislators and you find a hodgepodge.” (emphasis omitted ; John F. Manning, Continuity and the Legislative Design, 79 NOTRE DAME L. REV. 1863, 1864 (2004) (“[M]odern formalists (qua textualists) doubt that intent or purpose gleaned from the legislative history offers a reliable way to resolve statutory indefiniteness.”); John F. Manning, Textualism as a Nondelegation Doctrine, 97 COLUM. L. REV. 673, 684–89 (1997) [hereinafter Manning, Nondelegation Doctrine] (describing textualist arguments against “genuine legislative intent”). See generally Daniel A. Farber & Philip P. Frickey, Legislative Intent and Public Choice, 74 VA. L. REV. 423, 453–57 (1988) (describing public choice insight behind textualism); Kenneth A. Shepsle, Congress is a “They,” Not an “It”: Legislative Intent as Oxymoron, 12 INT’L REV. L. & ECON. 239, 239 (1992) (“Legislative intent is an internally inconsistent, self-contradictory expression.”). On textualism generally, see William N. Eskridge, Jr., The New Textualism, 37 UCLA L. REV. 621 (1989); John F. Manning, The Absurdity Doctrine, 116 HARV. L. REV. 2387, 2390, 2408–19 (2002). For a contrast between textualism and intentionalism, see John F. Manning, Textualism and Legislative Intent, 91 VA. L. REV. 419 (2005) [hereinafter Manning, Legislative Intent]. For a contrast between textualism and purposivism, see Michael C. Dorf, The Supreme Court, 1997 Term—Foreword: The Limits of Socratic Dialogue, 112 HARV. L. REV. 4, 14–26 (1997); John F. Manning, What Divides Textualists from Purposivists, 106 COLUM. L. REV. 70 (2006).

6. E.g., Eskridge, supra note 5, at 646–50 (reviewing traditional textualist arguments that the use of legislative history distorts the separation of powers between Congress and the judiciary); see also, e.g., Easterbrook, supra note 5, at 66 (“[The] text prevails over intent because only the text went through the constitutional process.”); Farber & Frickey, supra note 5, at 454–56 (describing Justice Scalia and Judge Easterbrook’s constitutional objection that using lawmaking process); Manning, Nondelegation Doctrine, supra note 5, at 711–19 (arguing that the use of legislative history cedes constitutional lawmaking authority from Congress as a whole to individual committees or members of Congress).

legislative history circumvents the constitutional

7. See, e.g., David Epstein & Sharyn O’Halloran, The Nondelegation Doctrine and the Separation of Powers: A Political Science Approach, 20 CARDOZO L. REV. 947, 950 (1998)

(“[L]egislators will delegate in those issue areas where the normal legislative process is the least efficient relative to regulatory policymaking by executive agencies.”).

8. See id. at 967. 9. See, e.g., WILLIAM N. ESKRIDGE, JR., DYNAMIC STATUTORY INTERPRETATION 38 (1994) (asserting that legislators “create rather than avoid ambiguity” when necessary to avoid making choices that are unpopular with their constituents); Joseph A. Grundfest & A.C. Pritchard, Statutes with Multiple Personality Disorders: The Value of Ambiguity in Statutory Design and Interpretation, 54 STAN. L. REV. 627, 641 (2002) (“[It] is not unusual for competing factions of Congress to ‘agree to disagree’ in the drafting of a statute.”); Victoria F. Nourse & Jane S. Schacter, The Politics of Legislative Drafting: A Congressional Case Study, 77 N.Y.U. L. REV. 575, 596 (2002) (interviewing legislative staffers who confirm that legislators use deliberate ambiguity to obtain consensus); Richard A. Posner, Statutory Interpretation in the Classroom and the Courtroom, 50 U. CHI. L. REV. 800, 811–12 (1983) (identifying the failure to agree as a cause of statutory ambiguity).

10. See Mathew D. McCubbins, Roger G. Noll & Barry R. Weingast, Administrative Procedures as Instruments of Political Control, 3 J.L. ECON. & ORG. 243, 246 (1987) [hereinafter McCubbins et al., Administrative Procedures]; Mathew D. McCubbins, Roger G. Noll & Barry R. Weingast, Structure and Process, Politics and Policy: Administrative Arrangements and the Political Control of Agencies, 75 VA. L. REV. 431, 442 (1989) [hereinafter McCubbins et al., Structure and Process].

11. See DAVID EPSTEIN & SHARYN O’HALLORAN, DELEGATING POWERS 131–33 (1999). 12. See Nourse & Schacter, supra note 9, at 596–97. 13. See Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., , 865 (1984).

14. See id. at 842, 843 n.9; see also Jacob E. Gersen & Adrian Vermeule, Chevron as a Voting Rule, 116 YALE. L.J. 676, 690–91 (2007) (noting that the traditional interpretive theories ask courts to construct a best meaning for ambiguous statutory language); Grundfest & Pritchard, supra note 9, at 628 (“[J]udges and scholars have developed an arsenal of interpretive techniques that are designed to extract functional meaning from ambiguous statutory text and conflicting legislative history.”); Nourse & Schacter, supra note 9, at 617 (“[T]o the extent that intentionalism assumes a legislative intent always exists or that textualism assumes a deliberate and precise legislative word choice, these theories may rely on questionable empirical assumptions across the run of cases.” (emphasis omitted .

15. See, e.g., Epstein & O’Halloran, supra note 7, at 950 (arguing that legislators delegate when the legislative process produces less efficient outcomes than the administrative process); Grundfest & Pritchard, supra note 9, at 640–42 (describing legislative incentives to delegate authority); Nourse & Schacter, supra note 9, at 596–97 (quoting legislative staffers discussing the political-consensus-building purposes of delegation).

16. United States v. Mead Corp., 533 U.S. 218 (2001). 17. Id. at 231–33.

18. Zuni Pub. Sch. Dist. No. 89 v. Dep’t of Educ., 127 S. Ct. 1534 (2007). 19. Id. at 1541. 20. See, e.g., id. at 1551 (Kennedy, J., concurring) (worrying because the “opinion of the

Court . . . inverts Chevron’s logical progression”).

21. Gonzales v. Oregon, 126 S. Ct. 904 (2006). 22. Id. at 925.

23. See id. at 917–22 (considering the relevant expertise of the agency, the moral nature of

the question, and the procedures for issuing the interpretation).

24. Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., , 865 (1984). 25. Administrative law has long provided a basis for such review, for example, in the requirement of reasoned decisionmaking. See Motor Vehicles Mfrs. Ass’n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 41 (1983) (elaborating the requirement of reasoned decisionmaking).

26. See Lisa Schultz Bressman, How Mead Has Muddled Judicial Review of Agency Action, 58 VAND. L. REV. 1443, 1446–81 (2005) (providing an empirical study of lower court cases); Adrian Vermeule, Introduction: Mead in the Trenches, 71 GEO. WASH. L. REV. 347 passim (2003) (showing confusion in D.C. Circuit cases trying to apply Mead).

27. See William N. Eskridge, Jr., No Frills Textualism, 119 HARV. L. REV. 2041, 2044 (2006) (book review) (arguing against an institutionally simple interpretive proposal that generates normative costs).

28. United States v. Mead Corp., , 239 (2001) (Scalia, J., dissenting). 29. The nondelegation doctrine reflects constitutional limits on Congress’s ability to delegate regulatory authority to agencies. See Lisa Schultz Bressman, Schechter Poultry at the Millennium: A Delegation Doctrine for the Administrative State, 109 YALE L.J. 1399, 1403–06 (2000) [hereinafter Bressman, Schechter Poultry at the Millennium] (encapsulating the history of the nondelegation doctrine). With the exception of two 1935 decisions, the Court has never applied the doctrine to ban a statutory delegation. See A.L.A. Schechter Poultry Corp. v. United

States, 295 U.S. 495, 529 (1935) (invalidating a statute under nondelegation doctrine); Panama Refining Co. v. Ryan, 293 U.S. 388, 433 (1935) (same). It has merely insisted that Congress supply an “intelligible principle” in the statute guiding the agency’s exercise of delegated authority. See J.W. Hampton, Jr., & Co. v. United States, 276 U.S. 394, 409 (1928) (articulating the intelligible principle requirement for the first time). Finding this requirement easily satisfied, the Court has relied on statutory interpretation to vindicate its abiding worries about broad delegations. See Lisa Schultz Bressman, Beyond Accountability: Arbitrariness and Legitimacy in the Administrative State, 78 N.Y.U. L. REV. 461, 515–52 (2008) (describing constitutional law law as efforts to combat worries about arbitrary administrative and administrative decisionmaking that stem from broad delegations). More specifically, it has used various interpretive norms to narrowly construed broad statutory delegations. See Bressman, Schechter Poultry at the Millennium, supra, at 1408–15 (describing the use of interpretive norms to constrain broad delegations); Cass R. Sunstein, Law and Administration After Chevron, 90 COLUM. L. REV. 2071, 2110–15 (1990) (discussing the emergence of clear statement principles as a surrogate for the nondelegation doctrine).

I. THE DOMINANT THEORIES AND THEIR ADMINISTRATIVE LAW DOCTRINE

30. See, e.g., Aleinikoff, supra note 3, at 23 (“Intentionalism . . . claims that textualism inappropriately ignores contextual elements in statutory interpretation. Contextual analysis is necessary as a matter of semantics (words have no ‘plain meaning’; meaning depends on context and usage).”); Sunstein, supra note 3, at 416 (“The central problem is that the meaning of words (whether ‘plain’ or not) depends on both culture and context. Statutory terms are not selfdefining, and words have no meaning before or without interpretation.”); see also John F. Manning, Competing Presumptions About Statutory Coherence, 74 FORDHAM L. REV. 2009, 2013–16 (2006) (describing the interpretive approaches of many thinkers, including Justice Stevens and Judge Posner, as reflecting this view of the legislative process).

31. See, e.g., Redish & Chung, supra note 4, at 816 (“[Purposivists] would also have their judges explore the entire legal landscape to determine how the statute at hand can best be made to fit within its greater legal context.”); Sunstein, supra note 3, at 430 (arguing that intentionalism can provide useful “context and purpose”); see also STEPHEN BREYER, ACTIVE LIBERTY 85 (1994) (describing purposivism as an approach under which “judges should pay primary attention to a statute’s purpose”); Abner S. Greene, The Missing Step of Textualism, 74

FORDHAM L. REV. 1913, 1916 (2006) (“Purposivists, or intentionalists, look at . . . legislative history and other background social understandings[]in an effort to figure out what Congress was up to.”).

32. See, e.g., HART & SACKS, supra note 4, at 1124 (“Every statute must be conclusively

presumed to be a purposive act.”).

33. See supra note 5 and accompanying text. 34. See Manning, Legislative Intent, supra note 5, at 441 (“[Textualists] believe that smoothing over the rough edges in a statute threatens to upset whatever complicated bargaining led to its being cast in the terms that it was.”).

35. See supra note 5 and accompanying text. 36. See Manning, Nondelegation Doctrine, supra note 5, at 695–705 (arguing that the legislative history bypasses the Article I requirements of bicameralism and presentment); John F. Manning, Textualism and the Equity of the Statute, 101 COLUM. L. REV. 1, 71–74 (2001) (arguing that Article I requirements of bicameralism and presentment compel textualism).

37. Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., , 842, 843 n.9

(1984).

38. Id. at 843–44. 39. Id. at 865. 40. See id. 41. Id. 42. Id. 43. Id. 44. United States v. Mead Corp., , 233 (2001). 45. Id. 46. Id.

47. Id. at 230, 232. 48. Chevron, 467 U.S. at 842–43. 49. Id. at 843 n.9. 50. Id. at 842. 51. See, e.g., Einer Elhauge, Preference-Estimating Statutory Default Rules, 102 COLUM. L. REV. 2027, 2132 (2002) (“[E]ven adherents to [Chevron] theory . . . acknowledge that the evidence of such enacting congressional intent is ‘weak’ and even ‘fictional’ . . . .”); Gersen & Vermeule, supra note 14, at 689 (noting that Justices and commentators have recognized that the presumption of legislative intent underlying Chevron is a “fiction” and agreeing with this view); Thomas W. Merrill & Kristin E. Hickman, Chevron’s Domain, 89 GEO. L.J. 833, 871–72 (2001) (noting that the presumption of legislative intent underlying Chevron deference “has been described by even [Chevron’s] strongest defender [Justice Scalia] as ‘fictional’”).

52. See Chevron, 467 U.S. at 865. Many have expressed other views about Chevron, including the view that it is premised on congressional allocation of interpretive authority between agencies and courts. See Merrill & Hickman, supra note 51, at 837; Kevin M. Stack, The

Statutory President, 90 IOWA L. REV. 539, 587 (2005); cf. Thomas W. Merrill & Kathryn Tongue Watts, Agency Rules with the Force of Law: The Original Convention, 116 HARV. L. REV. 467, 472–74 (2002) (arguing that Congress once used a legislative convention to confer on agencies the authority to act with the force of law and that, when this convention fell out of use, courts adopted a deferential approach to determining whether Congress authorized an agency to act with the force of law). Although the literature conceptualizing Chevron is too vast to cite, other excellent examples include Michael Herz, Deference Running Riot: Separating Interpretation and Lawmaking Under Chevron, 6 ADMIN. L.J. AM. U. 187 (1992); Ronald M. Levin, The Anatomy of Chevron: Step Two Reconsidered, 72 CHI.-KENT L. REV. 1253 (1997); Richard J. Pierce, Jr., Reconciling Chevron and Stare Decisis, 85 GEO. L.J. 2225 (1997); Mark Seidenfeld, A Syncopated Chevron: Emphasizing Reasoned Decisionmaking in Reviewing Agency Interpretations of Statutes, 73 TEX. L. REV. 83 (1994); Sunstein, supra note 29.

53. See United States v. Mead Corp., , 218 (2001) (holding that congressional intent to delegate the authority to make rules with the force of law “may be shown in a variety of ways, as by an agency’s power to engage in adjudication or notice-and-comment rulemaking”).

Textualism . . . seems to transform statutory interpretation into a kind of exercise in judicial ingenuity. The textualist judge treats questions of interpretation like a puzzle to which it is assumed there is one right answer. The task is to assemble the various pieces of linguistic data, dictionary definitions, and canons into the best (most coherent, most explanatory) account of the meaning of the statute . . . .

interpretation

is subtly This active, creative approach to incompatible with an attitude of deference toward other institutions—whether the other institution is Congress or an administrative agency. In effect, the textualist interpreter does not find the meaning of the statute so much as construct the meaning. Such a person will very likely experience some difficulty in deferring to the meanings that other institutions have developed.54

54. Thomas W. Merrill, Textualism and the Future of the Chevron Doctrine, 72 WASH. U.

L.Q. 351, 372 (1994).

55. See id. 56. See Gersen & Vermeule, supra note 14, at 697–98.

It is difficult, after having examined a legal question in depth with the object of deciding it correctly, to believe both that the agency’s interpretation is legally wrong, and that its interpretation is reasonable. More often one concludes that there is a “better” view of the statute for example, and that the “better” view is “correct,” and the alternative view is “erroneous.”59

57. See id. at 693–97 (discussing the conceptual problems that arise when attempting to

discern the limits of “reasonable” interpretation).

58. See id. at 697–98 (making use of Justice Scalia’s opinion in MCI Telecommunications Corp. v. AT&T Co., 512 U.S. 218 (1994), to illustrate the tendency of courts to commit to one “right” interpretation when other reasonable interpretations are available).

59. Gersen & Vermeule, supra note 14, at 697 (quoting Stephen Breyer, Judicial Review of

Questions of Law and Policy, 38 ADMIN. L. REV. 363, 379 (1986 .

60. See Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., , 865 (1984) (“In these cases the Administrator’s interpretation represents a reasonable accommodation of manifestly competing interests and is entitled to deference: the regulatory scheme is technical and complex, the agency considered the matter in a detailed and reasoned fashion, and the decision involves reconciling conflicting policies.” (footnotes omitted .

61. See id. at 863–64 (“[T]he agency, to engage in informed rulemaking, must consider

varying interpretations and the wisdom of its policy on a continuing basis.”).

62. See Nat’l Cable & Telecomms. Ass’n v. Brand X Internet Servs., 545 U.S. 967, 982–83 (2005) (holding that an agency interpretation may overrule a prior judicial interpretation only if the underlying statutory language is ambiguous).

63. See SUSAN ROSE-ACKERMAN, ECONOMICS OF ADMINISTRATIVE LAW, ECONOMICS

OF LAW vi–xvii (2007) (summarizing the literature).

64. See, e.g., Epstein & O’Halloran, supra note 7, at 961 (“We assume the preferences of

legislators and the President to be, first and foremost, reelection.”).

65. Id.

The first and most obvious reason is that the executive branch is filled (or can be filled) with policy experts who can run tests and experiments, gather data, and otherwise determine the wisest course of policy, much more so than can 535 members of Congress and their staff. The second, less obvious reason has to do with the fact that expertise garnered legislative committees cannot be transformed directly into policy outcomes. Rather, it must first pass through the floor, which may decide to make some alterations to the committee’s proposals. The existence of the floor as a policy middleman gives committees less incentive to gather information in the first place. Executive agencies, on the other hand, are not hampered by the need to obtain congressional approval; their rulings become law directly.70

66. Id. (“We assume that political actors who seek reelection will, on any given policy, attempt to bring final outcomes as close as possible to the median voter in their politically relevant constituency.”).

67. Id. at 962; see also David Epstein & Sharyn O’Halloran, Administrative Procedures, Information, and Agency Discretion, 38 AM. J. POL. SCI. 697, 699 (1994) (noting that a direct method of circumscribing agency influence is “explicitly limiting the discretion of an agency to move outcomes from the status quo”); Terry M. Moe, Political Institutions: The Neglected Side of the Story, 6 J.L. ECON. & ORG. (SPECIAL ISSUE) 213, 228 (1990) (“The most direct way is for today’s authorities to specify, in excruciating detail, precisely what the agency is to do and how it is to do it, leaving as little as possible to the discretionary judgment of bureaucrats . . . .”).

68. See Epstein & O’Halloran, supra note 67, at 701; David B. Spence & Frank Cross, A Public Choice Case for the Administrative State, 89 GEO. L.J. 97, 106–12 (2000); B. Dan Wood & John Bohte, Political Transaction Costs and the Politics of Administrative Design, 66 J. POL. 176, 176 (2004).

69. Epstein & O’Halloran, supra note 7, at 962. 70. Id. at 967 (footnote omitted).

71. Id. 72. See DAVID SCHOENBROD, POWER WITHOUT RESPONSIBILITY: HOW CONGRESS ABUSES THE PEOPLE THROUGH DELEGATION 9–12, 49–57 (1993); see also JAMES Q. WILSON, BUREAUCRACY: WHAT GOVERNMENT AGENCIES DO AND WHY THEY DO IT 235–256 (1989) (exploring how Congress exerts control over regulatory agencies through delegation and the legislative process); Peter H. Aranson, Ernest Gellhorn & Glen O. Robinson, A Theory of Legislative Delegation, 68 CORNELL L. REV. 1, 55–63 (1982) (dissecting and comparing several hypotheses that try to explain why Congress delegates power to agencies); Morris P. Fiorina, Legislative Choice of Regulatory Forms: Legal Process or Administrative Process?, 39 PUB. CHOICE 33, 46–52 (1982) (examining claims that Congress delegates power to regulatory agencies to shift both decisionmaking and political costs to the agency).

73. See Morris P. Fiorina, Legislator Uncertainty, Legislative Control, and the Delegation of Legislative Power, 2 J.L. ECON. & ORG. 33, 44–55 (1986) (arguing that model legislators choose between two lotteries: one involving courts, whose decisions are expected to reflect the preferences of the median legislator, and one involving agencies, whose decisions are expected to be biased away from the intent of that legislator).

74. See Matthew C. Stephenson, Legislative Allocation of Delegated Power: Uncertainty, Risk, and the Choice Between Agencies and Courts, 119 HARV. L. REV. 1035, 1038, 1047–49 (2006).

75. See id. at 1049–58. 76. See EPSTEIN & O’HALLORAN, supra note 11, at 131–33. 77. See id. 78. See Epstein & O’Halloran, supra note 7, at 966. 79. See EPSTEIN & O’HALLORAN, supra note 11, at 154–55. 80. See generally Murray J. Horn & Kenneth A. Shepsle, Commentary on “Administrative Arrangements and the Political Control of Agencies”: Administrative Process and Organizational Form as Legislative Responses to Agency Costs, 75 VA. L. REV. 499, 501–04 (1989) (describing bureaucratic drift, which is the difference between policy passed in legislation and policy implemented by an agency, and coalitional drift, during which legislative and executive preferences change over time); Kenneth A. Shepsle, Bureaucratic Drift, Coalitional Drift, and Time Consistency: A Comment on Macey, 8 J.L. ECON. & ORG. 111, 113–16 (1992) (same).

81. See JOHN D. HUBER & CHARLES R. SHIPAN, DELIBERATE DISCRETION? THE INSTITUTIONAL FOUNDATIONS OF BUREAUCRATIC AUTONOMY 26 (2002) (“The principalagent framework from economics has played an extremely prominent and powerful role in [the] institutional approach to relations between politicians and bureaucrats.”).

82. See Terry M. Moe, The Politics of Bureaucratic Structure, in CAN THE GOVERNMENT GOVERN? 267, 271 (John E. Chubb & Paul E. Peterson eds., 1989) (“Experts have their own interests—in career, in autonomy—that may conflict with those of [legislators].”); Epstein & O’Halloran, supra note 7, at 963.

83. Mathew D. McCubbins & Thomas Schwartz, Congressional Oversight Overlooked:

Police Patrols Versus Fire Alarms, 28 AM. J. POL. SCI. 165, 166 (1984).

84. Id. at 166. 85. Id. 86. Id. at 173–74. 87. See McCubbins et al., Administrative Procedures, supra note 10, at 248; McCubbins et al., Structure and Process, supra note 10, at 442; see also Rui J.P. de Figueiredo, Jr., Pablo T. Spiller & Santiago Urbiztondo, An Informational Perspective on Administrative Procedures, 15 J.L. ECON. & ORG. 238, 301 (1999) (modeling the functions of administrative procedures).

88. See McCubbins et al., Structure and Process, supra note 10, at 442. 89. Id. at 442, 444 (italics omitted). 90. See McCubbins et al., Administrative Procedures, supra note 10, at 254 (“[B]ecause policy is controlled by participants in administrative processes, political officials can use procedures to control policy without bearing costs themselves . . . .”).

91. Epstein & O’Halloran, supra note 67, at 716 (emphasis omitted). 92. Kathleen Bawn, Choosing Strategies to Control the Bureaucracy: Statutory Constraints,

Oversight, and the Committee System, 13 J.L. ECON. & ORG. 101, 101 (1997).

93. Id. 94. See, e.g., ESKRIDGE, supra note 9, at 38 (noting that the simple “plain meaning” rule of textualism overlooks that “the goals of at least some of the authors are to create rather than avoid ambiguity”); Grundfest & Pritchard, supra note 9, at 641 (“[I]t is not unusual for competing factions of Congress to ‘agree to disagree’ in the drafting of a statute.”); Mathew D. McCubbins, The Legislative Design of Regulatory Structure, 29 AM. J. POL. SCI. 721, 742–43 (1985) (asserting that legislators delegate when necessary to obtain consensus on policy); Nourse & Schacter, supra note 9, at 596 (interviewing legislative staffers who confirm the use of deliberate ambiguity to obtain consensus); Posner, supra note 9, at 806–07 (identifying failure to agree as a cause of statutory ambiguity).

95. See Grundfest & Pritchard, supra note 9, at 641 (“Each constituency can hope that its position will ultimately prevail, and ambiguity thereby expands the circle of winners in legislative battles, at least temporarily.”); Robert A. Katzmann, The American Legislative Process as a Signal, 9 J. PUB. POL’Y 287, 290 (1989) (“A problem may be defined in general terms . . . because it is easier to secure support for ambiguously worded statutes that mean all things to all people.”).

96. See Katzmann, supra note 95, at 290–91 (“Once a problem is identified or recognized, interests within and without Congress push for an interpretation consistent with their policy preferences, regardless of whether it is faithful to the original legislative intent.”).

97. Id. at 290.

98. See Nourse & Schacter, supra note 9, at 596 (interviewing sixteen staffers working on the Senate Judiciary Committee or one of its subcommittees, who reported that legislative drafting involves “willful lack of clarity”).

99. Id. at 596–97.

100. Id. at 596 (“This is . . . a political process. Sometimes one cannot allow the perfect to be the enemy of the good.” (quoting a staffer discussing deliberate ambiguity . 101. Id. at 596–97. 102. See Grundfest & Pritchard, supra note 9, at 650, 650–66 (describing the “legislative evolution” of the pleading standard used in the PSLRA). 103. See infra Part II.C. 104. Grundfest & Pritchard, supra note 9, at 650. 105. Id. at 652. 106. Id. 107. Id. (quoting Shields v. Citytrust BanCorp, Inc., 25 F.3d 1124, 1128 (2d Cir. 1994 .

This formulation has the air of a compromise, suggesting that neither proponents nor opponents of recklessness were capable of garnering a majority (much less a supermajority) for their view. By

108. Id. (quoting In re Glenfeld, Inc. Sec. Litig., 42 F.3d 1541, 1547 (9th Cir. 1994) (en banc . 109. Id. 110. See id. at 652–53 (“The bill eventually passed by the House . . . codified recklessness as ‘[d]eliberately refraining from taking steps to discover whether one’s statements are false or misleading.’” (quoting Securities Litigation Reform Act, H.R. 1058, 104th Cong. § 10A(a)(4) (as passed by the House, Mar. 8, 1995 . 111. See id. at 653 (explaining that the Senate banking committee report clearly stated an intent to adopt the Second Circuit standard). 112. See id. at 654 (asserting that Senator Specter’s proposed amendment would have made the law rely more explicitly on the Second Circuit tests). 113. Id. 114. See id. at 655–57 (explaining that the language of the statement of managers section conflicts with a reasonable interpretation of the concededly vague language of footnote 23 in that same section). 115. Id. at 658 (internal quotation marks omitted).

leaving the question unanswered in the statute, both sides could hope that the Supreme Court would eventually rule in their favor.116

116. Id. 117. See id. at 659. 118. Id. 119. Id. 120. Id. at 660. 121. See, e.g., id. at 667–77 (explaining that the final construction of the PSLRA prompted the appellate courts to interpret the strong inference standard using three different approaches); Nourse & Schacter, supra note 9, at 616–21 (describing potential flaws in judicial assumptions about legislative intent that belie the leading theories of statutory interpretation); Sunstein, supra note 3, at 411 (“[D]ebates about statutory interpretation, in and out of the judiciary, often dissolve into fruitless and unilluminating disputes about the constraints supplied by language ‘itself’ (as if such a thing could be imagined).”).

II. A DELEGATION-RESPECTING THEORY

122. Nourse & Schacter, supra note 9, at 619. Grundfest and Pritchard do not challenge judicial practice but note that it thrusts legislators into a game with courts over whose interpretive strategy ultimately will prevail. See Grundfest & Pritchard, supra note 9, at 670–72.

123. See supra Parts I.B–C. 124. Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., , 865 (1984). 125. Id.

126. Cf. W. Kip Viscusi, Toward a Diminished Role for Tort Liability: Social Insurance, Government Regulation, and Contemporary Risks to Health and Safety, 6 YALE J. ON REG. 65, 76 (1989) (noting that the benefit of agency regulation (compared to tort litigation) increases “with the degree of specialization and complexity of data and with the greater general need for the information in the economy”). 127. Cf. id. at 105 (arguing that “courts lack the scientific expertise to develop the necessary doctrines and causation rules relevant to multiple and probabilistic causation” in toxic tort cases). 128. Chevron, 467 U.S. at 865. 129. See supra text accompanying notes 68–71. 130. See supra text accompanying notes 72–73.

131. Kristin E. Hickman, Of Lenity, Chevron, and KPMG, 26 VA. TAX REV. 905, 923 (2007) (“As designated representatives of the people, members of Congress are both more in touch with communal perceptions of ‘right’ and ‘wrong’ and more accountable to the public for the moral judgments they make than agencies are. While the Supreme Court has not explicitly made this link, other courts and scholars have highlighted the moral element of criminalization as a further reason for not extending judicial deference to Justice Department interpretations of the criminal code.”); see also BREYER, supra note 31, at 107 (arguing that Congress is capable of making value judgments and so courts should not defer to agencies’ judgments instead). Another possible example is private rights of action. The Court has held that courts, not agencies, must determine whether to imply such rights of action. See Adams Fruit Co. v. Barrett, 494 U.S. 638, 649–50 (1990) (finding that Congress expressly delegated authority to courts to determine whether statutes confer private rights of action). Yet, even here, Congress can be deliberately ambiguous on the issue. See Marc I. Steinberg, Implied Private Rights of Action Under Federal Law, 55 NOTRE DAME L. REV. 33, 40 (1979). Furthermore, Professor Matthew Stephenson has argued that determining whether to imply a private right of action involves the sort of complex judgment that Congress might seek to avoid. See Matthew C. Stephenson, Public Regulation of Private Enforcement: The Case for Expanding the Role of Administrative Agencies, 91 VA. L. REV. 93, 127–28 (2005). Thus, private rights of action may or may not be complex. Cf. id. at 148–70 (arguing that Chevron should automatically apply to agency determinations on this issue). 132. Chevron, 467 U.S. at 865.

133. See Grundfest & Pritchard, supra note 9, at 650–66; see also supra Part I.C. (explaining the academic hypothesis that the opposing factions on the “strong inference” issue agreed to ambiguous language, each hoping the Supreme Court would rule in its favor). 134. Grundfest & Pritchard, supra note 9, at 667–77. 135. See Samuel Issacharoff, Regulating After the Fact, 56 DEPAUL L. REV. 375, 381 (2007) (“There is little dispute about the centrality of private actions in enforcing the complex web of securities law.”); Harvey L. Pitt & Karen L. Shapiro, Securities Regulation by Enforcement: A Look Ahead at the Next Decade, 7 YALE J. ON REG. 149, 206–08 (1990) (noting that the SEC’s approach to insider trading relies on enforcement actions brought by the SEC and managed by lower courts). But see James J. Park, The Competing Paradigms of Security Regulation, 57 DUKE L.J. 625, 630 (2007) (noting that the SEC has issued rules in certain areas).

136. See supra text accompanying notes 80–82. 137. See supra text accompanying notes 111–15. 138. See supra text accompanying notes 83–90. 139. United States v. Mead Corp., , 230 (2001). 140. Id. at 230, 232. 141. Barnhart v. Walton, 535 U.S. 212 (2002).

142. Id. at 222. 143. Mead, 533 U.S. at 230. 144. Barnhart, 535 U.S. at 222. 145. An agency cannot bootstrap itself into a delegation by using relatively formal procedures when Congress has not authorized those procedures. If the linchpin is legislative

intent to delegate, then Congress must intend to provide the agency with interpretive authority. The agency’s decision to afford Congress practical monitoring opportunities via procedures, even if beneficial on this ground or others, cannot cure the delegation deficiency. 146. See supra text accompanying notes 83–90. 147. Barnhart, 535 U.S. at 222. 148. Id.

149. See Henry P. Monaghan, Marbury and the Administrative State, 83 COLUM. L. REV. 1, 33 (1983) (“[T]he judicial duty is to ensure that the administrative agency stays within the zone of discretion committed to it by its organic act.”).

150. See Epstein & O’Halloran, supra note 67, at 701 (discussing statutes that, without dictating a single best policy, “did eliminate certain policies from consideration”). 151. Zuni Pub. Sch. Dist. No. 89 v. Dep’t of Educ., 127 S. Ct. 1534, 1543 (2007) (“A customs statute that imposes a tariff on ‘clothing’ does not impose a tariff on automobiles, no matter how strong the policy arguments for treating the two kinds of goods alike.”). 152. See J.W. Hampton, Jr., & Co. v. United States, 276 U.S. 394, 409 (1928) (understanding the nondelegation doctrine as entailing a statutory limit on delegations of authority to agencies). 153. Id. 154. See Whitman v. Am. Trucking Ass’ns, 531 U.S. 457, 474 (2001) (acknowledging that although Congress, when delegating decisionmaking, must establish intelligible principles for agencies to follow, those principles only must provide minimal guidance); Mistretta v. United States, 488 U.S. 361, 415–16 (1989) (Scalia, J., dissenting) (same). 155. See A.L.A. Schechter Poultry Corp. v. United States, 295 U.S. 495, 542, 541–42 (1935) (invalidating a statutory delegation that conferred “virtually unfettered” discretion on an agency); Panama Refining Co. v. Ryan, 293 U.S. 388, 430 (1935) (invalidating a statutory delegation for lacking standards).

156. Relatedly, when Congress instructs an agency to regulate x, it cannot decline to regulate one type of x. See Massachusetts v. EPA, 127 S. Ct. 1438, 1460 (2007) (finding that carbon dioxide is an “air pollutant” within the meaning of the Clean Air Act and therefore that the statute “forecloses” the EPA’s contrary reading). 157. See Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., , 844 (1984) (explaining that agency interpretations are valid unless “arbitrary, capricious, or manifestly contrary to the statute”). 158. See 5 U.S.C § 706 (2006) (“The reviewing court shall . . . set aside agency action, findings, and conclusions found to be arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law . . . .”). 159. See M. Elizabeth Magill, Step Two of Chevron v. Natural Resources Defense Council, in A GUIDE TO JUDICIAL AND POLITICAL REVIEW OF FEDERAL AGENCIES 85, 96–102 (John F. Duffy & Michael Herz eds., 2005) (noting confusion among courts concerning the relationship between Chevron Step Two and APA arbitrary and capricious analysis and arguing that the two should mirror one another); Kevin M. Stack, The Constitutional Foundations of Chenery, 116 YALE L.J. 952, 1005–07 (2007) (demonstrating that Chevron Step Two incorporates the reasoned decisionmaking requirement). 160. See Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 41, 43 (1983) (“[T]he agency must examine the relevant data and articulate a satisfactory explanation for its action including a ‘rational connection between the facts found and the choice made.’” (quoting Burlington Truck Lines, Inc. v. United States, 371 U.S. 156, 168 (1962 .

161. See Lisa Schultz Bressman, Procedures as Politics in Administrative Law, 107 COLUM. L. REV. 1749, 1777–78 (2007) (tracing the pre-APA roots of the reasoned decisionmaking requirement). 162. Id. at 1778–79 (describing the function of the reasoned decisionmaking requirement); see also State Farm, 463 U.S. at 41 (listing considerations agencies might neglect). 163. See, e.g., State Farm, 463 U.S. at 42 (requiring the agency to explain the basis for its decision); Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 415, 419–20 (1971) (same). 164. See Bressman, supra note 161, at 1778–79. 165. See id. at 1779.

166. Skidmore v. Swift & Co., 323 U.S. 134 (1944). 167. Id. at 140.

168. See Matthew C. Stephenson, The Strategic Substitution Effect: Textual Plausibility, Procedural Formality, and Judicial Review of Agency Statutory Interpretations, 120 HARV. L. REV. 528, 531 (2006) (arguing that agencies, to secure approval for their interpretations from courts, choose textually plausible interpretations when they want to avoid the costs of issuing more aggressive interpretations through formal procedures that also satisfy courts). 169. Nat’l Cable & Telecomms. Ass’n v. Brand X Internet Servs., , 983–84 (2005).

III. ILLUSTRATIONS

Department officials listed each of New Mexico’s 89 local school districts in order of per-pupil spending . . . . After ranking the districts, Department officials excluded 17 school districts at the top of the list because those districts contained (cumulatively) less than 5 percent of the student population; for the same reason, they excluded an additional 6 school districts at the bottom of the list.

The remaining 66 districts accounted for approximately 90 percent of the State’s student population. Of those, the highest ranked district spent $3,259 per student; the lowest ranked district spent $2,848 per student. The difference, $411, was less than 25 percent of the lowest per-pupil figure, namely $2,848. Hence, the officials found that New Mexico’s local aid program qualifies as a program that “equalizes expenditures.”176

170. See 20 U.S.C. § 7709(b) (2006). 171. Id. 172. Id. § 7709(b)(2)(A). 173. Id. § 7709(b)(2)(B)(i). 174. Zuni Pub. Sch. Dist. No. 89 v. Dep’t of Educ., 127 S. Ct. 1534, 1538−39 (2007). 175. Id. at 1538. 176. Id. at 1540 (quoting 20 U.S.C. § 7709(b)(1) (2000 .

177. Id. 178. Id. 179. Id. 180. Id. 181. Id. 182. Id. at 1537. 183. Id. at 1540. 184. Id. at 1541 (citation omitted). 185. Id.

186. Id. (quoting 20 U.S.C. § 7709(b)(1) (2000 . 187. Id. 188. Id. 189. Id. 190. Id. 191. Id. 192. Id. at 1541–42. 193. Id. at 1541. 194. Id. at 1543 (“[N]ormally neither the legislative history nor the reasonableness of the Secretary’s method would be determinative if the plain language of the statute unambiguously indicated that Congress sought to foreclose the Secretary’s interpretation.”). 195. Id.

196. Id. at 1543–44. 197. Id. at 1545. 198. Id. at 1545–46. 199. Id. at 1546. 200. Id. 201. Id. 202. Id. 203. Id. at 1549–50 (Stevens, J., concurring); id. at 1550–51 (Kennedy, J., concurring); id. at 1551–59 (Scalia, J., dissenting); id. at 1559 (Souter, J., dissenting). 204. Id. at 1549–50 (Stevens, J., concurring). 205. Id. at 1550. 206. Id. 207. Id.

208. Id. (quoting Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., , 843 n.9 (1984 . 209. Church of the Holy Trinity v. United States, 143 U.S. 457 (1892). 210. Zuni, 127 S. Ct. at 1550 (citing Church of the Holy Trinity, 143 U.S. at 459). 211. Id. at 1550–51 (Kennedy, J., concurring). 212. Id. at 1551. 213. Id. 214. Id. 215. Id. 216. Id. at 1551 (Scalia, J., dissenting). 217. Id. at 1553 (citation omitted). 218. Id. at 1554. 219. Id.

220. Id. (emphasis omitted) (quoting 20 U.S.C. § 7709(b)(2)(B)(i) (2000 . 221. Id. at 1554–55. 222. Id. at 1551. 223. Id. at 1556 (quoting id. at 1549 (Stevens, J., concurring . 224. Id. at 1556–59. Justice Souter felt similarly constrained by the strict meaning of the text and joined that part of Justice Scalia’s dissent, though not the broader defense of textualism. See id. at 1559 (Souter, J., dissenting) (agreeing with the Court that “Congress probably intended, or at least understood, that the Secretary would continue to follow the methodology devised prior to passage of the current statute in 1994” but finding the language “unambiguous and inapt to authorize that methodology”).

225. The parallels have some limits: Justice Breyer has stated that he would not defer to agency interpretations on “question[s] of national importance,” including those that concern the scope of a statute, because Congress (or a reasonable member of Congress) would not intend to delegate these issues. BREYER, supra note 31, at 107. He has also stated that reasonable legislators would decide how to allocate interpretive authority to “help[] the statute work better . . . in both the functional and the democratic sense of the term,” rather than for more strategic reasons. Id. at 108. 226. Thomas J. Miles & Cass R. Sunstein, Do Judges Make Regulatory Policy? An Empirical Investigation of Chevron, 73 U. CHI. L. REV. 823, 826 (2006) (reporting that Justice Breyer is “the most deferential justice in practice,” whereas Justice Scalia is the least deferential); see also Daniel J. Gifford, The Emerging Outlines of a Revised Chevron Doctrine: Congressional Intent, Judicial Judgment, and Administrative Autonomy, 59 ADMIN. L. REV. 783, 785 (2007) (observing that Justice Breyer has “long favored . . . . mandatory deference in the more routine or interstitial interpretations, but not necessarily in matters at the core of the statutory design”). 227. See BREYER, supra note 31, at 106 (arguing that courts should examine considerations beyond statutory language to determine what a “reasonable member of Congress” would intend when allocating interpretive authority); Stephen Breyer, Lecture, Our Democratic Constitution, 77 N.Y.U. L. REV. 245, 267 (2002) (arguing that judges should vary how much deference they show to an agency interpretation based on how much deference Congress wanted courts to show). In addition to teaching and writing in administrative law before becoming a judge, Justice Breyer served as special counsel of the U.S. Senate Committee on the Judiciary and as chief counsel of the committee. The Justices of the Supreme Court, http://www.supremecourtus. gov/about/biographiescurrent.pdf (last visited Dec. 3, 2008).

228. Gonzales v. Oregon, 126 S. Ct. 904, 911 (2006). 229. Dispensing of Controlled Substances to Assist Suicide, 66 Fed. Reg. 56,607, 56,608 (Nov. 9, 2001), invalidated by Gonzales, 126 S. Ct. at 925–26. 230. Gonzales, 126 S. Ct. at 913. 231. Dispensing of Controlled Substances to Assist Suicide, 66 Fed. Reg. at 56,608 (quoting 21 C.F.R. § 1306.04(a) (2001 ; see also Gonzales, 126 S. Ct. at 913–14 (quoting Attorney General Ashcroft’s ruling in Dispensing of Controlled Substances to Assist Suicide, 66 Fed. Reg. at 56,608). 232. Gonzales, 126 S. Ct. at 914 (quoting Dispensing of Controlled Substances to Assist Suicide, 66 Fed. Reg. at 56,608). 233. Id. at 916–22. 234. Id. at 910. 235. Id. at 916 (“All would agree, we should think, that the statutory phrase ‘legitimate medical purpose’ is a generality, susceptible to more precise definition and open to varying constructions, and thus ambiguous in the relevant sense.”). 236. Id. at 921. 237. Id. (quoting Whitman v. Am. Trucking Ass’ns, 531 U.S. 457, 468 (2001 .

238. Id. (citation omitted) (quoting Washington v. Glucksberg, 521 U.S. 702, 735 (1997 . 239. See id. at 918 (“[T]he Attorney General claims extraordinary authority.”). 240. Id. at 911. 241. Id. 242. See id. at 923 (“Oregon’s regime is an example of the state regulation of medical practice that the CSA presupposes.”). 243. Id. 244. Id. at 917. 245. See id. at 913 (noting that the attorney general did not consult the state or “anyone outside his Department”).

246. See id. at 922 (deciding only that the attorney general may not “issue the Interpretative Rule as a statement with the force of law”). 247. Id. at 921. 248. See id. (“[T]he authority claimed by the Attorney General is both beyond his expertise and incongruous with the statutory purpose and design.”). 249. Id. at 917 (quoting 21 U.S.C.A. § 821 (Supp. V 2005 . 250. Id. at 921. 251. Id. (quoting Martin v. Occupational Safety & Health Review Comm’n, 499 U.S. 144, 153 (1991 . 252. See id. (“The structure of the CSA . . . conveys unwillingness to cede medical judgments to an Executive official who lacks medical expertise.”). 253. See id. (“The importance of the issue of physician-assisted suicide . . . makes the oblique form of the claimed delegation all the more suspect.”).

254. See id. at 913 (noting that the attorney general issued the interpretative rule without consulting legislators). 255. See id. at 916 (finding that Auer and Chevron deference was unwarranted). 256. Id. at 922, 922–25 (applying Skidmore v. Swift & Co., 323 U.S. 134 (1944 . 257. Id. at 922. 258. Id. at 923. 259. Id. at 912, 924. 260. Id. at 925. 261. Id. 262. Id. at 931 (Scalia, J., dissenting). 263. Id. at 931, 935.

264. Id. at 931 (quoting Dispensing of Controlled Substances to Assist Suicide, 66 Fed. Reg. 56,607, 56,608 (Nov. 9, 2001 . 265. Id.

IV. OBJECTIONS AND RESPONSES

266. See, e.g., William N. Eskridge, Jr. & Philip P. Frickey, The Supreme Court, 1993 Term— Foreword: Law as Equilibrium, 108 HARV. L. REV. 26, 71–72 (1994) (“Because of their place in governance, agencies are both knowledgeable about and responsive to presidential and congressional preferences.”); Elizabeth V. Foote, Statutory Interpretation or Public Administration: How Chevron Misconceives the Function of Agencies and Why it Matters, 59 ADMIN. L. REV. 673, 693 (2007) (noting that agencies interpret statutes in light of many nontextual considerations, including “the goals of public administration”); Herz, supra note 52, at 194 (observing that agencies may have participated in drafting, have an “institutional memory,” and have more familiarity with statutory purposes); Jerry L. Mashaw, Agency Statutory Interpretation, ISSUES IN LEGAL SCHOLARSHIP, Nov. 2002, art. 9, at 8–9,

http://www.bepress.com/ils/iss3/art9 (noting that agencies consult legislative history and current political context, engaging in an interpretive process that is dynamic, because “[i]t is precisely their job as agents of past congresses and sitting politicians to synthesize the past with the present” (citing Ed Rubin, Dynamic Statutory Interpretation in the Administrative State, ISSUES IN LEGAL SCHOLARSHIP, Nov. 2002, art. 2, at 1, http://www.bepress.com/ils/iss3/art2 ; Trevor W. Morrison, Constitutional Avoidance in the Executive Branch, 106 COLUM. L. REV. 1189, 1240 (2006) (“Because they have ‘programmatic responsibility for implementing statutory regimes,’ and because they interact frequently with Congress in the course of discharging that responsibility, agencies often have a very nuanced sense of congressional aims and statutory purpose.” (footnotes omitted) (quoting Peter L. Strauss, When the Judge is Not the Primary Official with Responsibility to Read: Agency Interpretation and the Problem of Legislative History, 66 CHI.-KENT L. REV. 321, 321 (1990 ; Edward L. Rubin, Modern Statutes, Loose Canons, and the Limits of Practical Reason: A Response to Farber and Ross, 45 VAND. L. REV. 579, 586 (1992) (“[A] legislature and the administrative agencies within the same jurisdiction are linked by an incredibly dense network of relationships and shared activities.”). 267. See Jack M. Beermann, Congressional Administration, 43 SAN DIEGO L. REV. 61, 150 (2006) (“[A]gencies may be better than courts at discerning and applying congressional intent.”); cf. Cass R. Sunstein & Adrian Vermeule, Interpretation and Institutions, 101 MICH. L. REV. 885, 931 (2003) (advocating an institutional approach to Chevron that would envision agencies as choosing purposivism and courts as deferring to that judgment). 268. See Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., , 865 (1984) (“[T]he Administrator’s interpretation represents a reasonable accommodation of manifestly competing interests and is entitled to deference . . . .”). 269. See id. (“While agencies are not directly accountable to the people, the Chief Executive is . . . .”). 270. See Bressman, supra note 161, at 1804–13 (discussing political involvement in the administrative process).

271. See United States v. Mead Corp., , 230, 231–33 (2001) (“Congress contemplates administrative action . . . when it provides for a relatively formal administrative procedure tending to foster [] fairness and deliberation . . . .”). 272. See id. at 241, 245 (Scalia, J., dissenting) (“The principle effect [of the majority’s decision] will be protracted confusion.”). See generally Bressman, supra note 26 (suggesting confusion over Mead was worse than Justice Scalia had predicted).

273. Bressman, supra note 26, at 1444 (quoting Mead, 533 U.S. at 239, 261 (Scalia, J., dissenting . 274. Mead, 533 U.S. at 239 (Scalia, J., dissenting). 275. Id. (citations omitted). 276. See Bressman, supra note 26, at 1458–64 (“[T]he courts can be sorted into two groups: those that consider Mead-inspired factors and those that consider Barnhart-inspired factors.”). 277. See id. at 1459–61.

278. See id. at 1464–69 (describing how some courts “simply determine that lower-level Skidmore deference supports the agency’s interpretation”). 279. See Nat’l Cable & Telecomms. Ass’n v. Brand X Internet Servs., , 982–83 (2005). 280. See Bressman, supra note 26, at 1469–74 (observing that courts are confusing explicit and implicit delegation questions under Mead).

281. See Bressman, supra note 161, at 1788–96 (drawing a connection between administrative procedures and legislative monitoring in Mead). 282. United States v. Mead Corp., , 233 (2001). 283. See Bressman, supra note 161, at 1792; see also McCubbins et al., Administrative Procedures, supra note 10, at 257–59 (discussing how procedural requirements impact accessibility). 284. See Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 415, 419–20 (1971) (considering an interpretation under a statute requiring “a public hearing conducted by local officials for the purpose of informing the community about the proposed project and eliciting community views on the design and route,” followed by an administrative record containing an explanation of the decision). Interestingly, a harder case is a procedure that Mead approved for force-of-law purposes: formal adjudication. Formal adjudication does not permit legislative monitoring because neither constituents nor Congress may freely intervene. Yet it is possible to embrace Mead on this point. Congress, in designing the adjudicatory agencies, might have traded away the potential for legislative monitoring in favor of the sort of expertise that the adjudicatory model delivers—individualized consideration. Cf. Bawn, supra note 92, at 105 (“Members of Congress choose not whether to use statutory provisions to control agencies but how much control to build into agency procedures at the possible expense of other goals like technical expertise, due process, and optimal use of information.” (citations omitted . Or Congress may have intended that the adjudicatory agencies would, to the extent possible, resolve interpretive questions through notice-and-comment rulemaking in advance of formal adjudication. The major adjudicatory agencies like the National Labor Relations Board and the Federal Trade Commission possess notice-and-comment rulemaking authority. See M. Elizabeth Magill, Agency Choice of Policymaking Form, 71 U. CHI. L. REV. 1383, 1399 n.48 (2004). Those agencies have not always used their rulemaking power. See id. at 1399 & n.48 (“The NLRB and the FTC are known for their heavy reliance on adjudication . . . .”). Furthermore, Congress has seen fit to restrict their choice of procedures. But, as the Court has recognized, agencies cannot always foresee the need to issue interpretations in advance of adjudications. See SEC v. Chenery Corp., 332 U.S. 194, 203 (1947) (“[T]he choice made between proceeding by general rule or by individual, ad hoc litigation is one that lies primarily in the informed discretion of the administrative agency.”).

285. See Merrill & Hickman, supra note 51, at 845, 912–13 (arguing that the Court in Brown & Williamson addressed a question about the scope of the agency’s jurisdiction and suggesting that the Court address such questions in future cases at Chevron Step Zero).

286. Professor Adrian Vermeule has argued that courts should look for stripped-down plain meaning and, finding none, accord deference to the agency interpretation because courts gain little benefit from probing other interpretive sources, including most textualist sources. ADRIAN VERMEULE, JUDGING UNDER UNCERTAINTY: AN INSTITUTIONAL THEORY OF LEGAL INTERPRETATION 183–229 (2006); see also Gersen & Vermeule, supra note 14, at 688–708 (advocating an approach that casts “Chevron as a voting rule,” requiring consensus among the relevant decisionmakers as to the proper interpretation of a statute). My approach would not reintroduce the search for meaning even in limited form because it still asks a question that risks judicial misuse.

287. See Bressman, Schechter Poultry at the Millennium, supra note 29, at 1415–18 (collecting tools and principles that enable courts to address concerns about broad delegations as a matter of statutory construction rather than constitutional law); Cass R. Sunstein, Nondelegation Canons, 67 U. CHI. L. REV. 315, 338 (2000) (noting the judicial turn to nondelegation canons, which require a clear statement from Congress on policy issues). 288. See SCHOENBROD, supra note 72, at 10. 289. See id. at 14–15 (noting that broad delegation permits agencies to regulate in ways that restrict individual liberty without a sufficiently public purpose); Bressman, Schechter Poultry at the Millennium, supra note 29, at 1416 (noting that a lack of statutory standards permits arbitrariness). 290. See, e.g., Ernest A. Young, Two Cheers for Process Federalism, 46 VILL. L. REV. 1349, 1385 (2001) (stating that a presumption against preemption forces Congress to provide states with notice of when their interests are at stake, enabling them to fight for protection in the legislative process). But see Jack Goldsmith, Statutory Foreign Affairs Preemption, 2000 SUP. CT. REV. 175, 182–87 (arguing that any clear rule—including the opposite presumption in favor of preemption—would provide states with notice, and arguing more generally that a presumption against preemption cannot be justified on any of the asserted grounds). 291. See Bressman, Schechter Poultry at the Millennium, supra note 29, at 1415–18 (collecting interpretive principles for narrowing broad delegations); Sunstein, supra note 287, at 338 (identifying certain canons of construction as means for narrowing broad delegations). 292. See, e.g., Manning, Nondelegation Doctrine, supra note 5, at 702–25.

293. See id. at 690–95 (arguing that the ultimate concern about legislative history is legislative self-dealing); John F. Manning, Putting Legislative History to a Vote: A Response to Professor Siegel, 53 VAND. L. REV. 1529, 1529 (2000) (“If the judiciary accepts certain types of legislative history (committee reports and sponsors’ statements) as ‘authoritative’ evidence of legislative intent in cases of ambiguity, then the particular legislators who write that history (the committees and sponsors) effectively settle statutory meaning for Congress as a whole.”). 294. Antonin Scalia, Judicial Deference to Administrative Interpretations of Law, 1989 DUKE L.J. 511, 521. 295. See Elhauge, supra note 51, at 2051–55 (describing many canons as useful for eliciting congressional responses); Sunstein, supra note 287, at 338 (characterizing certain exceptions to Chevron as nondelegation canons that require a clear statement from Congress on policy issues); see also Williams N. Eskridge, Jr., Overriding Supreme Court Statutory Interpretation Decisions, 101 YALE L.J. 331, 388–89 (1991) (arguing that the rule of lenity helps to elicit congressional responses); Roderick M. Hills, Jr., Against Preemption: How Federalism Can Include the Legislative Process, 82 N.Y.U. L. REV. 1, 18–37 (2007) (arguing that a presumption against preemption would improve the legislative process in general because state laws often bring critical issues to national attention that Congress might otherwise be inclined to avoid); Dan M. Kahan, Lenity and Federal Common Law Crimes, 1994 SUP. CT. REV. 345, 346–47 (noting the argument that the rule of lenity serves a nondelegation function). 296. MCI Telecomms. Corp. v. AT&T Co., 512 U.S. 218 (1994). 297. Id. at 224 (quoting 47 U.S.C. § 203(b)(2) (2000 .

298. See id. at 231 (“It is highly unlikely that Congress would leave the determination of whether an industry will be entirely, or even substantially, rate-regulated to agency discretion— and even more unlikely that it would achieve that through such a subtle device as permission to ‘modify’ rate-filing requirements.”). 299. FDA v. Brown & Williamson Tobacco Corp., 529 U.S. 120 (2000). 300. See id. at 160 (“[W]e are confident that Congress could not have intended to delegate a decision of such economic and political significance to an agency in so cryptic a fashion.”). 301. Whitman v. Am. Trucking Ass’ns., 531 U.S. 457, 468 (2001). Some scholars have agreed that certain questions are too significant for Congress to delegate through ambiguity. See, e.g., Thomas W. Merrill, Rethinking Article I, Section 1: From Nondelegation to Exclusive Delegation, 104 COLUM. L. REV. 2097, 2173–75 (2004) (arguing that agency interpretations of the scope of their own regulatory power should be given Skidmore, rather than Chevron, deference). 302. See Catherine M. Sharkey, Preemption by Preamble: Federal Agencies and the Federalization of Tort Law, 56 DEPAUL L. REV. 227, 243 (2007) (“The regulatory preemption debate centers on the extent to which [agency interpretations] go beyond simply reciting the preemptive effect of the governing statute or regulation promulgated within the agency’s delegated authority, and instead attempt to discern the proper scope of preemption with little or no direction from Congress.”). 303. See, e.g., Medtronic, Inc. v. Lohr, 518 U.S. 470, 485 (1996) (“[W]e ‘start with the assumption that the historic police powers of the States were not to be superseded by the Federal Act unless that is the clear and manifest purpose of Congress.’” (quoting Hillsborough County v. Automated Med. Labs., Inc., 471 U.S. 707, 715 (1985 ; Nina A. Mendelson, Chevron and Preemption, 102 MICH. L. REV. 737, 738–40 (2004) (describing cases). 304. See Medtronic, 518 U.S. at 485.

305. Solid Waste Agency of N. Cook County v. U.S. Army Corps of Eng’rs, 531 U.S. 159 (2001). 306. Id. at 172–73. For discussion and examples of the doctrine of avoidance, see Adrian Vermeule, Savings Constructions, 85 GEO. L.J. 1945, 1948–49 (1997). 307. See, e.g., Nat’l Cable & Telecomms. Ass’n v. Brand X Internet Servs., , 985 (2005) (suggesting that a court might employ the rule of lenity to declare a statute unambiguous, thereby precluding deference to the agency). 308. See Rachel E. Barkow, Originalists, Politics, and Criminal Law on the Rehnquist Court, 74 GEO. WASH. L. REV. 1043, 1067 (2006) (characterizing the rule of lenity as a “common-law tradition”); Hickman, supra note 131, at 935 (“Lenity’s status as an absolute constitutional requirement rather than a quasi-constitutional canon of construction is questionable.”); Kahan, supra note 295, at 346–47 (describing the rule of lenity as a “quasi-constitutional” doctrine). 309. See Kahan, supra note 295, at 346–47 (making these observations about the rule of lenity). 310. Many scholars do not rely on the association with legislative intent in defending the nondelegation canons. See, e.g., Eskridge, supra note 27, at 2052–53 (noting the importance of

“constitutional guideposts,” including substantive and procedural fairness, when judges interpret criminal statutes); Ernest A. Young, Making Federalism Doctrine: Fidelity, Institutional Competence, and Compensating Adjustments, 46 WM. & MARY L. REV. 1733, 1848– 49 (2005) (defending a presumption against preemption as “a normative rule of construction, which means it cannot be grounded in some descriptive judgment about Congress’s intent in enacting the relevant statute”). 311. Congress may not even be able to distinguish between extraordinary and routine questions. See Cass R. Sunstein, Beyond Marbury: The Executive’s Power to Say What the Law Is, 115 YALE L.J. 2580, 2606 (2006) (“[N]o simple line separates minor or interstitial from major questions.”); id. at 2604 (noting that the line between jurisdictional questions and others is “far from clear”). 312. See MCI Telecomms. Corp. v. AT&T Co., 512 U.S. 218, 220–21 (1994) (cataloguing the development of FCC regulations restricting the monopoly power of dominant carriers). 313. See 21 U.S.C. § 371(e) (2006) (authorizing the FDA specifically to conduct notice-andcomment rulemaking); MCI Telecomms. Corp., 512 U.S. at 220–37 (acknowledging the authority of the FCC to conduct notice-and-comment rulemaking). 314. See FDA v. Brown & Williamson Tobacco Co., 529 U.S. 120, 137–38, 144–61 (2000) (consulting later-enacted statutes to confirm interpretation of a prior statute). Such evidence is arguably relevant to the reasonableness of the agency’s interpretation. See Lisa Schultz Bressman, Deference and Democracy, 75 GEO. WASH. L. REV. 761, 779–80 (2007); Elhauge, supra note 51, at 2148 (arguing that agency interpretations are not reasonable if they conflict with current enactable congressional preferences).

315. 12 U.S.C. § 371(a) (2006). 316. Id. § 24. 317. Ernest T. Patrikis & Glen R. Cuccinello, Supreme Court Extends Federal Preemption to National Bank Operating Subsidiaries, 124 BANKING L.J. 512, 515 (2007). 318. 12 U.S.C. § 484(a) (2006). 319. 12 C.F.R. § 7.4006 (2006). 320. See 12 C.F.R. § 5.34(e)(1) (2006); 31 Fed. Reg. 11,459, 11,459–60 (Aug. 31, 1966). 321. See Watters v. Wachovia Bank, N.A., 127 S. Ct. 1559, 1572–73 (2007). 322. Watters v. Wachovia Bank, N.A., 127 S. Ct. 1559 (2007). 323. See id. at 1572. 324. See id.

325. Id. at 1570. 326. See Energy Independence and Security Act of 2007, Pub. L. No. 110-140, § 102(a)(2), 121 Stat. 1492, 1499. 327. See John M. Broder & Micheline Maynard, Deal in Congress on Plan to Raise Fuel Efficiency, N.Y. TIMES, Dec. 1, 2007, at A1. 328. See id. 329. See id. 330. See John M. Broder & Felicity Barringer, E.P.A. Says 17 States Can’t Set Greenhouse Gas Rules for Cars, N.Y. TIMES, Dec. 20, 2007, at A1 (noting that, on the heels of the new fuel economy legislation, the EPA blocked California and other states from imposing their own standards and that the states plan to file a federal lawsuit to reverse that decision). 331. See, e.g., Rust v. Sullivan, 500 U.S. 173, 190–91 (1991) (refusing to invalidate an agency interpretation—the so-called abortion “gag rule”—even though it raised a possible First Amendment question); Mendelson, supra note 303, at 740 (noting inconsistency in the preemption context); Caleb Nelson, Preemption, 86 VA. L. REV. 225, 232 (2000) (“Most commentators who write about preemption agree on at least one thing: Modern preemption jurisprudence is a muddle.”); Sharkey, supra note 302, at 243 (stating that Congress has provided “little or no direction” in the preemption debate).

332. See Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 41 (1983) (requiring the agency to consider the relevant factors and remanding a rule that failed to do so). 333. See Bressman, Schechter Poultry at the Millennium, supra note 29, at 1415–31 (arguing that courts might withhold deference from agency interpretations that fail to articulate limiting standards); see also Stack, supra note 159, at 958 (arguing that the reasoned decisionmaking requirement may serve nondelegation norms). 334. See Merrill & Hickman, supra note 51, at 915 (arguing that the avoidance of questions canon expands the Court’s authority beyond its constitutional limits and quoting Judge Posner, who says that the canon is therefore likely to “sharpen the tensions between the legislative and judicial branches” (quoting RICHARD A. POSNER, THE FEDERAL COURTS: CRISIS AND REFORM 285 (1985 ; Frederick Schauer, Ashwander Revisited, 1995 SUP. CT. REV. 71, 90–97 (advocating abandoning the avoidance of questions canon).

335. See Note, New Evidence on the Presumption Against Preemption: An Empirical Study of Congressional Responses to Supreme Court Preemption Decisions, 120 HARV. L. REV. 1604, 1605 (2007) (examining congressional responses to Supreme Court preemption decisions during the 1993 through 2003 Terms and concluding that “Congress almost never responds to the Court’s preemption decisions, so mistaken interpretations for or against preemption are unlikely to be corrected”). 336. See id. 337. See Lisa Schultz Bressman, Disciplining Delegation After Whitman v. American Trucking Ass’ns, 87 CORNELL L. REV. 452, 469–81 (2002) (describing the Court’s angry response to the efforts of a D.C. Circuit judge to alter the nondelegation doctrine). 338. See supra text accompanying notes 268–69. 339. See supra text accompanying notes 270–71. 340. See supra text accompanying note 72.

CONCLUSION