Statutory interpretation from the inside - an empirical study of congressional drafting, delegation, and the canons.
INTRODUCTION I. THE STUDY AND WHY IT MATTERS A. Faithful Agency, Fictions, and Empirics in Statutory Interpretation 1. Fictions 2. Previous empirical work B. Methodology: Our Survey of 137 Congressional Staffers II. CONGRESS AND THE CANONS A. Overview of the Findings: A Spectrum of Canon Knowledge and Use B. Textual Canons: More Familiarity by Concept than by Name 1. Concepts in use: expressio, noscitur, and ejusdem 2. Canons known, but rejected: superfluities, consistent usage, and dictionaries a. Superfluities: redundancy to satisfy political stakeholders b. Committee jurisdiction and "unorthodox lawmaking" as barriers to the whole act and whole code rules c. "No one uses a freaking dictionary" C. Substantive Canons 1. Federalism, preemption, and clear statement rules a. A partial courts-Congress feedback loop for federalism and preemption b. The irrelevance of clear statement rules 2. Lenity unknown by name 3. Constitutional avoidance unknown but assumed D. Do the Data Matter? Linking the Findings with the Normative Justifications for the Canons 1. Textual rules as approximation canons and rejected canons a. Expressio, noscitur, and ejusdem as approximation canons b. Superfluities, consistent usage, and dictionaries as rejected canons 2. Substantive canons: disconnected, approximation, and feedback canons a. Different types of disconnected rules: lenity and clear statement rules b. Avoidance as approximation c. Federalism, preemption, and a "due process of lawmaking" feedback loop E. Alternative "Rule of Law" Justifications: Coherence, Coordination, and "Interpretive Activism" III. LEGISLATIVE HISTORY A. Legislative History-Specific Delegation Concerns About Staff and Committees Appear Unfounded B. A Textualist Approach Is Not Likely to Diminish the Production or Importance of Legislative History Because Legislative History Plays Many Other Roles 1. Legislative history as a tool for congressional oversight of agencies 2. Legislative history as intracongressional communication 3. Legislative history as political communication with the public 4. Legislative history as a vehicle for details that are inappropriate for statutory text C. Legislative History as Evidence of Congressional Intent 1. The centrality of committee-produced legislative history 2. Unorthodox lawmaking's relevance: distinguishing party leader, omnibus, and appropriations legislative history 3. Other factors: staffer involvement, timing, opposition, and centrality to the bargain a. Staff drafts everything but is not unaccountable b. Timing affects the reliability of legislative history c. Additional considerations volunteered by our respondents: personal and reputational factors, colloquies, and markups D. Conclusion: Smarter Judicial Use of Legislative History IV. THE ADMINISTRATIVE LAW CANONS A. Chevron and the Presumption of Delegation 1. Chevron is a feedback canon 2. Chevron is not a reason for ambiguity B. Mead and Other Signals of Delegation as Reasonable Approximations 1. Barnhart, agency participation in drafting, and divided government: substantiated but with qualifications 2. Different subject matters also validated with nuances 3. Major questions, preemption questions, and the obligation not to punt C. More than One Federal Agency and State Implementers of Federal Law 1. Multiple federal agencies and new linguistic conventions of delegation 2. Delegation to state agencies implementing federal statutes D. Theoretical and Doctrinal Interventions 1. Chevron Step One 2. The obligation not to punt difficult questions in broader context CONCLUSION: COMPARING THE CANONS, THE ENDURING ALLURE OF FAITHFUL AGENCY, AND CONGRESS AS "FAITHFUL PRINCIPAL" A. Comparing Canons B. The Allure of Faithful Agency and Judicial Reluctance to "Make Law". C. Canons as Collective Knowledge, the Shifting Effect of Legal Education, and the Potential for a Dynamic Interpretive Regime D. A Normative Framework for Congress's Side of the Relationship: Congress as "Faithful Principal"? METHODS APPENDIX http://www.stanfordlawreview.org/sites/default/files/Gluck_Bressman Stan._L._Rev._Methods_Appendix.pdf.
Judge Jones, a strict textualist, is interpreting an appropriations statute. He knows that, unlike other statutes, appropriations bills place most key directives in the legislative history rather than in the enacted text. Should the judge depart from his normal practice and consider legislative history?
Judge Smith is interpreting the term "work," which appears several times throughout a statute that she has learned was drafted in different parts by seven different congressional committees. Should the judge apply the usual "whole act rule" of interpretation, which presumes that words are used consistently throughout statutes?
Judge Jacobs is reviewing Agency A's interpretation of an ambiguous statute. He has information that insiders tell him counsels against deferring to A's interpretation, namely, that the Secretary of A has a bad reputation inside of Congress. Should the judge take this factor into account?
What role should the realities of the legislative drafting process play in the theories and doctrines of statutory interpretation and administrative law? Although the past several decades have seen exhaustive debates about how courts and agencies should interpret federal statutes, almost no empirical work has been done to shed light on the relationship, if any, between the theories and doctrines of the fields and the actual statute-creating process.
From a theoretical perspective, the relevance of the realities of legislative drafting depends entirely on the answer to another fundamental question: namely, what is, or should be, the objective of the so-called "canons" of statutory interpretation, the default presumptions that judges apply to interpret ambiguous statutes? There are many possible normative frameworks judges could use to answer this question. Judges might believe that the canons reflect how Congress actually drafts, and therefore that applying them effectuates legislative supremacy. Or judges might use the canons for more dialogical reasons, such as to encourage Congress to draft more precisely or in other ways that judges think would be preferable. Or the canons might be understood to effectuate judicial responsibilities that are essentially external to the legislative process--such as advancing constitutional values or furthering the "rule of law" by coordinating systemic behavior or imposing coherence on the U.S. Code.
Deciding which (or how many) of these objectives should be the goal is a foundational inquiry that goes to the nature of the courts-Congress relationship and the scope of the judicial power. But there has been some profound imprecision with respect to how this inquiry has been addressed. Most practicing judges claim allegiance to an exceedingly general model of the judge as a "faithful agent" of the legislature, and that model has been deployed to justify an enormous number of canons that seem to be doing very different types of work. There is arguably a major difference, for instance, between a theory of the judicial role in statutory interpretation that grounds its legitimacy in whether it is accurately reflective of congressional practice and one that, instead, aims to change how Congress itself deliberates and drafts. And there is perhaps an even greater difference between those visions and one grounded in the view that judges have an obligation to impose coherence on the U.S. Code, even where imposing such coherence achieves results never intended by its drafters. Allegiance to the faithful-agent model also often translates to claims that interpretive methods reflect actual congressional practice---claims at odds with the admission by most judges and scholars that many of the canons on which they rely are "fictions." (1)
This Article offers the most extensive empirical study to date about this intersection of statutory interpretation, administrative law doctrine, and the process of legislative drafting. Over five months in 2011 and 2012, we interviewed 137 congressional counsels with responsibilities over drafting legislation. (2) We surveyed counsels, rather than elected members of Congress, for several reasons, which we elaborate in Part I. Most importantly, current doctrine makes assumptions about what legislative drafters know, and it is widely acknowledged (and our study confirms) that members do not do the actual drafting. Interpretive doctrines designed to reflect how members actually participate in the drafting process would look very different, and certainly less text oriented, than the ones that we currently have. Moreover, doctrine rarely grapples with the role of staff, and judges often make assumptions about staff accountability to members in the drafting process that have never been empirically verified.
Our respondent-counsels were approximately equally divided between the House and the Senate, both political parties, and whether they worked for members in the majority or the minority in each legislative body. (3) They worked on twenty-six different committees, as well in as the professional drafting offices known as the Offices of House and Senate Legislative Counsel. Every survey consisted of the same 171 questions, (4) which covered topics ranging from the role of canons such as the presumption against preemption, expressio unius, and Chevron (5) deference, to legislative history, the legislative process, and the way that staffers perceive the responsibilities of courts and agencies in statutory interpretation. In addition, our survey provided unlimited opportunities for qualitative explanations. Our respondents used those opportunities not only to provide more texture to their responses, but also to highlight important influences in the drafting process not captured by our questions or legal doctrine.
Our findings shed light on some of the key debates of both fields. They also allow us to categorize the canons in ways that reveal many still-unanswered questions about the normative frameworks that underlie them. Contrary to the prevailing wisdom, a majority of our respondents were not only aware of some of the interpretive rules that courts employ--including the presumption against preemption and Chevron--but told us that these legal rules affect how they draft, although not always in ways that courts expect. We call these rules "feedback canons," as they at least partially substantiate the existence of an interpretive conversation between the Supreme Court and Congress that many have assumed impossible. For other canons, such as many textual canons like noscitur a sociis and a surprising number of the administrative delegation doctrines--including Mead, (6) Barnhart, (7) and the major questions doctrine--our respondents displayed unfamiliarity with them as legal doctrines but told us that the assumptions underlying those rules accurately reflect how they draft legislation. We call these rules "approximation canons"--rules in which the Court seems to be correctly intuiting how Congress signals its intent even as Congress remains unaware of the rules' existence.
At the same time, however, there were a host of canons that our respondents told us that they do not use, either because they were unaware that the courts relied on them or despite known judicial reliance. For example, our respondents were mostly unaware of and do not use "clear statement rules"--an example of a rule that we therefore call a "disconnected canon." And although they were well aware of other rules, including the rule against superfluities, the Court's penchant for dictionary consultation, and some Justices' distaste for legislative history, our respondents told us that they nevertheless do not generally draft in accordance with the rule against superfluities, that they do not consult dictionaries when drafting, and that legislative history remains a critical tool regardless of whether courts use it. Indeed, despite the decades of judicial squabbling over it, legislative history was overwhelmingly viewed by our Democratic and Republican respondents alike as the most important tool of interpretation after statutory text. We call this last set of rules, collectively, "rejected canons," because our drafters knowingly reject judicial preferences relating to their application in favor of institutional or other pragmatic considerations.
Our aim in thus disaggregating and typologizing the canons is not to say that certain rules are necessarily illegitimate. Rather, our aim is to illustrate how undertheorized the canons have been and to highlight the kinds of normative questions that arise from testing the connection between legal doctrine and legislative drafting practice. What model of the judicial role justifies the use of canons that legislative drafters know but consciously do not employ? Is this the same model that justifies the use of canons that depend on an interbranch interpretive feedback loop? The faithful-agent model has had remarkable staying power as the "umbrella" justificatory model of most interpretive approaches, even though it offers little specific assistance in answering questions at this level of detail. Indeed, in light of our findings, the faithful-agent model seems incapable of bearing the full weight of modern interpretive practice.
Our interviews also elicited a treasure trove of information about key influences on the actual drafting process that courts and scholars rarely, if ever, consider. For example, our drafters highlighted the importance of congressional-committee jurisdiction; the type of statute being interpreted (e.g., single-subject versus omnibus legislation); the specific path that legislation takes through Congress; the personal reputation of the relevant drafter or agency; the various audiences for and the type and timing of legislative history; and the centrality of the nonpartisan Offices of Legislative Counsel in drafting statutory text--all as critical to understanding how statutes should be interpreted.
We detail these findings, and many more, in this Article and its companion piece, which will appear in the following volume of the Stanford Law Review. (8) But by way of making the stakes clear, the findings have potential relevance for virtually all of the major interpretive debates, both at the canon-specific level and also more broadly at the theoretical level. At the canon-specific level, for example, understanding that statutes are drafted by congressional committees that generally do not communicate with one another pulls the rug out from under the bases of many interpretive rules, beloved by textualists, that presume that statutory terms are used consistently within and across statutes. Understanding that legislative history plays an entirely different role in ordinary statutes than it does in omnibus or appropriations statutes arguably should affect how it is used by courts, but even purposivist proponents of legislative history do not make such distinctions. Realizing that Congress uses certain signaling conventions--for example, the words "in consultation with" to indicate that it wishes one agency to take the lead in a multiagency statute--might resolve continuing interpretive disputes.
At the broader level, for instance, our findings have direct relevance for ongoing debates about the administrative law doctrines, which many have charged are too disconnected from congressional practice to be legitimate; according to our respondents, the Court has actually done a surprisingly good job at approximating how Congress delegates. So, too, understanding that members and staff focus more on policy, while the nonpartisan, professional Offices of Legislative Counsel draft much of the actual statutory text, should change entirely the contours of the debate over legislative history. As it turns out, much enacted statutory text is not drafted by the staff most accountable to the members, but legislative history is.
Our findings have relevance not only for faithful-agent-based theories of interpretation, but also for theories that additionally rely on rule of law arguments, such as the idea that judges should interpret statutes in ways that are predictable for systemic actors or in ways that impose coherence on the corpus juris. Few practicing judges justify their interpretive approach solely on the basis of such rule of law arguments--most claim that the rules that they apply capture intended or ordinary meaning and so also are consistent with the faithful-agent paradigm. Our study calls those faithful-agent justifications into question for many canons and so raises the question whether the rule of law model, alone, can justify the continuing application of those doctrines. For example, if coherence or predictability were really the goal, one might expect interpretive doctrine to be applied much more consistently than it has been and in a less complex manner. Moreover, without a link to congressional practice, it becomes clear that these role of law canons allow judges to shape statutes in ways that may diverge from congressional expectations as much as do more openly pragmatic approaches to statutory interpretation that are more frequently attacked as improper exercises of judicial activism. The pervasive modern discomfort with federal judicial "lawmaking," we believe, has led judges to take shelter behind seemingly neutral interpretive roles whose use is bolstered by the assumption that they also reflect how Congress works or understands statutory language. To the extent that our study undermines that empirical claim for some roles, those roles are not necessarily illegitimate, but should be acknowledged and assessed for what they are.
Not all of the findings that we relay are amenable to incorporation by legal doctrine. Some of the key factors that our respondents consider--for example, the personal reputation or sophistication of the staffer responsible for a piece of legislation--seem impenetrable by courts. Even those factors that courts may be able to discern themselves--for instance, whether multiple committees participated in drafting a single piece of legislation or the timing of legislative history--may prove too costly for use in everyday legal practice. We also recognize that building a typology of the canons around their link to the legislative drafting process is just one of many possible organizing frameworks that might be employed. (9) We focus on this organizing principle because drafting "reality" has been a central component of virtually all of the theoretical and doctrinal debates, but another unanswered question is whether a more tailored set of interpretive rules is desirable in the first place and, if so, tailored along what dimension. Yet another question is whether a partially tailored set of legal doctrines is better than none at all.
To date, there has been almost no other empirical research of this kind. Only one previous article, an important 2002 study by Victoria Nourse and Jane Schacter, began to make inquiries of the nature that this Article undertakes, but was admittedly limited in its sample and its methodology. (10) Despite those limitations, that study has had major influence on legal scholarship, (11) offering proof of the hunger for empirical data about legislative drafting and also the difficulty of ignoring such data once they are discovered.
In this publication, we offer the first half of our results, focusing on our respondents' awareness and use of the interpretive rules that courts routinely employ. In the companion Article, we elaborate on the various influences on the drafting process that our respondents told us are central but are rarely considered by courts. We also relay some rather startling findings about our respondents' view of courts as interpreters. We were not surprised that our respondents saw agencies as the primary statutory interpreters, but we were surprised to learn that our respondents generally did not view courts as delegates or as welcome "partners" in their work, a finding that may pose challenges for broader theories of interpretation that advocate a more engaged judicial role. The second Article also considers more specifically how interpretive theory and doctrine might change in light of our findings.
The discussion of this first Article proceeds in four Parts. Part I provides the theoretical background and an explanation of our methodology. Part II relays our findings concerning our respondents' awareness, use, and perceptions of the textual and substantive canons, while Part III relays those findings with respect to legislative history and Part IV does so with respect to the administrative law doctrines. We conclude by comparing our findings across canons. Interestingly, the canons that have provoked the most scholarly and judicial debate--legislative history and the administrative law doctrines--find the greatest support in our respondents' drafting practices, while less controversial canons, including many textual rules and clear statement rules, were more disconnected from how our respondents described the statute-creating process. The Conclusion also presses the question of the utility of the faithful-agent model and calls attention to the lack of a parallel model on the congressional side. Although judges and scholars often refer to the courts-Congress interpretive dialogue in statutory interpretation, little has been said about what normative model should drive congressional behavior in that relationship. Does Congress, for example, have an obligation to pay more attention to, or try to affect, how courts interpret statutes? Is Congress, or should it be, a "faithful principal"? (12)
Ultimately, we believe that our findings, even if only suggestive, demonstrate the need for a more nuanced account of how statutes are produced in the modern regulatory state. The foundational scholarship of federal legislation has, for the most part, been based on a generic and stylized account of statutory drafting--an understandable focus for a field that is still in its relative infancy. However, there is great variety that exists across drafters, types of statutes, the reasons why and ways in which Congress delegates, and countless other aspects of the drafting process. A mature theoretical account will have to contend with that variety or else come up with better justifications for ignoring it. In this first outing, we can only begin to develop these ideas, but we hope that this Article will encourage more work in a similar vein.
I. THE STUDY AND WHY IT MATTERS
What is the relevance of a study of this nature to the theories and doctrines of statutory interpretation and administrative law? Should we care whether Congress is aware of the interpretive rules that courts employ or whether congressional drafting practice corresponds to the assumptions that legal doctrine makes?
The answer to this inquiry implicates some of the most important, and still unresolved, questions of both fields. It turns on the reasons that courts look to those interpretive tools in the first place, a question inextricably tied to one's views about the proper role of judges, the judicial power, and the courts-Congress relationship. Doctrine and theory have remained surprisingly vague about how exactly the various rules of interpretation effectuate different normative visions of the judicial role, even as judges and scholars continue to contest which of the rules are appropriate.
Some interpretive tools seem designed merely to reflect how Congress actually drafts; others seemed more proactively aimed at affecting how Congress should draft in the future. Still others do not seem related to Congress at all, but rather enable judges to layer policy preferences, or constitutional law norms, atop Congress's work product. A study of drafting "reality" has obvious significance for evaluating canons that are intended to reflect or affect Congress. It also exposes more clearly those canons that are unrelated to drafting practice and that, as a result, may require more explicit justifications on other grounds.
Examining the relationship between Congress and the canons also raises questions about courts-Congress communication. Many scholars and judges have argued that judicial interpretive practice has a salutary, "teaching" effect on legislative drafting, (13) or that consistently applied interpretive rules help courts and legislators coordinate their behavior. But those types of arguments depend entirely on whether Congress is aware of the rules that courts employ. New empirical work, moreover, shows that Congress rarely overrides judicial statutory decisions, (14) a finding that heightens the importance of an interbranch interpretive dialogue from a democracy-theory perspective: if ex post overrides are rare, the ex ante drafting process becomes Congress's central means of communicating with courts and shaping their interpretive behavior.
Profound disagreements exist with respect to most of these framework questions about statutory interpretation, and it is not the goal of this Article to resolve them. Our goal, rather, is to illustrate that--no matter where one comes down on these questions--investigating the realities of the congressional drafting process advances the debates. Our study permits us to intervene in rule-specific battles; for instance, the battle over whether judicial reliance on legislative history affects congressional production of it, or whether Congress really uses ambiguity to signal delegation of interpretive authority to administrative agencies. Our study also sheds light on larger jurisprudential divides, such as the question of which branch of government--judicial, legislative, or both--has power over interpretive rules. (15) The case for congressional power to change interpretive rules, for instance, might be more compelling for rules whose justifications turn on legislative reality than for more disconnected rules that courts employ for other purposes, such as coordinating judicial behavior or enforcing constitutional norms. Our efforts to disaggregate the various canons, their normative bases, and their connections to actual drafting practice help to reveal such distinctions.
The remainder of this Part offers a brief sketch of the relevant theoretical and empirical landscape. It concludes with an outline of our study's methodology.
A. Faithful Agency, Fictions, and Empirics in Statutory Interpretation
One need look no further than the furor over Justice Scalia and Bryan Garner's new book for proof that intense discord remains over the proper role of judges in statutory cases and which tools of interpretation support that role. (16) This is so despite the fact that the two leading theories, purposivism and textualism, both claim consistency with a "faithful-agent" vision of the judicial role. But for each set of interpreters, the faithful-agent concept provides an extremely broad umbrella for the application of many different kinds of interpretive rules.
Purposivists, for instance, make faithful-agent-based arguments that judicial reliance on legislative history helps to cabin judicial discretion and effectuate congressional intent, (17) but also argue, somewhat in conflict, that judges are legislative partners who should interpret statutes "in a manner that.., will produce a workable set of laws," (18) even if the outcome reached is not one specifically intended by Congress. (19) Textualists, in turn, argue that their version of faithful agency hews more closely to legislative supremacy and that "texts should be taken at face value," (20) but at the same time argue that judges should adopt particular linguistic conventions that, even if they do not reflect how Congress drafts, will either teach Congress how to draft better in the future or make the law more predictable for those outside of Congress. (21) Both sets of theorists also routinely apply many policy presumptions, such as the presumption that ambiguous statutes will not be construed to preempt state law, without consistently justifying them as reflective of congressional intent, common sense, actual drafting practice, constitutional values, judicial policy preferences, or something else. (22) Underpinning these interpretive approaches are obviously varied visions of how reflective or how proactive courts should be.
Some pragmatic theorists have argued explicitly for interpretive methodologies less tethered to congressional reality. Cass Sunstein and Adrian Vermeule, for example, focus on the limited institutional capacity of courts in arguing that faithful agency is best advanced through "non-ideal interpretive theory" (23) that relies on simple decision rules and/or transfers of decisionmaking authority to other institutions, such as agencies. (24) Still other theorists--although a minority, if any, among practicing judges--have moved away from faithful agency altogether and focus on matters such as the importance of the rules as coordinating devices for courts (25) or the duty of courts to impose coherence on the statutory landscape, even where such coherence is not intended by Congress. (26)
Of particular relevance to this Article, all of these theories make different assumptions about the realities of the legislative process. Textualists have argued, for instance, that most members of Congress do not read legislative history. (27) Purposivists, in response, claim that their understanding of statute-making is more realistic and that Congress welcomes the kind of judicial assistance that their methodology offers. Pragmatic theorists, in turn, often assume the existence of an interbranch dialogue--a "feedback loop" between the courts and Congress that puts interpreters and drafters on the same page with respect to the interpretive conventions that both will follow. Many others, however, have argued that such a dialogue simply does not exist. (28)
Somewhat paradoxically, interpreters of all stripes also have acknowledged that many of the rules that they routinely apply rest on "fictions." (29) Such admissions pose particular difficulties for those faithful-agent theorists who claim that their interpretive rules reflect how Congress actually drafts. (30)
For instance, the fiction of the unitary drafter--the idea that all laws are drafted by the same group of legislators--undergirds a huge number of interpretive rules applied by textualists and purposivists alike. But this principle, as even the Justices who use it admit, (31) is most certainly false. So, too, is the notion of a single "congressional intent," although purposivists continue to assert that such a fiction is useful nonetheless. (32) And, in the context of congressional delegations to agencies, judges on all sides have recognized the fictitious nature of using statutory ambiguity as a signal of congressional intent to delegate. (33) Justice Scalia views this fiction as a useful "background rule of law against which Congress can legislate," (34) and Justice Breyer believes it is a rule that Congress likely intends because it facilitates "a workable partnership" (35) with the courts. Each, of course, assumes something nonfictitious about which we, to date, have no proof. Justice Scalia's view assumes that Congress is aware of this allegedly fictitious rule and accordingly legislates in its shadow. Justice Breyer's approach, meanwhile, assumes both that Congress seeks such a partnership with the courts and that the rule effectuates that relationship.
A threshold question for any empirical study of Congress is why interpreters treat rules that they believe to be fictions as benign ones. Perhaps some rely on them as proxies for data they do not (but wish to) have about how Congress works. For those theorists, empirical research might affect whether those doctrines are utilized in the future. But others might rely on the fictions despite knowing how Congress works--perhaps because they assume that the legislative process is too complex ever to be captured by legal rules or because they wish to enforce external norms. For those theorists, the real-world data might force a more explicit recognition of the extralegislative nature of their approach and its implications for the judicial power and the interbranch relationship.
2. Previous empirical work
There has been only one empirical study that attempted to link the doctrines of statutory interpretation to the actual legislative drafting process. That work--the important case study by Nourse and Schacter (36)--has had much influence despite its narrow scope. (37) As the authors themselves acknowledged, the study's small sample size (eighteen counsels) and its confinement to the rather atypical Senate Judiciary Committee (38) limited the generalizability of its results. (39)
The Nourse and Schacter study also asked only several broad questions and, because it used open-ended interview questions rather than a standardized format, relayed its findings anecdotally rather than quantitatively. (40) Specifically, the study produced four principal findings: (1) that staff, and to a lesser extent, lobbyists, have central roles in the statutory drafting process and that legislation is drafted in a variety of ways, ranging from a deliberate multi-staffer mode of drafting to last-minute drafting on the Senate floor; (2) that although staffers value legislative clarity, the pressures of time and compromise make ambiguity inevitable; (3) that drafters do little legal research about interpretive tools; and (4) that legislative history continues to be produced despite the textualist critique of it. (41)
With some exceptions, as elaborated in the Parts that follow, we generally do not quarrel with these findings, and our study replicates most of them. Our interest is in deepening the Nourse and Schacter account and going beyond it. We inquired about more than twenty different interpretive doctrines and the different types of legislative history. We also investigated more specifically the ways in which and reasons why Congress delegates, and examined how the particulars of the legislative process--the path that statutes take from introduction to enactment--affect how statutes are drafted.
Part of our motivation stems from how much the scholarship has relied on the Nourse and Schacter study. For example, their argument that interpretive rules play little role in the drafting process has been cited as fatal to hopes of any interpretive dialogue between courts and Congress--a dialogue upon which much interpretive theory and doctrine depends. (42) The Supreme Court, for instance, has often said that the canons are background rules against which Congress presumptively legislates; (43) textualists argue that a text-centric approach will spur Congress to draft statutes more carefully; (44) and rule of law theorists claim that that judicial consistency will have coordination benefits for courts, litigants, and statutory drafters alike. (45) All of these arguments depend on some congressional awareness and responsiveness to the rules that courts employ.
At the same time, the Nourse and Schacter findings have not been cited by any federal court. Nor, to our surprise, has some important political science literature about congressional drafting. Although certain strands of the political science literature have had a major impact on interpretive doctrine, (46) courts largely seem to have overlooked other important work on how the congressional committee system affects how statutes are drafted (47) and how the "textbook" legislative process no longer exists. (48) On the administrative law side, there has been more empirical work, mostly focusing on the structure of delegation (49) and how consistently the Court employs its deference doctrines. (50) Apart from Nourse and Schacter, however, no one has addressed the precise question of whether Congress thinks about the delegation of interpretive authority in the same way that the Court does.
A variety of scholars have noted the distinct challenges of using empirics to evaluate and change statutory interpretation doctrine. (51) But even those scholars have recognized that "there is surely a nontrivial range of questions amenable to research, and having some data is usually better than having no data." (52) Judge Robert Katzmann (himself a political scientist) recently lamented that the ongoing interpretive debates have "taken place in a vacuum, largely removed from the reality of how Congress actually functions." (53)
The primary caution has been against assuming that empiricism can answer all questions in the field. We certainly agree and have tried to be clear about the linkages between the relevance of the data that we have collected and normative theories of the relationship between courts and Congress. We also fully acknowledge that there are countless other influences on the drafting process that our survey does not address. But given that the field has placed much reliance on the limited data that do exist, there is value in attempting to expand upon it.
B. Methodology: Our Survey of 137 Congressional Staffers
As elaborated in the accompanying Methods Appendix, (54) our study builds on what the Nourse and Schacter project began in several ways. First, we have engaged in a broader study, with quantifiable results. To that end, we asked 171 questions (55) of 137 staffers in Congress, and respondents were given the opportunity to offer additional qualitative comments to explain their answer to any question. Respondents were also asked at the end of most sections whether "we were asking the right questions" about the topic at issue, and were further asked if there was anything that we left out or that they wished to add at the end of the entire survey. We received more than 4000 comments, which we coded and quantified where possible. (56) Our goal in using this format was to combine the objectivity of a standardized survey format with the benefits of a more open-ended interview, although we recognize that there are inevitable dangers and biases associated with both formats. (57) All interviews were conducted orally. The majority were conducted in person, but forty-one (30%) were conducted by telephone due to scheduling difficulties.
The survey also engaged in a more specific set of inquiries than the Nourse and Schacter study: we asked about particular interpretive rules and different groups of rules rather than inquiring about them as a whole. We also asked questions that went beyond the canons, including questions about how the legislative process (for example, the use of omnibus versus single-subject legislative vehicles, or the relevance of committee consideration) affects how statutes are put together; the causes of statutory ambiguity; the role of states, agencies, and other actors in drafting and implementing legislation; and the role of Legislative Counsel (the nonpartisan, professional drafters in Congress). Finally, we asked more general questions about how the Supreme Court's current approach to statutory interpretation affects, or does not affect, how statutes are drafted.
Our project targeted a much broader population of staffers. Like Nourse and Schacter, we essentially confined our study to committee counsels with drafting responsibility. (58) But unlike their study, we interviewed counsels across both houses and across twenty-six different committees (thirteen in the House, twelve in the Senate, and one joint committee; many worked on subcommittees within these committees). We had a near perfect split between staffers working for Republicans and Democrats (fifty-two and fifty-four respectively out of 137) and between staffers working in the House and Senate (sixty-seven and seventy respectively out of 137), and a 58/48 split between staffers working in the majority and the minority. (59) The remaining staffers were nonpartisan. (60) We also included twenty-eight counsels from the nonpartisan drafting Offices of the House and Senate Legislative Counsel; (61) their role in the drafting process, though potentially significant, has generally been given little attention in the study of statutory interpretation.
That said, there are important limitations to our study, particularly as related to the ability to obtain a random sample. As detailed in the Methods Appendix, there are no sources available that designate which among all 6099 congressional staffers have drafting responsibility, and congressional staff perform a wide variety of other tasks as well. The bulk of legislative consideration occurs at the committee level, however, and although there are nonlawyer committee staffers, many do not draft legislation. We therefore limited our pool almost entirely to counsels serving on congressional committees or in the Offices of the House and Senate Legislative Counsel. In general, we surmised that counsels were likely to be the most aware of the Supreme Court's practices with respect to statutory interpretation, and so to the extent that counsels were unaware of, or uninterested in, those rules, such evidence would be particularly useful. (62)
Congressional staffers are a notoriously difficult population to study, both because of the difficulty of identifying the relevant staffers and also because of the strong culture of confidentiality and the fear of "leaks" that permeates Congress. Staffer schedules are also constantly in flux and the pace of the work is intense, making it difficult for staff to set aside time for projects like this one. As such, the level of response to our survey--137 out of the approximately 650 Committee and Legislative Counsels in Congress--greatly surpassed our expectations. (63) However, we recognize the limitations. For example, there is a danger of self-selection bias in surveys of this nature: those who responded might have been more interested in or more knowledgeable about these issues than those who did not. Unfortunately, such problems are unavoidable in a project like this one, with a generally reticent population that necessarily depends on volunteers.
We did our best to mitigate these concerns. Where possible, we have verified many of our respondents' observations with external sources. We also confirmed that our sample drew from many committees, both political parties, and both houses of Congress. In addition, we compared the publicly available data about our sample with publicly available data about a random selection of roughly 30% of the full committee-counsel population relating to factors including House/Senate employment, political party, majority/minority party status, committee assignment, age, law school graduation date, law school ranking, and years of experience. As detailed in the Methods Appendix, we did not find statistically significant differences (using a standard 95% confidence threshold) between our sample and the control group for House/Senate employment, political party, or majority/minority party status. (64) The committee assignments of the survey group were also reflective of the control group, except that the survey population had a greater share of Legislative Counsels than the control group, (65) as well as a greater share of members of some of the larger committees, although those differences were not significant. We did find statistically significant differences between the control and survey groups on the basis of age (based on college graduation date), experience (based on law school graduation date), and law school ranking, with the survey group consisting of older counsel as well as those who went to higher ranked schools. These findings, however, are limited due to difficulties with data collection for the age, experience, and law school information. (66)
Out of an abundance of caution, moreover, we have chosen to report our findings in a descriptive manner mostly using only the raw data rather than engaging in more sophisticated hypothesis testing to explore whether there were statistically significant drivers of certain answers. Even with our study's limitations, for many questions, our results were sufficiently lopsided at least to suggest that most counsels would be likely to respond in the same manner. For example, forty-six of our questions had more than 70% of respondents agreeing on a particular answer choice, and twenty-five had more than 90%. (67)
The fact that we limited the survey mostly to counsels is both suggestive and limiting, depending on the question at issue. For example, whereas legal rules unknown to our counsel respondents are unlikely to be known to other noncounsel drafters, it is difficult to draw inferences about the noncounsel population concerning those rules that our respondents did know. We address matters of this nature, where relevant, in the context of the specific findings discussed in the Parts that follow.
It is also difficult to draw inferences about the knowledge of the elected members of Congress themselves from our data, although we doubt that members know Latin canons of construction any better than do our counsel respondents. Our decision not to interview members was both pragmatic--we doubted that many would agree to be interviewed absent a personal connection--and theoretical. Members do not draft statutory text, but most interpretive doctrine is based on assumptions about the legislative drafting process. Moreover, judges rarely acknowledge the role of staff, except in conjunction with concerns about "sneaky" staffers unaccountable to members, but this accountability concern has never been explored empirically. Our decision to focus on counsels thus also highlights the question of which interpretive community within Congress (if any) is the relevant community for courts to focus on in fashioning interpretive doctrine.
To our knowledge, this is the most extensive survey of this nature ever conducted. (68) The survey questions and additional details about our methodology are included in the Article's accompanying Methods Appendix.
II. CONGRESS AND THE CANONS
Canons, as this Article uses the term, are simply the interpretive principles and sources that judges consult when resolving questions about statutory ambiguity. The canons are deployed in virtually every statutory interpretation case, and there are hundreds of them, but they do not all seem to be regarded in the same way. The canons that courts use to decide when to defer to administrative agencies' statutory interpretations, for instance--such as the Chevron deference rule--almost always are referred to as "doctrines" by litigants and administrative law scholars, whereas textual presumptions like the expressio unius rule (the inclusion of one statutory term implies the intentional exclusion of another) sometimes are referred to as mere "guidelines." But both types of rules, plus other interpretive tools like legislative history, (69) are relied upon by courts to resolve statutory ambiguities, which is why legislation scholars conceptualize all of these tools as canons of interpretation. (70)
That said, and despite their doctrinal centrality, the canons remain undertheorized. The main canons can roughly be divided into three categories: (1) the "textual canons," which are default rules about how text is drafted, such as expressio unius; (2) the "substantive canons," which are policy-based presumptions, like the rule of lenity or Chevron deference; and (3) the "extrinsic canons," which are outside sources, such as legislative history. Commentators have offered multiple justifications for each type of canon and, with each justification, a different take on the courts-Congress relationship.
Some justifications turn expressly on congressional awareness and use of the canons. For example, some canons have been justified as background presumptions in whose shadow Congress drafts; (71) as rules that teach Congress how to draft better; (72) as rules that force legislative deliberation; or as rules with such established common law pedigrees that it is assumed everyone knows them. (73) Others turn on an understanding of how Congress legislates, even if Congress need not be aware of them. Judges and scholars argue, for instance, that the textual canons reflect how ordinary people use language; (74) or how Congress signals intent to delegate to agencies; (75) or that legislative history reflects congressional intent. (76) All of these justifications turn on empirical knowledge about Congress that has been conspicuously absent and that, for many canons, our study calls into question.
Some justifications are less tethered to congressional practice. Some are institutional--for instance, that the canons further the judicial values of coherence and consistency; (77) that they are the easiest rules for judges to apply and so serve a judicial coordination function; (78) or that they transfer interpretive authority to more competent institutions. (79) Other justifications are normative--for example, that the canons make under-the-radar constitutional law (80) or advance underenforced (judicial or societal) policy preferences. (81) These types of justifications raise fewer questions about the reality of how Congress legislates but more questions about the judicial power to overlay Congress's work product with these rules. They also raise the question whether applying the canons for these reasons is really consistent with the faithful-agent paradigm that most practicing judges claim to espouse.
A. Overview of the Findings: A Spectrum of Canon Knowledge and Use
Our survey asked fifty-five questions about the textual and substantive canons, (82) thirty-seven about legislative history, (83) and forty-five more about the administrative law doctrines used by courts to resolve statutory ambiguity. (84) The variety of questions was designed to elicit information not only about whether the canons reflect how legislation is drafted, but also about whether anything resembling a feedback loop exists between the courts and Congress with respect to how the canons are utilized on both sides. We also surmised that there might be canons and doctrines whose assumptions drafters utilize without knowing their formal (often Latin or case-derived) names or without even realizing that they are rules that courts apply. (85) As a result, we inquired about almost all of the canons in several ways, typically first asking about the concept in layman's terms, (86) before inquiring about the canons by name or about their role.
The following Figures provide a bird's-eye summary of the key findings with respect to our respondents' familiarity with the canons and how those rules hold up to our respondents' descriptions of their actual drafting practices: whether question order affected the results, some respondents were asked about the deference doctrines by name first, while others were asked about them first by concept.
Our respondents displayed a much higher degree of familiarity with some of the canons and utilized many more of the concepts underlying them than we had expected or than the Nourse and Schacter study suggested. But not all of the canons were familiar to our respondents in the same way or used by them to the same degree. Nor did evidence for, or the prospect of, judicial-legislative dialogue seem the same for all of the canons.
We detail these findings in the following pages. This Part focuses on the textual and substantive canons, while the Parts that follow focus on legislative history and the administrative law canons, respectively. We divide our presentation in this way reluctantly, for convenience, and not because our respondents thought of the canons as belonging to separate categories. Indeed, one central finding of our study is that drafters treat text, legislative history, and agency implementation as integrated parts of a single process. Drafters do not think about statutory text without legislative history, and both text and history are drafted with agency implementation in mind and often with agencies at the table.
This Part also introduces three major institutional themes that emerged from our respondents' comments: the importance of (1) committee jurisdiction; (2) the legislative process; and (3) inside information to understanding how statutes are interpreted on the inside. These themes are major players in the companion Article, in which we detail central influences on the drafting process to which courts pay insufficient attention. Here, and in the Parts that follow, we merely identify these influences where relevant to our respondents' views of whether or how the canons influence or reflect the drafting process.
B. Textual Canons: More Familiarity by Concept than by Name
We asked first by concept and then by name about the six textual canons most commonly deployed by courts and scholars: (87)
* Noscitur a sociis (construe ambiguous terms in a list in reference to other terms on the list)
* Ejusdem generis (construe general, often catch-all, terms in a list in reference to other, more specific, terms in a list)
* Expressio/Inclusio unius est exclusio alterius (the inclusion of specific terms or exceptions indicates an intent to exclude terms or exceptions not included)
* The rule against superfluities (construe statutes to avoid redundancy; when there are two overlapping terms, construe to give an independent meaning to each)
* The whole act rule (statutory terms are presumed to have a consistent meaning throughout a statute)
* The whole code rule (statutory terms are presumed to have a consistent meaning throughout the U.S. Code)
We also inquired directly into the use of dictionaries (dictionaries should be consulted to determine the ordinary or plain meaning of statutory terms) and the in pari materia rule (similar statutory provisions should be interpreted similarly).
As the Figures below reveal, our respondents displayed a high degree of familiarity with the concepts underlying the textual canons, but much less familiarity with their formal names. Our respondents also appeared to regularly use several of these canons in the drafting process. But, of particular note, the concepts that our respondents indicated they used most often--for example, the concept underlying the expressio unius canon--are among the least consistently utilized textual canons by the courts, and they have come under criticism (even from textualist judges) about the extent to which they reflect drafting reality. (88) In contrast, the canons most commonly employed by courts, including the rule against superfluities, the whole act rule, and the use of dictionaries, appear to be used the least often by our drafters--despite our respondents' awareness that the courts use them--due to a host of political or institutional factors that courts rarely take into account.
1. Concepts in use: expressio, noscitur, and ejusdem
Approximately 33% of our respondents told us that the assumption underlying the expressio unius canon--that the inclusion of specific terms signifies the exclusion of terms not mentioned--always or often applies. (89) Five percent more agreed that the default rule is always exclusivity unless language indicates otherwise, and most of the remaining respondents likewise validated the assumption by explaining that they "signaled" whether they wished a list to be something other than exclusive, usually through the use of the word "including" or a catch-all term. Only 10% of respondents indicated that the presumption typically goes in the other direction, toward inclusivity. (90)
Expressio was also one of the most recognized textual canons by name (along with the rule against superfluities). But when asked about the rule by name, most of our respondents told us that they did not employ it (several respondents made statements such as "we don't know any Latin"), even though, when asked about the concept, they already had substantiated their use of the assumptions underlying it. A number of respondents (18%) got at this disconnect by describing the ideas embraced by the textual canons as "intuitive." As one stated: "We consider them not expressly but intuitively: how does this legislation interact with existing code? Is it inclusive, exclusive, are like things treated alike--those values are thought about here." (91)
With respect to the general concept underlying both the noscitur and ejusdem rules, 71% of respondents (ninety-seven) said that terms in a statutory list always or often relate to one another, and only two respondents said they rarely or never did. (92) The vast majority of respondents, however, did not know those rules when asked by name (85% did not know noscitur and 65% did not know ejusdem (93)).
2. Canons known, but rejected: superfluities, consistent usage, and dictionaries
Our respondents also were quite familiar with the concepts underlying the rule against superfluities (statutory words are intended to have independent meanings and are not intended to overlap with other terms) and the presumption of consistent usage, also known as the whole act or whole code rule (the presumption that terms are used consistently in multiple places in a single act or across the U.S. Code). They were also aware of the Court's frequent use of dictionaries in statutory interpretation. For each of these canons, however, respondents' awareness did not translate to routine use in the drafting process. Instead, we learned of institutional barriers to our drafters' frequent application of these rules.
a. Superfluities: redundancy to satisfy political stakeholders
For instance, even though 62% of our respondents knew the rule against superfluities by name, 18% of respondents told us it rarely applies, and 45% more told us it only sometimes does. Eighteen percent also explained the relative weakness of this rule's application by reference to two recurring reasons, one practical and one political. From a practical perspective, our respondents focused on the need to ensure that the statute covers the intended terrain. They told us that drafters intentionally err on the side of redundancy to "capture the universe" or "because you just want to be sure you hit it." (94)
These respondents also pointed out that the political interests of the audience often demand redundancy. They told us, for example, that "sometimes politically for compromise they must include certain words in the statute--that senator, that constituent, that lobbyist wants to see that word"; similarly, they said that "sometimes the lists are in there to satisfy groups, certain phrases are needed to satisfy political interests and they might overlap" or that "sometimes you have it in there because someone had to see their phrase in the bill to get it passed." (95) One example provided was a statute drafted to cover "medical service providers" that had to be amended to include a specific (and redundant) reference to "hospitals" to satisfy stakeholders. (96)
We were not surprised to see pragmatic considerations trumping application of the rule against superfluities. Common sense tells us that, despite the popularity of this rule with judges, (97) there is likely to be redundancy, especially in exceedingly long statutes. (We have seen no evidence, however, that judges take the length of statutes into account when applying the rule.) But what respondents told us was different from that common-sense assumption: namely, that even in short statutes--indeed, even within single sections of statutes--that terms are often purposefully redundant to satisfy audiences other than courts.
This is an argument that has been made in other contexts. Scholars have argued that the audience for legislative language or legislative history is much broader than judges, or even agencies, and that these statutory materials are sometimes expressly directed at noninterpreters, such as lobbyists and other stakeholders. (98) Whether this "audience" issue should have an effect on how courts interpret statutes is a different matter--after all, how will courts be able to discern when drafters are talking to them as opposed to other audiences? A fictitious interpretive rule may be required precisely because investigating the intended audience would be too difficult. But that has not been the main judicial justification for the rule against superfluities.
Our findings certainly call into question what has been the rule's primary justification: namely that, because it reflects how Congress drafts and also because Congress is aware of it, the rule helps faithful-agent judges effectuate congressional intent. (99) We note also that, in several recent cases, the Court has divided over application of the rule--with the majority relying on the rule to decide the case over the objection of dissenters who have argued, like some of our respondents, that Congress is often intentionally redundant to be certain that it has made its point. (100) We have seen no case, however, in which the Court acknowledged the political considerations, like satisfying stakeholders, that some of our respondents also mentioned. Our findings suggest that those considerations likewise may mean that judicial application of the rule does precisely the opposite of effectuating drafter intent. (101)
b. Committee jurisdiction and "unorthodox lawmaking" (102) as barriers to the whole act and whole code rules
The whole act and whole code rules seem to fall prey to even more pervasive institutional barriers. Although more than 93% of our respondents affirmed that the "goal" is for statutory terms to have consistent meanings throughout, (103) our respondents emphasized time and again the significant organizational barriers that the committee system, bundled legislative deals, and lengthy, multidrafter statutes pose to the realistic operation of those rules.
Our respondents told us that congressional committees are "islands" that limit communication between committees drafting different parts of the same statutes and that, because of the increasing tendency to legislate through omnibus or otherwise "unorthodox" legislative vehicles, most major statutes are now conglomerations of multiple committees' separate work. Later in the survey, in response to a different question, a large majority of our respondents (74%) said that omnibus bills are less likely to be internally consistent than single-subject bills. (104)
For the same reasons, our respondents also vigorously disputed that the first cousin of the whole act rule--the "whole code rule," under which courts construe terms across different statutes consistently--reflects how Congress drafts or even how it tries to draft. Specifically, only 9% of respondents told us that drafters often or always intend for terms to apply consistently across statutes that are unrelated by subject matter. (105)
This presumption of consistent usage, we would note, is widely accepted in the federal courts. Indeed, leading commentators have called it one of the most important and consistently applied textual default rules, (106) and it has been employed by textualists and purposivists alike. In the October 2011 Term of the Supreme Court alone, the whole act rule was used in at least three cases, (107) and the leading case for the principle has been cited in at least 118 federal cases since 1995. (108) To our knowledge, however, courts have never considered the role that committee jurisdiction plays when applying the rule, and courts have rarely focused on the type of statutory vehicle. (109)
We also note that, given the institutional factors that our respondents identified, application of the consistent-usage presumption is unlikely to exert any positive influence on the drafting process. This suggestion runs contrary to popular arguments that a strict textual approach may incentivize Congress to draft more carefully. Justice Scalia's new book offers a typical example of such an argument in support of the consistent-usage rule:
The canons ... promote better drafting. When it is widely understood in the legal community that, for example, a word used repeatedly in a document will be taken to have the same meaning throughout ... you can expect those who prepare legal documents competently to draft accordingly. (110)
Such arguments, however, depend on the absence of other barriers to such "better" drafting. Almost all of our respondents told us that consistent term usage was the "goal" or what "should be," but they still told us that the rule was unlikely to hold because of the way that Congress is organized.
c. "No one uses a freaking dictionary"
More than 50% of our respondents said that dictionaries are never or rarely used when drafting. (111) This finding stands in stark juxtaposition with the frequent and increasing use of dictionaries by the Supreme Court in statutory interpretation cases. (112) Although the Court has always looked to dictionaries in some statutory cases, scholars have documented that the Court's use of this interpretive tool recently has risen dramatically: the Court used dictionaries in 225 opinions from 2000 to 2010, compared to just sixteen opinions in the 1960s. (113)
Our respondents were aware of this judicial trend, but told us that it nevertheless did not affect their practice. Several specifically referenced Justice Scalia--acknowledging that the Court frequently uses dictionaries but noting that they remain mostly irrelevant to the drafting process. As one respondent put it (while laughing): "Scalia is a bright guy, but no one uses a freaking dictionary." (114) Another noted more delicately: "This question presumes that legislative staff have dictionaries. I have tried to get an OED but people over at finance say we aren't spending money to buy you a dictionary. And no Black's Law Dictionary either." (115)
The Court's rationale for dictionary consultation, however, may assume that Congress does use dictionaries or at least would welcome their use by judges. Individual Justices have stated that "[t]here is no cause to conclude that Congress [is] unaware of the ordinary definition" of words that are not otherwise defined in the statute, (116) a statement that implies that dictionaries are either a proxy for the "ordinary meaning" that the Court thinks that Congress intends or that the public understands, or that they are sources with which the Court presumes that Congress is familiar. Only the former explanation--that dictionaries accurately approximate ordinary meaning--even plausibly comports with our respondents' experiences, but that explanation likewise depends on other empirical evidence that is shaky at best and that faithful-agent theorists have done little work to confirm. (117) It also raises the question of precisely who is the audience that the Court is after--Congress or the public?
Given the accessibility of dictionaries on the Internet, moreover, it is implausible that the only reason our respondents do not consult dictionaries is their unavailability in print. Instead, our respondents simply seem to prefer other methods of defining terms. Mentioned far more often than dictionaries, for instance, were the definition sections that drafters themselves write into statutes. We did not inquire about definition sections directly; their importance was volunteered seventy times by forty-six different respondents in response to numerous different questions throughout the survey. However, as James Brudney and Lawrence Baum have recently pointed out, the Court at times has preferred dictionary definitions as evidence of "ordinary meaning" over those definitions drafted into the statutes themselves. (118) Likewise, dictionaries remain a far less controversial interpretive tool in the courts than does legislative history, even though, as elaborated in the next Part, our respondents also emphasized the utility of legislative history far more than dictionaries in resolving statutory ambiguities.
With respect to all of the textual canons discussed, the sophistication of our sample population seems informative. Regarding the three rejected canons, it is unlikely that nonlawyer drafters would place a greater emphasis than did our respondents on nonsurplusage, word use consistency, or dictionary definitions, or be more eager or able to overcome institutional obstacles to the application of those concepts. It also seems unlikely that nonlawyer drafters would be more familiar with expressio, noscitur, or ejusdem by name than were our respondents. Less predictable, however, is whether nonlawyer drafters would be as fluent in the concepts underlying those Latin rules.
It is a different question what, if anything, these findings tell us about the members of Congress themselves. The majority of respondents described their members' involvement as taking place at the more abstract level of policy rather than at the granular level of text. The Court, however, makes no such distinctions and, in fact, rarely acknowledges that staff, not members, are the primary drafters of enacted text (in strong contrast to judicial discussions of legislative history, in which the role of staff is often raised as a reason to disregard it). Moreover, whereas the Court seems almost entirely focused on members, the Court approaches statutory interpretation at a "close reading" level that seems quite different from the high level of abstraction with which elected officials actually seem to engage with statutory questions. The result is something of a doctrinal mismatch: a Supreme Court truly interested in crafting doctrines to reflect elected officials' approach to the drafting process would likely be a far less textual Court than ours.
C. Substantive Canons
There are more than 100 substantive canons, and they run the range from transsubstantive policy presumptions (e.g., ambiguous federal statutes will not be construed to intrude on traditional state functions); to subject-specific rules (e.g., ambiguous bankruptcy statutes shall be construed in favor of the debtor); to the dozen or so presumptions that concern delegation of interpretive authority to administrative agencies. (119) These canons are infamously conflicting, overlapping, and manipulable, (120) and have been described as everything from "judicial lawmaking" to "democracy protective" to "constitutional law." (121) We inquired about four of the most commonly deployed categories of substantive canons: (1) canons that advance federalism values, including clear statement rules; (2) the rule of lenity; (3) the canon of constitutional avoidance; and (4) the administrative law doctrines related to interpretation. As Figure 5 reveals, our respondents' awareness of the canons by name and their use of the concepts underlying them ran the range from high awareness and use of Chevron and the presumption against preemption to almost no awareness or use of clear statement rules. We also learned that, even when our respondents do draft in the shadow of the substantive canons, they do not always think about the canons in the same way that courts do.
This Subpart relays our findings with respect to all of these canons except for the administrative law rules, which we defer to their own separate discussion in Part IV.
1. Federalism, preemption, and clear statement rules
There are three basic iterations of the federalism-enforcing canons, and we inquired about all of them. Two function as presumptions: the eponymous "federalism canon," which counsels courts to interpret ambiguous federal statutes so as not to intrude on traditional state functions, and the "presumption against preemption," the default principle that courts should not interpret ambiguous federal statutes to preempt state law. The third class of principles are the so-called "clear statement rules," which are the "super-strong" (122) presumptions that the Supreme Court has articulated to enforce norms across a range of situations that typically implicate federalism, from the abrogation of sovereign immunity (123) to Congress's ability to impose conditions on federal grant money to the states. (124) Clear statement rules require drafters to use what are effectively "magic words" to achieve the result that runs contrary to the constitutional default rule--such as the rule requiring "unmistakably clear" language that Congress intends to abrogate the states' immunity from suits before a statute will be so construed. (125)
a. A partial courts-Congress feedback loop for federalism and preemption
Unlike in the case of the textual canons, most of our respondents said that they both knew of the federalism and preemption presumptions by name and that they also drafted with those rules in mind. Approximately 80% of our respondents told us that they were familiar with one of these rules by name and approximately 50% said they were familiar with both. (126) Of our respondents who were familiar with at least one of these presumptions, 65% said that at least one played a role when drafting.
These findings offer the first evidence that some kind of courts-Congress interpretive feedback loop does exist, at least with respect to certain interpretive rules. Knowing that the courts consider these federalism presumptions, many of our respondents told us, has an effect both on the substance of statutes and on how that substance is expressed. At the same time, however, our respondents did not understand these canons to function as interpretive "tiebreakers" in the same way that courts do.
Specifically, 14% of our respondents told us that the canons serve to tee up debate about the issue when conceptualizing a statutory scheme. The following comments were typical: "The presumptions help to highlight and remind us of the importance of the question whether this is the right role for the federal government," and "It gives us sensitivity to the issue when we legislate concerning the states." (127) These comments are consistent with academic and judicial arguments that canons such as these may have a salutary, deliberation-forcing effect on the legislative process. (128)
Another 19% (or 24% of those who knew at least one of these canons) told us that knowledge of these canons encouraged more specificity once pen was put to paper. As some put it: "If we are making a provision we think will preempt, it's part of the deliberation: it's a reminder to everyone to make sure we are clear," and "No one refers to them by name but they come up in discussion, through about how specific you want to be and how the clause will be interpreted." (129) These explanations are consistent with a different, more directly dialogical, kind of justification that has been offered for these canons: namely, that they teach Congress how better to communicate in general, and how specifically to telegraph its intentions to the courts. They also are consistent with a democracy-based justification; they make it more likely that important decisions will be made by elected officials rather than by courts.
At the same time, there was an important disconnect between common judicial understandings of these canons and the way that our respondents told us they understood them. Courts use these canons as "thumb-on-the-scales" (130) presumptions that tip interpretation of ambiguous statutory language in one direction, and courts assume that Congress knows the direction in which the scales will be tipped. Our respondents' answers, however, did not substantiate that assumption.
Our respondents told us that the canons serve as reminders of the importance of resolving the issue, but do not create expectations about how any ambiguities that remain in the text will later be resolved--that is, after the reminders are ignored for political reasons, lack of time, and so on. The majority of respondents (60%) told us that these presumptions did not necessarily cut in a particular direction. Instead, most respondents said that they expected federalism-related ambiguities to be resolved based on a variety of factors, including the particular court and area of law being interpreted. (131) Similarly, in response to other questions in the survey, we received thirty additional comments (from twenty-two different respondents) emphasizing that canons serve as "frameworks" that "focu[s] attention on certain issues that wouldn't be there if you were drafting from scratch" rather than as "dispositive rules."
Indeed, only 6% of our respondents said that ambiguities in federal statutes relating to preemption would be construed by courts in favor of the reach of state law. (132) But that is exactly the way the presumption usually functions in the federal courts. (133) And 12% predicted that the presumption would run in the opposite direction, with courts favoring the reach of federal law. (134)
The federal-law-oriented perspective of our respondents perhaps should have been unsurprising, given our respondents' vantage point and mission. The federal statute is their work product and, when ambiguous, these respondents expect their work product, not state law, to control. (135) The contrast with the Court's assumption--that the federalism canons reflect "an acknowledgment that the States retain substantial sovereign powers under our constitutional scheme, powers with which Congress does not readily interfere" (136)--is evident. In this sense, the courts-Congress feedback loop with respect to the federalism canons that we have identified appears to be only a partial one. (137)
b. The irrelevance of clear statement rules
Evidence of a feedback loop for clear statement rules, on the other hand, was almost entirely absent. These findings are problematic for clear statement rules on most normative justifications for them because clear statement rules--regardless of what purpose they serve---depend on the idea that the "rules" about those magic words that drafters must use are actually being transmitted to Congress. (138) They also are usually described as "super-strong" interpretation rules: canons that are more important and more consistently applied than ordinary presumptions like preemption. (139)
None of these assumptions was validated by our respondents. Even as most of our respondents told us that they would be more attuned to the courts' interpretive practices if the courts were more consistent, (140) only about 30% of respondents said they could name any clear statement rule that they thought was important in the drafting process (of any sort, not just federalism related) and, when asked to list such rules, of that number only six respondents (4% of 13 7) named a rule that actually was a clear statement rule. (141) Only 22% of our respondents thought that clear statement rules were relied upon more than other canons. (142)
These findings also lend support to some of the pragmatic criticisms that have been levied against the use of clear statement rules. Scholars have complained that these magic words requirements are unnecessary to enforce presumptions already effectuated by the ordinary canons; that there are too many such rules; and that the variety of rules makes it difficult for drafters to predict the courts' interpretive path. (143) Indeed, our respondents did not seem to know, use, or understand the broad landscape of clear statement rules, but did display much familiarity with the ordinary federalism presumptions.
2. Lenity unknown by name
We were surprised by our respondents' lack of familiarity with the "rule of lenity" by name. Lenity is one of the oldest interpretive rules; it provides that ambiguous criminal statutes are to be construed in favor of criminal defendants. The Court often uses it as a tiebreaker when statutes are ambiguous, and some Justices will turn to it to resolve a case before, or instead of, legislative history. (144) The rule is often justified precisely by virtue of its assumed familiarity based on its centuries-old pedigree (145)--most recently by Justice Scalia, who argued in his new book that "rules like these, so deeply ingrained, must be known to both drafter and reader alike so that they can be considered inseparable from the meaning of the text." (146) Even those who generally oppose application of the substantive canons on the ground that they improperly advance judicial policy preferences often create an exception for lenity on this basis: because it is assumed that all drafters know and draft in accordance with the rule, its application has been viewed as consistent with faithful agency.
Our findings may challenge this assumption. Of the sixty-five respondents in our survey who had participated in drafting criminal legislation, only 35% were familiar with the rule of lenity by name. (147) The nature of our sample seems informative here, as it is unlikely that the broader drafting population would possess more knowledge of this canon by name than did our counsel population. We recognize, however, that our findings do not address whether respondents know the rule by concept, as we did not inquire about lenity in that manner. (148)
Also worth noting is that, of the twenty-three respondents who knew of the rule by name, fifteen worked on the House or Senate Judiciary Committees--the committees generally charged with jurisdiction over criminal law. This result may lend some support to the idea that canon knowledge is subject-area specific, and thus perhaps congressional-committee specific. Judiciary staff might be more familiar with lenity because the Judiciary Committee works on criminal law bills more frequently than other committees. Or Judiciary staff may be a more expert set of drafters altogether. (149) In the companion Article, we elaborate on such intercommittee differences and their potential implications for doctrine. For present purposes, suffice it to say that the accuracy of judicial assumptions about canon awareness and use may depend on what kind of staffer is doing the drafting. Whether courts would be capable of making those distinctions, and doing so without excessive cost, is a separate question, and one that we also address in the companion Article.
We note that we would not have been surprised had our respondents told us that they disagreed with the rule of lenity, or that they did not draft in accordance with it. Many state legislatures have passed laws attempting to abrogate judicial application of the rule, (150) and its underlying presumption is at odds with elected officials' predisposition to appear "tough on crime." What was surprising was that, given how often the rule's widespread knowledge has been assumed by courts, most of our respondents had never heard of it, at least not by name.
3. Constitutional avoidance unknown but assumed
Lastly, our respondents were not familiar with the canon of constitutional avoidance, the oft-applied presumption that courts will construe ambiguous statutes to avoid constitutional issues. We asked several questions aimed at this rule. We initially asked whether staffers had any assumptions about judicial presumptions involving construction of statutes that raised constitutional questions. Forty-four percent of our respondents reported a judicial presumption in favor of upholding federal statutes (what many respondents called something on the order of "congressional deference" (151))--a federal-law-oriented position that seems similar to the assumption of those respondents who told us that ambiguous federal statutes are to be construed in favor of the reach of federal law rather than in favor of state law. Another 45% responded to this question with "it depends," with some respondents in that category explaining that their assumptions about judicial presumptions related to constitutionality depend on the subject, the court, or the drafter. (152) But none mentioned the idea that courts might avoid the constitutional question altogether. To reach that question more directly, after completing approximately half of our interviews, we added a question explicitly referring to the avoidance rule. Of the sixty-seven counsels asked that question, only 25% had heard of the rule. (153)
At the same time, that canon may nevertheless rest on accurate assumptions about congressional intent, at least with respect to our respondents. For example, although most of our respondents did not know the canon by name, 69% of the 137 said that that their expectations about how the courts would rule on the constitutionality of statutes played a significant role in the drafting process. (154) A number (18%) emphasized that it was their "job" to make sure that their statutes are upheld (an emphasis that may not be replicated among noncounsel staffers, who may have different conceptions of their "jobs"). Thirteen percent (including four respondents counted toward the 18% who emphasized their "job") said that that they examine prior case law or try to create a clear record for courts in anticipation of judicial ruling. (155) As such, it was clear that our respondents do at times focus rather closely on courts, particularly when dealing with constitutional issues. To the extent that the avoidance canon rests on the presumption that Congress tries to legislate within constitutional bounds, (156) our respondents' answers were consistent with it.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||part 1; Introduction through II. Congress and the Canons C. Substantive Canons, p. 901-948|
|Author:||Cluck, Abbe R.; Bressman, Lisa Schultz|
|Publication:||Stanford Law Review|
|Date:||May 1, 2013|
|Previous Article:||Federalism and the state civil jury rights.|
|Next Article:||Statutory interpretation from the inside - an empirical study of congressional drafting, delegation, and the canons.|