STATUTORY INTERPRETATION FROM THE INSIDE-AN EMPIRICAL STUDY OF CONGRESSIONAL DRAFTING, DELEGATION, AND THE CANONS: PART II
| By Gluck, Abbe R | |
| Proquest LLC |
This is the second of two Articles relaying the results of the most extensive survey to date of 137 congressional drafters about the doctrines of statutory interpretation and administrative delegation. The first Article focused on our respondents' knowledge and use of the interpretive principles that courts apply. This second Article moves away from the judicial perspective. Our findings here highlight the overlooked legislative underbelly: the personnel, structural, and process-related factors that, our respondents repeatedly volunteered, drive the details of the drafting process more than judicial rules of interpretation. These factors range from the fragmentation caused by the committee system, to the centrality of nonpartisan professional staffin the drafting of statutory text, to the use of increasingly unorthodox legislative procedures-each of which, our respondents told us, affects statutory consistency and use of legislative history in different and important ways. Our respondents also painted a picture of legislative staffers in a primary interpretive conversation with agencies, not with courts, and as using different kinds of signals for their communications with agencies than courts consider.
Most of the structural, personnel, and process-related influences that our respondents emphasized have not been recognized by courts or scholars, but understanding them calls into question almost every presumption of statutory interpretation in current deployment. The findings undermine the claims of both textualists and purposivists that their theories are most democracy enhancing, because neither makes satisfactory efforts to really reflect congressional expectations. Our findings challenge textualism's operating assumption that text is always the best evidence of the legislative bargain and suggest more relevant-but still formalist-structural features that might do better. Our findings further reveal that, although purposivists or eclectic theorists may have the right idea with a more contextual approach, many of the factors on which they focus are not the same ones that Congress utilizes. With respect to delegation, our findings suggest that, for both types of theorists, Chevron now seems too text- and court-centric to actually capture congressional intent to delegate, although that has been its asserted purpose.
In the end, our findings raise the question whether the kind of "faithful agent" approach to interpretation that most judges currently employ-one aimed at effectuating legislative deals and often focused on granular textual details- can ever be successful. We thus look to different paradigms less dependent on how Congress works, including rule-of-law and pragmatic approaches to interpretation. These alternatives respond to the problem of the sausage factory, but pose different challenges in light of the modern judicial sensibility's pronounced concern with legislative supremacy.
Is the goal of statutory interpretation to reflect how Congress actually drafts legislation? Is such an accomplishment possible? Would courts actually desire it? Where would the responsibility lie-with the Court, Congress, or both-to effectuate it?
The most common iterations of legislation and administrative-law theory view the goal of interpretive doctrine as reflecting congressional practice or expectations. Alternative theories have posited different roles for doctrine, less tethered to the details of how Congress works, including providing predictable coordinating rules for the legal system, or rules that assist judges, as partners of the legislature, in effectuating broad statutory purposes. Each of these theories, however, relies in different ways on empirical assumptions that appear mistaken. Each also presupposes the existence of some kind of
This is the second of two Articles relaying the results of the most extensive survey to date of 137 congressional drafters about the doctrines of statutory interpretation and administrative delegation. The first Article focused on our respondents' knowledge and use of the interpretive principles that courts apply.1 This second Article moves away from the judicial perspective. We focus here instead on the many other influences that our respondents told us have more relevance to the drafting process than most of the Court's interpretive rules. We also explore the implications of our respondents' views about delegation to courts and agencies, and their expectations of the
Our findings highlight the overlooked legislative underbelly: the personnel, structural, and process-related factors that, our respondents repeatedly volunteered, drive the details of legislative drafting. They also paint a picture of legislative staffers in a primary interpretive conversation with agencies, not with courts, and as using different kinds of signals for their communications with agencies than the signals courts consider. At the same time, our respondents were in some ways as court-centric as interpretive doctrine has been: they identified the need for more
Most of the personnel, structural, and process-related influences that our respondents emphasized have not been recognized by courts or scholars; but understanding them calls into question almost every presumption of statutory interpretation in current deployment. For example, our respondents told us that statutes are sometimes drafted in contorted ways to guard committee jurisdiction and agency oversight; that committee staff, leadership staff, Legislative Counsel, and personal staffall draftstatutes, but that each type of staffhas different goals and varied drafting practices; that legislative history plays different roles in omnibus, appropriations, and single-subject legislation; that the congressional budget score has an enormous impact on statutory language; and that whether a statute goes through committee or not-and which committee- should affect the interpretive presumptions applied.
The committee system and the varied roles and practices of different types of staffemerged as central organizing features. The division of Congress into committees causes deep fragmentation that not only defeats common presumptions of textual consistency, but also drives decisions about delegation to agencies. With respect to the importance of personnel differences, for example, we uncovered an unappreciated disconnect between those staffers who help elected members make policy and draftlegislative history, and the nonpartisan, professional staffers who are not directly accountable to members but nevertheless often take the primary role in drafting the actual text.
These findings have significance for textualism, purposivism, and beyond. They undermine the current democracy-based claims of each theory, because none makes satisfactory efforts to really reflect congressional expectations. Our findings challenge textualism's operating assumption that text is always the best evidence of the legislative bargain and suggest more relevant-but still formalist-structural features that might do better. Our findings further reveal that although purposivists or eclectic theorists may have the right idea with a more contextual approach, many of the factors on which purposivists and eclecticists focus are not the same ones that Congress utilizes. With respect to delegation, for both types of theorists, Chevron now seems too text- and courtcentric, in light of our findings, to actually capture congressional intent to delegate, although that has been its asserted purpose. And our respondents' strong resistance to any notion that Congress intends to delegate interpretation to courts raises the difficult question for partnership and pragmatic theorists about whether the democratic legitimacy of those approaches depends on Congress's assent.
Different obstacles present themselves from Congress's side. Foremost are the new questions that we raise about whether Congress has obligations in this interbranch conversation, and how those obligations could be effectuated or enforced. If the democratic legitimacy of the interpretive effort depends on Congress's engagement in some way with judicial statutory interpretation-be it by overriding erroneous decisions, standardizing drafting practices to make them more transparent to courts, or otherwise altering its drafting practices to respond to judicial assumptions-how can Congress be incentivized to address existing barriers to such efforts, especially given its deep internal structural fragmentation? Relatedly, can we really fault courts for coming up short in their efforts at statutory translation if Congress has not offered substantial assistance? At the same time, we have doubts, notwithstanding the ubiquitous judicial embrace of the concept, that courts would really want Congress to be, if it could, an engaged principal, partner, or system co-coordinator in the first place.
Part I summarizes the study design and the key findings of the first Article. Part II relates our findings about the personnel, structural, and process-related influences that our respondents emphasized. In Part III we move to delegation, exploring the implications, both for Chevron and for interpretive theory in general, of our respondents' assertions that their primary interpretive partners are agencies not courts, but that they nevertheless want the courts to coordinate with Congress.
Part IV explores the broader implications of the findings. We begin with the obvious question: namely, given that legal doctrine cannot possibly reflect all of the intricacies of the legislative process that our study uncovered, should such a reflective theory still be the goal if it must be partially incomplete? To that end, we explore the two most commonly offered alternatives to the reflective paradigm that are less trained on the details of congressional practice: a "rule-of-law" model that sees the goal of interpretive doctrine as facilitating systemic coordination, predictability, and coherence; and a broader version of purposivism that aims to pragmatically effectuate the broad goals of the statute rather than specific legislative understandings. Although such theories may be liberated from the complexity of the details that we uncovered, they face other challenges. Both depend on at least some kind of interbranch communication or consistency of interpretive practice that we have found utterly lacking, and both require more comfort with a broader lawmaking role for judges in statutory interpretation than many modern judges are willing to embrace.
In the end, however, we do not cry "sausage!" and abandon the effort. Regardless of which theoretical paradigm one prefers, many of the structural and process-related cues that our respondents identified-for instance, committee jurisdiction and type of statutory vehicle-are more objective and predicable factors than the menagerie of canons currently beloved by interpreters of all stripes, and would aid the goals of any of the mainstream theories. We also illustrate how this kind of circumstance-specific reorientation away from universal interpretive principles is not as radical a departure from current practice as it seems. The Court currently utilizes special interpretive rules for different subjects, interpreters, and types of statutes. And, in the agency-deference context, the Court already made the explicit move, in
This is no small order for statutory interpretation-and for a single article, or even two. We are limited by space constraints with respect to the depth in which we can explore all of these questions here. Our aim is to frame the debate, and with it, a research agenda for statutory interpretation in the modern regulatory state.
I. SUMMARY OF THE STUDY AND KEY FINDINGS IN THE FIRST ARTICLE
We conducted our survey of 137 congressional staffers over five months in 2011-2012. Each interview included the same 171 questions on topics ranging from the staffers' knowledge and use of the interpretive canons, legislative history, and the administrative law deference doctrines, to their views of the legislative process and the roles of courts and agencies in statutory interpretation. Almost all respondents were counsels on congressional committees or counsels in the understudied nonpartisan drafting Offices of House and Senate Legislative Counsel. Our counsels were drawn approximately equally from both parties, both chambers of Congress, and from twenty-six different congressional committees. Detailed elaboration of our methodology, including our decision to focus on committee counsels, is available in the first Article, which appeared in the previous volume of this journal, and in the online Methods Appendix.3
The first Article had two main goals. First, we aimed to probe empirically the assumptions that underlie the most common interpretive doctrines. Most of these canons-which are utilized by textualist and purposivist judges alike- explicitly depend on the notion that Congress knows or acts in accordance with their presumptions.4 Yet whether that courts-Congress interpretive dialogue exists, and what it looks like, had never been empirically tested.5
Second, we aimed to expose that the prevailing justification for the canons- that they effectuate a vision of judges as Congress's "faithful agents"-is both unhelpfully general and often inaccurate. We illustrated that courts use this faithful-agent paradigm to support what are actually many different types of interpretive rules doing many different types of work. Some canons aim to reflect how Congress drafts (e.g., the presumption that Congress does not use redundant language); others aim to affect how Congress drafts (e.g., clear statement rules, which require Congress to express its views extra clearly on high-salience issues); still other canons impose external norms on the legislative process (e.g., the due process values that the rule of lenity imposes on criminal legislation). It is not clear that the same-or a coherent-conception of the judicial role in statutory interpretation animates these various goals. Moreover, our findings revealed that many canons do not seem to be achieving their purported aims.
We did find some evidence of the kind of courts-Congress interpretive feedback loop that many have assumed impossible. For example, our respondents told us that they consider Chevron and the federalism canons when drafting precisely because they understand that courts use them. There were other canons of which our respondents were unaware as legal rules but whose underlying assumptions seemed to accurately reflect how they write statutes; Mead and a surprising number of the other administrative law deference doctrines exemplified this phenomenon. At the same time, there were commonly utilized canons, such as clear statement rules, which our drafters did not know and whose assumptions were not reflected in their drafting practices; even more significantly, our drafters knew of other popular canons-including the presumption that statutory terms are used consistently, the rule against superfluities, and the use of dictionaries-but consciously rejected them as unreasonable assumptions about the legislative process given trumping structural or political considerations. The following Figure summarizes the results:
These data slice through current theory in several ways. For instance, the discovery that our drafters have learned and do use some of these presumptions is major news for legislation theory: it indicates that interpretive doctrine actually could serve as a bridge for communication between the courts and Congress. But the findings also make clear that the current regime cannot be justified solely on faithful-agent grounds. Judicial application of seemingly neutral interpretive rules (like the presumption of consistent usage) that do not actually reflect how Congress drafts is not neutral at all, and must be understood instead as the imposition of external values on legislation. There are many conceptualizations of the judicial role in statutory interpretation that could support such an approach, most notably rule-of-law arguments that justify the canons as tools that help judges coordinate systemic behavior or cohere the corpus juris. But as we elaborated, those arguments cannot bear the full weight of the current regime. Judges rarely justify the canons as unrelated to congressional practice, and judges do not actually follow through with rule-oflaw goals-federal courts are notoriously inconsistent in their application of the canons and have created too complex a web of them to advance coherence or predictability.
Space does not permit a recapitulation of all of the first Article's other interventions. For present purposes, we note two more as especially relevant to the pages below. First, our exploration helped us identify a glaring omission in the theoretical debates about principal and agent in statutory interpretation. Despite the extensive dispute over how courts should effectuate their role as "faithful agents," there has been little discussion of Congress's obligations or preferences in the same interpretive conversation. Second, our findings raised the question of what, if anything, courts should do with the information that our study uncovered. There is an obvious tradeoffbetween rules and standards and complexity and simplicity in law that is relevant for the construction of all legal doctrine,6 but that seems particularly relevant to the question of how courts and scholars might use empirical work like ours.
Here, we go deeper. We learned much more from our respondents about the factors that shape the legislative process beyond the limited ones that judges consider. We also gathered the first substantial evidence of how at least some congressional staffers view the
II. UNAPPRECIATED STRUCTURAL INFLUENCES AND VARIETY-AND THEIR IMPLICATIONS FOR THE LEADING INTERPRETIVE PARADIGMS
Legislative drafting is driven by influences that seem largely invisible to courts. By this, we do not mean politics, which of course is a major factor and one that most everyone acknowledges. What we refer to is a variety of personnel- related and structural influences on the drafting process that our respondents volunteered, time and again, as more central to understanding statutes than are the drafting conventions that form the bases of the court-created interpretive presumptions that we probed in the first Article. Three points emerged with particular salience:
First, the common disconnect between staffers who draftstatutory text and staffers who craftthe policy that underlies it (and the related disconnect between text and legislative history drafters);
Second, the immense variety and fragmentation of types of staffinvolved in the legislative process and their different goals and drafting practices; and
Third, structural features like the centrality of committee jurisdiction, the type of statutory vehicle, the path the statute takes through Congress, and the requirement that statutes be "scored" for budgetary purposes-each of which affects how statutes are drafted and understood inside Congress.
These findings pose challenges for every mainstream theory of interpretation. For the most common kind of "faithful-agent"-based theory espoused by most practicing textualists and now even by most practicing purposivists7-by which we mean a theory under which the ostensible goal of interpretive doctrine is to reflect how Congress drafts-these details counsel toward major revisions in the dominant interpretive presumptions. For example, our respondents affirmatively rejected many popular textual rules, such as presumptions of consistent usage and even sometimes plain text, in favor of rules that reflect how the structural division of Congress into committees and the way that statutes are put together influence drafting. We make the case in Part IV why even formalists might be attracted to these structural factors.
A more radical move would be to abandon this common premise altogether. Textualists and most modern purposivists divide over which interpretive rules best effectuate how Congress works or what Congress knows, but not over whether the goal of tying interpretive rules to congressional practice is the right goal in the first place.8 These battles over individual interpretive presumptions now seem misguided-not only because the fights seem focused on the wrong kinds of cues, but also in light of the bigger question that our findings raise about how any legal doctrine could possibly reflect the kind of staff, process, and structural variety that our respondents told us affects statutory meaning. This question, however, is likely to cause panic in the many lawyers and judges who have insisted that the only democratically legitimate theory of interpretation for unelected federal judges is one that is linked to congressional intent or practice.9
The main alternatives to this version of the faithful-agent model face different obstacles in light of our findings-in addition to having to contend with charges (with which we do not agree) that they are in tension with democratic values.10 Both a rule-of-law theory and a more broadly purposive, pragmatic, or "partnership" model of judging might be viewed as responses to the legislative complexity that our study reveals: neither of those theories depends on the same kind of reflection of the details of dealmaking that the most common iterations of modern textualism and purposivism require. At the same time, both of these alternative approaches depend, at least in part, on some coordination with Congress. Our findings suggest that Congress may be not only uninterested in coordinating with courts, but also too fragmented to be able to do so without some radical changes.
This Part also offers the first sustained effort to differentiate among different types and roles of congressional staff. That effort reveals that the Court's current text- and canons-focused approach actually privileges the perspective of one particular kind of staffer: the professional, nonpartisan drafters in the offices of Legislative Counsel who focus on text and court cases. Privileging Legislative Counsel comes at the expense of the other staff-and elected members- who create the policy and make the deals, but often do not drafttext and are not as focused on the courts. So understood, the Court's approach empowers those who speak and understand this lawyerly language11-a perhaps understandable approach for the Court to take given that Congress has offered it little guidance. But our findings suggest that an approach designed to reflect the way that policymakers work would look different.
A. InterstaffDifferences and the Disconnect Between Text and Policy
The Figure below summarizes our respondents' answers to questions concerning who participates in the drafting process and where the first drafts of statutes originate. We note that, although most respondents told us that elected members participate, their answers were often qualified, here and elsewhere in the survey, by comments such as "only on a high level and it depends on the member"; and "they do conceptual, but don't participate in the language."14
Three implications of these findings have particular relevance for our study. First, they raise the question: which of these various contributors is drafting what? One of our most important findings is the centrality of the role that the nonpartisan drafters inside Congress-the Offices of the House and Senate Legislative Counsel-have in the drafting of statutory text. As we relate below, there is a disconnect that current doctrine and theory do not even attempt to engage between members and their immediate staff, who make the deals, craftpolicy, and write legislative history, and these professional drafters who often write the legislation.
Second, these data drive home our point that a theory based on how Congress drafts may be impossible to accomplish. At the most basic level, most of the canons are based on the assumption that there is some consistency of drafting practice, and all of the theories assume that the answer is roughly the same no matter who is drafting. Our data suggest that this assumption should vary depending on factors such as whether the staffer works for a committee as opposed to a member's personal staff, or what subject matter is at issue. This is a bigger point than the common notion that the "omniscient" drafter assumption is a fiction. Even assuming an omniscient drafter exists, there are simply too many categories of different types of omniscient drafters to make general assumptions across them.
Third, these findings make plain how difficult it would be for Congress to standardize drafting practices that courts, in turn, could incorporate into their own interpretive assumptions. The committee system divides policymakers into "silos" that do not communicate with one another, a fragmentation exacerbated by the separate and different roles that noncommittee leadership staffand personal staffplay in the drafting process. Nor do the Offices of Legislative Counsel have the formal or informal clout to perform a coordinating function that some assume they may already be performing.
1. The central role of Legislative Counsel
The current doctrinal regime, almost certainly unknowingly, privileges above the other approximately 12,000 congressional staffers, the drafting perspective of the eighty-three nonpartisan staffers in the Offices of House and Senate Legislative Counsel.15 The current regime accomplishes this through its focus on text and textual structure at the expense of general policy and (sometimes) legislative history, and also through its reliance on judge-made interpretive presumptions, which most staffers are not as focused on, instead of the kinds of interpretive cues that ordinary staffand members actually seem to use.
The Offices of House and Senate Legislative Counsel began their work in 1916 and 1919, respectively,16 but they are mostly invisible in the literature on statutory interpretation.17 One exception is
Across the first 104 surveys we conducted, 83% of respondents volunteered 183 different comments about the centrality of Legislative Counsel in response to numerous questions about the canons and drafting without any prompting from us.20 At that point, we added two explicit references to Legis- lative Counsel in the second version of the survey to better capture this story in the remaining interviews.
We also interviewed a broad cross-section of Legislative Counsels themselves- twenty-eight in total-working in both chambers, across multiple committees and with differing degrees of seniority. Both our Legislative Counsel and other staffer respondents showed remarkable agreement about the Offices' role, strengths, and weaknesses.
a. Legislative Counsel as the primary drafters of text; others as the primary makers of "policy"
Our non-Legislative Counsel respondents underscored that they rarely draftstatutes from "scratch," and most told us that the drafting of statutory text is often done by Legislative Counsel.21 We believe these findings would be magnified in the broader drafting population, given that committee counsels are more likely to draftthemselves than nonlawyers or even lawyers on personal staffs. The following comments were typical: "99% is drafted by Legislative Counsel. Most legislation is an amorphous concept given by member or staffer";22 "Staffers are more focused on the idea and execution . . . then talk to Legislative Counsel conceptually and have them execute the draft";23 and "No staffer drafts legislative language. Legislative Counsel drafts everything."24 The process by which this occurs, according to many of our respondents, is that staffers who work directly for members or committees provide Legislative Counsel with policy "bullet points"25 or sometimes rough outlines of statutory text, which Legislative Counsel then turns into legislative language.
These findings suggest an important disconnect not only between members and statutory text (a disconnect long suspected) but also between staffwho are policy experts and staffwho turn policy into legislative text. It is not uncommon to hear that a group of elected officials has reached a "deal" before pen is put to paper.26 But a main theoretical assumption has been that members, or at least their direct staffs, write up the details of those deals.
Moreover, the idea that legislative history is somehow more removed from this process than the statutory text also seems inaccurate. The textualist argument against legislative history is often couched as a members-versus-staffproblem-i.e., that legislative history is the product of unaccountable or "sneaky" staffor committees, as opposed to text, which is formally approved by all elected members. Instead, the real issue seems to be a staff-staffdistinction with respect to both legislative history and statutory text. Those policy "bullet points" may be the strongest evidence of congressional intent, but courts have never referenced them (nor are they public). And the kind of legislative history that courts do fight over, such as committee reports, as the first Article detailed, is drafted by those staffwith more policy expertise and greater direct accountability to the members than the staffwho may draftthe text. Unlike ordinary staff(who told us, "staffers don't keep their jobs if they disagree with members"27), Legislative Counsels (who also viewed themselves as accountable, telling us, "if you're not elected, you're replaceable"28) cannot be fired by individual members,29 and do not lose their jobs when control over Congress changes.
We qualify this observation with a finding from our first Article: not all legislative history is created equal. Our respondents were consistently more discerning than courts about which legislative history is reliable, for example, emphasizing consensus, bipartisan legislative history that reflects the views of more than one member. Although the same staffmay draftless reliable legislative history as well, it is still the more reliable type of legislative history that seems to provide the most important alternative, or supplement, to statutory text.
b. Implications of the Legislative Counsel story for a text-focused approach
We doubt that most judges would claim an intentional focus on the understandings of the unelected, nonpartisan congressional career staff. But
Our focus here is on instances in which statutory meaning is uncertain- the cases that implicate judicial interpretation or that pose particularly vexing problems for agency implementation. Text, several respondents told us, is always "the gold standard" or the "ideal,"31 but for both practical reasons (e.g., lack of time to draftclearly, unforeseen circumstances, mistakes, etc.) and political reasons (e.g., need to compromise, inability to reach consensus on an issue, desire to retain some flexibility), gaps and ambiguities are inevitable. The age-old question for statutory interpretation is what to do then.
We recognize the formalist argument for a text-centric approach in these circumstances. Some textualists argue that the impenetrability of the legislative process makes it necessary to rely on the only official act-the vote of the member on the statutory text-as the best evidence of statutory meaning.32 And, of course, Congress does vote on the text that Legislative Counsel produces. Some textualists also ground their formalist approach in the impossibility of discerning collective intent;33 our breakdown of the various staffresponsibilities and perspectives indeed may drive home that point. But a text-based response to these difficulties, we think, must find its justification in rule-of-law values, and not under the textualists' current, democracy-based, faithful-agent rubric that a text-based approach best respects legislative bargains.34 That premise seems fictitious. Our findings cast doubt on whether members or highlevel staffread, much less are able to decipher, all of the textual details and illustrate how the different players in the legislative process contribute to the final product in different ways, not all of them text focused. This does not mean that formalist justifications cannot sustain a vote-on-the-text-based approach; but it means its basis should be in such formalism, not in assumptions about legislative bargains.
Moreover, even under ideal drafting circumstances, the particularly granular level on which many text-focused debates now take place in the case law seems disconnected from the way in which members and congressional policy staffengage in the drafting process. This is not to say that members and staffdo not care about text, or that legislative history or the other influences we shall identify will always be able to answer disputes about particular statutory phrases. Rather, it is to say that micro-level legal disputes over what is often a single word in a lengthy statute may be improperly focused in the first place.
Another reason that we raise these concerns is that our non-Legislative Counsel staffers told us that they often are not capable of confirming that the text that Legislative Counsel drafts reflects their intentions. Our respondents repeatedly mentioned anxieties about the lack of interstaffaccountability attendant to this divided process of policymaking and drafting. Our ordinary counsels reported that the difficulties of understanding technical statutory language and of tracking the numerous statutory cross-references and amendments to preexisting legislation make penetrating the language that Legislative Counsel generates challenging even for staffers who are lawyers.35 Many expressed special concerns that younger or less experienced staffers cannot confirm that Legislative Counsel accurately translates their deals. The following comment was typical: "Legislative Counsel drafts, and the staffer doesn't have the law degree or expertise to evaluate what Legislative Counsel did."36 One of our respondents vividly described the complexity of the task: "Leg. Counsel rewrites it and sometimes changes it. It's kind of like translating the Bible."37
We do not suggest that Legislative Counsels are disloyal; to the contrary, we were impressed by their nonpartisan dedication to their work. But consider how this response from one of our Legislative Counsels-making the point that the disconnect between policymaking and drafting is not a problematic or unusual one for law-pinpoints the issue:
If
This description, although reasonable, is based on a starkly different set of assumptions than those that courts apply about how directly elected members, and even high-level staff, are involved in the details of statutory text.
c. The limitations of Legislative Counsel as a bridge between Congress and courts
One can imagine arguments for why interpretive doctrine should be aimed at Legislative Counsel. One set of arguments might go to the idea that statutes, as legal documents, should indeed be drafted in the language of lawyers-an argument we think open to serious challenge because statutes must also perform a democratic notice-giving function to the public. A more persuasive argument might be that, as Congress's professional drafters, Legislative Counsels are likely best situated to master the courts' doctrinal approach and incorporate it into drafting, and so, on the surface, they appear the best vehicle for Congress to use to facilitate communication with courts. Some scholars have recently argued that the courts should study and follow the interpretive presumptions listed in Legislative Counsel's drafting manuals for precisely this reason.39 Our findings, however, reveal serious limitations to the idea that Legislative Counsel is the best entity to bridge the courts-Congress gap.
i. Legislative Counsel lacks its assumed doctrinal expertise
Our respondents did report that they rely on Legislative Counsel's assumed knowledge of the rules of interpretation.40 Our Legislative Counsel respondents themselves were in fact more likely than other respondents to say that some of the canons they knew played a role in drafting.41 But we also found that, contrary to these expectations about Legislative Counsel's expertise, the Legislative Counsels we interviewed had no greater knowledge of most of the canons than the other respondents.42
In other words, our Legislative Counsels had more confidence in the idea that interpretive rules have a role to play in the drafting process, but they do not know all of those rules as well as everyone expects. These findings throw another wrench in the prospect of a courts-Congress interpretive dialogue. Across the many questions about the canons, 45% of respondents volunteered that it was the obligation of Legislative Counsel to raise the canons for their clients where the canons would be relevant to statutory drafting. Thus, to the extent staffers assume that they are in a conversation with courts simply by virtue of their reliance on Legislative Counsel, that assumption seems at least partially incorrect.
This is not to deny that Legislative Counsels may be better at flagging interpretive issues than other counsels. They also seem more focused on the courts than others and seem to have more faith in courts.43 These differences indicate that intense professional education might make Legislative Counsel a more effective bridge across the branches. And they do seem more expert when it comes to certain types of internal drafting practices. The process of identifying how certain statutes amend and cross-reference preexisting statutes, for instance, is a kind of drafting expertise, although one not focused on judicial doctrine.
ii. Legislative Counsel's inability to be the coordinating arm-"the OIRA of Congress"
Even putting aside doctrinal knowledge, Legislative Counsel does not appear to have the reach, the convening power, or even the consistency of practice to coordinate Congress's drafting process. By way of comparison, consider the
Within Legislative Counsel itself, coordination is lacking. We were told that "even Legislative Counsel is silo'd"46-the fragmenting effect of committee jurisdiction applies to Legislative Counsel, too, because each Legislative Counsel is assigned to a specific subject area(s).47 Eleven of our twenty-eight Legislative Counsels mentioned this point as an impediment to consistent usage across statutes involving different subjects or different committees, even when the professional drafters are involved; sixteen rejected the idea of consistent term usage across different statutes or different committees.48
Moreover, Legislative Counsel does not play a central role in every statute. Key amendment text is often drafted after hours, when Legislative Counsel may not be present, or under extreme time pressure, when there is no time to involve them. Still other drafters do not consult Legislative Counsel at all: there is no requirement that they do. And some present Legislative Counsel with text already drafted by others, such as lobbyists or agencies. Seventy-five percent of our Legislative Counsel respondents told us that they have less leeway to change such outside-prepared texts than when they are asked to draftfrom scratch.49
Politics also often trumps Legislative Counsel's advice. Legislative Counsel (unlike OIRA) is not permitted to intervene in political disputes.50 As one non-Legislative Counsel respondent observed, "the art of legislative dealmaking intercedes in their efforts."51
All of these limitations hamper the coordinating potential of Legislative Counsel much more so than even the professional legislative drafting bodies in other countries with which Legislative Counsel is sometimes compared.52 For these reasons, we suggest in Part IV that a more effective coordinator would be a new entity inside the offices of the congressional leadership.
2. Committee jurisdiction as a fundamental organizing and interpretive principle
Committee jurisdiction is another defining feature of the drafting process- and one that, likewise, is rarely mentioned by judges in statutory interpretation cases. Without inquiring about this topic, we received an overwhelming number of volunteered comments about how the division of Congress into committees creates drafting "silos" that exacerbate drafting fragmentation and also "turf" consciousness that incentivizes drafting to protect jurisdiction. Forty-five percent of respondents brought up the issue a total of eighty-three times, repeatedly emphasizing its importance across virtually every section of the survey. 53
a. Committees as drafting "silos"56
Congress is structurally divided into twenty-one standing committees in the House and sixteen in the
Across statutes, respondents were much more hesitant to say that statutory terms are even intended to mean the same thing. Forty-three percent of respondents told us that the presumption of consistent usage applies across statutes in related subject-matter areas precisely because the same committee is drafting.60 But only 9% of respondents told us the presumption should apply across statutes in areas overseen by different committees.61 Ten percent of respondents volunteered that, unless there was an explicit cross-reference, the presumption should actually be the opposite.62
i. Different drafting practices and manuals
Different committees deploy different drafting practices. Several committees, for example, engage in "conceptual drafting" or "conceptual markups," practices in which members debate and amend the measure using a narrative about what the text is supposed to accomplish rather than the actual text.63 For statutes passed by those committees and then reported to the full chamber, more judicial attention arguably should be paid to the narrative documents used for the committee vote, because other members often defer to that vote due to the committee's policy expertise. We also were told that regulatory committees draftdifferently, as a substantive matter, than appropriators, who are not as focused on policy issues as the subject-matter committees.64
Different drafters also have different style manuals or checklists for drafting. 65 The Appropriations Committee, as our respondents pointed out, uses a GAO Redbook, a two-thousand-page treatise that includes topics ranging from agency discretion to federalism presumptions.66 The
In addition, as noted, both the House and Senate Legislative Counsel offices have drafting checklists and style manuals68-but even those are not identical. The Senate Legislative Counsel checklist references one canon: "exclusion of surplusage."69 The Senate Manual references superfluities again and also consistent term usage.70 The House Manual recommends consultation of dictionaries and tells drafters to "[u]se [the] same word over and over,"71 but does not appear to reference any formal canons. Each manual also ends with miscellaneous rules, among them the use of the terms "may" and "shall,"72 and, in the case of the House Manual, encouraging the use of dictionary definitions.73 House Legislative Counsels also consult a five-hundred-page treatise written by the head of that office, which has one chapter that discusses statutory interpretation by courts and includes approximately a dozen canons.74
The GAO Redbook has been cited four times by the
Various commentators have recommended the use of standardized legislative drafting manuals and checklists by both legislators and courts for some time.77 All commentators have focused only on the Legislative Counsel manuals. But the proliferation and variation of these kinds of materials inside Congress would make suggestions about standardization exceedingly difficult to implement. Other committees may have resources similar to the manuals discussed, and it is possible that those resources contain conflicting advice. Indeed, the very canons mentioned in the Legislative Counsel manuals- superfluities, whole act, and dictionaries-were the same ones rejected by most of our respondents as unrealistic given other structural factors.78 These problems are compounded when multiple committees or types of drafters often work on the same bill; for instance, an appropriations statute that is reviewed by Legislative Counsel. Nor is it the case that Legislative Counsel reviews every enacted statute or that, when it does, its stylistic suggestions are followed. Given this variety, without a major reorientation of internal congressional practices, how could courts know which standardized manual to rely upon, and when?
ii. Different hiring practices: nonpartisan staff, lawyers, nonlawyers
Committees also diverge in the types of staffthey hire. Some committees have or have had nonpartisan drafting staffs, whom our respondents described as playing a different role from both ordinary committee counsel (because the nonpartisan drafting staff, like Legislative Counsel, are expert drafters) and Legislative Counsel (because the nonpartisan staff, unlike Legislative Counsel, are also policy experts).79 The
Other committees, we were told, prefer not to have lawyers working for them at all. Some respondents said the use of nonlawyers makes a difference from an interpretation standpoint. For instance: "There are different levels of sophistication in drafting-but some members don't use lawyers in drafting and it's more sloppy."81 As noted in the first Article, we did see differences between our lawyer counsels and our few nonlawyer respondents with respect to familiarity with some of the canons.82 Still other committees are composed almost exclusively of lawyers, like Judiciary. b. Committee turf guarding as a key interpretive presumption The importance of protecting committee jurisdiction and the ways in which committees specialize in particular subject-matter areas also emerged from our respondents' answers about what kinds of assumptions they make about the meaning of statutory text, ambiguity, and even delegation. The picture our respondents painted is consistent with the political science work emphasizing that the committee system incentivizes "turf guarding" at the same time that it facilitates expertise building.83 We were told that committees draftstatutes to keep matters within their own jurisdiction, even if doing so requires contorted language and not the "ordinary" language that courts presume drafters use. This turf-guarding point also has relevance for delegation. Fifteen percent of respondents volunteered comments that committees go out of their way to draftstatutes so that agencies within their jurisdiction will implement them.84 As noted in the first Article, our respondents also volunteered committee jurisdiction as an important signal of congressional intent to delegate interpretive authority when multiple agencies are given overlapping statutory duties.85 Our respondents told us, for example: "The committee wants the agency over which they have jurisdiction to have the lead,"86 and "[U]nless it's explicit it's presumed in the jurisdictional aspect. The committee that writes it has jurisdiction and the agency within its jurisdiction is the signal."87 Differences across committees seem to have other implications for delegation, too. Our respondents told us that different committees approach delegation differently. They told us that presumptions about delegation can depend on the committee's personal relationships with the agency in question (23%) or the particular subject matter of statutes under the committee's jurisdiction (60%).88it's more sloppy."81 As noted in the first Article, we did see differences between our lawyer counsels and our few nonlawyer respondents with respect to familiarity with some of the canons.82 Still other committees are composed almost exclusively of lawyers, like Judiciary. b. Committee turf guarding as a key interpretive presumption The importance of protecting committee jurisdiction and the ways in which committees specialize in particular subject-matter areas also emerged from our respondents' answers about what kinds of assumptions they make about the meaning of statutory text, ambiguity, and even delegation. The picture our respondents painted is consistent with the political science work emphasizing that the committee system incentivizes "turf guarding" at the same time that it facilitates expertise building.83 We were told that committees draftstatutes to keep matters within their own jurisdiction, even if doing so requires contorted language and not the "ordinary" language that courts presume drafters use. This turf-guarding point also has relevance for delegation. Fifteen percent of respondents volunteered comments that committees go out of their way to draftstatutes so that agencies within their jurisdiction will implement them.84 As noted in the first Article, our respondents also volunteered committee jurisdiction as an important signal of congressional intent to delegate interpretive authority when multiple agencies are given overlapping statutory duties.85 Our respondents told us, for example: "The committee wants the agency over which they have jurisdiction to have the lead,"86 and "[U]nless it's explicit it's presumed in the jurisdictional aspect. The committee that writes it has jurisdiction and the agency within its jurisdiction is the signal."87 Differences across committees seem to have other implications for delegation, too. Our respondents told us that different committees approach delegation differently. They told us that presumptions about delegation can depend on the committee's personal relationships with the agency in question (23%) or the particular subject matter of statutes under the committee's jurisdiction (60%).88it's more sloppy."81 As noted in the first Article, we did see differences between our lawyer counsels and our few nonlawyer respondents with respect to familiarity with some of the canons.82 Still other committees are composed almost exclusively of lawyers, like Judiciary. b. Committee turf guarding as a key interpretive presumption The importance of protecting committee jurisdiction and the ways in which committees specialize in particular subject-matter areas also emerged from our respondents' answers about what kinds of assumptions they make about the meaning of statutory text, ambiguity, and even delegation. The picture our respondents painted is consistent with the political science work emphasizing that the committee system incentivizes "turf guarding" at the same time that it facilitates expertise building.83 We were told that committees draftstatutes to keep matters within their own jurisdiction, even if doing so requires contorted language and not the "ordinary" language that courts presume drafters use. This turf-guarding point also has relevance for delegation. Fifteen percent of respondents volunteered comments that committees go out of their way to draftstatutes so that agencies within their jurisdiction will implement them.84 As noted in the first Article, our respondents also volunteered committee jurisdiction as an important signal of congressional intent to delegate interpretive authority when multiple agencies are given overlapping statutory duties.85 Our respondents told us, for example: "The committee wants the agency over which they have jurisdiction to have the lead,"86 and "[U]nless it's explicit it's presumed in the jurisdictional aspect. The committee that writes it has jurisdiction and the agency within its jurisdiction is the signal."87 Differences across committees seem to have other implications for delegation, too. Our respondents told us that different committees approach delegation differently. They told us that presumptions about delegation can depend on the committee's personal relationships with the agency in question (23%) or the particular subject matter of statutes under the committee's jurisdiction (60%).88it's more sloppy."81 As noted in the first Article, we did see differences between our lawyer counsels and our few nonlawyer respondents with respect to familiarity with some of the canons.82 Still other committees are composed almost exclusively of lawyers, like Judiciary.
b. Committee turf guarding as a key interpretive presumption
The importance of protecting committee jurisdiction and the ways in which committees specialize in particular subject-matter areas also emerged from our respondents' answers about what kinds of assumptions they make about the meaning of statutory text, ambiguity, and even delegation. The picture our respondents painted is consistent with the political science work emphasizing that the committee system incentivizes "turf guarding" at the same time that it facilitates expertise building.83
We were told that committees draftstatutes to keep matters within their own jurisdiction, even if doing so requires contorted language and not the "ordinary" language that courts presume drafters use. This turf-guarding point also has relevance for delegation. Fifteen percent of respondents volunteered comments that committees go out of their way to draftstatutes so that agencies within their jurisdiction will implement them.84 As noted in the first Article, our respondents also volunteered committee jurisdiction as an important signal of congressional intent to delegate interpretive authority when multiple agencies are given overlapping statutory duties.85 Our respondents told us, for example: "The committee wants the agency over which they have jurisdiction to have the lead,"86 and "[U]nless it's explicit it's presumed in the jurisdictional aspect. The committee that writes it has jurisdiction and the agency within its jurisdiction is the signal."87
Differences across committees seem to have other implications for delegation, too. Our respondents told us that different committees approach delegation differently. They told us that presumptions about delegation can depend on the committee's personal relationships with the agency in question (23%) or the particular subject matter of statutes under the committee's jurisdiction (60%).88 These differences, our respondents explained, result in varied levels of committee- agency trust and different norms for delegation to agencies depending on the committee and subject at issue.89
In the first Article, we illustrated how these findings relate to the Court's own implicit practice of according different levels of agency deference depending on the subject matter.90 Here, we go deeper. It is not only the case that different subject matters are accorded more or less interpretive room by drafters. It also has to do with whether the committee with jurisdiction over the agency is doing the drafting. As one respondent put it: "It depends on jurisdiction. If your committee has jurisdiction over [Homeland Security], you'd rather have your agency interpreting the statute."91 In other words, and unsurprisingly, drafters of statutes want control over statutory implementation. Our findings suggest that drafters will be more likely to intend delegation if the agency charged with implementing the statute is within their oversight authority.92
Other research substantiates this account. Political scientist
Committee jurisdiction is really important to how stuffgets drafted-I never learned that in law school . . . . It affects general policy approaches, leads to contorted ways of talking about things in legislation. For example, in national water policy or hydropower, the Clean Water Act is under Transportation's jurisdiction, Natural Resources has jurisdiction over oceans and fisheries. So you try to phrase the policy to keep it in your own committee.95
Our data also suggest that, because different committees have different areas of policy expertise, even well-informed drafters are likely to have varied knowledge of the policy-based canons. For example, even though 47% of our respondents said they had at some point participated in drafting a criminal statute, only twenty-three respondents from that group knew the "rule of lenity"- and fifteen of them were on Judiciary, the committee generally charged with jurisdiction over criminal law.96
One implication of this finding is that judges may wish to consider which committee is doing the drafting before applying a substantive canon. Our
3. Other staffdifferences: leadership vs. committee vs. personal staff
Legislative Counsel and committee staffare not the only statutory drafters. Congress is a big bureaucracy: there are more than 14,000 congressional staffers, 97 and Legislative Counsel and committee staffcomprise fewer than onethird of that number.98 The other main legislative drafting staffare personal staff-who work directly for elected members-and the staffwho work for the congressional leadership.
a. Personal staff: often young, nonlawyers, and with different goals
According to our respondents, personal staffhave a different job description than committee staff, which affects how they draft. Fifteen percent of respondents volunteered such differences, the most salient of which was that the work of personal staff, unlike leadership or committee staff, is targeted toward the positions and reelection of their individual members.99
These differences affect application of interpretive presumptions, like federalism. We were told, for example, "committee staffis more federal policy staff,"100 but personal staff"are drafting to their member's preferences/beliefs about federalism."101 It also affects legislative history; our respondents told us that personal staffdraftlegislative history more with an eye toward their own member's perspective or reelection needs rather than with a focus on statutory meaning.102
More generally, respondents emphasized the difference between a committee's need to find compromise and personal staff's ability to be more focused on a single member's preferences: "You are working for a specific member, not the whole committee. You focus on different things, things relevant to your district, a much narrower focus and you can be more ideologically pure."103 In the
b. Drafting by leadership-dealmaking over policy or clarity
Congressional leaders often take statutes out of the committee process to push major deals through Congress. These increasingly frequent deviations from the textbook drafting process-what
Twenty-six percent of respondents volunteered, at different points in the survey, that leadership involvement results in a less transparent legislative process, because leadership can bypass those steps of the committee process- such as hearings, markup, and committee report-that are visible to other members and the public.110 They also told us that statutes put together by leaders have less legislative history and a lower-quality version of it, because leadership both lacks policy expertise and usually gets involved when statutes become bundled deals-a negotiated process not conducive to the production of high-quality legislative history. Because omnibus bills are not usually under the jurisdiction of one committee, and because leadership often has a role in cobbling them together, the same quality reports or expert explanations do not always get produced.111 Consistent with their views that drafting by leadership leads to a different sort of legislative history, our respondents also told us that they discounted floor statements by party leaders as "spin," compared to what they viewed as reliable statements by committee leaders with more policy expertise. 112
We recognize our respondents' potential bias here, including with respect to their comments about personal staff. Our committee staffrespondents are the very people who stand to lose power in the face of greater leadership control. Other staffers in Congress might make different assessments of the value of leadership's involvement. The critical point, and one the political science literature and legal work by
These details, though deep in the trenches of the legislative bureaucracy, cannot be ignored under any interpretive theory that depends on either the idea of reflecting how Congress works, or the idea of being in communication with Congress-whether as partners, system coordinators, or otherwise-or even on the idea that there is a fictitious "reasonable drafter" on whom interpretive doctrine should be based. At a minimum, theories that ignore these details require some justification for doing so-some acknowledgment that the current presumptions are actually judicially imposed, rather than derived from congressional practice or expectations, and some defense of the judicial power to impose them.
Even assuming that the various categories of staffers have some consistency of practice, we have illustrated that the number of categories and diversity of practice across them bedevils the goals of interpretive projects based on congressional practice or expectations. We have more to say about this in Part IV, but there are other structural features that complement this account that we must first introduce below.
We also note that 25% and 34% of our respondents told us that first drafts are typically written by, respectively, the
B. Statutes Are a "They" and Not an "It," Too
We move beyond staffdifferences now to three structural points that are equally important to the themes of this Part: the type of statute, the path it takes, and its budgetary effects.
Ninety-three percent of our respondents told us the legislative process differs for different types of statutes,116 and ninety percent agreed that the process by which a statute is enacted affects how it is drafted.117 Twenty-nine percent of our respondents volunteered fifty-five separate comments over the course of the survey emphasizing, as Sinclair's work has,118 that the textbook statute and textbook legislative process, if they ever existed, are no more. The following Figure summarizes the findings:
Here, too, however, the Court does not make such distinctions,119 and scholars generally overlook them. Two important exceptions are
1. The type of statute matters: omnibus vs. appropriations vs. ordinary bills
Our findings suggest that there is an enormous difference between omnibus and non-omnibus legislation and between both kinds of legislation and appropriations legislation (which often are omnibus in nature although different in content from regulatory omnibus bills). These differences affect how our respondents read both text and legislative history.
Seventy-four percent of our respondents said that omnibus bills are less likely to be internally consistent than single-subject bills.122 More than half of that number elaborated in their comments with explanations like: "[I]n an omnibus statute, different committees put in language related to their subject matter, so the whole code rule doesn't apply."123 A little more than half of all respondents also said that legislative history plays a different role in omnibus statutes.124 Most of our respondents reported that omnibus history was often "confused" or nonexistent. Many also reported that the way in which omnibus bills are put together makes it "more likely the legislative history [that does exist] is erroneous,"125 because it often derives from the earlier-drafted, singlesubject bills that later are meshed together and sometimes changed for the conglomerate product. We were told, for example: "The omni is a kitchen sink, and it's very difficult to know with any assurance that the legislative history that comes along with it is coordinated with the statute."126
With respect to appropriations legislation, as the first Article discussed, the majority of respondents said that legislative history plays a totally different role,127 with 39% offering additional comments to explain it as more central and important. This, too, is due to an underappreciated structural feature, which Nourse has pointed out: both the
Our respondents also had the distinct impression that the use of these "unorthodox" vehicles is on the rise and that the rise is attributable to increased polarization. The political science data confirm these impressions, too.130 The following comment was typical: "Congress doesn't do single subjects much anymore because it's too hard to do legislation, so bundling is the only way to get things."131 As noted, many respondents also told us that leadership, not committees, controls these bills: "[t]he omnibus is a totally different process, it gets down to
2. The stages of the process matter
Even across statutes utilizing the same type of statutory vehicle, our respondents emphasized other process-related differences that affect the final product. Two of these differences, which should come as no surprise and have been noted by others,134 are the length of the statute and the time the drafter has to write it. As one staffer told us: "A lot of times you have one night to draftand the issue is time. Also, in the same section I should be able to pick up on [inconsistent usage], maybe not in a 200 page bill."135
But our respondents also highlighted less obvious distinctions. For example, they told us that the degree of cooperation varies in accordance with the process and affects the end product.136 Sixty-five percent of respondents emphasized that the committee process requires the most collaboration and results in "more vetting."137 We were told, for instance, that it is easier to draftalone than in a group: "If you know you don't have to endure a markup you may say things more bluntly and plainly."138 Nevertheless, 70% of respondents (although, again, we recognize their potential bias) told us that "more process" leads to a better final product.139
Our own research is consistent with our respondents' accounts of the increasingly unorthodox legislative process. As one of many possible examples, in the first year of the 112th Congress, only 7 (8%) of the 91 measures that passed went through the "textbook" process in both houses, passing through committees on each side, while 37 measures (41%) did not go through the committee process in either chamber before final passage.140 Only 3 of those 91 measures went through the conference committee process;141 the rest were worked out by leadership deals, special legislative processes such as reconciliation, or "preconference"-a process in which differences are negotiated behind the scenes by staff, and then each chamber passes the amendments necessary to make the bills identical without going through conference.142 Many of our respondents had never participated in a conference, even though most had worked on
One interesting takeaway from this set of questions is that even those working on the statutes themselves cannot always predict whether they will be able to touch, or fix, them later in the process. A statute that gets diverted through unorthodox processes deprives counsels of the chance to address interpretive issues, or errors, that may later arise. Numerous respondents mentioned health reform as an example. We were told, "[e]veryone expected the Affordable Care Act to go to conference and it didn't and so we were stuck."144 Courts rarely consider such possibilities when deciding how much weight to accord enacted text.
3. The
Our respondents also discussed other significant drafters. We do not dwell on most of these, including individual constituents and academics (each identified by 9% of respondents).145 We especially recognize that lobbyists and agencies, also mentioned by our respondents, are central external drafters who merit their own sustained treatment. We conclude this Part, however, by highlighting one overlooked actor that 15% of respondents identified without prompting: the
The picture that our respondents painted of the centrality of the CBO score offers an excellent example of how the "language" of legislative drafters differs from the language of courts, and not always in ways that would be inaccessible to lawyers if they chose to look. As we have argued, there are certainly aspects of the drafting process that courts could not capture. At the same time, Congress does have its own set of structural cues-some of them, like the CBO score, that are transparent and publicly available-that our drafters told us have a profound influence on the words they select.
Our respondents repeatedly told us that they routinely change the bill text to bring legislation within a budgetary goal: "In tax and spending programs you live and die by the score. We have a number in advance and we work back and retrofit the policy to the score. We send them draftafter draft";147 "Anything with a budget impact, we have to repeatedly go back to them to understand . . . their reading of the statute and then we have to go back and change it. This is extraordinarily widespread."148 Popular reports make similar observations. During debates over health reform, for instance, news outlets reported that "the bill's fate hinged on the results" of the CBO budget analysis and that the bill was continuously tweaked to change the score.149 Both our respondents and other commentators have observed that the centrality of the CBO score has increased since the passage of the statutory Pay-As-You-Go Act, which requires a budgetary estimate of a bill's effects to accompany all covered legislation.150
Our respondents also told us that the budget score affects decisions about how much detail to put in legislation-that is, how much ambiguity to include. This is a critical point from a doctrinal perspective because, in the courts, ambiguity triggers decisionmaking presumptions. For example, in response to our question about whether drafters use ambiguity to trigger deference to agencies under Chevron, one respondent disagreed and focused instead on the budget score: "Legislators have different incentives to leave language ambiguous either because they don't want to answer it or because it would affect the score."151 Another told us, in response to our question about the expressio unius canon (the presumption that an enumerated list is intended to include no additional elements), that the score affects how he drafts lists. He said he cannot use "catch-all" terms-which trump the expressio presumption-even if the list isn't intended to be exclusive because "catch-all terms cause us CBO problems" by inflating the score.152 Given that these counsel-respondents took this view, in many instances putting their concerns about the CBO score ahead of concerns about courts, it seems likely that noncounsel staffers would do the same.
It is a different question whether it would be normatively desirable to apply a presumption that, for instance, ambiguities in legislation be construed consistently with the assumptions underlying the CBO score, even if the score does reflect the congressional understanding of the bill. There may be democracypromoting reasons why courts might avoid interpretive rules that could further enhance what some consider the CBO's already disproportionate influence or what some view as its lack of neutrality.153 Our point is simply that within the context of current debates about which interpretive rules reflect congressional bargains, the CBO score has not even been mentioned.
III. DELEGATION AND DIALOGUE
Our findings on delegation likewise suggest that current theory and doctrine are focusing on the wrong cues and the wrong relationships. Our respondents viewed Congress's primary interpretive relationship as one with agencies, not with courts. Indeed, they strongly resisted one of the major premises of most current interpretive theories: namely, that Congress is in some kind of dialogue with courts-be it a principal-agent relationship, a partnership, or a ruleof- law relationship focused on shared, system-coordinating rules. Instead, they saw agencies as the everyday statutory interpreters, viewed interpretive rules as tools for agencies, too, and made no distinction, as some scholars have, between agency statutory "implementation" and agency statutory "interpretation."
We asked our respondents fifty-nine questions designed to elicit their impressions on these matters. Our inquiries ranged from questions about whether our respondents intend to delegate specific questions, like preemption questions, to courts or agencies, to questions about how the consistency of the Court's approach affects legislative behavior. The following Figure summarizes the central findings:
These findings have significance both for current theory and for moves away from it. As an initial matter, Chevron-the central deference doctrine- seems too court-centric. Our findings highlight the fact that Chevron, with its emphasis on text and legal presumptions, seems to assume that Congress is talking to the courts when Congress signals an intent to delegate. In fact, our respondents told us that Congress is communicating with agencies about delegation, and doing so utilizing internal and structural cues, like committee jurisdiction and legislative history directives, that Chevron's text- and court-focused inquiry does not privilege.
At a broader level, these findings put pressure on current theories of the interbranch relationships. Even the dominant alternatives to the kind of faithfulagent/ Congress-reflecting theory that our study challenges are still focused on the
A different approach might be for the courts to move the central weight of the regime to a different interbranch relationship entirely-namely, to the Congress- agency relationship. Some scholars already have advocated that courts allocate even more interpretive authority to agencies.154 But our study leaves us dissatisfied with such a wholesale delegation-based solution. Instead, we would shiftthe focus: we emphasize not court-driven transfers of power-which our findings suggest our respondents would resist-but rather how delegation theory and doctrine might look if they were situated in the very context of how Congress and agencies communicate.
A. Agencies as Statutory Interpreters
Our drafters saw their primary interpretive relationship as one with agencies, not courts. This theme emerged first in our questions about the canons- questions that we ourselves had designed as court-centric. Many respondents volunteered those interpretive tools as tools of Congress-agency communication. Thirty-seven percent of respondents told us, without prompting, that the canons are useful specifically because they help drafters predict whether and how agencies will interpret statutes.155 As an example of a typical comment: "If you know the agency will use these interpretive principles they matter absolutely because you want to know how they will be interpreted."156
We are not the first to highlight "agency statutory interpretation."
Our findings deepen this work. For instance, they lend support to Strauss's argument that agencies are important audiences for legislative history.159 More than 94% of our respondents said that the purpose of legislative history is to shape the way that agencies interpret statutory ambiguities (almost the exact same number said the same for courts and individuals).160 Moreover, as noted, 53% of respondents pointed out that legislative history is particularly important in the appropriations context, and 73% of that number explained that its importance derives from the fact that it directs how the appropriated money is to be spent.161
Another nuance we add is how legislative history helps Congress mediate its relationship with agencies after statutes are enacted. Without inquiry from us, 21% of respondents volunteered that legislative history is one means of congressional oversight of agency implementation.162 The political science literature frequently focuses on congressional hearings and administrative procedures as tools of oversight,163 but it rarely discusses legislative history for that purpose. Some legal scholars have indeed suggested that postenactment legislative materials might be useful to agencies in maintaining fidelity to Congress and the President over time,164 even as courts have generally rejected those materials. Some of our respondents specifically mentioned such "subsequent legislative history," even noting that courts and agencies might view that resource differently: "[I]t's still important after [passage] because agencies need to know what members want or expect from the legislation";165 "Agencies might use it too, but courts won't."166
We also note that our respondents did not distinguish between how agencies and courts might use the canons, with the exception of the subsequent legislative history and appropriations legislative history already mentioned. But we did not ask our respondents about these matters, and so cannot make strong claims.
B. Interpretation as Implementation-Implications for both Agencies and Courts
Our findings also support eliminating the common theoretical distinction between statutory implementation authority and statutory interpretation authority. The Mashaw and Strauss works are on point in this context, too: both authors view the types of authority as interrelated (as Mashaw puts it: "agency interpretation is part of agency policy development"167). Other administrative law scholars have tended to treat interpretive authority as something distinct: namely, the sort of authority that may belong only to courts, based on courts' unique attributes, such as independent judgment and stability, or the Administrative Procedure Act's special de novo review provision for questions of law.168
We did not hear from any respondent that drafters intend to reserve interpretive authority only for courts. Nor did our respondents distinguish between interpretive authority and implementation authority. Instead, they told us that drafters often have a desire for agencies to fill textual gaps and that such gap filling includes the details of implementation, or that legislative history is useful for influencing future interpretation and contains the details of implementation. Fourteen percent of our respondents also volunteered that statutory gaps are often necessary to keep statutes at a level of detail that is not overwhelming. 169 But all of those respondents also told us that they expected those kinds of gaps to be filled by agencies-and they did not distinguish between the agency as interpreter or implementer in that context. We were leftwith the impression that our respondents would resist the notion that Congress needs to grant agencies some "extra" authority to interpret statutory ambiguities when the power to implement the statute already exists.170
This finding also has implications for the judicial side of interpretation, where the same artificial divide persists, although not as explicitly. Theorists often distinguish between textual interpretation and substantive "lawmaking" (often in service of arguments about federal common lawmaking power).
C. Implications for Chevron
As should be evident, our respondents' narratives did not line up well with Chevron as currently constructed. But nor did their resistance to courts translate to arguments for broader delegation to or increased empowerment of agencies.
Our respondents told us that they speak to agencies in a variety of ways in addition to statutory text and that they intend to delegate with firm limitation. As discussed in the first Article, there were many kinds of questions that our respondents told us were inappropriate for agency resolution-even where statutes were leftambiguous-including major policy questions and questions involving particular subject matters.175 Furthermore, our drafters told us that "at the end of the day, you lay out the policy issues from which the agency crafts regs, but you don't want to create open-ended authority."176
These findings suggest that the Court has the right idea insofar as it has tried to narrow Chevron from applying wherever ambiguity exists to only when it is most likely that Congress actually intended to delegate.177 At the same time, they suggest that Chevron's Step One is far more textualist than the way in which our respondents described how Congress communicates with agencies. Relatedly, courts applying Chevron appear to view Step One's tools-like canons and legislative history-as judicial resources, rather than resources that Congress first and foremost may intend as cues for agencies.
Our respondents told us that Congress uses structural signals, like committee jurisdiction, to tell agencies when and to whom it intends to delegate.178 Indeed, they suggested that the link between delegation and oversight on which Chevron depends may be much more specific than the doctrine acknowledges- i.e., it may exist only when the committee with jurisdiction over the subject matter also oversees the implementing agency. Jurisdiction currently plays no role in the Chevron inquiry.
Congress also uses legislative history to communicate with agencies, but in ways that escape current doctrine. It is not that legislative history has been ignored by the Court in the Chevron context; the Court remains divided over its use in Step One as a tool for discerning ambiguity (just as it remains divided over the use of legislative history in general). But that debate seems misdirected. If drafters use legislative history to speak to agencies, then the current use of legislative history in Step One-as an aid to judges in determining textual clarity-may be too limited. The presence of instructions to agencies in the legislative history itself seems to be its own good signal of delegation, separate from textual ambiguity. Nor has there been much discussion of legislative history at
There has been a related battle over whether various canons of interpretation belong in Step One or Step Two.180 But, likewise, the real question seems to be which canons Congress-and the agencies it is talking to-knows in the first place.
Some judges, particularly textualists, might resist this more contextual, agency-focused approach to interpretive deference. But the Court itself has given us a reason to adopt it. In Mead, the Court made clear that Chevron turns on congressional intent, so how Congress signals that intent should matter.
Of course, this means that courts must better understand how Congress communicates with agencies. We do not know, for example, which canons agencies know, and we cannot verify what many scholars have suggested about agencies' superior competence to utilize legislative history. We also do not know how some of the other structural influences that we have identified affect agency implementation. For example, do agencies take the type of statute into account more than courts and so discount legislative history in omnibus vehicles? We wonder also how agency decisions relating to delegation change depending on the type of staffinvolved. For instance, personal staffare unlikely to approach drafting with the same focus on agency oversight as committee staffwhose job it is to oversee the agencies themselves. Do statutes drafted by personal staffreceive less agency input and delegate less frequently to them? With respect to unorthodox lawmaking, does the fact that leadership, rather than committees, manages those statutes reduce or increase the agency's role in the drafting process?185
We cannot offer answers to these questions, because we did not inquire about them and did not interview agency staff. What we can say is that Chevron, in its current form, displaces the idea of a Congress-agency conversation in favor of a court-centric approach that utilizes interpretive presumptions unknown to some drafters and ignores other signals that some drafters do employ. It also seems clear that the conversation on "agency statutory interpretation"186 has only just begun.
D.
Our respondents painted almost the opposite picture of Congress's interpretive relationship with courts. They viewed courts, at best, as interpreters of "last resort"-or worse, as interpreters whom Congress does not even think about or whose input was unwelcome. At the same time, our respondents told us that a shared interpretive language between the branches was desirable, but they blamed the courts for the lack of common rules.
1. "Congress never wants courts to decide"
Only 39% of our respondents said that drafters leave gaps in statutes out of a desire for courts to fill them, compared to 91% for agencies.187 Seventeen percent of respondents volunteered at various points in the survey that drafters do not even think about courts when drafting; and 23% of respondents volunteered that drafters affirmatively prefer that courts not interpret their statutes at all.188
Some of these comments were expressed through a preference for agency over judicial interpretation. For example: "The desire is more for the administration to fill the gap. If a court fills the gap, we're probably in trouble."189 But other times, the comments were directed at courts alone. For example, in response to questions about whether drafters were ever deliberately ambiguous about preemption, numerous respondents offered answers of this order: "Saying the courts will just figure it out, I've never really seen that";190 "The last thing we want is for courts to decide what your law means."191
We again acknowledge the potential bias of our respondents. As counsels for committees that typically oversee agencies, they may disproportionately prefer agency to judicial interpretation more than other staffers. Of the 39% of respondents who did say they leave ambiguities for courts to fill, it may not be surprising that most worked for entities without an oversight responsibility: 40% were Legislative Counsels, 23% were from Judiciary (which oversees the
2. "It's a dance as long as we all know the steps"194
At the same time, 20% of our respondents volunteered the importance of courts and Congress being on the "the same page." That is, their resistance to judicial interpretation was not also a resistance to interbranch dialogue.
The vast majority of our respondents told us that utility of the canons was directly related to how consistently courts apply them. They told us, "it's more an issue of providing some basis for a more consistent feedback loop between courts and Congress." Seventy-three percent explained that the ideal number of canons was "about the same" as the current number or "didn't matter" as long as their application is predictable.195 Typical comments included: "They should simply be followed! . . . One problem is the perception that the Court is selective in application rather than objective in application."196
In total, 81% of our respondents said that it would or does affect the way they draftif they knew that the Court applies certain interpretive rules consistently. 197 Thirty percent specifically accused the courts of inconsistency or being result-oriented.198 For example: "If I believed the Court wasn't political, and had rules for 200 years, I would try to write statutes that complied with those rules"; "If [the rules] were enduring, not just product of a 5-4 court."199 Seven percent of respondents invoked
3. Consistent rules do not have to reflect how Congress drafts
Finally, we were surprised by some respondents' suggestions-without mention of the topic from us-that the canons might have value as systemic coordinating devices even if they cannot be grounded in anything resembling the legislative process. Fifteen percent of respondents emphasized that "the idea there is an intent of Congress is crazy," but still urged judicial consistency in interpretation for rule-of-law reasons. In response to our question whether the canons help courts to effectuate congressional intent, many offered answers such as: "I'm not sure how to know congressional intent but [the canons] are helpful to create a systematic way to establish what courts will use."201
This series of findings is potentially big news for dialogue-based theories of interpretation. Understanding Congress as being in primary dialogue with agencies rather than with courts offers not only another normative-and democratically legitimate-justification for Chevron deference, but also powerful support to the idea that courts should interpret statutes in the context of that central agency-Congress relationship and in ways that facilitate that relationship. It also makes us wonder which, if any, interpretive conventions are actually shared among all three branches. A number of respondents specifically mentioned agencies, too, when talking about consistent interpretive rules. For example, "All I want is clear rules-I don't need a million but I want to know in advance how the statute will be interpreted first by agency, then by court."202
At the same time, great swaths of modern interpretive theory rely on the notion-which our respondents rejected-that Congress tends to delegate at least in some measure to courts. The most explicit instance of this assumption about delegation to courts exists in the common law statutes context, discussed above.203 But other interpretive theories implicitly rest on similar assumptions.
Our respondents rejected that kind of judicial help. This does not mean that they rejected the idea of a courts-Congress relationship. Indeed, our respondents' assertions that consistently applied interpretive rules could, in fact, coordinate interbranch behavior is a different kind of finding that should not be overlooked. Many legislation theorists have long claimed that Congress has neither the desire nor the ability to coordinate with courts.209 Our study is the first to offer some evidence to the contrary. The first Article relayed findings about a few canons that indeed are getting through to some drafters-our respondents told us they use those rules precisely because they know courts use them.210 The findings in this Part suggest the possibility that drafters might welcome even more coordinating rules-perhaps especially when agencies are not in the picture. But our respondents drew a distinction between coordinating rules and the idea of judges as interpretive "partners," even implying that one attraction of shared interpretive conventions is that they would limit judicial discretion.
A different question is whether it matters what Congress wants. It is another question why, even if shared conventions are the goal, our drafters put the full onus on the courts to establish them. And yet a third question is whether, given the fragmentation that we described in Part II, any set of interpretive conventions could be sufficiently incorporated and standardized across Congress in the first place. We turn to these and other theoretical and doctrinal implications of our study in the next and final Part.
IV. THEORETICAL AND DOCTRINAL IMPLICATIONS
Our findings have implications not only for what the doctrines should look like but also for what theory of the judicial role should underlie them in the first place. The study has particularly profound implications for those who wish to retain the theoretical structure of the current regime-under which the stated goals of interpretive rules have been to reflect or to affect how Congress drafts and interbranch communication is assumed. We have exposed a significant mismatch between the Court's doctrines and Congress's practices that, at a minimum, should change the way any theorist who wishes to reflect Congress reads a case. More fundamentally, the variety that we uncovered reveals that any such theory, at best, can be only partially complete.
Our findings pose different challenges for interpretive theories less trained on the intricacies of congressional practice, although such theories may be liberated from the complexity of the details that we uncovered. For example, a rule-of-law approach-one that aims for clear rules to coordinate systemic behavior- depends on communication and consistency of practice from both branches that we have found utterly lacking. A broadly purposive or partnership model-one that views judges as pragmatic partners who sometimes must go beyond the statutory text to effectuate a statute's broader purposes or evolving norms-may raise concerns about democratic legitimacy if Congress does not welcome the assistance or does not do its share as a "responsible partner" in the dialogue.
As a practical matter, we think courts are unlikely to abandon the most common version of the faithful-agent model-that the doctrines should reflect Congress and that judges should assume Congress drafts in the shadow of those rules-even though it will be impossible to fully effectuate. The pull of this faithful-agent premise derives from the persistent discomfort that judges have with admitting to "lawmaking" in the statutory context, a distinctly modern sensibility that we do not see vanishing anytime soon (although we think it is overstated).211 But if a move toward a more reflective interpretive regime is the goal, such a move will require a departure from the operative assumption that one set of interpretive rules can apply to all kinds of statutes. This departure would not be as novel as it may seem: as we shall illustrate, the Court already appears to be trending toward circumstance-specific "interpretive tailoring,"212 even if it has not explicitly acknowledged it.
We address these matters as they relate to Congress, too. Scholars and judges focus almost exclusively on how courts should fashion doctrine, but not on whether that answer turns on how Congress itself acts. If Congress is not a responsible partner, can courts be faulted for their efforts to translate congressional practice as best they can-even if the result is an overly legalistic, sometimes inconsistent, set of rules poorly matched with congressional practice?
Stated differently, given that a constitutional vision compels, or at a minimum legitimizes, each of the main theories' visions of how courts should exercise their interpretive responsibilities,213 why has that conceptualization not been extended, at least as a matter of theory, to Congress's side of the equation? We question whether courts could formally enforce such a congressional obligation, but we do think there are some efforts Congress itself could attempt, be it coordinating or standardizing some drafting practices or changing other internal drafting norms, to respond to courts' assumptions.
At the same time, we have serious doubts about whether courts really want the dialogue at all, despite the ubiquity of assertions that dialogue forms the basis of-and lends democratic credibility to-the dominant theories. As
This final Part offers our preliminary intuitions on these matters. Our goal is not to leave the reader dispirited. Rather, we wish to stimulate more honest and explicit consideration of what interpretive doctrine is supposed to do and where it comes from. Our goal is also to offer a set of alternative frameworks for analyzing the real-world details we have uncovered, at the same time acknowledging that we leave much for future work. As one important example, there is an overarching question about the value of empirics that we do not attempt to answer and that has not yet been deeply explored even in the empirical legal studies literature; that is, how valuable can such work be to legal theories ostensibly grounded in actual practice when much of what is uncovered is too complex or otherwise impossible for doctrine to absorb? For instance, courts obviously will not be able to incorporate the information about individual staffreputations that our respondents told us affect how they interpret statutes. Does that mean the rest of the game is not worth the candle?
A. Directions for the Courts
Putting aside for the moment any assistance that Congress could offer, in this Subpart we offer three possible theoretical responses to our findings for courts. First, we work from within the confines of the current regime, because as a practical matter we do not believe that a more fundamental theoretical change will happen overnight. We offer some low-hanging doctrinal fruit-in the form of a new focus on the structural and process-related cues that we have identified-that could better accomplish the stated goals of both textualism and purposivism. But then we move to alternatives and explore both a rule-of-law approach and a broader conceptualization of purposivism as responses to the difficulty of the task of having interpretive rules fully grounded in the details of congressional dealmaking.
1. For the current paradigm: a reorientation around structural and process-related influences
We begin with improvements for the current regime. Our findings about interstaffdifferences make clear that any set of interpretive doctrines based on how Congress now works, or what Congress now knows, will entail an unavoidable choice about which staffers or features of the drafting process the theory is going to privilege. This is a choice that courts are already making-we have shown that current doctrine privileges the lawyerly perspectives of the court- and text-focused Legislative Counsels-but not at a conscious level.
We would suggest a different choice, focusing instead on the structural cues that our respondents identified, including committee-related signals, agency relationships, type of statute, the path through Congress, and the CBO score. We defend this choice for reasons similar to those underlying our decision to focus on committee staffin our survey design in the first place: the centrality of those structural features to the actual policymaking decisions made by members and high-level staff.216 We read current textualists and purposivists alike as purportedly focused on those same decisions-textualists contend their approach best reflects legislative bargains; purposivists argue theirs captures congressional intent-but a canons-based approach, or one that eschews anything external to the text, does not seem to be the way to get there.
We recognize the limitations of our findings. More study is needed of other types of staffand elected members to determine whether our committeefocused account is generalizable. So, too, more research is needed into the role of committees themselves and the particulars of their relationships with Congress as a whole. For present purposes, however, we take some comfort in the lopsided nature of many of our findings (46 questions had more than 70% of respondents agreeing on a particular answer choice, and 25 had more than 90%)217 and the fact that the accounts offered by our Legislative Counsel respondents did not differ, on most matters, from the accounts offered by our committee staff.
a. Committee jurisdiction, type of statute, process, and the CBO-some examples
Let us now consider how some of the structural and process-related features we have identified might be incorporated into an interpretive approach that is still relatively formalist and within the bounds of judicial competence. With respect to committees, for example, courts might apply a presumption that ambiguities in statutes should be construed to retain jurisdiction with the drafting committee. Courts also might reject consistent-usage presumptions if statutory sections were drafted by multiple committees.
As one example of how a familiar decision might have been affected by such a committee-focused approach, consider
Courts similarly could incorporate the salience of committee jurisdiction into Chevron-Mead. Our respondents told us that the assumptions about delegation and oversight that justify Chevron are particular to the committee-agency relationship and that that relationship is an active and interpretive one. We take these comments as suggestions not only that presumptions of delegation should be stronger where the agency in question is overseen by the drafting committee, but also that courts seeking evidence of delegation should use the signals that committees themselves use to communicate with agencies.
Here, another well-known case, Gonzales v. Oregon-in which the Court rejected the Attorney General's interpretation of the Controlled Substances Act as criminalizing physician-assisted suicide-offers an example.219 The interpretive difficulties with which the Court grappled in concluding that the decision should lie with the Secretary of
Legislative process-related interpretive presumptions also seem ready for the taking. Omnibus, appropriations, and single-subject statutes are not alike, nor are their legislative histories. Consistent usage presumptions might be discarded for omnibus statutes, and courts might discount the value (or absence) of omnibus history and pay more attention to appropriations history. Courts might consider whether statutes went through committee or other stages of the "textbook" process before imputing too much intent to apparent errors, or to the presence or absence of legislative history. And we believe that construing legislation consistently with the CBO score could help courts reflect congressional expectations in resolving disputes that implicate the score, given the centrality of those calculations.223
Indulge us two more examples from familiar cases. In one of the most famous statutory interpretation cases involving an appropriations statute,
And with respect to our suggested "CBO canon" and unorthodox lawmaking, litigation is currently pending over whether certain tax subsidies were intended under some especially sloppy sections of the health reform statute. The answer significantly affects the budgetary impact of the law. Given the centrality of the CBO score to the drafting of that statute, construing the statutory ambiguity consistently with the assumed score-as opposed to focusing on possible errors in the text (which, recall, our respondents told us no one had the chance to fix because of the statute's unpredictable legislative path), seems an obvious, and more easily ascertainable, way for a court to reflect the legislative bargain.225
There are qualifications, of course. We recognize that some structural and procedural features would be more accessible to courts than others. Selective application of, for example, federalism canons only to committees engaged with federalism would no doubt be more difficult than disregarding presumptions of consistent usage where multiple committees are involved or giving legislative history different weight depending on whether an omnibus statute (less) or an appropriations statute (more) is under consideration. It is for that reason we have emphasized the most transparent structural and process features at the expense of the more subjective factors that our respondents also told us play a role-choices that necessarily make our suggestions only an improvement upon, not a "cure" for, the faithful-agent problem. We also have not suggested differentiations based on the type of staffer, even though those interstaffdifferences were major findings of our study. We have doubts about the feasibility of different interpretive presumptions for statutes drafted by different types of staff. Intriguingly, courts do make such distinctions with respect to contract interpretation- namely, between sophisticated and unsophisticated parties,226 but there are many more categories of different drafters in Congress, and identifying them is not always possible.
Nor should our findings about staffers' knowledge and use of the current canons, even if generalizable, be understood as the final word. As the first Article described, our respondents who had taken a legislation course in law school were more familiar with many of these rules, and so advances in legal education could change the baseline. So too, factors like the particulars of committees' personal relationships with agencies undoubtedly will change over time. Any empirically grounded theory of interpretation will face this problem of keeping up with changing circumstances (which itself may be another strike against such theories). The structural and process-based factors on which we focus, however, are long ingrained in congressional practice and are not likely to change without courts or litigants noticing-if they are looking. Further, we recognize that there may be normative reasons to reject doctrines that emphasize factors like committees, unorthodox lawmaking, and the CBO, perhaps for the reason that those factors already have disproportionate influence inside Congress, in ways that may undermine democracy. Such concerns, however, implicate a different set of justifications for the judicial role in statutory interpretation- justifications based on using legal doctrine to improve upon, not merely reflect, the legislative process. As we illustrated in the first Article, federal courts have been loath to openly embrace that role,227 and part of our goal here is to encourage more explicit acknowledgement by courts of such motivations, if they are actually driving judicial practice.
We also would reduce the canon clutter by eliminating the most extinctionworthy ones. In addition to presumptions of consistent usage, as we argued in the first Article, the rule against superfluities, dictionary use, clear statement rules, reliance on certain types of legislative history, and some of the administrative deference doctrines seem ripe for elimination if a drafting-based model is the goal. It does not seem a coincidence that a forthcoming study of congressional overrides by
Finally, it is not our claim that the structural presumptions we introduce will always be able to resolve the kinds of narrow textual disputes often at the center of the courts' attention. Part of our effort is to push courts back to a broader vantage point. But even in the context of such disputes, the factors we identified should have persuasive power at least to rule out certain arguments. For example, without countervailing evidence, we would give little import to the absence of confirmatory legislative history in omnibus bills, or to the use of similar or different terms in unrelated statutes, or to claims of delegation to one agency when another was within the jurisdiction of the drafting committee.
i. Limitations and implications for textualism and purposivism
Even these incomplete amendments are advances on what textualists and purposivists are doing now-and ones that we believe to be consistent with the stated aims of both theories. Textualists see as their goal the judicial "decod[ing of] legislative instructions"230 and emphasize a formalist approach in the face of the difficulties of discerning legislative intent. Textualism has trained all of its focus on one particular type of formalism-the vote on statutory text-and there are democratic legitimacy reasons (the vote) for doing so. But textualists also claim that their theory's democratic legitimacy rests on how well it effectuates the legislative bargain, and our study calls into question the conclusion that text is always the best evidence of the deal. The structural and process features that we have emphasized are clear, publicly observable criteria consistent with a formalist approach that may help to better effectuate textualism's goals.
Those textualists who instead argue that fair notice to the public justifies the theory might fare better. But we note the complete absence of empirical work substantiating the idea that average citizens interpret language in accordance with the canons. We have suspicions that many would suffer a similar fate as those in our study (although we would not expect the public to be aware of the structural influences we identified either). Does the average citizen not repeat herself for emphasis or to "cover all the bases"-our respondents' reasons for rejecting the rule against superfluities? Does the average citizen focus on grammatical rules, such as the "concessive subordinate clause," as one of the Court's recent decisions did,231 or the policy presumptions that even our lawyer- drafters did not know?
Purposivism is a more complex case, because the theory has more permutations. As a general matter, most purposivists today begin with text but also aim to use contextual evidence to discern congressional intent. Our study reveals how short purposivism falls of those goals, as most purposivist interpreters use few of the contextual signals that our respondents identified. Indeed, the purposivists on the Court (like the other Justices) have displayed a surprising unawareness of even simple contextual details about the tools they do use-like how conference reports are produced.232 (Consider that the composition of the Court has not included a former member of Congress since
But there is wide variety in what else animates purposivist theory. On the narrowest version of the textualism-versus-purposivism debate, the question is simply which tools best effectuate the shared goal of discerning textual meaning. The fight over legislative history use has been the most contentious, but the premise of the two theories is often shared: both assume that there is some way for legal doctrine to reflect the congressional bargain. We attribute this shared premise to
But there is a broader version of purposivism-which some might call a "partnership" theory-that diverges from this shared premise, moves away from a deal-specific focus, and sees the courts in the broader role of interpreting statutes to engage their overarching objectives.234 The kind of "pragmatism" or "realism" espoused by judges like
Our findings confound the theoretical assumptions of each of these interpretive models if (and this is a big if) it matters what Congress wants. Our respondents were not textualists. Eighty-one out of 103 respondents238 told us that they did not think that courts, when they must interpret statutes, should consider text alone (although eight of these emphasized that text must come first).239 Half of those respondents worked for Republicans, and so the answer cannot be attributed to politics, even though in the judicial realm the division between textualists and purposivists is often politicized. Indeed, there also was wide agreement-across 92% of respondents-that legislative history should be considered in addition to (but well ahead of) the other canons.240
But nor, as we have detailed, did our respondents embrace the model of courts as delegates or partners, in the broad purposivist sense. In the end, we were leftwith the impression that our respondents preferred some combination of textualist and purposivist values-a constrained, consistent judicial approach that was more contextual than either theory currently is.
ii. A defense of the current approach as a "best effort" without Congress's help
A different way to understand our findings in the context of the current paradigm is to see what is often taken as judicial interpretive inconsistency as, instead, the courts trying to do the best they can in the face of the kind of complexity we have identified, and in the absence of direction from Congress.
Seen in this light, our findings offer friendly amendments to the eclectic factors that Eskridge and Frickey, among others, identify as already utilized by the courts. Our findings not only suggest some additions (e.g., committee jurisdiction) and subtractions (e.g., a good number of the canons), but also might inform how such eclectic theorists weigh other factors, like text and legislative history, that they already consider. Perhaps one cannot fault the courts for being law-centric or inconsistent in these efforts at translation if they must go it on their own. But if courts are going to continue to insist that these efforts are an attempt to reflect congressional practice, our study offers evidence of a better set of cues.
2. Alternatives to a Congress-reflecting approach
In the alternative, courts could put the goal of mirroring congressional decisionmaking to the side. The broader purposivism that we have discussed is one version of such a response to the legislative complexity we have uncovered, although it is not usually conceptualized as a theory shorn of congressional engagement. Another is a rule-of-law approach-a set of clear rules designed to advance values such as coordination, reliance, and notice-rather than rules that necessarily reflect how Congress drafts. Still another response might be for courts to transfer authority to different actors who are in a better position to communicate with Congress or understand Congress's internal cues. Each of these alternatives might be responsive both to our findings and to the longappreciated difficulties of discerning collective legislative intent-difficulties that our findings make even more apparent.
a. Rule of law
The dominant theories each espouse rule-of-law values-textualists, in particular, argue that their theory advances legal coherence, fair notice, and predictability. 243 But both textualists and purposivists still connect the rule-of-law features of their theories to Congress. Textualists, for example, argue that the linguistic presumptions both reflect how Congress drafts and also serve rule-oflaw values.244 Our study breaks that link, and so forces consideration of a ruleof- law approach at least partially divorced from congressional practice.
We foresee an uphill battle in getting substantial support for such an approach. It all comes back to the question-and fear-of exercising independent judicial power. Acknowledging interpretive doctrine as judicial creations, rather than as reflections of Congress, means acknowledging a lawmaking role for courts with which many are uncomfortable. Modern federal judges tend to resist the idea of making any kind of "federal common law" and their concerns are generally heightened in the statutory context, where legislative supremacy looms large.
We do not share the concern with this kind of judicial lawmaking, but in any event we question whether a set of doctrines that aims, but fails, to reflect congressional practice really has superior democratic bona fides over a set of doctrines that makes no such reflective claims and instead advances the different, but still democratically important, goals of a rule-of-law regime. Moreover, textualism, as currently applied-because its doctrines cannot entirely be derived from Congress-is already a form of judge-made law, even though it is not acknowledged as such. We also believe that a set of interpretive rules, even if divorced from congressional practice, could gain democratic legitimacy if they were conventions that gave notice to Congress and of which Congress became aware.
Another challenge for the rule-of-law model is that, in practice, the Court has not been particularly preoccupied with those values despite its stated fidelity to them.
Scholars focusing more generally on questions about how to strike the right balance between simplicity and complexity, and rules and standards, in legal doctrine have emphasized the importance of this audience question.249 The tradeoffbetween realism and doctrinal complexity-as well as decisions about what interpretive assumptions make sense-will differ depending on whether the intended audience for doctrine is lawyers, nonlawyer staffers, agencies, or the everyday public.250 Any interpretive regime intent on coordinating systemic behavior needs to be clearer about who exactly it is coordinating.
Even in the face of these obstacles, our study offers some reasons to consider a rule-of-law regime. Our findings offer support for the long-held idea of the "law/politics" divide251 and the notion that courts and Congress may never be able to speak the same language. We saw this divide not only in our drafters' comments to this effect but also in how apoliticized interpretive methodology seems in the minds of our respondents. We saw almost no statistically significant differences across the 171 questions in our survey between respondents working for Democrats and those working for Republicans. This is not to say that our respondents did not discuss politics. To the contrary, politics was often foremost in their comments, but it was politics in its raw form-not politics mediated through methodology, as seems to be the case for courts, likely because courts cannot comfortably engage in politics directly. If this divide is unbridgeable, consistently applied and loudly communicated principles of interpretation may best be justified not on the ground of how well they capture congressional practice, but on the ground that some common language is necessary for the branches to interact.
b. Transferring authority to agencies
Finally, some scholars have suggested that courts respond to this divide by transferring more interpretive authority to agencies. Most provocatively, both
Our findings support the intuitions about the centrality of the Congressagency relationship. Nor do we quarrel with many of the competence and efficiency arguments for agency interpretation-although our study does point to the conclusion that agency competence depends on which agency, and which subject, is at issue. But to the extent these arguments are based on the different democratic value of congressional intent-the value the Court has most strongly embraced-our study suggests that this kind of presumptive allocation may not be the answer. Our respondents resisted the idea of broader delegations to agencies, emphasized the limitations that Congress puts on delegation, and even would have narrowed some of the deference doctrines currently in deployment. Our respondents also would likely resist transfers of authority to agencies that their committees do not oversee. Our impression was that our respondents want to be in control not only of the scope of the delegation but of the decision to delegate itself. A judicially imposed transfer of power would be inconsistent with that understanding, and so might raise democracy concerns, especially if one views the delegation decision as one that is Congress's to make.
Instead, our findings show more support for the conclusion that courts should focus on facilitating those relationships with agencies that Congress itself initiates. The Court has indeed moved recently toward an approach that gives more weight to an agency's own interpretation of the scope of its authority, but that deference still is triggered by textual clarity rather than by the signals that Congress actually uses to communicate with agencies and may go too far to the extent it relinquishes judicial power over those questions.254 Instead, courts might attempt to better understand those signals as part of the judicial role (which we believe should continue) in determining the scope of the delegation. Scholars have variously described this as an "umpireal" role for courts or a role for courts in defining the "space" or "zone" of delegation;255 we add to that account an argument that judges in that role should focus on different cues.
Having already offered numerous examples of the Congress-agency cues that courts have overlooked, we leave this discussion with just one more, from the tax context. The "Blue Book"-an explanation of the tax laws written by the staffof Congress's
Congress certainly could be a more responsible principal, partner, or system co-coordinator than it is right now. There was a "passing the buck" feel to virtually all of our respondents' comments about Congress's obligations; they repeatedly told us that they expect others to handle the coordination with courts. At the same time, they were not comfortable with the disconnect between judicial and congressional practice-but did not seem incentivized to act on it.
For example, most of our drafters told us that it was the job of Legislative Counsel and the Judiciary Committee to know the canons, and they assumed superior doctrinal knowledge on the part of those players. However, our findings did not corroborate this assumed doctrinal expertise. Typical comments included: (from a Judiciary staffer) "we probably consider the canons more than other committees on Judiciary";258 or (from a non-Judiciary staffer) "Leg. Counsel are good and they raise the [canons] issue if there is one-we aren't the Judiciary Committee."259 On this view, there is a basic accountability problem: drafters understand the importance of coordinating but are mistaken that someone else is getting the job done.
Even assuming these experts could be better educated, there are bigger obstacles to Congress's ability to communicate with courts-apart from the basic fact that elected members and congressional staffhave more pressing matters at the forefront of their thinking. One major concern is the decreased ability of Congress to override judicial statutory interpretation cases. The possibility of an override has always been understood as the ultimate safeguard against judicial misinterpretation and is what some scholars have argued lends democratic legitimacy to interpretive theories that go beyond reflecting narrow legislative deals-including the broad purposivist vision that we detailed above.260 New empirical work reveals that increased gridlock and polarization have dramatically reduced Congress's ability to speak to courts in this way.261
Overrides, of course, are an ex post remedy-they do little, from a rule-oflaw perspective, to coordinate systemic behavior in advance, and they are a costly and combative way of conducting an ongoing interbranch conversation. 262 Noted jurists for years have suggested other mechanisms of dialogue, but most have likewise been ex post remedies, such as new offices focused on bringing judicial decisions to congressional attention.263 Even those who have suggested ex ante remedies have mostly conceptualized those ideas as aids to the judicial effort rather than as independent arguments about Congress's own obligations. We focus here instead on those ex ante obligations.
1. Coordination and standardization through leadership
More internal coordination of congressional drafting practices would facilitate interbranch coordination even if Congress is not able to communicate directly with courts more than it already does. The few academic suggestions in this vein have focused on Legislative Counsel as the best hope of achieving such standardization. But as we detailed in Part II, our findings suggest that those hopes are misplaced.
Instead, the greatest potential seems to lie in new entities in the offices of the congressional leadership. Given the drafting fragmentation that we identified and the political considerations that affect drafting (such as committee turf guarding), an office that is centralized and has convening power and political clout-three things that Legislative Counsel lacks-is necessary to do any coordinating work.
A leadership-based congressional entity might take a holistic view of statutes that are the conglomeration of multiple committees' and different types of staffers' work. It might be charged with imposing consistency of drafting style and conventions, and resolving upfront, rather than leaving to inside cues, questions about turf and delegation that affect drafting in subtle ways often missed by courts.
We suspect that our respondents would resist such an entity, even if it existed in both the majority and minority leadership offices. Such an entity would continue the expansion of leadership power over committee power and would exacerbate the problem that our respondents identified about non-policyexperts having control over the ultimate details. If housed in both the majority and minority offices, the text-drafting process also would shiftfrom a nonpartisan process to a partisan one, removing what some viewed as a particular benefit of Legislative Counsel. These are significant concerns, but the fact that these aggregate statutes already are often taken out of committees' hands and shepherded through Congress by leadership "unorthodoxly" is a further reason to consider a coordinating entity centered in those offices.
As an alternative, we note that the
2. Change internal drafting norms to reflect judicial practice
Alternatively, Congress might more directly adjust its own internal structures to reflect the preferences of the Court's current text-focused regime. We were leftwith the strong impression that our respondents felt that legislation had to "look" a certain way. They often mentioned internal drafting norms of formal language, omitting explanatory examples from statutory text, and the ubiquity of cross-references that, in turn, make legislation difficult for even expert policy staffto understand and lead to the use of legislative history for important explanatory information.265
Changing those internal drafting norms, or perhaps even transferring more text-writing power to policymakers and away from non-policy-oriented drafting staff, as our respondents suggested, might remedy the disconnect we saw between text and policymaking. Closing that disconnect would be responsive to judicial doctrine because it would make the text a more reliable indicator of congressional intent.
There are precedents for such approaches. Several committees have or have had nonpartisan policy experts on staffto drafttheir legislation.266 Some states have laws prohibiting the use of internal cross-references in statutes, or requiring a description of the cross-referenced section.267 Some scholars already have suggested that elements of legislative history be incorporated into enacted text,268 or that measures be taken to ensure that members and high-level staffactually read the statutory text.269 Some states already have laws to that effect, including nonwaivable requirements that bills be read aloud to the full chamber and requirements that text in its final form be presented to members at least several days before the vote.270
3. Look to Congress for more direction when delegating to courts and agencies?
Finally, the Congress-side analogue to the suggestion that courts should transfer interpretive power to agencies is the argument made by some scholars that Congress itself should more aggressively transfer that power.271
It was not our sense that our respondents saw broader delegation as the solution to the
We question whether consensus could be reached on a set of "interpreter default rules" that Congress could give to courts.275 But we note that Congress already routinely legislates some statute-specific "rules of construction" (e.g., the directive that the federal racketeering statute be "liberally construed") and some default interpretive rules already appear in Title I of the U.S. Code (e.g., the directive that all references to "persons" in the U.S. Code be interpreted to include corporations).276 The stakes of an interpreter default rule would be much higher, especially if it were generally applicable rather than statute specific.
A harder question is how Congress might be pressed to make these efforts. It is one thing to suggest that Congress has a constitutional obligation to communicate more with courts. But it is another to say that such an obligation, if it does exist, is enforceable by courts (by, for example, invalidating statutes that lack clear interpretive guidance). Courts and scholars have almost universally resisted the idea of "due process of lawmaking"-the notion that courts have a role in making Congress more deliberative or otherwise improving the legislative process.277 Judicial enforcement of Congress's obligations as principal, partner, or co-coordinator would raise the same separation of powers, political question, and standing concerns.
4. Do courts really want an interpretive dialogue with Congress?
We cannot conclude this discussion without advancing our doubts that the courts would really welcome Congress into an interpretive conversation in the first place. Concluding that Congress has obligations to better communicate or to write standardized drafting manuals for courts to follow implies that Congress's actions could affect judicial interpretive practice. But other courts have resisted such efforts. Many state legislatures have passed statutes dictating rules of interpretation for courts to follow. State courts continue to ignore these rules; some have struck them down as a violation of separation of powers.278 Scholars remain in heated debate about whether Congress could legislate such rules for federal courts.279
We suspect that courts desire the democratic imprimatur of ostensibly utilizing rules that reflect congressional practice, but do not actually want direction over how those rules should be utilized or which should be deployed. It may be that interpretive methodology is just too personal to judges for Congress to influence. Or it may be that interpretation is so central to the idea of Article III power that courts are not willing to give it up, even in the Age of Statutes.282 But in reality, the Court may not want to be in a dialogue with an active principal, partner, or co-coordinator at all.
C. It's Happening: The Court Already Quietly Tailors Interpretive Rules to Particular Circumstances
Hopefully the reader is still with us, and not offto invest in a sausage factory. This final Subpart makes the case that the Court already has opened the door to a more circumstance-specific, sometimes even Congress-driven, approach to interpretation than is commonly acknowledged. So understood, our pragmatic doctrinal recommendations for better tailoring doctrine to the cues of congressional practice may be viewed as the next step in an evolution already underway.
The Court's moves are under the radar. It is true that statutory interpretation has been conceptualized in "universalist" terms: the mainstream theories generally assume that one set of presumptions applies to all statutory drafters, types of statutes, legislative processes, subject matters, and agencies.
But in practice, there has been movement toward tailoring interpretive doctrine in different ways.284 There is already a species of tailoring by subject matter: namely, the more than 100 subject-specific canons of interpretation currently in deployment (for instance, the rule that ambiguous bankruptcy legislation be construed in favor of the debtor).285 As noted, the Court also sometimes tailors agency deference doctrine by subject matter,286 and recent scholarship has observed the tendency of courts to rely on "agency-specific precedents," although administrative law is supposedly generally applicable.287 On the legislative- process side, four Justices recently suggested a new "anti-severability" rule especially for omnibus legislation.288
There also is a species of normative or functional tailoring-a few special canons of interpretation that apply only to certain types of statutes. For example, the canon that remedial statutes shall be liberally construed, or, in the agency context, lesser deference for major policy questions.289 And there already are canons that differentiate among interpreters, Chevron being the most prominent example.
Why the Court has failed to explicitly acknowledge these more circumstance- specific moves is a puzzle. (
Against this backdrop, it is of particular interest that there is an ongoing, much more explicit debate in the administrative law context over precisely this question of tailoring interpretive doctrine to how Congress works. This is the heated debate over Mead-in which the Court walked back Chevron's broad presumption of delegation whenever ambiguity exists and instead chose to "tailor deference to [the] variety" of ways in which Congress legislates.
Some Justices, along with scholars, also have been more conscious of differences across agency staffwith respect to fashioning interpretive doctrines for administrative law in a way that they have not engaged the staff-related implications of interpretive doctrines in other statutory cases.292 In his Mead dissent,
It may be that the Court and scholars have grappled more consciously with this tradeoffbetween legislative realism and doctrinal simplicity in the administrative law context because the question of delegation brings these matters to the fore. We also recognize that these doctrinal efforts on the administrative law side have imposed costs on lower courts that are not trivial,296 and that our recommendations may make the landscape even more complex, especially if some of the current canons that we would eliminate are not retired. But at least part of Mead's problem turns on the specifics of the doctrine that the Court has articulated. For example, Mead lacks bright lines and specifically admits of unpredictable exceptions,297 and the major questions doctrine requires a subjective judgment from the Court about what kinds of questions are too important to delegate. Moreover, none of those doctrines has been justified as an exercise in formalism.298 The factors we have emphasized-type of statute, committee jurisdiction, the CBO score, and so on-are intentionally more amenable to clear legal rules.
Thinking more broadly about the future of legislation and administrative law theory, we wonder whether these individual moves are part of a trend toward a set of even more tailored interpretive principles organized around subject matters or individual statutory schemes. Some of our most important findings, including canon knowledge, presumptions about delegation, and all of the different committee-jurisdiction-related presumptions can be brought together through this subject-matter lens. We are not the first to suggest this focus. There is a growing literature on the choice between intra- and trans-substantive doctrine in other areas of law,299 and, specifically in the legislation context,
Consider how a series of interpretive regimes constructed around statutory schemes might be the natural evolution for a mature theory of interpretation in the regulatory age. One way to understand the past seventy years of universalizing doctrinal and theoretical work is as the foundational work necessary to establish a field. Our findings raise the possibility that the recent focus on legislation as a separate subject is only a temporary stop along the way to a more nuanced and specific understanding of what it means to live and lawyer in a statutory era.302 Perhaps years from now, we will no longer have "legislation." Instead, we may just have "tax" or "environmental law," all individual statutory fields which themselves each will have incorporated whatever specific interpretive principles apply to the committees who work on those subjects, or the types of statutes typically utilized for them. Arguably, such an evolution might truly announce the arrival of the Age of Statutes. The irony, of course, would be that the greatest intellectual achievement of the field might be its own obsolescence.
CONCLUSION
The doctrines of statutory interpretation and agency deference rest on assumptions about how Congress works and how the three branches communicate that have never been empirically tested. This pair of Articles has begun that effort, while recognizing that much remains to be investigated. Our aims have been to expose more clearly the variety of jurisprudential bases for the various doctrines and the weaknesses of all of the dominant theoretical paradigms. We also have suggested alternatives to theories that depend on the details of congressional practice, highlighted Congress's own obligations in the interpretive conversation, and identified what may already be a trend toward more context-specific interpretive presumptions.
But our study also makes clear the challenges of using empirical studies of Congress as a tool for improving legal doctrine, given that many of the on-theground details that we uncovered could never be used by courts, or at least not without excessive cost. One of the best known empirical studies of Congress- political scientist
We lean toward a similar view. If the democratic legitimacy of courts rests on at least a partial dialogue with Congress, then we need more study not only of Congress, but also of agencies and lobbyists' interpretive practices, not a throwing up of hands in anticipation of the difficulty of dealing with the complexities that will emerge from such work. The current model-one grounded in assumptions that do not survive empirical testing and deployed without a true effort to communicate by either side-does not seem sustainable. Nor has it done enough to foster an interbranch interpretive relationship.
1.
2.
3.
4. See, e.g., Lockhart v.
5. One important study preceded ours.
6. See ADRIAN VERMEULE, JUDGING UNDER UNCERTAINTY 1, 2, 109 (2009);
7. See infra notes 233-36 and accompanying text (discussing how purposivism has moved from broader legal process traditions toward this narrower focus in modern times).
8.
9.
10. See id. at 320-21.
11. Cf.
12.
13. See, e.g.,
14. Q11; see
15. See Methods Appendix, supra note 3, at 6.
16. See 2 U.S.C. § 281 (2012) (establishing the
17. For a few exceptions, see
18.
19. This is likely because the Judiciary Committee seems to use Legislative Counsel less frequently than other committees. See id. at 581.
20. The only reference to Legislative Counsel in the first version of the survey was in the first substantive question (Q11) about groups that participate in drafting.
21. Universal Comment Code 1. For an explanation of comment codes, see Methods Appendix, supra note 3, at 49-50.
22. Q49.
23. Id.
24. Id.
25. Id.
26. See, e.g.,
27.
28. Confidential Telephone Interview with staffer (
29. See 2 U.S.C. § 272 (2012) (providing that the
30.
31. E-mail from confidential respondent to author (
32.
33.
34.
35. See Q78.36. Q49.
37. Q50.
38. Confidential Telephone Interview with staffer, supra note 28.
39. See BJ Ard, Comment, Interpreting by the Book: Legislative Drafting Manuals and Statutory Interpretation, 120 YALE L.J. 185 (2010).
40. Q49.
41. See Q18 (federalism or preemption); Q23 (Mead); Q46 (superfluities, whole act, whole code).
42. The exceptions were the whole act rule (by name), Q45f (95% confidence), the whole code rule (by concept), Q44c (99% confidence; 95% using super population assumption), the rule against superfluities (by name), Q45c (99% confidence; 95% using super population assumption), and the rule against superfluities (by concept), Q43 (95% confidence). All calculations are the same using a baseline population of 650 counsels or a super population assumption unless noted. See Methods Appendix, supra note 3, at 7.
43. See, e.g., Q65 (finding that Legislative Counsels are more likely to say courts apply consistent interpretive rules) (95% confidence).
44. See Exec. Order No. 13,563, 3 C.F.R. 215 (2011), reprinted in 5 U.S.C. § 601 app. at 102-03; Exec. Order No. 12,866, 3 C.F.R. 638 (1993), reprinted in 5 U.S.C. § 601 app. at 88-92 (2012);
45.
46. Q44d.
47. See E-mail from
48. Q44c-d.
49. See Q81.
50. See 2 U.S.C. § 281a (2012) (requiring impartiality of House Legislative Counsel); Policies Governing the Performance of Duties, OFF. OF THE LEGIS. COUNS.,
51. Q49.
52.
53. There were seventy mentions of the importance of committee jurisdiction before any question mentioning committees was asked. See Methods Appendix, supra note 3; Universal Comment Code 4.
54. A search of the
55. See generally
56. Q44c.
57. See Committees,
58. See Q44a-d (comment code).
59. Q44c.
60. Id.
61. Q44d.
62. See Q44a-d (comment codes).
63. Q49; Q60 ("[C]onceptual marks are what we vote on, not really on language, and someone works on the language after that."); Q78.
64. Q71; cf.
65. Cf.
66. Q28; see U.S. GEN. ACCOUNTING OFFICE, GAO-04-261SP, 1 PRINCIPLES OF FEDERAL APPROPRIATIONS LAW 2-113, 3-40 (3d ed. 2004), available at http://www.gao.gov/ assets/210/202437.pdf.
67. STAFF OF S. COMM. ON COMMERCE, SCI. & TRANSP., 111TH CONG., GUIDE FOR PREPARATION OF COMMITTEE REPORTS 2-3 (Comm. Print 2009).
68. Q35.
69. OFFICE OF THE LEGISLATIVE COUNSEL,
70. OFFICE OF THE LEGISLATIVE COUNSEL,
71. OFFICE OF THE LEGISLATIVE COUNSEL,
72. Id. at 61-62; SENATE DRAFTING MANUAL, supra note 70, at 76.
73. HOUSE DRAFTING MANUAL, supra note 71, at 3.
74.
75.
76.
77. See ROBERT A. KATZMANN, COURTS AND CONGRESS 65-66, 138 n.93 (1997); cf.
78.
79. Q44a.
80.
81. Q49.
82.
83. See KEITH KREHBIEL, INFORMATION AND LEGISLATIVE ORGANIZATION 254-57 (1991);
84. Q23; Q24.
85. Q23.
86. Id.
87. Id.
88. Q24.
89. Id.
90. Gluck & Bressman, supra note 1, at 1001-02; see
91. Q24.
92. Cf.
93.
[t]o lay claim to Internet legislation and avoid referral of their bill to the Commerce Committee, two
Id. at 83.
94. Kaiser, supra note 14, at 88.
95. Q83.
96. Q3; Q29; Q30. Judiciary respondents were more likely than others to know the rule of lenity (99% confidence).
97. See R. ERIC PETERSEN ET AL., CONG. RESEARCH SERV., R41366, HOUSE OF REPRESENTATIVES AND SENATE STAFF LEVELS IN MEMBER, COMMITTEE, LEADERSHIP, AND OTHER OFFICES, 1977-2010, at 19, 24 (2010).
98. Methods Appendix, supra note 3, at 2-3.
99. Universal Comment Code 9 (selected); Universal Comment Code 15 (all).
100. Q28.
101. Q19.
102. Q64.
103. Q83.
104. See OLESZEK, supra note 93, at 187; see also Unanimous Consent,
105. Q70.
106. Q49 (reporting, for example, that "on the personal office side they often have only a vague idea of what they want and Leg. Counsel has to do a lot more").
107. Id.
108.
109. See id. at 111.
110. Universal Comment Code 20.
111. Q72; cf.
112. Q64; see Q61.
113.
114. Q77b-c.
115. Q81.
116. Q70.
117. Q74; Universal Comment Code 16; accord OLESZEK, supra note 93, at 299.
118. See SINCLAIR, supra note 108.
119. Apart from the occasional acknowledgement of a statute's omnibus nature, see, e.g., Reno v. Am.-Arab Anti-Discrimination Comm.,
120.
121.
122. Q71. Seven of the remaining thirty-five respondents who did not answer the initial question in this manner offered comments to the same effect. Id.
123. Id.; see Q44b.
124. Q72; Q73.
125. Q72.
126. Id.
127. Q73.
128. Id.; see CONSTITUTION, JEFFERSON'S MANUAL, AND RULES OF THE
129. Q58.
130. SINCLAIR, supra note 108, at 147; see OLESZEK, supra note 93, at 15.
131. Q72.
132. Q71.
133. Q70.
134.
135. Q44c.
136. Q74.
137. Id.; Q75.
138. Q74.
139. Q70; Q74; Q75; Q76. Forty-seven percent also said that the committee process makes statutes more internally consistent.
140. See Memorandum from
141. Id. at 8; see
142. Cf. OLESZEK, supra note 93, at 247.
143. Q74.
144. Q74.
145. Q11i.
146. The CBO was created in 1974, as part of the Congressional Budget and Impoundment Control Act of 1974, Pub. L. No. 93-344, §§ 201-203, 88 Stat. 297, 302-05 (codified as amended at 2 U.S.C. §§ 601-603 (2012)).
147. Q11.
148. Id.
149.
150. Q83; see
151. Q22.
152. Q42.
153. See, e.g.,
154. See ADRIAN VERMEULE, supra note 6, at 206;
155. Q12; Q21; Q33; Q37-Q40; Q44b; Q48-Q49; Q57; Q59-Q61; Q64; Q67-Q69; Q73.
156. Q37; see Q25.
157.
158. Mashaw, supra note 157, at 508.
159. Strauss, supra note 157, at 329-35.
160. Q60f-g.
161. Q73.
162. Q57.
163.
164.
165. Q63.
166. Id.
167. Mashaw, supra note 157, at 519; accord
168.
169. Q50f; Universal Comment Code 8.
170. Cf.
171. For a summary of such arguments, see
172. Cf.
173. See, e.g.,
174. Cf.
175. Gluck & Bressman, supra note 1, at 1002-05.
176. Q55.
177. See
178.
179. Cf.
180.
181.
182. Strauss, supra note 179, at 1169-71; see also
183.
184. Mashaw, supra note 157, at 537-38.
185. Cf. SINCLAIR, supra note 108, at 111 (arguing that the rise of unorthodox lawmaking has increased the strength of the President relative to committees).
186. Mashaw, supra note 157, at 501; see Strauss, supra note 157, at 342.
187. Q50.
188. In total, forty-eight unique respondents (35%) offered comments of this nature.
189. Q50.
190. Q13.
191. Id.
192.
193. Q50e.
194. Q39; accord Q68.
195. Q38.
196. Id.
197. Q67.
198. Q37; Q67; Universal Comment Code 10.
199. Q67.
200.
201. Q39.
202. Q38.
203. See supra note 172 and accompanying text.
204. STEPHEN BREYER,MAKING OUR DEMOCRACY WORK: A JUDGE'S VIEW 96 (2010).
205.
206. See generally
207.
208.
209. Gluck & Bressman, supra note 1, at 914.
210. Id. at 906-07, 949 (calling Chevron and the federalism canons "feedback canons"); id. at 959-89 (discussing feedback canons).
211.
212.
213.
214.
215. See ANTONIN SCALIA &
216. See STANDING RULES OF THE
217. See Methods Appendix, supra note 3, at 19.
218.
219.
220. See id. at 265-68.
221. Id. at 265 (citing 42 U.S.C. § 290bb-2a).
222. For more detail see Gluck & Bressman, supra note 1, at 1010.
223.
224.
225.
226.
227.
228.
229. We also recognize that our focus on committee counsels is a limitation. We cannot know whether personal staffwould leave us with the same impressions, but all of the elected members for whom personal staffwork do sit on committees.
230. Manning, supra note 8, at 16.
231. Bruesewitz v.
232. See, e.g., Transcript of Oral Argument at 15-17, Arlington Cent. Sch. Dist. Bd. of Educ. v. Murphy,
233. For examples from October Term 2012, see Bullock v.
234. See, e.g., BREYER, supra note 204, at 96; CALABRESI, supra note 206, at 164-65; ESKRIDGE, supra note 206, at 6;
235. POSNER, supra note 205, at 198-99;
236. See HART & SACKS, supra note 234, at 1415.
237. See BREYER, supra note 234, at 102; ESKRIDGE, supra note 206, at 132, 151; Strauss, supra note 208, at 247.
238. This question was added to the survey after the first few days of interviews.
239. Q77A.
240. Q68.
241.
242.
243.
244.
245. Finley v.
246. Federal courts give "super" stare decisis to their substantive interpretations of statutes but do not give precedential weight to the interpretive decisions used to construe them- for example, which legislative history is reliable or which canon trumps which-even where the same statute is being construed.
247.
248.
249. See supra note 6.
250.
251.
252. VERMEULE, supra note 154, at 205-15; Eskridge, supra note 154, at 416-26; Sunstein & Vermeule, supra note 6, at 925-32.
253. VERMEULE, supra note 154, at 209.
254. Compare, e.g.,
255.
256. See, e.g., Hard Cider, Semi-Generic Wine Designations and Wholesale Liquor Dealers' Signs, 66 Fed. Reg. 58,938, 58,940 (
257. See, e.g., Fed. Nat'l Mortg. Ass'n v.
258. Q37.
259. Id.
260. Eskridge & Frickey, supra note 241, at 332-33.
261.
262. But cf. Christiansen & Eskridge, supra note 228, at 80 (noting cost of overrides but also arguing they advance rule-of-law values).
263. See, e.g., JUDICIAL CONFERENCE OF
264. Amendments, HOUSE COMMITTEE ON RULES, http://rules.house.gov/amendments (last visited
265. Q60; Universal Comment Code 8. Twenty-four percent of respondents volunteered cross-references, without prompting, as obstacles to understanding statutory text. 266.
267. See KY. REV. STAT. ANN. § 446.140 (West 2013); 1 PA. CONS. STAT. § 303 (2013) (requiring a description of the cross-referenced section); see also, e.g., Fawbush v. Louisville & Jefferson Cnty. Metro. Sewer Dist., 240 S.W.2d 622, 624 (Ky. 1951) (holding that crossreferences cannot be used for statutory construction).
268.
269.
270. See, e.g., MICH. CONST. art. IV, § 26 (requiring bills to be "printed or reproduced and in the possession of each house for at least five days"); OKLA. CONST. art. V, § 34 ("Every bill shall be read on three different days in each House, and no bill shall become a law unless, on its final passage, it be read at length . . . ."); see also Read the Bills Act,
271.
272.
273. Manning, supra note 207.
274. Q50e.
275. Compare, e.g.,
276. 1 U.S.C. §§ 1-8 (2012); 21 U.S.C. § 854(d); Organized Crime Control Act of 1970, Pub. L. No. 91-452, § 904(a), 84 Stat. 922, 947, reprinted as amended in 18 U.S.C. § 1961 note; Rosenkranz, supra note 215, at 2110 (describing interpretive instructions in Title I); cf. Christiansen & Eskridge, supra note 228, at 6 (arguing Congress could codify or negate certain canons in Title I).
277.
278. Gluck, supra note 215, at 1782-98, 1824-29.
279. See generally
280. SCALIA & GARNER, supra note 215, at 245. It is a puzzle that this debate generally overlooks the existence of the thousands of rules of construction that already exist in the U.S. Code.
281.
282.
283. SCALIA & GARNER, supra note 215, at 49, 51.
284. See generally Gluck, supra note 212 (describing subject-related, procedural, institutional, and normative tailoring in statutory interpretation).
285. For a full list, see
286. See supra note 90.
287.
288. See Nat'l Fed'n of Indep. Bus. v. Sebelius, 132
289. See
290.
291. Id.
292. Thanks to
293. Mead, 533 U.S. at 258 n.6 (Scalia, J., dissenting).
294.
295.
296.
297.
298. The Court's most recent administrative law decision,
299. See generally
300.
301. We recognize that subject-specific interpretive regimes could pose problems for areas where oversight is split among multiple committees.
302. See generally Gluck, supra note 212 (introducing and elaborating this point).
303. FENNO, supra note 55, at xiv.
* Associate Dean for Academic Affairs and David Daniels Allen Distinguished Professor of Law,
** Associate Professor of Law,
Many thanks to
| Copyright: | (c) 2014 Stanford University, Stanford Law School |
| Wordcount: | 31025 |
Best Buy says hot new products are needed to spur sales growth
Advisor News
- Flourish brings private-bank-like cash solution to MassMutual’s network
- Majority of Americans concerned recent market highs are unsustainable
- GLP-1 users choose between medication and retirement saving
- Gen X and millennials seek new retirement model
- Are families ready for the costs of aging at home?
More Advisor NewsAnnuity News
- Corebridge Financial Introduces Progressive Trigger, a First-of-Its-Kind Index Annuity Strategy Designed to Turn Missed Years into Future Opportunity
- Insurer proposes $1.1M deal to resolve CA teachers’ annuity fee lawsuit
- IMOs get frustrated with slow carrier adoption of faster technology
- New class-action lawsuit targets Delaware Life over annuity disclosures
- A client remarried: Does their annuity still fit?
More Annuity NewsHealth/Employee Benefits News
Life Insurance News