The Yale Law Journal

November 2013

Leviathan and Interpretive Revolution: The Administrative State, the Judiciary, and the Rise of Legislative History, 1890-1950

abstract. A generation ago, it was common and uncontroversial for federal judges to rely upon legislative history when interpreting a statute. But since the 1980s, the textualist movement, led by Justice Scalia, has urged the banishment of legislative history from the judicial system. The resulting debate between textualists and their opponents—a debate that has dominated statutory interpretation for a generation—cannot be truly understood unless we know how legislative history came to be such a common tool of interpretation to begin with. This question is not answered by the scholarly literature, which focuses on how reliance on legislative history became permissible as a matter of doctrine (in the Holy Trinity Church case in 1892), not on how it became normal, routine, and expected as a matter of judicial and lawyerly practice. The question of normalization is key, for legislative history has long been considered more difficult and costly to research than other interpretive sources. What kind of judge or lawyer would routinize the use of a source often considered intractable?

Drawing upon new citation data and archival research, this Article reveals that judicial use of legislative history became routine quite suddenly, in about 1940. The key player in pushing legislative history on the judiciary was the newly expanded New Deal administrative state. By reason of its unprecedented manpower and its intimacy with Congress (which often meant congressmen depended on agency personnel to help draft bills and write legislative history), the administrative state was the first institution in American history capable of systematically researching and briefing legislative discourse and rendering it tractable and legible to judges on a wholesale basis. By embracing legislative history circa 1940, judges were taking up a source of which the bureaucracy was a privileged producer and user—a development integral to judges’ larger acceptance of agency-centered governance. Legislative history was, at least in its origin, a statist tool of interpretation.

author. Associate Professor of Law, Yale Law School. For valuable conversations about the project, I thank Bruce Ackerman, Ian Ayres, James Q. Barrett, Joseph Blocher, James Brudney, Aaron-Andrew Bruhl, Josh Chafetz, Robert Ellickson, Daniel Ernst, William Eskridge, Heather Gerken, Abbe Gluck, Bob Gordon, Oona Hathaway, Daniel Ho, Christine Jolls, Dan Kahan, John Langbein, Robert Lieberman, Yair Listokin, John Manning, David Marcus, Jerry Mashaw, Tracey Meares, Thomas Merrill, Robert Post, Edward Purcell, Cristina Rodríguez, Susan Rose-Ackerman, Ted Ruger, Reuel Schiller, Alan Schwartz, Peter Strauss, Adrian Vermeule, Jim Whitman, and John Witt; and audiences for talks at Duke, Stanford, Yale, and the annual meeting of the American Society for Legal History. The quantitative aspect of the project was made possible by the work of several excellent and dedicated research assistants: Allyson Bennett, Glenn Bridgman, Halley Epstein, Miles Farmer, Andrew Hammond, Tian Huang, Steven Kochevar, Stephen Petrany, Emily Rock, Clare Ryan, and Karun Tilak. For aid in obtaining sources, I thank the staffs of the Yale Law Library (particularly Sarah Kraus), the Harvard Law Library, the Columbia University Center for Oral History, the Library of Congress, the Franklin D. Roosevelt Presidential Library, the Harry S. Truman Presidential Library, the Margaret I. King Library at the University of Kentucky, the William L. Clements Library at the University of Michigan, and the National Archives; and Kim Dixon, Karen Needles, Doug Norwood, and Susan Strange. Glenn Bridgman generously shared data with me from his own research project. Alex Hemmer and his fellow members of the Yale Law Journal performed valuable work in editing and publishing the piece. I am grateful to Yale Law School for financial support. All errors are my own.


When a legislature enacts a statute, it leaves behind a history: the revisions that lawmakers made to the bill, the things they said about it during committee deliberations and floor debates, and the public input they officially received on it from experts and other witnesses. Should a court, when interpreting the act, consider that history?

For a generation, the field of American statutory interpretation has burned with controversy over this question. The controversy is a novelty of the last twenty-five years. In the 1980s, legislative history was uncontroversial and very common. It appeared in more than half the U.S. Supreme Court’s opinions on federal statutes.1 In the high courts of leading states like New York, it likewise appeared frequently.2 Using this material meant that judges were accustomed to engaging actively and openly with legislators’ discourse and policy reasoning. Beginning in the late 1980s, however, a movement of judges and lawyers—led by Antonin Scalia—began to argue that this familiar interpretive resource was pernicious and should be banished from the judicial system. They urged a textualist method of statutory interpretation that would ignore an act’s legislative history and focus more narrowly on its words. The legislative history of an act, warned Scalia and his allies, was a devil’s playground: it contained such a huge number of assertions about the act’s meaning, and those assertions were so contradictory and so easily inserted by manipulative politicians or lobbyists, that willful judges could always find support for whatever personal preferences they wished to impose. Adherence to the ordinary meaning of the text—the words on which lawmakers formally voted according to constitutional procedures—would do better at keeping judges accountable to the democratic will. Critics responded that Scalia was a false prophet. His method, they said, would not deliver the determinacy he promised, for text was often ambiguous (or became ambiguous when overtaken by events unforeseen by lawmakers), so textualist judges could just as easily impose their preferences, and all the more insidiously, since they would do so under the apolitical cloak of “ordinary meaning.” Besides, added the critics, legislation was meaningless without reference to policy, and what better source for understanding a statute’s policy than legislative history?3

Whichever side is right, there is no doubt that judicial practice has moved dramatically in Scalia’s direction (even if his colleagues have not formally converted to his principle of complete exclusion). The proportion of U.S. Supreme Court opinions citing legislative history in statutory cases has fallen by more than half since the 1980s.4 The number of citations per statutory case has fallen even more steeply.5 At the state level—where legislatures have recently begun publishing legislative history far more extensively than in the past—courts have drawn upon textualist ideas to limit the new material’s use.6 Even leading defenders of legislative history concede parts of the textualist critique and look back with some embarrassment on how freely and easily the federal courts used such material twenty-five years ago.7 “[M]any contemporary courts,” when they do cite legislative history, seem to “apologize” for doing so.8

We are living through what appears to be an interpretive revolution. But we do not fully understand what (for better or worse) we are losing by it. That is because we do not have an adequate account of how and why judicial reliance upon legislative history—the once-dominant method that is now under attack—came to dominance in the first place. Further, an account of legislative history’s rise would provide us with a better understanding of how interpretive revolutions happen, which may be useful knowledge for those who want to consummate the present revolution and for those who want to prevent its consummation. The task of this Article is to provide that account.

To be sure, we do have a partial sense of the story, in some of its general outlines. Originally, English and American judges cited no legislative history, for none was available. In the early modern era, the English Houses of Parliament not only refrained from recording or publishing their debates but actually prohibited their publication, for the members feared the scrutiny of the crown and viewed themselves as an elite body that should be insulated from immediate popular pressure.9 With no tradition of legislative openness in the mother country, the American colonial assemblies also did not publish their deliberations.10

Eventually, however, English and American judges did have to confront the question of whether to use legislative history, for as both societies became more democratic, such material came to be published copiously. In 1771, Parliament began permitting publication of its debates, and by the mid-1800s, the famous Hansard company received public subsidies to publish them in great detail.11In the United States, Congress published its journal (a skeletal procedural record)12 from 1789 onward and soon began allowing private printers to publish its floor debates. These early reports were sometimes incomplete, irregular, and even biased, but this changed in 1850 when the Congressional Globe became semi-official, publishing a complete and timely verbatim transcript, supported by a congressional subsidy.13Eventually Congress started doing the same thing itself, establishing the in-house Congressional Record in 1873.14 The Record of the 46th Congress (1879-81) ran to 10,000 pages; of the 56th (1899-1901), to 10,000 again; of the 66th (1919-21), to 22,000; and of the 76th (1939-41), to 38,000. Meanwhile, Congress’s committees began publishing reports on all the public bills they sent to the floor. The House made such reports mandatory in 1880, and Senate committees were issuing them on the majority of bills by about 1900.15 These committee reports, combined with informative documents that committees ordered printed, made up a “Serial Set” that totaled 2,000 volumes by 1880, added another 2,000 by 1900, another 4,000 by 1920, and another 2,500 by 1940. And shortly after 1900, the committees began frequently publishing their hearings: the annual number stood at about 100 in 1900 and jumped into the range of 500-650 per year for 1910, 1920, 1930, and 1940.16 (The penchant for publishing did not extend to the state legislatures: even in the 1940s, almost none of them had ever recorded their floor debates or committee hearings, or issued substantive reports of standing committees.17)

Thus, English courts and U.S. federal courts had to decide whether to use the legislative history that was proliferating. The English judges shut their eyes to it. They had imposed an exclusionary rule in the era before 1771, when publication of parliamentary deliberations had been illegal, and they maintained that rule even as reports of the debates became lawful and available.18 American judges and treatise-writers followed the English rule from the late 1700s through the mid-1800s.19 Then, in the 1870s and 1880s, U.S. federal judges in a handful of cases made small, incremental departures from the rule. Next, in 1892, the U.S. Supreme Court in Church of the Holy Trinity v. United States20 relied upon committee reports (among other interpretive tools) to override what it admitted to be the literal meaning of an act of Congress.21 Holy Trinity would become the leading case for a new American rule that legislative history was permissible in statutory interpretation.22 Meanwhile, English judges—plus those of Commonwealth countries like Canada and Australia—held fast to the exclusionary rule. They relaxed it only in the 1980s or later, and even now, they do not use legislative history as much as American courts.23

So much we know. What we lack is an account of how legislative history went from being a permissible tool of American statutory interpretation to being the normal, routine, and expected tool that it had become by the time of Scalia’s attack. One might assume that, in an adversary system, permissibility automatically leads to normalization. That is, once the judiciary says it will consider a certain kind of source, lawyers will instantly compete with each other to cite ever more of that source. But nobody has proven that assumption with respect to legislative history. And it may well be wrong. The opinion in Holy Trinity itself did not consciously reject the English rule; it simply ignored it without comment. And, as some leading present-day scholars point out, legislative history is more (perhaps far more) difficult and costly for lawyers to research than other legal sources, such as case law and statutory text.24 Given the offhand quality of Holy Trinity and the headaches of using the source it made permissible, we must ask: why did this case become a landmark—a license for a method that has cost millions of billable hours—instead of being forgotten as an obscurity?

What we need is an account of legislative history’s rise as a matter of practical, workaday lawyering and opinion-writing. Or, to speak in more academic terms, we need an institutional account. Who were the lawyers and judges who started the pattern of routinely using a source often considered intractable? How did they manage to do it? What drove them?

The existing literature does not answer this question, though it does provide some valuable background and a few pieces of the puzzle. There has been major work on the intellectual history (as distinct from the institutional history) of judicial use of legislative materials. This work focuses on leading cases, treatises, and law review articles from the first half of the 1900s. It shows how the federal judiciary widened the scope of permissibility on an incremental basis, to include more categories of legislative documents, and it traces changes in high theory about the value of legislative history in interpretive thinking.25 These intellectual developments play a role in the institutional story that I tell below.

In addition, there have been several quantitative studies of legislative history usage, though mostly confined to the period from the 1950s to the present.26 One of the most careful of these studies finds that the percentage of Supreme Court statutory opinions using legislative history rose between the 1950s and 1980s and plummeted thereafter. Crucially, the level in the 1950s was substantially higher than it is today.27 Thus, our threshold question is how the Court even got to the level of the 1950s. By then, legislative history had already become normal. (Indeed, the first edition of a standard manual on appellate advocacy in 1950 said that “an advocate in the federal courts . . . cannot afford to ignore legislative history” and “should routinely check” it.28) There have been four quantitative studies of U.S. Supreme Court citations that reach back into the pre-1950 era and include legislative history.29 Taken together, these studies suggest a surge—perhaps a radical one—in the Court’s usage sometime in the 1930s or 1940s. But they do not allow us to state the point with confidence or precision. All four have major limitations in method or scope: two fail to control for the number of statutory cases,30 a third draws cases from only four years within the whole period 1925-50,31 and the fourth focuses exclusively on tax law.32 Also, only one of the four studies focuses mainly on legislative history, and only one (not the same one) focuses mainly on the pre-1950 period, so none of the four pay much attention to the pre-1950 shift toward legislative history that seems to appear in their datasets.33 Further, all four studies focus mainly on the Court’s opinions, saying little to nothing about the role of lawyers in bringing this hard-to-use source to the fore.34 A satisfying institutional account demands more work.

I begin that work in Part I, setting forth quantitative data on Supreme Court use of legislative history in every fifth year from 1900 to 1950, including (1) the number of cites per statutory case and (2) the proportion of statutory cases with any cites. The first metric is low through 1935, then suddenly increases four-fold in 1940, entering a range that becomes the “new normal” through 1950. Similarly, the second metric creeps up to 20% by 1935 but then more than doubles (to 44%) in 1940, again entering a range that becomes the “new normal” (and is comparable to the data from prior studies on the 1950s). In other words, the use of legislative history became normalized very suddenly, sometime in 1935-40. I then present comparable data on every single year from 1930 through 1945, finding that the radical break can be pinpointed to the year 1940 almost exactly. I then assemble qualitative evidence confirming a transformative shift toward legislative history as a matter of everyday practice just around 1940.

This finding has three implications. First, it shows that legislative history remained rare for decades after Holy Trinity. Permissibility did not automatically lead to normalization. Second, it establishes that the American turn to legislative history as a routine interpretive source originated at the federal level, not the state level, for scholars in the 1940s universally found that state courts’ use of legislative history was minimal, largely because the state legislatures at the time published so little of it.35 Third, identification of the years 1940-45 as the opening years of the “new normal” points us to the next step in our inquiry: we must analyze those years intensively to see which judges and lawyers were most responsible for the advent of the new regime.

In the remainder of the Article, I argue that normalization was caused simultaneously by two interrelated factors. First was the appointment of new Justices with a new openness to legislative history. Second was the rapid growth of a federal administrative state with unprecedented institutional capacity to use legislative history to its advantage.

Part II discusses the appointment of new Justices with new intellectual attitudes. The surge of legislative history in 1940 coincided fairly closely with the appointment of several new progressive Justices by President Franklin D. Roosevelt. Progressive Justices generally cited legislative history at a much higher rate than Justices with more classical views. This makes sense: as the literature on the intellectual history of statutory interpretation indicates, progressive legal ideology counseled that judges should recognize the inevitable indeterminacy of traditional legal sources—whether common-law doctrinal formulas or statutory texts—and engage consciously and transparently in the policy choices that were inevitable in legal decision-making. As progressive academics began arguing around 1940, legislative history could be of great value in guiding and deepening a judge’s policy reasoning, particularly if the judge used the history to reason at a high level of generality—discerning the legislature’s overall objective and then reasoning “downward” to find a disposition of the specific case that best implemented that objective. Based on my reading of numerous cases from the period, I do think these progressive academic views informed the Justices’ increasingly copious use of legislative history during the early 1940s.

Still, the period’s theoretical innovations are not a complete explanation for the change in workaday judicial practice. While there were many opinions that conformed to the academic focus on general intent, there were many others that went in a different direction, reflecting a more old-fashioned tendency to consult the history in search of lawmakers’ specific preference for how to dispose of the fact situation at issue, without much policy reasoning. Another difficulty with attributing the change entirely to the fresh ideas of new Justices is that even the old Justices greatly increased their citations to legislative history between the pre-1940 and post-1940 stages of their careers.

For a true understanding of normalization, we must appreciate the crucial role of the newly expanded federal administrative state—the leviathan—in providing legislative history to the Court. Congress was such a complex body that neither the Justices nor ordinary lawyers were capable of analyzing its discourse routinely and systematically, no matter how intellectually committed they might have been to doing so. Only the federal bureaucracy had the intimacy with Congress and the institutional capacity to drive the interpretive revolution. Certain present-day scholars posit that the federal bureaucracy, when it litigates, uses legislative history more, and better, than does any other litigant,36 but this Article is the first to confirm that idea empirically.

Part III traces the administrative state’s leading role in normalization. I begin by setting forth data, for a large sample of Supreme Court statutory opinions in the transformative years 1940-45, on “matches” between the legislative history citations appearing in the Court’s opinions and those appearing in the corresponding briefs.37 Of the Supreme Court’s statutory cases in 1940-45, 78% were briefed by the federal government, and these accounted for 86% of the legislative history citations. In the federally briefed statutory cases, the “matches” indicate that 33% of the legislative history citations came only from the federal brief, 22% from the federal brief and at least one non-federal brief, 10% from at least one non-federal brief and not from a federal brief, and 34% from no brief. In other words, it was 50% more common for the Justices to take a citation exclusively from the federal lawyers than to take one jointly from federal lawyers and non-federal lawyers. And it was more than three times more common for the Justices to take a citation exclusively from federal lawyers than to take one exclusively from non-federal lawyers. (Consistent with this, the data from the cases without federal briefs indicate that the Court had to do more work by itself in those cases. Only 45% of the citations matched any brief at all; the remaining 55% matched no brief.)

The dominance of the federal government in providing legislative history to the Justices makes sense when we consider some background. First, researching legislative history was indeed more difficult, by a wide margin, than researching case law or statutory text. Telling which documents were relevant required sophisticated knowledge of the legislative process; the documents were often hard to obtain; and once obtained, they were voluminous, internally disorganized, and indexed poorly, if at all. In finding this material and translating it into legal argument, federal agencies and the Department of Justice—whose research underlay the briefs submitted by the Solicitor General to the Supreme Court—had major advantages. Their manpower (both in lawyers and law librarians) dwarfed all other organizations in the country. They each devoted themselves full-time to implementing one or a few statutes, so they could amortize the cost of researching the history of those statutes over many cases. Further, they had uses for legislative history beyond litigation (e.g., to guide program implementation). Most important, they were constantly communicating and negotiating with Congress over how to administer (and whether to amend) the statutes in their charge, which meant they were intimately engaged with the congressional process. Indeed, agency officials and Justice Department lawyers actually authored a large amount of legislative history—either in their own name (such as testimony at hearings) or as ghostwriters (of congressmen’s committee reports and floor speeches). Executive branch ghostwriting of congressional discourse has always been a feature of American government, but it appears that the 1930s and 1940s were the high point of this phenomenon: the volume of legislative activity was unprecedented, but Congress did not seriously begin building up its own internal staff until the late 1940s. Until then, congressmen were remarkably dependent on agency personnel for their words. Thus, the judiciary took up legislative discourse as a normal tool of interpretation at the very moment when the executive branch’s dominance of the legislature was at its height.

Part III also provides data on changes in the federal government’s briefing practices across the period 1930-45. The government briefed little legislative history up to the mid-1930s but then suddenly leaped into using such material. This timing is consistent with the idea that the government’s routine use of legislative history was a precondition for the Court’s routine use of it. Because the roots of the interpretive revolution seem to be found in the government’s dramatic increase in use, I consider how that increase happened. There were two stages. The first came in about 1936, when federal lawyers adopted legislative history as a weapon to defend New Deal statutes against constitutional challenges, citing congressional documents to show that Congress had made the factual findings requisite to the exercise of its enumerated powers (e.g., that the problem targeted by the statute truly affected interstate commerce). Interestingly, the Court picked up very few of these citations: conservative Justices eschewed legislative history because they clung to abstract, formalistic tests of constitutionality, while liberal Justices eschewed it because they considered judicial scrutiny of legislative findings to be improper judicial activism. Then came the second stage of the federal government’s leap into legislative history: in about 1939-40, federal lawyers began copiously using such material to argue “pure” questions of statutory interpretation, unconnected to constitutional issues. It seems that this second and more lasting shift in briefing practice arose from (1) the accumulation of a critical mass of institutional capacity by the agencies and DOJ and the consequent peak in Congress’s dependence on those organizations; (2) the great familiarity that agencies and DOJ had with legislative history, arising from their engagement with Congress; and (3) the incentive created by the adversary system to wield a weapon that is less available to one’s opponent.

In addition, Part III considers the capacity (such as it was) of non-federal lawyers to brief legislative history. Non-federal lawyers generally briefed legislative material less than did the federal government, but a few of them used such material heavily. Who were these lawyers? For the most part, they were congressional lobbyists, and often they had lobbied Congress on the very bill that had become the statute at issue in the case. This makes sense: on any given statute, the relevant private lobbyists were the only people outside the federal bureaucracy who had knowledge of congressional activity rivaling that of the bureaucracy.

Further, Part III examines the Court’s own internal capacity to research legislative history. As noted above, my sample indicates that 34% of the citations in federally briefed cases (plus 55% of the citations in other cases) matched no brief, suggesting the Court did the research itself. Consistent with this, the Court, in the years leading up to 1940, had rapidly acquired institutional capacity far exceeding that of any judicial tribunal in American history. Though the Justices had long employed one clerk each, it was only in the 1930s that all of them began filling those positions with elite recent graduates capable of advanced research. Another reason for the Court’s high capacity was the paucity of its cases: the Judiciary Act of 1925 had empowered the Justices to decide what cases they would hear, and they immediately began hearing a lot fewer. This was a precondition for the 1940 surge in legislative history, though not the immediate trigger. That said, it should be noted that some part of the 34% figure represents not the Court’s internal capacity but the hidden influence of the federal leviathan: citations that matched no brief sometimes came from the administrative state through back channels (e.g., ex parte judicial requests for agency research, or agency donations of legislative history scrapbooks to the Court’s library).

Altogether, the litigation process that normalized legislative history was radically modern. Whereas the Anglo-American judge was traditionally conceived as a passive referee choosing which of two formally equal parties had the better argument, the Supreme Court’s normalization of legislative history depended significantly on the Court’s independent research capacity. And when the Court did rely on the lawyers, it did not draw equally from “Doe” and “Roe” but instead depended heavily on an extraordinary kind of party: the now-fully-grown federal leviathan, which in 1940 had far more lawyers and legal researchers than any organization America had ever seen—plus an intense, ongoing relationship with Congress that gave it a uniquely systematic knowledge of legislative history. The extraordinary nature of the process helps explain why legislative history was never a major part of American statutory interpretation until far into the twentieth century. This is consistent with my finding, noted at the end of Part III, that the lower courts did not play a significant role in providing legislative history to the Supreme Court. Judicial use of legislative history did not “bubble up” from below. It was invented at the high court and then trickled down.

In Part IV, I argue that legislative history was—at least in its origin—a statist tool of statutory interpretation. This was true not only because the bureaucracy was a privileged provider of such material to the judiciary, but also because the bureaucracy was a privileged participant in congressional discourse—a status that it could leverage in its briefs. Indeed, many of the government’s invocations of legislative history before the Court aimed to show that (1) congressmen intended for a bill to mean what the agency had told them it meant, or (2) congressmen, having been apprised of how an agency was interpreting a particular statute, acquiesced in the agency’s view. And these were only the explicit briefs; many other governmental invocations of legislative history rested on congressional utterances that had been ghost-written (or at least influenced) by agencies and thus reflected the bureaucratic agenda. Thus, legislative history frequently served to deliver the bureaucracy’s views to the judiciary clothed in the mantle of congressional authority, right at the moment when the Court had recognized such authority as virtually unlimited in the economic sphere. The Court’s acceptance of legislative history as a normal interpretive source was a key element of its larger acceptance of an agency-centered vision of governance.

Though fitting, the “statist” label requires two qualifications. First, the Supreme Court did acquire some capacity to research and evaluate legislative history unilaterally, even though such material was outside the traditional judicial “comfort zone.” By acquiring this capacity, the Roosevelt Justices—even as they acquiesced in an agency-centered regime—preserved a measure of the independence that has characterized the attitude of the federal judiciary toward the federal bureaucracy over the long run of U.S. history. Second, a small, emergent group of lawyer-lobbyists became highly proficient at influencing legislative history in the Capitol and briefing it at the Court. Through these lawyer-lobbyists, a select few corporate litigants could take advantage of legislative history to a degree that rivaled that of the agencies. The emergent corps of lawyer-lobbyists was not so much an alternative to the administrative state as an outgrowth and adjunct of that state: as the great “Washington law firms” became fully established in the postwar era, they staffed themselves with veterans of the bureaucracy and acquired libraries of legislative history that imitated those of the agencies.

Part IV closes by considering the statist nature of legislative history through the eyes of Felix Frankfurter and Robert Jackson—progressive Justices who initially used legislative history with enthusiasm but then, by the late 1940s, worried that they might have created a monster. To Frankfurter, it seemed that lobbying and litigation—particularly when done by federal bureaucrats and their lawyers—were collapsing into each other in a way that advantaged governmental litigants and others with privileged access to the legislative process. To Jackson, judicial reliance on legislative history made litigation more of an “insider’s game,” privileging the bureaucratic state and the few law firms able to approach that level of administrative capacity. He warned that legislative history put “knowledge of the law practically out of reach of all except the Government and a few law offices.”38


More than twenty years ago, Peter Strauss wrote an illuminating essay suggesting that agencies are peculiarly competent to use legislative history when interpreting statutes.39 His argument has not been tested or developed by other scholars as much as it deserves to be,40 particularly the possibility that the use of legislative history in litigation—a deeply controversial practice—is something the government does more, and more effectively, than anybody else.41 This Article is the first study to confirm that possibility empirically and to flesh it out through primary research. I see this as part of the scholarly effort, recently begun by others, to map the largely unknown territory of administrative interpretation of statutes.42 I also see it as a response to the fervent call by Cass Sunstein and Adrian Vermeule for scholars of statutory interpretation to pay more attention to the varying institutional capacities of interpreters, particularly at the empirical level.43 In particular, I think the Article deepens our understanding of how the interpretive methods of agencies and of courts mutually constitute each other. The legal academy is accustomed to think that judges shape agencies’ methods,44 but this Article suggests that, at a crucial turning point in our history, it was more the other way round.

I. the timing of normalization

A. Quantitative Evidence

Here I set forth quantitative evidence that legislative history became a normal and routine interpretive source in federal statutory interpretation just around 1940. In tracing the process of normalization, I focus on the Supreme Court. Obviously the Supreme Court is only one part of a much larger federal judiciary, but my research and that of others has given me reason to think that the Court played a very dominant role in the federal judiciary’s shift toward legislative history—more dominant than in most other aspects of federal judicial practice.45

To discern the general shape of the Supreme Court’s transition toward routine reliance on legislative history, I begin by measuring the Court’s use of legislative history in every fifth year from 1900 through 1950. I use two metrics. I calculate both metrics by focusing, for a given calendar year, on all federal statutory cases decided by the Court in that year. I define a federal statutory case as one that was decided with a full, signed opinion and involved the interpretation of a federal statute. I count any case involving a federal constitutional challenge to a federal statute as falling within this category. Figure 1 gives the number of federal statutory cases in every fifth year from 1900 to 1950. The number fluctuates but almost always falls between 70 and 125.46

Figure 1.

federal statutory cases in the supreme court, 1900-50


Working within this universe of federal statutory cases, let us consider the two metrics of legislative history use. The first is the ratio, for a given year, between (1) the total number of legislative history citations in all opinions (majority and separate) in federal statutory cases and (2) the number of federal statutory cases. Figure 2 graphs this ratio for every fifth year from 1900 to 1950. It indicates a huge and lasting increase between 1935 and 1940. The lowest ratio after the break (in 1940) is 4.6 times the highest ratio before the break (in 1935). In general, the ratios before the break have their ups and downs, moving between zero (in 1905 and 1910) and 0.7 (in 1935); one might discern a vague, but hardly consistent, upward trend, at least before 1920. But that is minor compared to what happens between 1935 and 1940. There also appears to be a continuing upward trend after 1940, from 3.3 to 4.6, but the most striking feature is the initial jump in 1940.

Figure 2.

supreme court legislative history citations per statutory case, 1900-50


The second metric of legislative history usage is the percentage of federal statutory cases, for a given year, in which at least one opinion (majority or separate) contained at least one citation to legislative history. Figure 3 plots this percentage for the same years as in the preceding figure. Again, the most dramatic change is between 1935 and 1940, from 20% of the cases to 44%. This graph indicates a more pronounced (if not entirely consistent) upward trend prior to 1935 than Figure 2. But the real break occurs between 1935 and 1940.

Figure 3.

percentage of supreme court statutory cases citing legislative history, 1900-50


To get a closer look at the Court’s big transition, I now present the same two metrics, but this time for every year from 1930 through 1945. Figure 4 presents the number of federal statutory cases per year during that period. The number is fairly constant, always between 76 and 116.

Figure 4.

federal statutory cases in the supreme court, 1930-45


The first metric—legislative history citations per statutory case—exhibits a radical and permanent increase between the years 1939 and 1940. As plotted in Figure 5, this metric averages 0.9 for the years 1930-39, ending with a value of 1.2 in 1939. It then shoots up to 3.3 in 1940, and it averages 3.5 for the years 1940-45. In other words, the number of citations per case in 1940-45 was nearly quadruple the number in 1930-39. There is some fluctuation within both periods, but the highest value in 1930-39 (that of 1.7, in 1933) is still lower than the lowest value in 1940-45 (that of 2.2, in 1942).

Figure 5.

supreme court legislative history citations per statutory case, 1930-45


The second metric—the percentage of statutory cases containing at least some legislative history in at least one opinion—also exhibits a big increase leading up to 1940, though for this metric, the ramp-up begins slightly earlier. As plotted in Figure 6, the metric averages 22% for the years 1930-39, but it ends with a big uptick to 35% in 1939 (which is 11 percentage points higher than the next-highest year in 1930-39, that being 1934, at 24%). The percentage then shoots up to 44% in 1940, and it averages 45% for the years 1940-45. Thus, the average percentage for 1940-45 is more than double that for 1930-39. The lowest annual percentage in 1940-45 (that of 38%, in 1942) is still higher than the highest in 1930-39 (that of 35%, in 1939) and much higher than the next-highest percentage in 1930-39 (that of 24%, in 1934).

Figure 6.

percentage of supreme court statutory cases citing legislative history, 1930-45


Together, the metrics indicate a sudden break around 1940, perhaps beginning in 1939. The Court rapidly began to cite legislative history in more of its statutory cases and, in the cases where it did so, to cite it more copiously.47

The rise of legislative history occurred broadly across several areas of federal statutory law rather than being concentrated in any single area, or even any couple of areas. On this point, consider the 139 opinions (majority or separate) during 1940-45 that each contained five or more legislative history citations. These opinions, together, accounted for 79% of the legislative history citations during the period. Of these, the most numerous opinions were those involving labor (most commonly the Fair Labor Standards Act, plus the Wagner Act, Railway Labor Act, and others), but these accounted for only 17% of the opinions. The next-most prevalent area was tax, at 11%. Next, arguably, was national security (at 10%), but only if one cobbles together several different subjects that are really quite diverse, including military appropriations, bonuses and relief for soldiers, trading with the enemy, foreign agents, espionage, and internment. After that, the next-most prevalent areas were antitrust (7%), Indian law (6%), electricity and natural gas regulation (5%), bankruptcy (4%), communications regulation (4%), and transportation regulation (4%). Even these nine categories do not cover the remaining 32% of the opinions, which ranged broadly across public lands, food and drugs, agriculture, immigration and naturalization, and violent crime (kidnapping, bank robbery, etc.). One might say legislative history was associated with the “economic-regulatory” area, but that area encompassed most of the federal government’s responsibilities in the 1940s.48

The Court’s use of legislative history generally pertained to questions of “pure” statutory interpretation, unconnected to constitutional issues. To be sure, my definition of “statutory case” included all cases involving a constitutional challenge to a federal statute, but when I read all seventy of the Court’s cases containing at least one opinion with 10 or more citations to legislative history in 1940-45, I found only one case in which the legislative history pertained mainly to a constitutional issue.49

One may wonder if the increase in citations to legislative history truly reflected an increase in the relative importance of such material. Perhaps the Court simply began including more citations to all kinds of legal sources at this point in time? If that were true, the surge in legislative history would merely reflect a general surge in citations, not anything about legislative history itself.

I find this doubtful, largely because of the findings in a prior study by Nicholas Zeppos. Zeppos randomly selected twenty annual Terms of the Court from the period 1890 to 1990, which happened to include the 1944 Term, plus six Terms prior to 1944, the latest of those being 1932. He identified the federal statutory cases in all these Terms and, from that universe of cases, drew a random sample of 413 (about twenty per Term, on average). Zeppos’s team then counted, in each case, every source of authority cited in support of the decision (statutory texts, constitutional texts, cases, canons of construction, legislative history, treatises, etc.). He found that citations to legislative history—as a percentage of the total citations to all kinds of sources—exceeded 10% in 1944 but were below 2% in 1932 and in all previous Terms covered.50 This suggests that the big increase in legislative history that I find in 1940 was an increase relative to other sources.

Nor can the proliferation of legislative history be explained as a function of the proliferation of separate opinions during this period. To be sure, the 1940s did witness a large increase in the tendency of the Justices to write separately.51 And the proportion of legislative history citations appearing in majority opinions did fall. But the fall was not very big: from 86% in 1930-39 to 81% in 1940-45. Even if we focus solely on majority opinions and ignore all separate opinions, the number of legislative history citations per case is still 3.6 times greater in 1940-45 than in 1930-39, and the percentage of cases citing at least some legislative history is still more than double. The increase in separate opinions may have contributed to an “arms race,” in which the Justices piled up ever more citations to refute one another,52 but again, the Zeppos study indicates that legislative history grew in importance even relative to a more general increase in all kinds of citations. It should also be noted that the dramatic decline in unanimity began only in the winter of 1941-42,53 after the surge in legislative history was already quite evident.

B. Qualitative Evidence

Qualitative evidence confirms what the numbers suggest: legislative history went from obscure to routine around 1940. Observers who lived through the period noted the change. Berkeley professor Max Radin, commenting on the Supreme Court in 1945, referred to “the now orthodox procedure of examining the legislative history.”54 Harvard professor Archibald Cox wrote in 1947:

Whereas formerly the [federal] courts, shut off from reliable indicia of the legislative intent, sometimes substituted their own conceptions of sound policy, they are now so overwhelmed by arguments from legislative history that the problem has become one of managing the mass of materials available and winnowing out the unreliable.55

Justice Robert H. Jackson announced in 1948 that the “custom of remaking statutes to fit their histories has gone so far that a formal Act . . . is no longer a safe basis on which a lawyer may advise his client or a lower Court decide a case.”56 A 1952 Columbia Law Review note observed that since 1940, “the use of legislative history in the interpretation of statutes has become standard practice in the federal courts.”57 In their classic The Legal Process (1958), Harvard professors Henry Hart and Albert Sacks noted that—over “the last fifteen years” and particularly since a key 1940 decision which I shall discuss below—there had been a “sudden efflorescence of citations of legislative materials in [Supreme Court] briefs and opinions,” providing the Justices with a “mass of available material” so large that they could not fully process it all.58

This chronology is further confirmed in the writings of Elizabeth Finley, the era’s best-known expert on legislative history research. From 1921 to 1942, she was the librarian of Root, Clark, Buckner, and Ballantine, a top Wall Street firm.59 In 1943, she became the librarian of Covington, Burling, Rublee, Acheson, and Shorb, which was then the largest firm in Washington, D.C.60 She and another woman, hired that same year at another D.C. firm, may have been the first private-firm librarians in the nation’s capital.61 In the course of twenty years at Covington, Finley would become the first private-firm librarian ever to be president of the Association of American Law Libraries (AALL).62 In particular, she became “[w]ithout question, . . . the leading national authority on the ways and means of collecting and compiling U.S. legislative history.”63

Finley’s writings indicate that legislative history emerged from obscurity around the late 1930s or early 1940s. Giving a paper at the AALL conference in 1946, she said:

Something new has been added to our profession, and whether we like it or not, it appears to be here to stay. I speak as the practicing lawyer’s librarian. . . . [F]rom the practical angle, I assure you the most frequent request these days is “Can you give me the legislative history of this act.”64

When How to Find the Law added a chapter on legislative history in its fourth edition of 1949, Finley wrote it and opened by saying: “Probably the most startling development in legal research within the past fifteen years, and certainly one with the least literature, has been the increasing importance of legislative histories.”65 In a 1959 essay, she noted that “many law libraries” had been compiling scrapbooks of legislative history for particular statutes for “the past twenty years,” adding that “only in rare cases will you discover a compiled history of any federal law more than twenty-five years old.”66 She also referred to a certain statute that had been “passed in 1920, long before legislative histories were being cited extensively.”67 Having been a big-firm librarian since 1921, Finley knew what she was talking about.

Consistent with this, legal research guides began paying serious attention to legislative history only in the 1940s. In 1926, the fifth and final edition of Roger W. Cooley’s standard work, Brief Making and the Use of Law Books, said nothing about how to research legislative history and devoted only two sentences specifically to such material (plus three other sentences of more general discussion about extrinsic evidence of historical context).68To gauge the treatment of legislative history in subsequent works of this kind, I located all major legal research guides that went through two or more editions from 1930 through 1950.69 Of the five guides in this category, four discussed legislative history for the first time (or went beyond negligible discussion for the first time) in their first post-1940 editions.70

In the same vein, the years 1943-46 saw three big institutional initiatives aimed at routinizing the use of legislative history, likely in reaction to the Supreme Court surge that began in 1940. One of these came in December 1943—near the end of the fourth and (as yet) biggest year of the Court’s post-1940 surge—when Chief Justice Harlan F. Stone wrote to Archibald MacLeish, the Librarian of Congress. Stone had learned that the Library of Congress had a “spare” copy of the U.S. Congressional Serial Set, which then ran to 10,000 volumes. Stone wanted MacLeish to transfer the “spare” 10,000 volumes to the Supreme Court’s own library. “We often have occasion to use the Congressional Serials in looking up matters of legislative history,” explained Stone, “and it would be of great assistance” if the spare set “could be placed in our Library where it would be available at any time to the members of the Court.”71 MacLeish agreed,72 and these 10,000 volumes became, de facto, part of the Court’s permanent collection.73

Another institutional initiative came from the nation’s leading law publisher, West. In 1941, the company inaugurated the U.S. Code Congressional Service, a regular series of pamphlets (consolidated at the end of each year into a hardbound volume) containing the latest congressional acts and federal agency regulations, plus small selections of legislative history, scattered throughout. But then, in 1943, the Service began paying much more attention to legislative history: 329 pages’ worth during that year (consisting of a few selected documents on each major statute), with its own dedicated section of the annual volume.74

The third institutional initiative that occurred in apparent response to the Court’s surge came from the Law Librarians’ Society of the District of Columbia. Because so few D.C. law firms had librarians in the mid-1940s, the Society consisted mostly of federal agency librarians.75 These librarians had, over the preceding years, compiled a large number of what I call “scrapbook” legislative histories—books compiled by hand that each pertained to a specific statute, often with an index. For lawyers specializing in the implementation of one or a few statutes, such scrapbooks were far more useful and economical than having to rely on whole runs of the Serial Set or Congressional Record. Compiling even one scrapbook took much time and research, making it a precious item when complete.76 The federal agency librarians realized they collectively held a plethora of such scrapbooks, and knew that such histories were “more and more frequently referred to by courts and lawyers.”77 They therefore formed a committee in 1946 to assemble and publish a “Union List of Legislative Histories.”78 From its nucleus in D.C.-based federal agencies, this list would gradually expand in subsequent editions (there have now been seven, the most recent in 2000).79

Consistent with the idea that use of legislative history became normal in the 1940s, the practice began, in those years, to attract a new kind of criticism. For the first time, people began to speak of using legislative history as a dominant practice that was going too far. Frankfurter in 1947 mentioned “the quip that only when legislative history is doubtful do you go to [the text of] the statute.”80 Jackson, having enthusiastically used such material in the early 1940s, reversed himself in 1948: “I am coming to think it is a badly overdone practice . . . .”81 In 1948, Charles Curtis, the popular legal writer, recalled that the “courts used to be fastidious as to where they looked for the legislative intention,” but now they were “fumbling about in the ashcans of the legislative process for the shoddiest unenacted expressions of intention.”82 The first of several editions of Effective Appellate Advocacy concluded in 1950 that the surge of legislative history over the last decade had caused lawyers “infinite grief.”83

II. explaining normalization: new justices with new ideas

The rise of legislative history was caused in part by a rapid turnover in the personnel and therefore the ideology of the Court. I devote this part to explaining how the appointment of new progressive-minded Justices by President Roosevelt helped bring about normalization (before turning to the other major driving force—the administrative state—in Part III).

To appreciate the ideological turnover and its connection to legislative history, consider Table 1. It includes every Justice who wrote more than three opinions in federal statutory cases in 1930-45. It lists, for each Justice, the number of legislative history citations per opinion in federal statutory cases during that period, the percentage of the Justice’s opinions that contained any legislative history, and a breakdown of the citation/opinion ratio as between majority and separate opinions. The Justices are listed in descending order of their citation/opinion ratio for all opinions (which roughly correlates with the percentage of each Justice’s opinions that contained any legislative history).84

Table 1

As Table 1 demonstrates, there is a strong correlation between the likelihood that a given Justice would cite legislative history and his association with progressive ideology and President Roosevelt. Of the nine Justices in the top half of the order, seven were appointed by Roosevelt (in 1937-43) and the other two, Brandeis and Stone, had reputations as leading progressives on the pre-Roosevelt Court.85 Of the nine Justices in the bottom half, eight had joined the Court before Roosevelt took office in 1933. Those eight include all of the “Four Horsemen” (the Justices who voted most frequently to undermine the New Deal) and the moderates Hughes and Roberts (who sometimes voted against the New Deal). The Four Horsemen and Hughes all left the Court in 1937-41.

Further evidence of the connection between progressive ideology and legislative history appears in United States v. American Trucking Ass’ns,86 decided in May 1940, near the very start of the surge. In that case, Roosevelt’s appointees formed a five-to-four majority for an opinion expressing great openness to legislative material. In fact, some commentators have briefly suggested that American Trucking played something like a causal role in the rise of legislative history.87 A close examination of the opinion is thus warranted to determine its exact place in the process.

First, some background: in Anglo-American statutory interpretation, there had long been a maxim known as the “plain meaning rule,” which said that a judge ought to follow statutory text that was unambiguous on its face. The rule had originated as a counter to the ancient idea of equitable interpretation,88 but after Holy Trinity in 1892, some U.S. federal judges repurposed it as an instrument to limit the use of legislative history: in the updated formulation, legislative history was inadmissible if the text was facially clear.89 From the 1890s to the 1930s, the Court had sometimes applied the plain meaning rule and sometimes ignored it.90

The plain meaning rule was at the center of controversy in American Trucking. In 1935, Congress passed the Motor Carrier Act (MCA), which brought the trucking companies under the regulation of the industry-friendly Interstate Commerce Commission (ICC).91 The MCA empowered the ICC to set maximum hours for “employees” of the trucking companies. But the ICC construed this power to cover only safety-related employees (such as drivers), not all employees. This became a high-stakes matter in 1938, when Congress passed the Fair Labor Standards Act (FLSA), empowering the Department of Labor (DOL), which was not industry-friendly, to set maximum hours for all industrial employees with an exception for those already covered by the MCA. The trucking companies, hoping to escape the FLSA, argued that the MCA meant to cover all their employees. DOL disagreed (as did the ICC, which wanted to stick to its core competence of regulating safety).

The lower court held that the word “employees” unambiguously covered all the trucking companies’ employees and refused to consider legislative history indicating that the provision was aimed at preserving safety.92 But the Supreme Court sided with the government, five-to-four, with the five Roosevelt appointees in the majority and the four pre-Roosevelt Justices in the minority.93 Justice Reed, writing for the majority, selectively read the Court’s prior cases so as to effectively repudiate the plain meaning rule:

When [the text’s] meaning has led to absurd or futile results, . . . this Court has looked beyond the words to the purpose of the act. Frequently, . . . even when the plain meaning did not produce absurd results but merely an unreasonable one “plainly at variance with the policy of the legislation as a whole” this Court has followed that purpose, rather than the literal words. When aid to construction of the meaning of words, as used in the statute, is available, there certainly can be no “rule of law” which forbids its use, however clear the words may appear on “superficial examination.”94

The four dissenters wrote no opinion but simply endorsed “the reasons stated” by the lower court.95

Reed’s pronouncement in American Trucking was significant, but I do not think it predetermined the surge in legislative history that occurred over the next few years. Neither before 1940, nor since, has the Supreme Court treated statements on statutory interpretation methodology as having much precedential value. In the years up to 1940, the Justices, progressive and conservative alike, had been quite inconsistent in their adherence to the plain meaning rule.96 They had also been inconsistent on other methodological points regarding legislative history, such as the admissibility of floor debates.97 Likewise, in the first several years after 1940, the Court did not follow “consistently accepted principles of interpretation,”98 and this casualness about method encompassed legislative history.99 As Hart and Sacks wrote in the 1950s, American Trucking “was decided by a closely divided Court, and no one could then [in 1940] have been sure” that its statement about the universal admissibility of legislative history “really meant what it said.”100

Still, even if American Trucking did not control the Court’s post-1940 interpretive practice as a matter of precedent, it did reflect the tendency of progressive judges to be receptive to legislative history. That tendency makes sense in light of the progressive legal thought of the period.

The bête noire of progressive legal thinkers—such as Oliver Wendell Holmes, Louis Brandeis, Robert Hale, and Karl Llewellyn—was the classical notion that common-law doctrines in fields like property and contract were neutral and apolitical. Such doctrines, according to the progressive critique, were not determinate enough to decide particular cases. When judges claimed to make decisions on the basis of these doctrines, they in fact exercised discretion and made policy choices under a cloak of neutrality. Judges, counseled the progressives, ought to be explicit about the policy choices that inevitably underlay their decisions. Conscious, transparent articulation of the policies behind the common law was the best way to ensure the wise resolution of disputes. Although absolute determinacy was impossible, explicit policy reasoning would bring legal decisions nearer to that ideal than would the empty verbal formulas of doctrine.101

Because they did not revere common-law rights of property and contract as neutral and apolitical, progressives were receptive to regulatory and social-welfare legislation. But progressive scholars who studied legislation—such as James Landis and Frederick de Sloovère—found that judicial interpretation of statutes often exhibited pathologies similar to those of conventional common-law reasoning. In particular, judges often made a pretense of reasoning in a neutral and objective manner from the statute’s literal words, without reference to its policies. This often had the effect of undermining the legislature’s policy choices and substituting the judge’s own, under the cloak of literalism. By the early to mid-1930s, Landis, de Sloovère, and other scholars were pushing legislative history as an interpretive tool to ensure that judges would consciously confront legislators’ intent, rather than hide behind the false determinacy and objectivity of the text (as they did when invoking the plain meaning rule).102 But here the progressives ran into a problem: the charge of indeterminacy that they leveled against common-law doctrinal reasoning and statutory text-reading could also apply to the process of discerning the legislature’s “intent.”103

As William Eskridge shows, several progressive scholars of statutory interpretation in the late 1930s and 1940s converged on an answer to this question. They began by conceding that statutory interpretation, even if couched as the search for legislative intent, frequently required the judge to make discretionary choices.104 But they insisted that the judge could productively cooperate with the legislature by (1) seeking, in good faith, to identify the legislation’s overall objective and (2) making discretionary choices in a way that creatively furthered that objective.

The practical implication of this new thinking, as Eskridge points out, was that judges should reason from legislative history at a relatively high level of generality. Legislative history was a source from which the judge could infer the overall objective—the general intent—of the legislation. Having gleaned the general intent from the legislative material, the judge would then reason “downward” to figure out how best to achieve that objective on the specific facts of the case. This approach contrasted with an older academic conception of legislative history as useful for discerning the legislature’s specific intent—that is, the exact result the legislature would have wanted on the particular issue presented to the court, without reference to the statute’s larger objective or to the implementational creativity of the judge.105

The focus on general intent reached its culmination in Hart and Sacks’s 1958 classic The Legal Process.106 In their view, judges should use legislative history only insofar as it spoke to general intent, and general intent should be defined to include only rational, public-regarding goals. Since legislators’ duty was “to think in general terms,” the “probative force of materials from the internal legislative history of a statute varies in proportion to the generality of its bearing upon the purpose of the statute or provision in question.”107 If a congressman predicted in the legislative record that the bill would produce a certain result when applied to a certain situation, then a judge facing that situation should not automatically impose that result. Instead the prediction “should be given weight only to the extent that [it] fits rationally with other indicia of general purpose.”108 The judge was to attribute a public-regarding rationality to the legislature and take account of legislative history only insofar as it fit into that vision of rationality. Although “a court should try to put itself in imagination in the position of the legislature which enacted the measure,” it “should not do this in the mood of a cynical political observer, taking account of all the short-run currents of political expedience that swirl around any legislative session.”109 Rather the court “should assume, unless the contrary unmistakably appears, that the legislature was made up of reasonable persons pursuing reasonable purposes reasonably.”110 As Richard Posner later observed, Hart and Sacks apparently believed “that the judge should ignore interest groups, popular ignorance and prejudices, and other things that deflect legislators from the single-minded pursuit of the public interest as the judge would conceive it.”111 Hart and Sacks exhibited “a reluctance to recognize that many statutes are the product of compromise between opposing groups.”112

How much can the Court’s surge in legislative history during the early 1940s be understood as a manifestation of the interpretive theory that was just then emerging in progressive academic circles (and later reached its culmination in The Legal Process)? Arguably, the general-intent theory required an especially comprehensive analysis of legislative history, which would perhaps increase the rate of citation.113

There is surely an important connection between the theoretical developments and the surge, though not so strong as to provide a complete explanation for the latter. With the exception of a very short article by Landis in 1930,114 the earliest (and, until after World War II, the clearest) articulation of general-intent theory came from Harry Willmer Jones, in a series of three articles published in 1939-40.115 At the time, Jones was a young lecturer at Columbia and would start a professorship at Berkeley in 1940.116 Only the first of his three articles was published before May 1940; it was cited in Reed’s American Trucking opinion, while his other two articles were not.117 The first of Jones’s articles (the one cited by Reed) documented the Court’s frequent failure to adhere to the plain meaning rule and argued that textual meaning could never be plain without context, but it said almost nothing about using legislative history to discern general intent—a theme that Jones would elaborate only in the second and third articles. Consistent with this, American Trucking itself was hardly a learned treatise on general intent. It discussed interpretive methodology only briefly, quoting a 1922 opinion about the need to honor “the policy of the legislation as a whole” but saying nothing more about general intent.118

Hart and Sacks in The Legal Process briefly noted the surge in legislative history that followed American Trucking, but—significantly—they believed the surge had occurred without much of a theory behind it. Amid the “sudden efflorescence of citations of legislative materials in briefs and opinions” in the years after American Trucking, neither the Justices nor the lawyers “had a sufficiently articulated and disciplined theory of interpretation to know how to appraise the mass of available material properly and prevent abuses.”119

To be sure, there were many opinions in 1940-45 that used legislative history to discern intent at a high level of generality, then reasoned downward to the meaning of specific provisions.120 These opinions indicate that the emergent theory of general intent played some practical role in the surge. For example, in a 1943 case on whether a federal statute exempted Indian lands from state taxation, the state invoked the remarks of the bill’s sponsor, but Justice Murphy, writing for the Court, declared that, even if those remarks did support the state’s interpretation, “we do not accept them as definitive,” for they were “opposed to” other indicia of the act’s meaning, including “the reasons for its enactment,” as evidenced by the committee reports describing “the problem at which the . . . Act was aimed.”121 Similarly, in a 1945 case on whether the FLSA allowed backpay claimants to settle their suits, Reed wrote for the Court that

[n]either the statutory language, the legislative reports nor the debates indicates that the question at issue was specifically considered and resolved by Congress. In the absence of evidence of specific Congressional intent, it becomes necessary to resort to a broader consideration of the legislative policy behind this provision as evidenced by its legislative history and the provisions in and structure of the Act.122

Several other opinions followed this pattern.123

But at the same time—in keeping with Hart and Sacks’s view that the surge in legislative history did not conform to a particular theory—there were also many opinions that used such material to discern legislative intent at a high level of specificity, without self-consciously elaborating a public-regarding policy. The persistence of this approach is consistent with the intellectual history told by Eskridge, who traces the interest in general intent at the academic-intellectual level and notes its resonance with American Trucking but adds that the older interest in specific intent survived and coexisted with the new thinking.124 For instance, in 1941, when deciding whether the Chandler Act of 1938 repealed an emergency 1933 statute establishing a broad definition of “farmer” for the purpose of certain bankruptcy protections, Murphy wrote an opinion for the Court citing copious legislative history indicating congressmen’s intention to leave the 1933 provision alone, with virtually no policy discussion.125 As another example, Reed in 1943 wrote an opinion for the Court holding that the statutory provisions on extra compensation for customs inspectors covered work at night but not Sundays and holidays. He cited a mountain of legislative history about the understandings of various congressmen and stakeholders but without weaving them into any broader purposive thinking.126 Several more opinions took this approach in using legislative history.127

At times, a single opinion relied upon legislative history both to discern general intent and to confirm the correctness of the specific application. Thus, Frankfurter in 1941 wrote that the Wagner Act’s purpose (evident from its legislative history) was to encourage worker self-organization for industrial peace, and he concluded that this purpose required the Court to read the Act’s anti-discrimination provision to cover not only firing but also hiring. But he also added that a committee report contemplated this exact application.128 In 1943, Black wrote an opinion giving a very broad reading to Congress’s 1939 abolition of assumption of risk under the Federal Employers’ Liability Act, basing his conclusion both on the enactment’s general policy and on the fact that congressmen during the hearings had criticized a particular case (whose holding, the opinion therefore reasoned, the statute must be read to override).129 In 1944, the Carolene Products Company argued that its milk compounds did not come within a statutory ban on products “in imitation or semblance of milk,” since the act (said the Company) required the imitation to be deliberate. The Court, per Reed, rejected this argument because the committee reports named specific products that were meant to be covered, which included “compounds of this innocent character.” Having said this, Reed could not resist adding that Carolene Products’ own “compounds were themselves so named.”130

The absence of a simple causal relationship between general-intent thinking and the surge in legislative history is especially evident in two further cases: Viereck v. United States (1943) and Gemsco, Inc. v. Walling (1945). In Viereck, the Court had to decide whether a statute requiring a foreign agent to disclose propaganda activities encompassed situations in which the agent spread propaganda on his own behalf, rather than on behalf of his foreign principals. On the basis of legislative history, Black argued that because the “general intent of the Act was to prevent secrecy as to any kind of political propaganda activity by foreign agents,” the act should be read broadly.131 But Black commanded only one additional vote for his position.132 Opposing him and writing for the majority was Stone, who acknowledged Black’s argument (and the government’s) that the act, if not read broadly, would be a “halfway measure.” But Stone marshaled post-enactment legislative history to show that “Congress itself . . . recognized that the legislation was in this sense a halfway measure.”133 Further, Stone (who showed himself capable of reasoning about general intent elsewhere134) expressed anxiety about this method:

While Congress undoubtedly had a general purpose to regulate agents of foreign principals . . . we cannot add to [the Act’s] provisions other requirements merely because we think they might more successfully have effectuated that purpose. And we find nothing in the legislative history of the Act to indicate that anyone concerned in its adoption had any thought of [the application urged by the government].135

Here, Stone gathered his own history to counter Black’s history-based invocation of general intent.

In Gemsco, the Court faced the question of whether the FLSA authorized the DOL to ban home-work across the entire embroideries industry if the agency concluded that doing so was necessary to enforce the minimum wage in that industry. Employers cited copious legislative history implying a political deal not to ban home-work outright, but Rutledge, writing for the Court, cast the history as ambiguous and insisted that, in any event, the industry’s reading would “deprive the [DOL] of the only means available to make [the statute’s] mandate effective” and “would make the statute a dead letter for this industry.”136 The “necessity to avoid self-nullification” was enough to refute the industry’s reading.137 In this case, legislative history was the enemy of general intent.

The Justices, when using legislative history, did not confine themselves to an idealized, rational vision of the legislature. On the contrary, they could, at times, be brutally frank about the realities of Congress. This was most true of Black, the former U.S. Senator. In one opinion, he emphasized that a certain interpretation of a bill had been “accepted both by the unions and the railroads” at the hearings.138 In another opinion, he pointed out that a federal statute alleged to override Oklahoma state taxes had been “sponsored by Oklahoma Congressmen who said nothing which supports the imputation that they intended to deprive their State of this income.”139 In a third, concerning the ICC’s treatment of railroad and barge lines, he fulminated that the “railway-minded” ICC was taking the very anti-barge actions that senators “from the midwestern states where the barge lines . . . were operating” had feared it would take—and which the statute’s cheerleaders had assured them it could not take.140 Similarly Reed, in a case about an excise tax on the processing of imported oils, based his reading on the special-interest nature of the statute: “the legislative history cannot be read without reaching a conviction that the advantages which would result to American vegetable oil producers from the heavy tax on oils not produced in the continental United States played a leading part in promoting the legislation.”141 For his part, Roberts argued that one of the provisions of the Federal Power Act, having been amended to meet the demands of state regulators, should not be construed to encroach on their turf.142 Frankfurter, in one instance, pointed out the unusually long “continuity of membership” on the House and Senate Commerce Committees, and he suggested the Court should assume that bills from those committees contained “skilled language” whose variation from other relevant statutes must be meaningful (apparently implying that other committees were less “skilled”).143

All in all, the coincidence between Roosevelt’s appointments and the surge in legislative history—combined with the intellectual resonance between progressive legal thought and legislative history—indicates that the ideologically-charged turnover of the Justices was surely a cause of the normalization of this interpretive source. Yet the turnover is not a complete explanation. We know it is incomplete because of the tensions (just discussed) between the emergent progressive legal theory of legislative history and the Justices’ actual use of that material.

And we know it is incomplete for another reason: even the pre-Roosevelt Justices (the dissenters in American Trucking) greatly increased their use of legislative history in the 1940s. On this point, consider the examples of two Justices who joined the Court well before 1940 and served till well after 1940: the progressive Harlan F. Stone and the moderate Owen Roberts. In the years 1930-39, Stone had a citation/opinion ratio of 1.5, but in 1940-45, it was double that (3.0). Similarly, Roberts in 1932-39 had a ratio of 0.5, but in 1940-45, it was more than double (1.4). A similar pattern obtains for the moderate Hughes: his ratio grew from 0.9 in 1930-39 to 1.7 in 1940-41 (though we should not assign too much weight to this, since Hughes’s track record in 1940-41 was short). The two early Roosevelt appointees who acquired more-than-negligible track records during the 1930s also exhibited a big increase in their citation rates in 1940 and later. Black in 1937-39 had a citation/opinion ratio of 0.8, which more than doubled to 1.9 in 1940-45. Reed in 1938-39 had a ratio of 0.9, which quintupled to 4.7 in 1940-45. The experiences of all these Justices suggest that we cannot explain the transformation simply by saying that (1) progressive Justices were more inclined to cite legislative history and (2) Roosevelt appointed more progressive Justices. Something more was happening.

III. explaining normalization: the new administrative state

The progressive ideology of the Roosevelt appointees was a major but not sufficient cause of the interpretive revolution. Its insufficiency, as noted above, is evident from the tensions between progressive theory and judicial practice and from the pre-Roosevelt Justices’ change in behavior.

But even if the shift toward legislative history had occurred exclusively in the opinions of the Roosevelt Justices, and even if those Justices’ use of legislative history had corresponded perfectly with cutting-edge progressive thought, that still would not be enough to explain the interpretive revolution, for this was not the kind of revolution that judges, however intellectually committed, could effectuate on their own. Congress was a remarkably complex body that produced a rich, sophisticated, confusing, and voluminous discourse. Serious, sustained, and sympathetic engagement with that discourse—especially in the holistic manner urged by the progressive jurists in their focus on general intent—required an intimacy with Congress, a sophistication about its processes, and an institutional capacity to monitor its sheer output that had rarely, if ever, been seen in American litigation.

The initiative for systematically briefing congressional discourse did not come—could never have come—from ordinary lawyers. It originated instead from the New Deal administrative state, vested with unprecedented capability to process and analyze congressional discourse and translate it into legal argument. The federal leviathan provided the indispensable institutional basis for what might seem, on the surface, to be simply a change in judges’ ideas.

A. The Court’s Reliance on the Federal Government for Legislative History

To appreciate the importance of the federal government in the Court’s use of legislative history, we must first recognize that the federal government was an overwhelmingly familiar presence before the Court in federal statutory cases during the late 1930s and especially the early 1940s. As my research team and I found, in 1936-39, 68% of federal statutory cases were briefed by the federal government, usually as a party, but sometimes as an amicus curiae. In 1940-45, during the surge, the proportion was even higher, at 78%.144

Further, the Court, especially during the surge, was far more inclined to cite legislative history in federal statutory cases that were briefed by the federal government than in those without a federal brief. In 1940-45, the number of legislative history citations per statutory case was 3.9 for those with federal briefs and only 2.2 for those without, and the proportion of statutory cases citing any legislative history was 50% for those with federal briefs and only 31% for those without. Notably, the divergence between cases with federal briefs and those without was less clear prior to 1940. In 1936-39, the number of legislative history citations per case was 0.9 for statutory cases with federal briefs and 1.0 for those without (though there was a larger divergence in the proportion of statutory cases citing any legislative history: 28% for those with federal briefs and 14% for those without).

Thus, during the surge of legislative history in 1940-45, the Court was hearing from the federal government in a large and increasing majority of statutory cases, and the Justices’ citations were becoming more concentrated in those federally briefed cases. Of all the legislative history citations appearing in federal statutory cases in 1940-45, 86% were in cases with federal briefs.

This evidence suggests that the Court was leaning on the federal government to provide it with legislative history research. To test that possibility, I took a sample of all federal statutory cases citing legislative history in 1940-45,145 and I had research assistants determine whether the Justices’ citations in those cases “matched” the citations appearing in the briefs that were submitted.146

The results are as follows. First consider the federally briefed cases (which, to reiterate, were 78% of all federal statutory cases and accounted for 86% of all legislative history citations). In these cases, I found, through my sample, 22% of the citations matched both the federal brief and at least one non-federal brief (such as the brief of a private party or state government); 33% of the citations matched the federal brief and no other brief; 10% of the citations matched at least one non-federal brief but not the federal brief; and 34% of the citations matched no brief (suggesting they arose from the Court’s own research).

Thus, the Justices relied far more on federal government lawyers for legislative history than on any other lawyers arguing before them. It was 50% more common for the Justices to take a citation exclusively from the federal lawyers than to take one jointly from federal lawyers and non-federal lawyers. And it was more than three times more common for the Justices to take a citation exclusively from federal lawyers than to take one exclusively from non-federal lawyers.147 (The high proportion of citations apparently arising from the Court’s own research is also interesting; I shall have more to say about that later.148)

For further confirmation of federal lawyers’ dominance over other lawyers in providing legislative history, let us now consider the cases without federal briefs. These, again, represented only 22% of federal statutory cases and accounted for only 14% of all legislative history citations. In these cases, I found, based on my sample, that 45% of the citations appeared in at least one of the briefs, while 55% did not (suggesting they came from the Court’s own research). That is, the Justices themselves provided a much higher proportion of the citations (55%) than they did in federally briefed cases (34%), presumably because they did not have the federal government to help them. This helps explain the fact (noted earlier) that the Justices cited legislative history only about half as much in non-federally-briefed cases.

Though the surviving case files of the Justices from this period are mostly thin, there is one archival document confirming that at least one member of the Court viewed the federal government as the “go-to” source of legislative history. In December 1941—two full years into the surge—Justice Reed and his clerk were working on a case between Alabama food-safety officials and a butter-making company. The question was whether certain federal food-safety statutes should be construed to preempt Alabama laws on the subject. Lawyers for Alabama and the company briefed the case, but the federal government was not a party and submitted no brief. Reed wrote to his clerk:

Please check the reports [i.e., committee reports] and Congressional debates to see what was the purpose of the adoption by the [federal] Renovated Butter Act of the details of the [federal] Meat Inspection Act. It may be that this was in the briefs but as there have been no Government briefs, it is not so likely that it will be.149

By “Government briefs,” Reed clearly meant federal government briefs, for the Alabama attorney general had briefed the case.

B. The Difficulty of Briefing Legislative History

To understand why the federal government was so dominant in providing legislative history to the Court, we must first appreciate the extraordinary difficulty of employing that material in legal argument, especially as of the 1940s. Compared to case law and statutory text, documents of legislative history were much harder to identify, obtain, and use. The people who wrote judicial opinions and statutory text wrote those documents with the primary purpose of making them usable in future legal decision-making, and the West Company in the 1880s had built an infrastructure to increase that usability. By contrast, the people who produced legislative history had far more diverse purposes; there was no infrastructure remotely comparable to the West Digest, and building any such thing would have been a Herculean task.150

Begin by considering the textbook version of the process of enacting a statute. In one chamber, a member introduced a bill. The chamber’s parliamentarian gave the bill a number (such as S. 123 or H.R. 456) and referred it to a committee. The committee might hold hearings on the bill and print documents relevant to it, such as executive branch recommendations. The committee then sent the bill back to the chamber, with a report. The whole chamber then took up the bill, debated it, and (possibly) amended it, before finally passing it. The bill then went to the other chamber, where the textbook process required all the same steps. If the second chamber passed a version of the bill that was not identical to the first chamber’s, the two chambers could appoint a conference committee, which then issued a report of its own. Thus, even the textbook process, applied to a single bill, could produce many documents: committee reports in both chambers, committee hearings and documents in both chambers, floor debates and amendments in both chambers, and a conference report, not to mention all the successive versions of the bill in both chambers.

Identifying all the relevant documents in the textbook process for a single bill was mostly easy, though not entirely. Starting with the statute, one could quickly find the number of the bill that had become that statute. One could then look up the bill number in the index to the Congressional Record for the Congress that enacted the statute and find references to all committee reports, floor proceedings, and conference reports on that bill. But the Record index did not cover hearings. Locating hearings required a more cumbersome process using a different (and less widely available) index, though it was still doable.151

The main problem in identifying documents was that the textbook process for a single bill was dwarfed, in both scale and complexity, by the usual “real-life” process for an important statute. It was common for the text of a single statute to be pasted together, through the amendment process, from several different bills (or pieces of bills), each of which had its own distinct number and might, in itself, have gone through some (or nearly all) of the steps in the textbook process. That is, the bill that officially became law might encompass amendments that occasioned little comment when offered but whose texts began life as separate bills on which extensive committee hearings, reports, or floor debates had occurred,152 often in Congresses prior to the one that actually passed the law.153 Crucially, the indexing was not remotely adequate to identify all the bills that had been sewn together into the final product. Only in the late 1930s did the bill-number index of the Congressional Record begin including complete substitutionsof one numbered bill for a different numbered bill,154 but complete substitutions were merely the tip of the iceberg: many statutes involved more complicated cut-and-paste jobs involving several bills, some from the present Congress and others from prior Congresses.155 Even if a legislative proposal never underwent combination with other proposals, its legislative history could still be fragmented if (as often happened) deliberations extended over multiple Congresses, since unenacted bills expired at the end of every Congress and had to be introduced at the start of the next Congress as “new” bills with new numbers.156 There was no index linking the old bill to the new one, even if their texts were the same. One had to read the reports, hearings, and debates on the enacted bill in hope that somebody would mention the existence of earlier bills, which was not guaranteed.157 Thus, merely identifying all the documents composing the legislative history of all the bills relevant to an eventual statute “may be long and laborious,” as a standard research guide warned in 1947.158 There was “no entirely adequate reference aid published by the government to cover the period between the introduction of a bill in Congress and its printing . . . in the ‘Statutes at Large.’”159

Assuming the researcher identified all the documents on all the relevant bills, the trouble had only begun. Obtaining them could be very hard. “Committee reports and hearings,” noted Finley in 1946, “go out of print in a very few years.” The only sure way to obtain such documents was to purchase them during the enactment process, hot off the press. Otherwise, warned Finley, it was a nightmare: “even if you have the citations to all the material you need, you will be unable to find most of it.”160 The nation’s two leading tax professors—Erwin Griswold at Harvard and Roswell Magill at Columbia—attested to this problem. Committee reports on internal revenue, complained Griswold in 1939, “have been very difficult to find, and sometimes almost unobtainable.”161 That same year, Magill agreed that the reports were “highly fugitive, for the editions of the reports are not large, and even a law librarian must exercise more vigilance than many do in order to acquire copies in time.”162 “Committee hearings,” noted a standard research guide in 1951, “especially older ones, are often difficult to obtain.”163 As to the Congressional Record, it was so vast that “few law offices would have shelf-room for” it, as Magill said.164 In general, concluded Finley, “[i]t is next to impossible to collect extra copies of the necessary material [for a legislative history] after several years have passed [since enactment].”165

One also needed to obtain the changing versions of the bill as it had moved through the process, but this, too, might be tough. The bill’s successive versions were essential, not only in themselves, but also because, in their absence, a researcher could fall into disastrous misunderstandings of the reports, hearings, and debates. A committee, a congressman, or a witness might make a remark about a particular section of the bill, but it was dangerous for a lawyer to cite that remark without confirming what the wording of the section had been at the time of the remark.166 Successive versions of the bill were printed for the use of congressmen during the process, but beyond Capitol Hill, they were not generally available.167 To be sure, it was common for committee reports, hearing transcripts, or floor debates to include the current text of the bill, but there was no fixed rule about this,168 and finding the successive versions might require scanning through several documents.

One might obtain documents from federal depository libraries, but even those institutions fell short. In the late 1930s, about 500 libraries were designated as federal depositories,169 but they were not required to receive all the congressional documents offered them, so “many” did not have “complete collections” of offered materials.170 This was especially true for hearings and bill prints: up to 1938, these two categories of material were not even offered,171 and even after 1938, depositories normally did not accept them (unless they were state libraries or big university libraries).172 Committee reports and floor debates were somewhat easier: as of the 1950s, Finley thought it “most unlikely that any sizeable city is without a set [i.e., at least one set] of the Congressional Record and the Congressional Serial Set, which collects the committee reports,”173 although Griswold and Magill’s complaints about committee reports in 1939 may indicate that Finley was too sanguine about the completeness of such collections, or that collections had been thinner prior to the 1950s. In any event, lawyers not near a large city were clearly out of luck. “In most cases,” wrote two attorneys in 1948, “particularly when the lawyer does not practice in or near a large city, there is no convenient access to Government publications.”174

Even if the researcher identified all documents and obtained them, it would often prove extremely time-consuming to use them. For a major statute, all the relevant documents on all the relevant bills could add up to thousands of pages. For example, when some law-firm librarians in the 1970s retrospectively compiled the complete legislative history of the Securities Exchange Act of 1934, it ran to about 4,000 pages.175 When in 1949 the National Labor Relations Board (NLRB) published the legislative history of the Wagner Act, it ran over 3,200 pages.176 When a drug-industry lobbyist compiled a legislative history of the Food, Drug, and Cosmetic Act of 1938, it ran to over 1,300 pages, even though it omitted 2,000 pages of hearings.177

Facing such a vastness of material, the researcher’s instinct would be to read selectively. That was easier said than done.178 Hearings and floor debates often involved desultory discussion, in which any topic might arise at any time. For a hearing, there was typically an index of the names of witnesses but not subjects. For the Congressional Record, there was a subject index for each entire two-year Congress, but this was far too general to find discussion of a particular section or issue within the debate on a particular bill. J.S. Seidman, an expert in tax law, wrote in 1940:

[The researcher]’s interest at any time is likely to be in only a particular provision or segment of the law. Nevertheless, each time he wanted to check its legislative history, he had to plow through virtually every page of all of the material to see what, if anything, it contained on the subject. Why this search for the needle in the haystack? For the reason that, with the exception of bill prints, the material is generally not marked or arranged in any particular order. To the contrary, it is in a state of disorder. Suppose we are looking up section 23(c). In the hearings, discussion of section 23(c) may be preceded by consideration of section 102 and followed by section 12(b). Later on, they may be back on 23(c) again. Or there may be nothing on 23(c) at all, but we would never know it until we went completely through the hearings. So, also, with the Congressional [floor] debates.179

Similarly, in 1955, the chief librarian of the U.S. Supreme Court sent a memo to the Justices’ law clerks, warning: “The Library Staff cannot ordinarily undertake the responsibility of selecting material pertinent to a particular [statutory] section from often lengthy debates and hearings.”180

The lack of indexing in the Record occasioned particular comment. In a 1930 opinion, U.S. District Judge William Clark quoted a floor speech on the Bankruptcy Code of 1898 that contradicted an opinion the Supreme Court had handed down in 1913 without benefit of the legislative history.181 Clark noted that he found out about the 1898 floor speech only because, in writing a scholarly book on bankruptcy, he had assumed the “arduous task” of reading the floor proceedings.182 Even then, the indexing of the Congressional Record was so inadequate that Clark feared there might be some other snippet of legislative history saying something different about the provision at issue. On this point, he lamented that the West Key Number System for case law had no analogue in legislative history: “Without the help of some such trained experts as are employed by the West Publishing Company, it can be only reasonably certain,” not fully certain, that the quotation he had found was “the only reference” to the Code section at issue.183 Clark stressed the crying need “for some sort of digest of the Congressional Record for the use of the United States judges and others required to interpret Federal legislation.”184 In 1939, Magill lamented that the “discussions in the Congressional Record . . . are quite inaccessible, due to inadequate indexing.”185 Committee reports were less problematic in this respect, since they were relatively short and might be organized according to the bill’s section numbers, though even then, the section numbers might have changed after the issuance of the report, so the researcher would have to scan the whole report.186

Even if a researcher waded into the history and found some nugget that appeared dispositive, it was dangerous to stop there and declare victory. The history encompassed the utterances of so many different congressmen and witnesses that one snippet might be contradicted by some other snippet. “On particular occasions,” warned the first edition of Effective Appellate Advocacy in 1950, “the committee reports will be valueless—because later disavowed on the floor of Congress.”187 Altogether, said Jackson in 1948, “[i]t is a poor cause that cannot find some plausible support in legislative history.”188 The whole matter was summed up by Frederick C. Hicks, the librarian of Yale Law School, in 1942: “Even though [legislative] material is available in libraries, it is of large volume, and, in use, very time-consuming, in spite of the existing lists and indexes.”189

At the Supreme Court itself, law clerks in the 1950s singled out legislative history as the one kind of source that the Court’s library staff was incapable of handling. As Frankfurter’s outgoing law clerks, in a message to their successors, put it:

The library staff is generally excellent in digging things out. But experience in the 1954 Term indicated that it is unwise to rely on their research on such matters as legislative history. If, for example, you wish all the hearings on a bill, you should check to make sure that they have found all materials.190

As lawyers and librarians increasingly recognized, the ideal method for rendering legislative history usable was for somebody to collect all the relevant documents for a single statute, bind them in a “scrapbook,” and draw up an index for the scrapbook.191 But this method was extremely costly. Assembling such a scrapbook was nearly impossible for a past statute; rather, the work needed to be done in “real time” as the enactment process was unfolding, so the compiler could purchase the documents before they went out of print. Gathering the documents in “real time” required a skilled researcher who could follow any and all congressional activity that might lead to a statute of interest. Since nobody knew in advance whether any particular bill would pass, the time expended for a single statute might be huge.192 “Tracing or compiling a legislative history,” wrote Finley in 1949, “may seem like a devious [i.e., tortuous] and laborious task. It certainly, to say the least, is troublesome.”193 “If complete compiled histories of all [federal] laws were to be undertaken,” she observed in the 1950s, “there would probably have to be a whole new agency to do the job.”194

Another possibility was that a researcher might compile all documents for a statute, not in the form of a scrapbook, but as a publication to be sold commercially. Theoretically, commercial sale could cover the large investment necessary to compile and index the documents. But so long as legislative history was not a routinely used source, the market would never be large enough for this purpose. A few published volumes appeared on the legislative histories of particular statutes before 1945, but (with four arguable exceptions) they were really just glorified bibliographies or highly selective collections of excerpts, not remotely comprehensive collections of the primary documents.195 The commercial publication of full-blown legislative histories (which finally took off in the 1950s and especially the 1970s) was a result of the normalization of legislative history, not its cause.

Even Harry Willmer Jones—who wrote far more scholarship arguing in favor of legislative history than anyone before 1945—admitted, as of 1940, that lawyers were not practically able to brief such material: “The unsatisfactory nature of the typical judicial reference to extrinsic aids is due, it would seem, to the incomplete presentation of relevant materials by counsel who would gain by their introduction.” Tellingly, Jones analogized the briefing of legislative history for statutory interpretation to the briefing of legislative facts in constitutional litigation (which were often drawn from legislative sources).196 Briefing legislative facts in constitutional litigation was labor-intensive and was peculiarly the competence of federal government lawyers—a profoundly significant fact, as we shall see in the next section.

C. The Federal Government’s Unique Capacity to Brief Legislative History

The federal government was uniquely suited to meet the challenge of briefing legislative history. To understand this, we must first recognize that when the federal government submitted a brief to the Court, that brief was a product of the federal bureaucracy writ large. In the 1930s and early 1940s, the Department of Justice (DOJ) in Washington, D.C., had five main divisions that focused on litigation. These were the Antitrust Division (which covered not only antitrust but business regulation generally),197 the Tax Division, the Claims Division, the Criminal Division, and the Public Lands Division.198 The DOJ also included the Office of the Solicitor General (OSG), which throughout the period consisted of around a dozen lawyers, far fewer than any one division had.199 The OSG was charged with representing the federal government before the Supreme Court,200 and although some independent agencies had authority to litigate on their own, the OSG could and normally did serve as their official representative,201 and so the SG’s signature appeared on nearly all federal briefs filed at the Court.202 But in reality, the OSG possessed only a tiny fraction of the manpower necessary to brief the numerous Supreme Court cases of which it was nominally in charge. The standard practice was for the brief to be drafted outside the OSG, most commonly by one of the DOJ divisions, but also, in a large minority of cases, by a separate federal agency such as the NLRB, ICC, SEC, FCC or DOL.203 (When a DOJ division drafted the brief, it would also consult with whatever agency was involved in the case.204) After the brief was drafted, one of the OSG lawyers would then revise it, to a greater or lesser extent, usually over a period of one or two weeks just before filing it with the Court.205 In this process, the “spade work” of research normally came mainly from the DOJ division or agency.206 This did not mean that the DOJ division or agency simply recycled whatever research it had already done in the lower courts; rather it appears that the divisions and agencies did special preparation for cases that went to the high Court.207

The agencies and DOJ divisions—the institutions that collectively did most of the work of briefing Supreme Court cases for the federal government—were better suited to research legislative history than any other institutions in America. First, they simply had more personnel to do legal research. To get a sense of this, consider that in 1938 the five DOJ divisions together had 421 lawyers in Washington, D.C.208 Federal agencies besides DOJ in 1939 totaled more than 4,000 lawyers nationwide.209 Assuming very conservatively that non-DOJ federal lawyers lived in the Washington area at the same rate as federal employees more generally, approximately 550 additional federal lawyers would have resided in and around the nation’s capital.210 Numbers like this dwarfed all other organizations. In 1935, the largest law firm in Chicago—the nation’s second-biggest city—had 43 lawyers.211 Even in 1950, according to the best estimates, no law firm in America had more than 92 lawyers, and only eleven had more than 60 (of which nine were in New York City).212 The New York City Law Department, largest of all government law offices below the federal level, had only about 100 lawyers in 1940.213

The federal government’s sheer number of lawyers allowed for a unique degree of specialization. In an era when private-sector appellate practice groups did not yet exist,214 the five DOJ divisions all had their own appellate sections by the late 1930s and early 1940s,215 and some had their own Supreme Court sections.216 Not only had the federal government pioneered appellate practice, but it sub-specialized appellate practice by substantive area.

The manpower advantage extended not only to lawyers but also (it seems) to law librarians. Regional societies of law librarians first began to form in 1938-39, and two of the first three were in New York and D.C.217 The New York society was apparently a diverse coalition from courts, bar associations, universities, and law firms.218 In D.C., however, it appears no law firms had librarians before 1943,219 and the society was dominated by federal agency librarians, with a minority from courts and universities.220

Beyond raw manpower, there were even deeper reasons why the agencies and DOJ divisions were uniquely suited to research legislative history. For one, they each devoted themselves full-time to implementing one or a few statutes, so they could amortize the cost of researching the background of those statutes over many cases. Moreover, they had uses for legislative history far beyond litigation. Program officers needed guidance on implementation, which legislative history could help provide. Further, the agencies and DOJ divisions constantly communicated and negotiated with Congress on how to administer (and whether to amend) the statutes in their charge, and legislative history was essential to that work—something that drove the agencies and DOJ not merely to collect legislative history, but to collect it in “real time,” which was the most efficient way.221

Take for example the Treasury Department. Its Legal Division, established in 1934, had three missions, for all of which legislative history was valuable. First was litigation: the Division “cooperated with the Department of Justice in preparing a large number of cases in which the Treasury was interested.” Second was advising Congress: the Division “prepared analyses and recommendations [in a single year] with respect to approximately 650 congressional bills affecting the Department” and “furnished technical assistance to congressional committees.” Third was program implementation: the Division produced “formal legal opinions and numerous informal opinions” for “the information and guidance of the administrative officers of the Department.”222 To help with all these tasks, the Division compiled legislative history on a systematic and ongoing basis:

[The] library compiles and binds in folders the legislative histories of current bills in which the Treasury is interested. The folders contain the debates taken from the daily Congressional Record, the reports and bills. The folders are filed by Congress and bill number. The hearings are also listed by Congress and bill number, as well as being shelf-listed and cataloged.223

In a testament to the role of legislative history in program implementation, Treasury’s Internal Revenue Bureau in 1939 distributed an 800-page volume of committee reports on all revenue acts going back to 1913, “in order that they may be available to officers and employees of the Bureau both in Washington and in the field divisions in construing the provisions of the various” statutes.224

As another example, consider the Interstate Commerce Commission (ICC). Founded as a tiny agency in 1887, the Commission expanded into a behemoth in the 1910s, establishing a Bureau of Law.225 As of the 1930s, that Bureau had the usual three missions: to represent the ICC in litigation, to advise on bills the ICC might propose, and to advise internally on applying the existing statutes.226 The Bureau accordingly kept “itself continually informed concerning legislative . . . developments.”227 Indeed, it was in charge of the ICC’s library, which maintained “a large file of congressional bills and resolutions from the creation of the Commission in 1887 to date” and was responsible for “indexing the same,” and also for “[r]eading the Congressional Record daily and indexing same.”228

As a third example, take the DOJ Tax Division. Founded in 1919 and greatly expanded in the 1930s, the division in 1931 hired an attorney named Carlton Fox,229 who became “a pioneer in the field of legislative histories.”230 By the late 1930s, Fox had assembled scrapbooks of primary material on the internal revenue laws—apparently consisting largely of legislative history—that totaled forty-five volumes.231 When a private tax attorney published a one-volume “legislative history” of the internal revenue laws in 1938, Griswold criticized itin the Harvard Law Review, arguing that it was not “much more than a digest or partial index,” declaring that the “only place where the complete sources may now be found is in the compilations which have been made by Carlton Fox, Esq., of the Department of Justice,” and concluding that “[t]he final story on the legislative history of the Revenue Acts will not be told until Mr. Fox’s compilations can be made more readily available.”232 The Tax Division apparently took advantage of its legislative history to guide the DOJ field service, sending circulars to the U.S. Attorneys’ Offices with legislative history on commonly litigated questions.233

Legislative history operations were common across the federal administrative state. Thus, although many of the congressional materials on the Securities Act of 1933 and the Securities Exchange Act of 1934 became rare and almost unobtainable shortly after enactment, the SEC was the great exception: it possessed the “almost one and only official history” of the two statutes, from which librarians “charted [their] course” in compiling a commercially published history in the 1970s.234 At the Federal Works Agency (a conglomeration of the Works Progress Administration and other agencies), the librarian in 1941 explained that

one of our chief services is to compile these so-called legislative histories....We get the material together by following the Congressional Record proceedings very closely, and we found that by binding all of that material together, . . . it has been of invaluable help to our attorneys in getting the interpretation of the law and also in drafting subsequent legislation.235

The dominance of the federal leviathan in using legislative history was a foregone conclusion among experts in legal research. Speaking before the AALL in 1947, Finley said, “I know the librarians here from government agencies will bear me out when I say that legislative histories are the most important non-routine part of a practising lawyer’s library.”236 An article in 1948 noted that the best approach to compiling legislative history—“[k]eeping a current file on all bills before Congress which may become important laws”—was “a job requiring the full attention of an experienced librarian and is consequently a method employed mainly by Government agencies.”237 It added that there was “only one private law firm at the present time [presumably Covington] that has such a primary interest in Congressional legislation as to justify such a method.”238 In 1953, the first edition of Effective Legal Research, by the Columbia law librarians, held up federal agencies as the gold standard for legislative history research. The book set forth an elaborate procedure for tracking bills and gathering congressional material in real time, which it characterized as “typical” for a federal agency but “perhaps somewhat more elaborate than most law offices would follow.” The method could “be simplified to the extent desired.”239

The advantage that the government consequently enjoyed in litigation was obvious. When two private attorneys in 1948 published a compiled legislative history of the Revenue Act of that year, they advertised it as part of a private effort to catch up with the government. The “Government attorney in the Nation’s Capital,” they warned,

has a decided advantage over the non-government practitioner who in most instances does not work in Washington and who, even though he may have the time at his disposal to make a search of voluminous legislative history material, invariably cannot find all the documents that need to be examined. The result of this disparity in the availability of source material is that the attorney for the Government generally is prepared to meet the problem while the private practitioner usually is not.240

One last point: agencies and DOJ divisions watched Congress carefully and recorded what it did, but in truth, their relationship to legislative history was even more intimate. They were watching and recording a drama in which they themselves were leading players—perhaps the leading players. Agency and DOJ employees drafted bills, provided congressmen with analyses, testified at hearings, and even served as ghostwriters for committee reports and floor speeches.241 Such practices have been a constant in U.S. history, to some degree.242

That said, it seems probable that the New Deal and World War II saw the greatest executive-branch dominance of congressional drafting and deliberations that America has ever witnessed. The volume and gravity of legislation in those years was unprecedented, but Congress did not seriously begin to build up its own internal administrative capacity until the Legislative Reorganization Act of 1946 and after. Thus, up to 1946, the House and Senate standing committees had no regular professional staff. “When such staff assistance was needed,” noted one political scientist, “it was borrowed on detail from an executive agency,” still on the executive payroll.243 Though personal staff for individual members had crept upward since the 1890s, the number stood in 1944 at only three for each Representative and six for each Senator (less than one-fifth what it is today).244 The 1946 Act provided four professional staffers per committee,245 but this was so few that they generally would merely review agencies’ proposals, not take initiative themselves.246 Only in the 1960s and 1970s would congressional committee staff grow into a major force.247 Likewise for the House and Senate Offices of Legislative Counsel. In the early 1940s, these offices, which helped congressmen with drafting upon request, had only five lawyers each.248 With such tiny numbers, they could not rival the administrative state.249 The Legislative Counsel Offices would expand after 1946 and become important only later.250 Today they have more than seven times as many lawyers as they had before 1946.251

With such a crushing demand for legislation and so little internal capacity to meet it, Congress leaned heavily on the administrative state in drafting and deliberations in 1933-45. On this point, a key observer was Edwin Witte. An economist at the University of Wisconsin, Witte had worked for a congressman and a federal investigatory commission in 1912-15; then as executive director of FDR’s Committee on Economic Security in 1934-35, in which capacity he was the chief lobbyist for the Social Security Act; and then as a consultant to the Social Security Board and to leading liberal congressmen in the succeeding years, visiting D.C. frequently.252 In a 1942 article, Witte attested that

[b]eyond question many important statutes enacted by Congress have their origin in administrative departments and congressional action is profoundly influenced by the wishes of these departments. To my personal knowledge this has been the situation as to substantially all social security legislation and also, I believe, as to most of the agricultural, banking, credit, defense, housing, insurance, public utility, securities, tax, and much other legislation of the last five or eight years.253

From the mid-1930s to the early 1940s, administrators’ “influence in legislation grew and reached an importance unique in the history of this country.”254 Administrators were especially important in shaping the content of legislative history:

The role of administrators and administrative agencies in statute lawmaking is not confined to the preparation of bills introduced in Congress. Their influence is important also in the congressional consideration of these measures. Administrators are constantly appearing before congressional committees to give testimony in public hearings both on measures which originated in their departments and on other bills.

. . . Departmental representatives work with legislative draftsmen and committee clerks [i.e., non-professional staff] in preparing committee amendments, substitute bills, and [committee] reports. Commonly also the interested departments supply the committee chairmen and other members who champion their bills with material for their speeches, and their representatives are within beck and call to supply needed information during the course of the debate in the houses themselves. [Here Witte dropped a footnote: “In the Senate, departmental representatives are often on the floor during debates. In the House they are barred strictly from the floor but are in the galleries or in near-by rooms to be called for as needed.”] Almost always representatives of the interested departments work and sit with conference committees in the final stages of important congressional enactments.255

To be sure, added Witte, “[m]embers of Congress who have served long on committees . . . often become real specialists in these fields, quite well able to hold their own with any one,” but “[o]n any particular measure . . . few members of Congress can become as expert as able civil servants who devote most of their lives to a narrow specialty.”256 In an article published shortly before Witte’s, political scientist O. Douglas Weeks similarly concluded that agency influence on congressional drafting and deliberations had reached a new peak.257 And in 1949, Harry Willmer Jones expressed a similar view of agencies as the principal originators and expositors of bills “in matters of government regulation and administration.”258

Often, an individual attorney would (1) work for an agency or DOJ, (2) draft or lobby for a statute on the administration’s behalf, thus helping to author its legislative history, and then (3) help write the government briefs in Supreme Court cases about the statute, citing that very legislative history. Take for example FDR’s crime package of 1934. Gordon Dean, of the DOJ Criminal Division, shouldered the “principal load” of drafting, testified on Capitol Hill, forcefully lobbied the House Judiciary Chair “behind the scenes,” and later briefed multiple Supreme Court cases on the package, citing its history extensively.259 Years later, Dean looked back on the opportunities that federal lawyering had offered him:

I think it is a rare thing for a lawyer to be able to get into the position where he has something to do with writing the statute, then guiding it through the district [court] and the [Court of Appeals], and then finally making the final argument for his work in the United States Supreme Court. That was the situation in which I found myself in those days. It was a lot of fun.260

Several other lawyers also played each of these multiple roles.261

D. The Federal Government’s Turn Toward Briefing Legislative History

My discussion so far has been static. As of about 1940, the federal government had extraordinary capacity to meet the challenge of providing legislative history to the Supreme Court, and it did so. But what was happening up to 1940? Had the federal government always been providing legislative history, only to have the Court ignore it? Had the federal government only recently begun to provide it? If so, what triggered this shift in briefing practice?

To investigate these questions, I now present data on the use of legislative history in federal Supreme Court briefs over time (irrespective of whether the Court’s opinions picked up on that history). Figure 7 gives, for each calendar year in 1930-45, the ratio between (1) the number of legislative history citations appearing in federal briefs filed in Supreme Court statutory cases decided during that calendar year and (2) the number of federally briefed Supreme Court statutory cases decided during that calendar year.

A clear break occurs in 1936. In 1930-35, the average of the annual ratios is 2.8. In 1936-45, it is 10.1, nearly four times higher. The lowest ratio in 1936-45 (of 5.1 in 1938) is still higher than the highest ratio in 1930-35 (of 4.5 in 1935). Thus, the government enormously increased its briefing of legislative history around 1936 and never returned to the initial baseline. (An alternative metric oriented toward the regularity of citation rather than its intensity—the annual proportion of federally briefed cases in which federal briefs cited any legislative history—likewise reveals a strong upward trend, but a smoother one, averaging 43% in 1930-35, 60% in 1936-40, and 70% in 1941-45.262)

Figure 7.

legislative history citations in federal supreme court briefs per federally briefed statutory case, 1930-45


These figures are consistent with the idea that heavy government briefing of legislative history was a necessary condition for heavy Court citation of such history. The Court’s process of normalization began (in 1940) only after the government’s process of normalization was in full swing. In a general way, the Court followed the government.

Still, there is an interesting lag between the government’s shift (around 1936) and the Court’s shift (around 1940). Why did the Justices wait four years before intensively using the material that the government was copiously providing? In part, it may be that the Justices on the Court at the time were less intellectually receptive to legislative history in general (as discussed in Part II).

But my research suggests there is also another reason. Within the period 1936-45, there was a change in the predominant purpose for which the government used legislative history. The initial spike centering on 1936 reflected the government’s use of legislative history to defend the constitutionality of federal statutes, whereas the more sustained upward trend from about 1939 onward reflected the government’s routinization of legislative history as a tool for arguing “pure” questions of statutory interpretation unconnected to constitutional challenges.263

This change in the government’s purpose fits with the sweep of events. The years 1935-37 saw a constitutional crisis, in which the Court initially threatened to block much of the New Deal but, in spring 1937, retreated to a more cooperative position. In the years after 1937, as scholars observed at the time, the Court shifted its focus away from judicial review of federal legislation and began focusing more on the interpretation and implementation of that legislation.264

The lag between the government’s shift and the Court’s shift can be explained not only by new progressive Justices’ general receptivity to legislative history, but also by the fact that (1) the government began intensively using legislative history for pure statutory interpretation only in 1939-40, immediately before the Court’s shift, and (2) the Justices were apparently more reluctant to shovel the government’s research into their opinions in cases on constitutionality than in cases on pure statutory interpretation. Why this reluctance? When defending the constitutionality of a federal act, the government’s usual aim in citing legislative history was to assure the Court that Congress had found the legislative facts requisite to the exercise of its enumerated powers.265 But conservative and progressive Justices each had reason to avoid confronting such material: conservatives wanted to preserve formalist doctrines immune from the functional considerations to which legislative facts pertained, while progressives thought the judiciary should defer to congressional fact-findings and therefore wished to avoid a searching judicial examination of those findings. For example, in defending the Bituminous Coal Conservation Act of 1935, the government cited an avalanche of congressional reports and hearings to support Congress’s finding that regulation of coal mining was necessary to redress the unregulated industry’s enormous deleterious effects on interstate commerce.266 In his majority opinion striking down the Act, however, the conservative Justice Sutherland ignored this congressional evidence. He insisted that the magnitude of the evil effect was irrelevant, for the effect still was not “direct” in the doctrinal sense required to legalize congressional action: “[T]he extent of the effect bears no logical relation to its character.”267

As a converse example, consider the government’s brief in defense of the Fair Labor Standards Act of 1938, which cited copious reports and hearings in support of Congress’s view that substandard working conditions created problems for interstate commerce that individual states could not solve.268 Writing for a unanimous Court, however, the progressive Justice Stone cited legislative history only sparingly, for he believed the examination of such findings was not appropriate in constitutional review: “The motive and purpose of a regulation of interstate commerce are matters for the legislative judgment upon the exercise of which the Constitution places no restriction, and over which the courts are given no control.”269 Stone and his progressive colleagues adopted a different attitude when the government briefed legislative history for the purpose of pure statutory interpretation, for then the judicial task was to carry out Congress’s intent, not to judge the legality of that intent.

Now that we have a picture of the government’s briefing practices across the period 1930-45, we can ask why the government began intensively briefing legislative history when it did, rather than earlier. At one level, the answer is obvious: the government’s initial leap into legislative history occurred in constitutional cases around 1936 simply because federal lawyers at that time faced the most acute constitutional crisis in living memory. Extraordinary challenges called for creative solutions. But this explanation is not entirely satisfying. Most importantly, the government’s intensive use of legislative history over the long run pertained more to pure statutory interpretation than to constitutionality, and the government began heavily briefing legislative history on purely statutory questions only in about 1939-40, even though statutory questions had been part of its workload for generations. Why did the government routinize the use of legislative history for statutory questions in 1939-40 and not before?

The answer appears to be that the federal government’s unique capacity to brief legislative history—which I documented in Section III.C—was of recent origin. It came into being only in the 1930s. The unrivaled manpower of the DOJ divisions resulted from remarkably rapid growth amid the New Deal. In Table 2, I give the available data on the number of attorneys located in Washington, D.C., at the five principal DOJ divisions270:


This is an explosion. In 1938, the divisions had nearly four times as many lawyers as in 1930. In 1942, they had more than five times. Back in 1930, it is possible that the D.C. headquarters of DOJ constituted the largest law office in America, but not by much. By the late 1930s, there was no comparison: “Main Justice” dwarfed any other American organization of lawyers. It was this unprecedented size that allowed DOJ to develop an appellate section and sometimes a Supreme Court section within each division—an amazing degree of specialization for the period.271

Agencies outside DOJ also played a major role in researching legislative history before the Court, and those agencies, too, exploded in the 1930s.272 Many of the agencies assigned to draft Supreme Court briefs in the early 1940s had come into existence in the preceding decade: the Reconstruction Finance Corporation (1932), the Securities and Exchange Commission (1934), the Federal Communications Commission (1934), the National Labor Relations Board (1935), and the Wage and Hour Division of the Labor Department (1938). New Deal statutes also gave vast new missions and personnel to preexisting agencies that would brief Supreme Court cases. The ICC acquired the power to direct railroad bankruptcies (1933) and regulate trucking (1935); its personnel grew by about 70% in 1935-41.273 The missions of the Treasury Department and Internal Revenue Bureau expanded with new taxes on rich individuals and corporations (1935-37) and Social Security taxes (1935); the Department’s new Legal Division, with its huge legislative history operation, was founded in 1934. The Federal Power Commission acquired the authority to regulate electricity (1935) and natural gas (1938); its “original organization” was therefore “greatly enlarged.”274

This explosion of DOJ and agency capacity during the 1930s occurred simultaneously with an increased volume of legislation and a failure by Congress to beef up its own internal staff. This led to the upsurge in the administrative state’s influence over congressional drafting and deliberations described by Witte and Weeks.

As the agencies and DOJ divisions grew bigger, and as Congress became more dependent on them, they became better integrated with the OSG, their pipeline to the Court. After the ineffectual tenure of James Crawford Biggs (1933-35), Stanley Reed became SG in 1935. He deliberately hired lawyers with experience in agencies: Wyzanski from DOL, Alger Hiss from the Agricultural Adjustment Administration,275 Paul Freund from the Reconstruction Finance Corporation, Robert Stern from the Interior Department, Arnold Raum from the DOJ Tax Division, etc. And he made a conscious effort to engage more with the agencies “to get the best background” in producing briefs and making arguments.276 This cleared the pipeline from the administrative state to the Court.

Thus, the administrative state by the late 1930s was in a far better institutional position to provide legislative history to the Court than any organization in American history to that time. This timing explains why the federal government did not leap into briefing legislative history before the late 1930s. But there remains the converse question: why did the government leap so quickly into briefing such history once it was positioned to do so?

I think that, once the critical mass of institutional capacity was reached, briefing legislative history was simply the natural thing for federal lawyers to do, because the agencies and DOJ were so familiar with the material (given their engagement with Congress) and because the adversary system provided an incentive to wield any weapon that was less available to one’s opponent. These causal factors operated across most, perhaps all, agencies and DOJ divisions. Consistent with this, federal briefs heavily citing legislative history during the upsurge of 1939-40 were not concentrated in any particular field.277

Another possible reason the government began wielding legislative history so enthusiastically in 1939-40 was that federal lawyers—confident in their ability to use such material insofar as the Court was open to it—perceived an emergent Supreme Court majority that was more open than any of its predecessors. Roosevelt made his third and fourth appointments in 1939 and his fifth in early 1940. Government briefing of legislative history in pure statutory interpretation was on the rise in 1939 and early 1940. The American Trucking opinion in May 1940 announced the new Roosevelt appointees’ openness, and it was an announcement that the federal government—as the petitioner and victorious party in that case—had actively sought and helped bring about. The government’s American Trucking brief, in contending that the Motor Carrier Act’s use of “employees” excepted only safety-related employees from the Fair Labor Standards Act, argued for a partial rollback of the plain meaning rule. Citing Holy Trinity, it declared that “general terms” like “employee” could never have a plain meaning but always had to be “interpreted in the light of the purposes sought by Congress,” allowing resort to legislative history.278 The government pushed a similar approach in another high-profile case simultaneously with American Trucking (which it also won),279 and after American Trucking came down, the government embraced Reed’s language in future briefs.280

The federal government’s adoption of legislative history arose mainly from its institutional position, not from any particular theory of interpretation. Indeed, the government’s institutional position was so well-suited to the briefing of legislative history that federal lawyers proved capable of using that material in widely disparate ways, which reflected divergent theories of its interpretive usefulness. On the one hand, many federal briefs in 1940-45 used legislative history in a clear effort to discern the statute’s general intent, then worked “downward” from there to the particular application.281 One can understand the attractiveness of general intent for agency thinking: in the view of many scholars, agency personnel are distinctly oriented toward program implementation, so they want to find and defend the interpretation that best furthers their organization’s overall objective.282 But on the other hand, federal lawyers in 1940-45 were quite capable of using legislative history to argue for the specific intent of a provision. Agencies had as much of an advantage in digging up specific references in legislative history as in using it to understand general intent, for agency personnel were so deeply immersed in congressional deliberations that they could remember the nitty-gritty specifics better than anybody. Thus, in a 1940 brief on whether a federal statute criminalized false affidavits submitted under the Hot Oil Act, the government cited a committee report indicating that Hot Oil affidavits were “one of the two specific instances which prompted the enactment of” the statute, so that a construction failing to cover them would contravene “not only the general purpose of the [provision] but one of its specific purposes as well.”283 In a 1943 brief on whether the Anti-Kickback Act applied to foremen, the federal lawyers dug up complaints from the hearings about foremen receiving kickbacks.284 In a 1944 brief arguing that the Federal Power Commission had the authority to force Connecticut Power and Light to fulfill certain accounting requirements, the government pointed out that the “proponents of the [Federal Power] Act, by way of illustrating the accounting abuses sought to be remedied thereby, specifically cited during the debates in the House alleged write-ups aggregating $21,000,000 in the accounts of petitioner [Connecticut Power and Light].”285

E. Non-Federal Lawyers Briefing Legislative History: The Importance of Lobbyists

While the federal government was formidable in its capacity to use legislative history, this did not mean that lawyers outside were entirely incapable of doing so. Here I discuss the non-federal lawyers who briefed legislative history. Their backgrounds deepen our understanding of what made lawyers capable of using this material.

Having counted legislative history citations in all federal briefs in Supreme Court statutory cases in 1930-45, my research team also counted such citations in all non-federal briefs in such cases, but only for the years 1938-41. The reason we focused on this narrow range of years was time and money: non-federal briefs were more numerous than federal ones, so processing a year’s worth of them took us longer. I selected the years 1938-41 because they encompassed the federal government’s normalization of legislative history for pure statutory interpretation, for which a comparison to non-federal briefing would be of interest.

We can begin our analysis by comparing (1) the number of legislative history citations in all federal briefs in federally briefed cases for a given year and (2) the number of legislative history citations in all non-federal briefs in federally briefed cases for that same year. In 1938, the federal-brief citations were 3% less than the non-federal-brief citations; in 1939, 39% less; in 1940, 21% less; and in 1941, 111% more. The near-parity in 1938 is not surprising, since Figure 7 indicates that 1938 saw a brief dip in federal briefing, after the constitutional crisis but before the routine use of legislative history for pure statutory interpretation. However, the ratios in 1939 and 1940—when federal briefs contained (respectively) 39% and 21% fewer citations than did non-federal briefs—are surprising. How could the non-federal briefs have been ahead? Importantly, the non-federal citations were concentrated in relatively few cases. For example, across the 61 federally briefed cases in 1939, 78% of the citations appearing in non-federal briefs were packed into two cases; by comparison, the two cases that year in which federal briefs contained the most legislative history accounted for only 34% of that year’s federal-brief citations. Similarly, across the 64 federally briefed cases in 1940, 59% of the citations appearing in non-federal briefs were packed into two cases; by comparison, the two cases that year in which federal briefs contained the most legislative history accounted for only 32% of that year’s federal-brief citations.

Because the citation total for a year’s worth of briefs (especially non-federal briefs) was often driven by just a few cases—much more so than was the case with judicial opinions286—I decided that my comparison of federal and non-federal briefs needed to go beyond aggregate annual numbers. Therefore, I made an ordered list, for each year, of all the federally briefed cases, starting with the case in which the federal brief(s) contained the most legislative history citations and moving downward to the case in which such brief(s) contained the fewest. I then made an analogous ordered list, for the same year and the same cases, starting with the case in which the non-federal brief(s) contained the most legislative history citations and moving downward to the case in which such brief(s) contained the fewest. Then, for each year, I placed the two lists side-by-side, so it was possible to compare the median number of federal citations with the median number of non-federal citations, or the 75th percentile of federal citations with the 75th percentile of non-federal citations, etc.

In Table 3, I report comparisons between numbers of federal and non-federal citations at percentiles ranging from the 95th to the 50th (below which the numbers become small or zero). I also report, for each year, the percentage of cases in which the federal brief(s) contained any legislative history and the percentage of cases in which the non-federal brief(s) contained any legislative history.287

Looking across the run of cases (rather than at aggregate annual numbers), we see that, starting in 1939, federal lawyers were generally ahead of non-federal lawyers when it came to legislative history, usually by a wide margin. The federal advantage is more impressive considering that it was not uncommon, in a single case, for there to be two or more non-federal briefs each of which came from a distinct non-federal party (or amicus) with its own set of lawyers. That is, the federal government might be double-teamed (or triple-teamed, etc.). This occurred in 26% of cases in 1938, 38% in 1939, 14% in 1940, and 26% in 1941.288

The biggest exception to the federal advantage is that, at least in 1939-40, the federal lawyers fell behind at the very top of the scale (the 95th percentile and sometimes the 90th). This is consistent with the high aggregate numbers for non-federal briefs in those years. Thus, although non-federal lawyers usually failed to match their federal counterparts across the board, there was apparently a small group of non-federal advocates who used legislative history very heavily.


To investigate how non-federal lawyers acquired this capability, I identified all non-federal briefs in cases decided in 1938-41 that contained 25 or more citations to legislative history. There were 25 such briefs. I then researched the backgrounds of the lawyers and litigants on those briefs.

Before I discuss common factors among the non-federal lawyers and litigants, the first thing to note is that there were no “repeat players.” No lawyer or litigant appeared more than once across the 25 briefs. This is consistent with what we know about the “Supreme Court Bar” during the 1930s and 1940s: no such bar existed. Back in the early republic, when travel to D.C. had been difficult, there had been a small cadre of private lawyers who argued frequently before the Court, but as travel eased, that cadre disappeared.289 The proportion of non-federal advocates arguing more than a single case before the Court fell from 25% in 1840 to 17% in 1900 to 7% in 1940. Thus, nearly all non-federal lawyers in our period were one-timers.290

While no factor was universal among all of the lawyers or litigants on the 25 briefs in our selection, it appears that the most important common factor was congressional lobbying. For eight of the briefs, there is direct evidence or strong circumstantial evidence that an attorney or litigant on the brief personally lobbied Congress on the passage of the statute at issue.291 For another two of the briefs, the legislative history in the brief consisted mostly of post-enactment congressional deliberations related to the statute at issue, and the attorney on the brief began lobbying Congress on the specific subject of the statute before that post-enactment history was produced.292 For one additional brief, about one-quarter of the legislative history in the brief consisted of post-enactment deliberations related to the statute at issue, and the attorney on the brief began lobbying Congress on the specific subject of the statute before that post-enactment history was produced.293 For two more of the briefs, there is direct evidence that an attorney or litigant on the brief knew and worked with a principal lobbyist on the statute at issue.294 For three more of the briefs, there is direct evidence that an attorney on the brief, long before submitting it, acquired general experience in lobbying Congress (e.g., testifying or submitting an analysis at a hearing, drafting bills and committee reports), albeit on statutes separate from those in the case.295 For one last brief, the attorneys were not exactly lobbyists, but they were closely connected to the Senator who produced the congressional documents cited in their brief, which he personally gave to them.296

The briefs covered in the preceding paragraph tally up to seventeen, all with evidence that attorneys or litigants had lobbied Congress (or were connected to those who lobbied Congress), almost always on the statute at issue or on other deliberations that produced the legislative history cited in the brief.

In light of this evidence, it seems that non-federal briefs heavily citing legislative history are the exceptions that prove the rule: the best way for an attorney outside the federal government to become expert at legislative history was to have direct experience in lobbying Congress, usually on the very statute being litigated, thus gaining the same kind of workaday familiarity with congressional deliberations that the agencies and DOJ divisions acquired on a systematic basis.

That lawyer-lobbyists played a role in normalizing legislative history makes sense. Lobbyists have been extremely important in the past century of congressional deliberations, and a high proportion of congressional lobbyists have been lawyers. In most Western democracies, national political power has been concentrated in strong political parties, but in the United States, especially in the early to mid-twentieth century, the political parties were weak, creating a vacuum that was filled by organized interest groups unbeholden to the parties. “American politics,” wrote E.E. Schattschneider in 1942, “is remarkable for the exaggerated role played by pressure groups. These organizations are found in all free countries, but they play a unique role in American politics.”297 Congress was “prodigal in its concessions to organized minorities because the parties [were] too decentralized to impose an effective discipline on their congressional representation.”298 Just as the federal bureaucratic establishment came into its own during the New Deal, so the corps of D.C. interest-group lobbyists came into its own around the 1920s and 1930s. In his 1929 book Group Representation Before Congress, Pendleton Herring explained that, although Congress had always been lobbied for individual monetary claims or company-specific subsidies, the period since World War I had seen an unprecedented “descent of Washington representatives of national associations upon the capital,” each encompassing a whole industry or other mass group.299 Indeed, many of the attorneys on the briefs in our selection represented just such mass groups.300 In keeping with the general importance of lawyers in American governance, a large portion of these lobbyists were attorneys. Quoting a congressman, Herring said, “The ‘general practice of law’ in Washington is coming to be synonymous with ‘general lobbying.’”301

What about the remaining eight briefs in our selection of 25? Interestingly, five can be understood as arising from the labor movement’s sudden leap in organizational capacity and its confrontation with the DOJ Antitrust Division.302 These briefs lend some limited support to my thesis about congressional lobbying, and they reinforce larger themes about the importance of the federal bureaucracy and interest groups in the normalization of legislative history.

The story behind those five briefs is as follows. Historically, litigators for the American Anti-Boycott Association (an anti-union group) had been among the rare early users of legislative history, citing the material to great effect in Loewe v. Lawlor (1908), which held that unions were subject to the Sherman Act,303 and in Duplex Printing Press Co. v. Deering (1921), which held that the ambiguous Clayton Act of 1914 did not provide the antitrust exemption that unions thought it did.304 In both these cases, the labor attorneys were caught flat-footed, providing almost no answer to their opponents’ extensive discussion of legislative history.305 Only in the 1930s did the labor movement acquire a dedicated national lawyering apparatus. The Congress of Industrial Organizations was founded in 1935, and in 1937 it acquired a full-time, three-man Legal Department, headed by former WPA general counsel Lee Pressman.306 Meanwhile, the traditionally decentralized American Federation of Labor in 1938 hired Joseph Padway as its first full-time general counsel.307 Both the CIO and AFL lawyers were intimately familiar with Congress. Pressman, a veteran agency lawyer, was lobbying Congress extensively by 1938, aided by junior CIO attorneys who wrote his testimony.308 Padway, too, was lobbying by 1939.309 Thus, both attorneys had a workaday familiarity with Congress.

Shortly after the labor movement acquired these lawyer-lobbyists, it faced one of its greatest legal challenges. In 1938, the DOJ Antitrust Division, headed by the ferocious Thurman Arnold, initiated the biggest campaign of antitrust enforcement in U.S. history, frequently targeting unions.310 With Arnold gunning for them, the AFL and CIO lawyers invested in enormous research to try to overturn Duplex Printing and immunize unions from antitrust prosecution. Briefing the history of the Sherman Act was relatively doable, given that Congress in 1903 had published the Bills and Debates in Congress Relating to Trusts, which included a 300-page selection of debates from 1890,311 and political scientist Edward Berman in 1930 had written a monograph on the 1890 debates.312 But the Clayton Act’s 1914 history remained largely uncharted. When one of Arnold’s enforcement suits reached the Supreme Court in United States v. Borden Co. (1939), Padway represented one of the defendants and filed a brief setting forth much new research on the Clayton Act.313 Some additional defendants were represented by Daniel Carmell, who (aided by the Bills and Debates and Berman’s book) briefed the Sherman Act’s history, in a kind of sideshow to Padway’s brief.314 Ultimately, the Court’s decision in Borden Co. did not reach the question of whether to overturn Duplex Printing.

Then, in 1940, the Justices agreed to hear Apex Hosiery Co. v. Leader.315 In that case, a private employer (rather than DOJ) sued a union for antitrust violations, but the case raised the union-exemption issue just the same. This time, Pressman and his CIO colleagues had their shot at the Clayton Act’s history: they filed an amicus brief recounting the history even more extensively than Padway had,316 and they assisted the attorney for the union (Isidore Katz), who likewise briefed it, albeit less deeply.317 The attorneys for the employer (Apex) had ignored legislative history in their opening brief but now faced an onslaught of it from their labor opponents. Incapable of briefing such material on demand, they went scurrying to the institution that briefed it best: the federal government. The Antitrust Division was then prosecuting a big case against unions in the U.S. District Court in Missouri, where it was fighting off union claims about the Clayton Act’s legislative history, answering with legislative history of its own. Apex’s attorneys obtained the Antitrust Division’s brief in the Missouri case and submitted a reply that block-quoted forty-five pages of it.318 Ultimately, the Justices in Apex ruled for labor, but on narrow grounds.319

This leaves only three briefs out of our selection of 25. In these, I find no evidence of lobbying by the lawyers or litigants. Perhaps some of them briefed legislative history in response to other parties in the case doing so. For one of the briefs, the opposing party was the federal government, which briefed legislative history extensively.320 For one other, the opposing party was represented by a lawyer-lobbyist who also briefed much legislative history.321

F. The Court’s Own Internal Research

As noted in Section III.A, 34% of the citations in the sample of federally briefed statutory cases in 1940-45 matched no brief, and that figure rose to 55% for non-federally-briefed cases. These citations apparently came from the Court’s own internal research. In some of the most striking instances of this internal research, the Justices analyzed a statute’s legislative history extensively even though the briefs said nothing about it at all.322 Given the difficulty of researching legislative history, how did the Justices succeed in doing so much on their own?

For one thing, the Court had acquired, by about 1940, an internal research apparatus that was quite extraordinary for an American judicial tribunal: the emergent corps of Supreme Court clerks. Historically, the Justices had frequently employed clerks with merely stenographic duties. But increasingly, in the early 1900s, they began adopting the modern model of the law clerk: a top recent graduate of an elite law school, suited to perform advanced research. Roosevelt’s appointees all adopted this model, and their doing so increased the Court’s internal research powers in a way that coincided with—and presumably helped cause—the surge in legislative history. To be sure, we should not exaggerate the rapidity of the change. At least three Justices who preceded Roosevelt—Brandeis, Cardozo, and Stone—had already adopted the modern model and followed it throughout the 1930s. Hughes apparently adopted it by 1938 and perhaps earlier. And even among the Justices who did not hire elite graduates on a rotating basis, the clerks were still often lawyers, as in the cases of Sutherland, Butler, and McReynolds.323 Nevertheless, the completion of the changeover to the modern hiring model around 1940 must have contributed to legislative history’s triumph.

The Court’s internal research capacity was extraordinary not only because of the number and skill of its clerks but also because of the paucity of its cases, allowing a large amount of manpower per case. Whereas the Court had often issued full opinions in 200 to 250 cases per year during the early 1900s, the annual number was much lower in 1940-45 (160 or less).324 This reduction in caseload resulted from the Judiciary Act of 1925, which had empowered the Court (with few exceptions) to choose which cases it would hear—a great leap in the centralization and rationalization of the federal judiciary.325 But while the reduced caseload may have been a precondition for the Court’s intensive use of legislative history in 1940-45, it surely was not the immediate trigger. The real reduction had occurred very soon after the Act, in about 1925-28. In 1940-45, the Court’s average number of full-opinion cases per year was virtually identical to that of the late 1930s.326 Still, the manpower that the Supreme Court could devote to each individual case—as of 1928 and continuing into the 1940s—was very large by the standards of American courts.

To some extent, however, the high proportion of legislative history citations that matched no brief does not simply reflect the Court’s internal capacity. It appears that a substantial number of these unmatched citations actually arose indirectly from federal agencies and DOJ divisions. Most importantly, those organizations played an important role in building up the Court’s library. Up to 1935, the Court met in the basement of the Capitol building, and most of the Justices and their clerks did their work in the Justices’ private homes, though the clerks could use the Capitol’s law library or venture to the Library of Congress building.327 Only in 1935 did the Court acquire its own building and independent library.328 By 1946, this library had acquired the second-largest collection of statute-specific legislative history scrapbooks in D.C. (bested only by the Treasury Department).329 But 82% of those scrapbooks were in tax or bankruptcy, and they arose from two gifts to the Court from the federal bureaucracy. The first gift was a 45-volume set on the internal revenue statutes, which the DOJ Tax Division gave the Court in 1937.330 The second was a multi-volume legislative history of the bankruptcy statutes, compiled by a DOJ attorney who worked on a federal investigation of the bankruptcy system in 1939-40.331

Further, it appears that the Justices sometimes obtained legislative history research from federal agencies ex parte. (Recall that the Court’s own library staff was considered incompetent to do original research in legislative history as late as 1955.332) Frankfurter’s papers contain several memos in this vein. For example, in February 1946, he asked the Court librarian to locate all materials bearing on the meaning of a certain phrase in the Social Security Act, apropos of a pending case, adding that the librarians of the Labor Department and Social Security Board “are the most hopeful sources for getting such materials, including of course legislative hearings and reports.”333 Other memos indicate that Frankfurter made similar requests to the Patent Office in 1945,334 to the Immigration and Naturalization Service in 1945,335 to the Patent Office again in 1946,336 and to the Treasury Department in 1949.337 It is possible that Frankfurter made other such requests that left no trace in his papers. Archival documents also indicate that Hughes, Reed, and Douglas communicated with agencies ex parte about cases before the Court, though none of the surviving memos have to do with legislative history.338

The Court’s familiarity with the agencies and with legislative history may also have arisen from the career patterns of the clerks, which were oriented more toward the administrative state, and less toward the judiciary, than today. In 1943-44, according to one recollection, several clerks had done stints at DOJ or in a federal agency, more than had served as clerks on lower courts.339

G. The Minimal Role of the Lower Courts

The lower courts were not an important source of the Supreme Court’s citations to legislative history during the period of normalization. Legislative history did not “bubble up” from below. In my sample of federally briefed statutory cases in 1940-45, only 7% of the legislative history citations could be matched to citations in the corresponding lower-court cases. For my sample of non-federally-briefed statutory cases in 1940-45, the figure was again 7%.340

This suggests that the normalization of legislative history was a top-down phenomenon, originating in the Supreme Court. This is understandable. The Supreme Court itself had extraordinary manpower per case, and the federal bureaucracy was most likely to invest in legislative history research when the stakes were high, as in Supreme Court litigation.341 Relatedly, a crude search of the Westlaw database indicates that, during the early 1940s, the U.S. circuit courts were about one-fourth as likely to cite at least some legislative history in a case on a federal statute as was the Supreme Court.342 More refined data, for a somewhat later period, has been collected by Glenn Bridgman, pursuant to his important recent study of the use of legislative history in the U.S. Courts of Appeals in 1950-2006, which employs innovative electronic counting methods.343 Bridgman’s data indicate that, over the period 1950-60, the number of legislative history citations in all published decisions of the Supreme Court was 37% greater than in all published decisions of the U.S. circuit courts combined,344 despite the fact that the circuit courts issued more than twenty times as many published decisions as did the Supreme Court during those years,345 with little difference in the proportion of decisions that interpreted federal statutes.346 Consistent with this, Ray Stringham wrote in 1961 that the “United States Supreme Court has been the open leader in the movement” for using legislative history.347 Stringham criticized the Court’s enthusiasm for legislative history, in part because the Justices too often decided cases on the basis of such material when it had not been briefed in the lower courts.348

IV. legislative history as a statist tool of interpretation

The federal administrative state was the dominant force in the normalization of legislative history as a tool of statutory interpretation. Federal lawyers almost always briefed more legislative history than other lawyers before the Court; provided a very disproportionate share of the citations in the Justices’ opinions; and even served as the hidden source for some of the Court’s “internal” research capacity, via library donations and responses to ex parte inquiries. Legislative history was therefore a statist tool of interpretation, in the sense that the administrative state enjoyed privileged access to such material and was a privileged provider of it to the Court, more than was true of other interpretive sources, such as statutory text.

But legislative history was also statist in a deeper sense. The federal bureaucracy was a privileged participant in congressional discourse, and legislative history had the potential to extend that privilege into the courtroom. As I discuss in Section IV.A below, the federal government in 1940-45 frequently invoked legislative history in order to show that (1) congressmen intended for a bill to mean what the agency had told them it meant, or (2) congressmen, having been apprised of how an agency was interpreting a particular statute, acquiesced in the agency’s view. And these were only the explicit briefs; surely numerous other invocations of legislative history by the federal government rested on congressional utterances that had been ghost-written (or at least influenced) by agencies and thus reflected the bureaucratic agenda. By invoking legislative history, the bureaucracy advertised and leveraged its constant, intense engagement with the body that officially originated the norms to be implemented. At times, this may have been a ventriloquist’s performance, with Congress as the puppet, though in other instances it presumably relayed a serious dialogue between lawmakers and bureaucrats (even if the government expertly presented the dialogue as favorably to itself as the record would allow).

In any event, legislative history served to deliver the bureaucracy’s views to the judiciary clothed in the mantle of congressional authority, and this at the very moment when the Court had recognized such authority as virtually unlimited in the economic sphere. The Court’s acceptance of legislative history as a normal interpretive source in 1940-45 was of a piece with its larger acceptance (at that same moment) of an agency-centered vision of governance, with a diminished role for the judiciary and for the sources farthest inside the “comfort zone” of judges.349

Though “statist” is a fitting label for legislative history, the term requires two qualifications, each of which has to do with a peculiar aspect of the American bureaucratic state. The first qualification, which I discuss in Section IV.B, concerns judges. Though legislative history was outside the traditional judicial “comfort zone,” the Supreme Court in 1940-45 did acquire some capacity to use it independently, as if learning a second language—one that the Roosevelt Court, given its extraordinary institutional resources and the administrative experience of several of its members, was more suited to acquire than were courts generally. By learning this second language, the Court—even as it acquiesced in an agency-centered view of governance—preserved a degree of autonomy for itself, ensuring that judicial review of administrative action would still have some bite. American governance has long been marked by a judiciary that is both separate from the bureaucratic state and willing to scrutinize that state. Despite their general friendliness to administrative governance, the Roosevelt Justices, by becoming skilled in legislative history, ensured that such separateness and scrutiny would not be extinguished.

The second qualification, which I discuss in Section IV.C, concerns non-federal lawyers. In congressional discourse, the most privileged participants (apart from congressmen themselves) were the agencies, but there was also a role for lobbyists, many of them lawyers. This reflected the peculiar openness of the legislative process in America (given its weak parties) to industry-specific and cause-specific advocacy. Judicial reliance on legislative history privileged lawyer-lobbyists above the general population of lawyers (if not to the same degree as it privileged the agencies). And another American peculiarity—the tendency of government to exchange personnel with law firms and lobbying firms through the proverbial “revolving door”—meant that advocates who produced and briefed legislative history in the agencies would frequently migrate to the private sector, where they would do the same thing (if on a smaller scale). As lawyers exited the government after the New Deal and World War II, they created a new kind of law firm—the “Washington law firm”—staffed by veterans of the administrative state and dedicated to constant lobbying of that state and of Congress. Accordingly, these firms would assemble libraries of legislative history that imitated the libraries of the agencies. Such firms were like adjuncts of the bureaucracy, and their clients—a select group of big corporations—benefited from a competence in legislative history approaching that of the leviathan.

A. Legislative History to Bless the Bureaucratic Agenda

In many instances, the federal government argued for its preferred interpretation of a statute by citing legislative history to show that the agency, in proposing the measure, had put congressmen on notice that it would have that meaning. In a case on the Walsh-Healey Act, which empowered DOL to define prevailing wages required of federal contractors in each “locality,” the government contended that DOL had wide discretion to define “locality,” partly because similar acts had been construed that way in state cases, one of which had been “specifically called to the attention” of the House Subcommittee by the Acting Solicitor of DOL.350 In American Trucking, the government sought an expansive reading of the FLSA by arguing that the carve-out for trucking employees to be covered by the ICC was narrow: the ICC had asked the Senate Committee for power to regulate only a particular subset of trucking employees, and the lower court had erred in holding that the ICC “had been granted a power for which it had not asked.”351 In a fight over whether a 1934 statute criminalizing false statements to the government covered oil-men’s lies to the Interior Department about illegal interstate shipments, the government invoked the Department’s pleas to Congress to enact the statute, which congressmen had acknowledged by stating the measure was “proposed by the Department of the Interior for the purpose of reaching” those very oil-shipment deceptions.352 Later, when defending the Grazing Service’s use of its rulemaking power to impose licensing and fee requirements on stockmen during the Taylor Grazing Act’s phase-in period, the federal lawyers cited statements of agency officials at committee hearings that (it was said) made clear to congressmen that such interim measures would be necessary.353 In a case on bank robberies, the government explained that the DOJ had initially proposed a bill with broad liability, only to have Congress cut it back before enacting it. This led to bad results, showing that DOJ’s original proposal had been wise, and Congress, realizing its error, acceded to DOJ’s recommendation for an amendment. The legislative history (said the government’s lawyers) showed that “the prime purpose of the amendment . . . was the same as that of the [DOJ] in recommending it,” so it should be read broadly, in line with DOJ’s explanations of its original and later proposals.354 In another case, on whether a recent statute empowered the Food and Drug Administration (FDA) to impose standard definitions of food products that manufacturers had to follow, the government noted that the FDA, prior to the statute, had enjoyed the power to impose such definitions for a select category of products, which the FDA Chief had held up as a model when urging Congress to pass the statute at issue.355 In yet another case, defending the decision of the Federal Power Commission (FPC) to impose a cost-based accounting method on utilities, the government pointed to statements of FPC officials during deliberations on a predecessor bill to the relevant statute, announcing their plan to follow the cost-based method.356 In a separate case under the same statute, the government, arguing for a broad reading of FPC jurisdiction, cited the explanations that FPC officials had given Congress on how the bill (which the officials themselves had drafted) would operate.357 In a Federal Trade Commission (FTC) prosecution against certain chain-store practices in the confectionary industry, the government urged the Court to construe the Robinson-Patman Act to prohibit such practices, since a major study by the FTC—“[f]oremost among the studies relied on by the sponsoring committees of the bill”—had concluded that those practices were major causes of the problems Congress wished to solve, particularly among confectioners.358 In all these cases, federal lawyers contended that Congress, in passing a bill, had been won over by the agency’s view of that bill, which the Court should therefore honor.

This pattern of argument went to the extreme in Interstate Commerce Commission v. Railway Labor Executives Ass’n (1942), on whether the Transportation Act of 1920 permitted the ICC to require that a railroad preserve its employees’ jobs when abandoning lines. The ICC wanted a free hand in fostering railroad reorganizations, without the bother of protecting employees, and it construed the ambiguous Act to confer no such authority. When Congress in the late 1930s considered amending the statute to confer such authority expressly, the ICC lobbied hard against the proposals and won. In its brief, the government essentially argued that, since Congress had bowed to the ICC’s insistence against expressly conferring such authority, the existing 1920 Act should not be read to confer it, either. The brief gave an epic 32-page narrative of Congress’s failure to enact the proposals of the late 1930s, prominently featuring the ICC’s victorious lobbying (through hearings and letters to congressmen). The brief contended that the congressmen who supported employee protection, by proposing the bills they did, had implicitly conceded that the 1920 Act did not allow for such protection.359

It was also common in 1940-45 for the federal government to argue that the agency had construed a statute in a certain way and that Congress, apprised of the agency’s practice (as shown by legislative history), had consciously gone along. A key example is the regulation of labor organizing. Under the National Industrial Recovery Act (NIRA) of 1933, President Roosevelt had created a National Labor Board and then (under further statutory authority added in 1934) a National Labor Relations Board. This second agency proved quite aggressive and creative but had only shaky authority for its actions, leading to proposals for a new statute.360 This push became especially urgent after the Court struck down the NIRA in spring 1935, and so Congress enacted the Wagner Act that summer, creating a second NLRB (superseding the “old NLRB” established in 1934). Importantly, the staff of the old NLRB worked closely with Wagner and his allies in designing the new-and-improved statute,361 and once the new NLRB was created, the staff generally stayed on to work for it.362 In several cases on the new agency’s interpretations of the Wagner Act in 1940-41, federal lawyers argued that Congress in 1935 had been closely engaged with the old NLRB, had learned about its practices and experiences under the NIRA, and had embraced many of its views about the implementation of analogous provisions. “Congress had before it and relied upon the administrative practice under the prior legislation,” as one such brief said.363 Besides the NLRB, another group of bureaucrats who employed legislative history to legitimate their own practices were those at DOL, who administered the FLSA. When angry congressmen tried to amend the statute to override various DOL interpretations but failed to get their amendments passed, the government would then argue, in subsequent cases, that the failure effectively endorsed DOL’s view.364 In another case, involving the Shipping Board, the government made the striking admission that the Board had illegally deposited money in private banks that was required to go into the Treasury, but it then argued that Congress had been apprised of this fact (as evidenced in hearings and floor debates) and had effectively ratified it through subsequent appropriations acts.365

Plus, in many of the cases (discussed earlier) in which the government invoked agency explanations of proposals to Congress, it also invoked subsequent congressional awareness and approval of agency practice. Thus, in American Trucking, federal lawyers cited floor debates on the FLSA indicating that, when Congress voted on the exception for ICC-covered employees, lawmakers knew the ICC was construing that coverage narrowly.366 In the case on oil shipments, the brief noted that, when Congress extended the oil regulation statute in 1937, it “had before it a report of the Department of Justice showing that prosecutions had been instituted under” the 1934 act that was presently being litigated.367 In the case on DOI regulations during the Taylor Grazing Act’s phase-in period, the brief emphasized that the agency had made the regulations known to congressmen, who had effectively approved them by subsequently appropriating money.368

A chilling use of legislative history to legitimate a bureaucratic agenda came in the Japanese-American internment cases. In this litigation, federal lawyers cited both kinds of evidence described above: agency explanations of proposals before passage and congressional acquiescence in agency practice afterward. In Hirabayashi v. United States, the government invoked a textually vague 1942 statute criminalizing disobedience to military orders as authority to require Japanese-Americans to register for relocation and comply with a curfew. The brief emphasized that Congress had been informed by letters from the War Department (printed in the Congressional Record) of exactly what the military planned to do with the power requested.369 Later, in Ex parte Endo, the military conceded that certain Japanese-Americans were loyal but insisted on continuing to detain them, since releasing them into the interior West might result in interracial violence. Insisting that it was authorized to hold concededly loyal persons, the government explained that “Congressional awareness of the nature of the War Relocation Authority’s regulations and procedures is evidenced generally by the reports and hearings,” and that “[t]he prompt release of the loyal evacuees . . . ha[s] at times been an object of concern in Congress. In the light of this awareness, Congressional ratification of the detention of loyal evacuees . . . may be judged.”370 Prior to the 1944 appropriations, the government noted, such detention had been “mentioned in both House and Senate committee hearings and on the floor of the House” and in a Senate committee report.371

The several briefs described in this Section were only the most explicit efforts to use legislative history to bless the bureaucratic agenda. Given the thinness of congressional staff during this period and the consequent importance of agency personnel in helping congressmen to write bills, committee reports, and floor speeches,372 we can presume that a large portion of the legislative history appearing in federal briefs repeated the view of the bureaucracy without designating it as such, or at least reflected bureaucratic influence.

Note that federal lawyers’ invocations of legislative history aimed primarily to legitimate bureaucratic power, not necessarily presidential power. As Witte noted, the “initiation of legislation in administrative departments is by no means the same thing as executive leadership in legislation,” for insofar as the President officially proposed legislation, he was often merely reacting to the initiative of bureaucrats, and not always in a very proactive or independent way.373 The President was practically dependent upon agency personnel, much as congressmen were. FDR’s professional White House staff numbered only three (increased to six in 1939), and he relied mainly on people detailed ad hoc from agencies.374 That said, at least one federal brief did begin to theorize the use of legislative history to further explicitly presidential views, anticipating an idea pushed aggressively by the Reagan Administration.375 FDR issued a statement on signing a 1940 act about alien registration, construing it broadly to preempt all state laws. Federal lawyers cited it in arguing for preemption: a “formal statement issued by the President upon completing his share in the legislative process is, we believe, of importance in determining the legislative intention to preclude or permit state action.”376

Though it may be unknowable whether the administrative state’s use of legislative history actually succeeded in bending the Justices to governmental will, there is a good deal of evidence to suggest it might have. For one thing, there were several cases in 1940-45 in which (1) the federal government briefed far more legislative history than did any other lawyers and (2) the Court, citing a large amount of legislative history, ruled for the government.377 Further, it seems that in 1940-41—the first two years in which both the government and the Court cited large amounts of legislative history—the government had excellent reason to associate legislative history with victory. Of the 21 federally briefed cases in which the majority opinion had eight or more citations to legislative history in those years, the government won 19. At a more granular level, there were cases in which the Court followed the exact government arguments described earlier in this Section, citing not merely congressional statements that matched the government’s preferences, but congressmen’s acceptance of agency proposals and of agency understandings designated as such. This occurred in American Trucking,378 in one of the key NLRB cases on pre-1935 practice,379 in the case on whether the Interior Department could regulate grazing during the Taylor Act’s long phase-in,380 in the case on the FDA’s power to impose standard definitions of food products,381 and in one of the Japanese-American internment cases.382 (As noted earlier, such explicit references were only a subset of the larger universe of legislative history that was agency-influenced.383)

B. The Court Learns to Fight Fire with Fire

Despite legislative history’s major contribution to a more agency-centered regime of statutory interpretation, it did not lead to wholesale government dominance of the judiciary. At least on the Supreme Court, the Justices retained a degree of autonomy by taking up legislative history themselves, learning to use the weapon that administrators had brought to the arena. The Justices could do this in part because of the Supreme Court’s institutional features—the clerkocracy, the small docket, the big library. Moreover, many of the Justices were veterans of the administrative state and thus were personally familiar with legislative history. In this way, the Justices of the 1940s, despite being friendlier to bureaucratic power than their predecessors, preserved the U.S. judiciary’s characteristic tendency to stand apart from the federal bureaucracy and sit in judgment upon it.384

The most spectacular example of the Court’s independent scrutiny of the federal government’s use of legislative history was Helvering v. Griffiths, a 1943 tax case written by Justice Jackson, the former tax administrator.385 Back in 1920, the Supreme Court had held, in Eisner v. Macomber,386 that stock dividends were not “income” within the meaning of the Sixteenth Amendment, so Congress could not tax them. Liberals hated this decision, and as Roosevelt’s appointees filled the Supreme Court bench, it seemed ripe for overruling. The federal government asked for just that in Griffiths. But the most recent Revenue Act, of 1936, was ambiguous as to whether Congress meant to tax stock dividends (and challenge Macomber) to begin with. Indeed, the Treasury Department, from 1936 to the time of the case, had construed the Act not to cover such dividends.387 Nevertheless, the government briefed a formidable amount of the 1936 legislative history—far more than any other party—in an effort to show that Congress indeed wanted to challenge Macomber.388 The opinion was assigned to Jackson, whose experience placed him in a unique position to question the government’s use of the history. He had been Chief Counsel of the Internal Revenue Bureau in 1934-36 and head of the DOJ Tax Division in 1936. Like any liberal, he hated Macomber, telling his clerk that “whenever it is properly reconsidered it should be overruled pronto.”389 But the idea of upsetting seven years of taxpayer reliance made him blench.390 Further, he believed the government was being opportunistic in its use of history. Jackson had been there in 1936, and he did not think there had been a plan to challenge Macomber.391 Because none of the private lawyers in Griffiths answered the government’s use of history, Jackson took it upon himself to answer it. In his majority opinion rejecting the government’s reading of the statute, he cited a record-breaking amount of legislative history, more than half of it his own (and his clerk’s) original research. Much of it clearly drew on his own memory. One of his most extensive supporting citations was to the 1936 testimony of Arthur Kent, acting Chief Counsel of the Internal Revenue Bureau, before the House Committee.392 One can guess how Jackson knew about Kent’s testimony: Kent had been Jackson’s assistant at the Bureau; Jackson had left the Bureau only a month before the House hearing and was probably the man originally scheduled to testify;393 and Jackson had continued to work closely with Kent upon moving to the Tax Division.394 Administrators might try to leverage their special knowledge of legislative history, but if the judge possessed the knowledge of an administrator, he could fight fire with fire.

Another example of the Court attempting to check government opportunism was Viereck v. United States, decided that same year. A statute of 1939 required every foreign agent to disclose his or her propaganda activities “as agent,” suggesting that propaganda activities on one’s own account did not have to be disclosed, though the government contended they did. A subsequent statute of 1942 clearly did cover activities on one’s own account, but it was unclear whether this later act declared the meaning of the prior act, or changed it. When a case under the 1939 act reached the Court, the government extensively briefed the legislative history, while the other side briefed almost none.395 The committee reports on the 1942 act suggested (at least arguably) that the new language was declaratory. This helped the government. However, Chief Justice Stone wrote a majority opinion rejecting the government’s case. He thought the 1939 text was clearly against the government. And he discounted the 1942 committee reports, in part because he discovered that their language was taken verbatim from an analysis (reprinted in a hearing transcript) by an official at the DOJ Special Defense Unit who drafted the 1942 bill. The DOJ analysis seemed to tell the congressmen that the new language was merely declaratory, but (as Stone said) it “did not point out to the committee the explicit [textual] limitation” of the 1939 act “to the registrant’s activities ‘as agent.’”396 As Stone apparently saw it, the DOJ had tried to have it both ways. For the 1942 bill, DOJ drafted clearly broad text, but it ghost-wrote a committee report stating that the old 1939 act was just as broad as the new—and it assumed that congressmen would neglect to read the old act and so would not realize how implausible the ghostwritten committee report seemed in light of the textual change.

In a few opinions, the Justices marshaled legislative history to show that an agency, by its action, was flouting the understandings that it had reached with congressmen. In one case, a 1940 statute granted the ICC jurisdiction over barge lines, and the agency proceeded to issue an order that, as the barge lines alleged, unfairly favored the railroads. In dissent, Black indignantly catalogued all the assurances that proponents of the bill had given to pro-barge congressmen, promising the ICC would never issue such a discriminatory order. He particularly emphasized the guarantee offered by one ICC commissioner, which the proponents quoted “as a closing, clinching argument” on the House floor, that amendments to protect the barge lines were unnecessary.397 In another case, on whether the FLSA empowered DOL to ban home-work across a whole industry if necessary to enforce the minimum wage, Roberts argued, in dissent, that the congressmen who passed the FLSA had reached a deal that precluded a ban on home-work—which the DOL itself had recognized, after FLSA’s passage, by helping to draft an amendment to authorize such bans, which failed.398

C. The “Washington Lawyer” as Adjunct of the State and User of Legislative History

U.S. institutions of governance have long been characterized by an extraordinary degree of porousness with civil society. Given the historic weakness of the political parties, industry-specific and cause-specific “pressure groups” have exerted greater influence in U.S. legislative and administrative decision-making than in other wealthy democracies.399 Also, U.S. agencies, more than in other developed countries, have tended to trade personnel back and forth—through the proverbial “revolving door”—with private organizations that lobby the government and advocate before it.400

Even though the U.S. administrative state was primarily responsible for the rise of legislative history, the porousness of that state (and of the U.S. legislative process) meant that no institution could maintain a monopoly on this source. As noted in Section III.E, lawyer-lobbyists of the late 1930s and early 1940s demonstrated some capacity to brief legislative history, usually on statutes for which they themselves had lobbied. This provided an opportunity for private interest-group representatives to claim that Congress had blessed their programs, much as the agencies often claimed. A striking example was J. Ninian Beall, general counsel of the American Trucking Associations. In 1937, during the joint House-Senate committee hearings on the FLSA, Beall testified that the trucking industry should be regulated by one agency, not two, for two might impose conflicting mandates. Shortly after his testimony, the committees adopted the exception to the FLSA for “employees” already covered by the ICC. A few years later, Beall wrote the brief for the Associations in American Trucking, and he argued that the committees’ adoption of the exception had been a ratification of his testimony, which he quoted for a full three pages of the brief. The word “employees” should be construed to cover all employees of the trucking industry, since that would produce unified regulation of the industry by a single agency—the very thing Beall had recommended in the remarks that (he said) so deeply moved the committees.401

The practice of lawyer-lobbyists for particular industries or causes helping to shape legislative history and then quoting it during subsequent litigation would continue, but it would become more institutionalized and concentrated.402 The private lawyer-lobbyists who briefed legislative history circa 1940 had relatively narrow bailiwicks: each would lobby for a particular interest on a certain statute, then brief that statute’s legislative history if it happened to reach the Supreme Court. The way of the future was to be found in the emergent “Washington law firms” that came into their own during the 1940s and 1950s: Covington & Burling; Shea & Gardner; Arnold, Fortas & Porter; and the like. Staffed by veterans of the federal administrative state, these firms became lobbying powerhouses for big regulated companies, both with Congress and with the agencies. Their growing size allowed them to collectivize and systematize lobbying as never before. Their size also provided them with the economies of scale for a closely related task: maintaining large, trans-substantive collections of legislative history—the kind of collections that once had existed only inside the federal government. The biggest such private operation was at Covington, run by Finley. Having arrived at the firm in 1943, she and her assistant systematized the gathering of legislative history and assembled scrapbook volumes on more than 600 federal statutes over the next twelve years.403 At Covington, explained Finley in 1947, “counsel is expected to keep in such close contact with Congress that one of the library assistants devotes almost full time to attending hearings on the ‘Hill’ and securing bills and reports as soon as they are released.”404 The Covington approach to collecting legislative history would become the model for other large Washington firms.405

The word lobbying may be too narrow and mundane to describe what firms like Covington did: they were, in some sense, adjuncts of the administrative state. Charles Horsky—a onetime OSG staff lawyer and Covington partner—wrote in 1952 of “The Washington Lawyer,” a type of advocate who had recently emerged to serve as “principal interpreter” between the business world and the federal government, “explaining to each the needs, desires and demands of the other,” thus promoting compliance by businesses and “interpreting . . . the client’s problems to the government.”406 This was the lawyer as public/private mediator—a duality in keeping with the importance of the “revolving door” between business and government, through which Horsky passed multiple times, and which he stoutly defended as an institution that served the interests of both.407

D. Critics of Legislative History Statism: Frankfurter and Jackson

By the mid-1940s, the Court, driven primarily by the agencies and DOJ, had normalized the use of legislative history. Once that was done, two leading progressive Justices who had played a part in the process, Frankfurter and Jackson, began to have misgivings about the approach they had fostered, particularly its potential for abuse within the context of a fully-matured federal bureaucracy and the private lawyer-lobbyist apparatus that had grown up with it.

Of the two Justices, Frankfurter made the more mild critique. It came in 1948 in Shapiro v. United States.408 The case concerned one of the most intrusive statutes in U.S. history: the World War II Emergency Price Control Act, passed in 1942 and administered by the giant Office of Price Administration (OPA). The Act empowered the OPA to force people to disclose information about their business, but it gave them immunity from prosecution for any “transaction, matter or thing” about which they were forced to make disclosure. The question was whether the immunity provision applied to forced disclosure of records that OPA required businesses to keep. This made a huge difference, since OPA required businesses to keep all records that they would “customarily” keep, so long as the records related to prices.409

On this fraught question of bureaucratic power over millions of private entities, the Court ruled for the government 5-4, relying heavily on legislative history that was created exclusively by government lawyers and then briefed exclusively by government lawyers. Back in December 1941, Congress had been considering the bill that became the 1942 Act. OPA was already in existence and was deeply involved in the deliberations. The House had sought to weaken OPA investigative powers. But then came Pearl Harbor, after which OPA persuaded the Senate committee to strengthen those powers. In so doing, the general counsel of the agency gave the Senate committee a twenty-page memo analyzing the bill, which was reprinted in the middle of a 2,300-page hearing transcript. The OPA memo included two pages on required record-keeping that supported the narrow view of the immunity provision,410 although nobody at the hearing said anything about the issue orally.411 Seven years later, in Shapiro, the government’s lawyers cited this 1941 OPA memo to argue that Congress meant for the required records to serve as a basis for prosecution. (The attorneys for the accused businessman did not cite the memo, nor even the hearing.412) Chief Justice Vinson, writing for the majority, was quite taken with the OPA memo, emphasizing it heavily in his opinion.413

In dissent, Frankfurter expressed his fear that lobbying and litigation—particularly when done by federal bureaucrats and their lawyers—would collapse into each other if the judiciary were not careful. Vinson’s reliance on the OPA memo was “so attenuated . . . as to discredit the whole paraphernalia of legislative history.”414 There was not “the slenderest ground for assuming that members of the Committee read [OPA] counsel’s submission.”415 The Court’s “task is to determine, as best we can, what Congress meant—not what counsel sponsoring legislation, however disinterestedly, hoped Congress would mean.”416 Noting that “it is the common practice to allow memoranda to be submitted to a committee of Congress by interests, public and private, often high-minded enough but with their own axes to grind,” Frankfurter worried that judicial reliance on such memos would give “great encouragement” to “the temptations of administrative officials and others to provide self-serving ‘proof’ of congressional confirmation for their private views through incorporation of such materials.”417

A stronger reaction against legislative history came from Justice Jackson. To him, the problem was more sweeping than Frankfurter thought it, for judicial reliance on legislative history inevitably reinforced the concentration of power in the federal administrative state and in the private lawyers most closely attached to it.

To understand Jackson’s critique, we must begin with some background on the man himself. He spent his whole life in rural New York state until 1934, when he moved to Washington to become Chief Counsel of the Internal Revenue Bureau and then rose through the DOJ before joining the Court in 1941. In the course of his rise, Jackson became a renowned advocate for the New Deal, but he supported its policies only selectively. At heart, he was a Jeffersonian and a civic republican, averse to any concentration of power. He supported the New Deal policies that broke up such concentrations (like antitrust enforcement and taxing the rich) but opposed those that had the opposite effect (like the NRA and other corporatist regulatory initiatives). Insofar as Jackson supported the rise of the federal administrative state, he viewed it as a necessary evil—a concentration of public power to counter the more dangerous concentration of private power in big business.418

At first, Jackson had embraced legislative history as much as any progressive, presiding over its increasing use by federal lawyers during his tenure as SG in 1938-40 and citing it copiously during his early years on the Court. But his most extensive early use of it was in Griffiths, where (as we saw) he leveraged his own personal administrative knowledge of the Revenue Act of 1936 to check the government’s opportunism. This experience may well have disturbed Jackson. He must have realized that not every poor litigant facing a barrage of legislative history from the federal leviathan would be so fortunate as to have a judge who knew the history from his own prior job.

Within five years, Jackson began denouncing legislative history as a weapon of the strong and well-connected—a tool usable only by the federal government or by law firms with the largest resources. As he said to the American Law Institute in 1948:

Only the lawyers of the capital or the most prosperous offices in the large cities can have all the necessary legislative material available. The average law office cannot afford to collect, house and index all this material. Its use by the Court puts knowledge of the law practically out of reach of all except the Government and a few law offices.419

Knowledge of legislative history, originally concentrated in the administrative state, was now somewhat more widespread, thanks to the expansion of firms like Covington, with systematic operations for monitoring Congress and assembling the relevant documents. But firms like Covington were no help to individuals or small businesses unconnected to big organized interests.

Jackson also worried that legislative history dissolved the distinction between judicial and legislative processes. He recognized that judges interpreting statutes had to engage in some degree of policy choice.420 Yet he was also a skeptic of judicial power and keenly sensitive to the limits of judicial competence, so he wanted as many value choices as possible to be made in the political branches, not the judiciary.421 Further, Jackson disliked what American politics had become—a pluralist struggle of organized and concentrated interests that excluded the unorganized and unconcentrated—and he was appalled to see that Supreme Court litigation now resembled that pluralist political struggle. In a handwritten draft of his 1948 speech, he included this passage, which he deleted from the public version:

Important cases no longer are between John Doe and Richard Roe, two individuals in controversy. . . . If such appears from the title the individuals usually are but symbols of groups in conflict. All fights in court become fights between classes, or races, or interests, usually with a government agency on one or the other side. Then come the pressure briefs usually labeled quite falsely “Amicus Curia[e].” These “friends of the Court,” who recently have been allowed to file their pleas with little regard to whether they can contribute to the case tend to convert a litigation into a political issue.422

Given the prevalence of legislative history in briefs of the federal government and of lawyer-lobbyists—and the tendency of those briefs to quote legislative statements made, ghostwritten, or influenced by those interests—one can perhaps understand why Jackson felt the courtroom was beginning to look like a committee hearing room. Spectacles like the one in American Trucking—of the trade association’s general counsel briefing lengthy quotations from his own congressional testimony—must have made it seem that litigation had become the continuation of congressional politics by other means.

Jackson reiterated his critique of legislative history in two concurrences on the Court. The first came in Schwegmann Bros. v. Calvert Distillers Corp. (1951), on how far the Miller-Tydings Act of 1937 permitted state legislatures to go in insulating certain kinds of contracts from antitrust liability. This was exactly the kind of case that disturbed Jackson: legislative history was copiously briefed by the federal government (as amicus) and by numerous private entities (including a few dozen trade associations as amici), many of them represented by Samuel Rosenman, a close advisor to FDR and Truman and a powerful lobbyist,423 and by Herbert Bergson, who had worked in DOJ’s legislative office (possibly on Miller-Tydings itself) and as head of its Antitrust Division before leaving to start a private antitrust practice (which would soon lead to his indictment for violating conflict-of-interest laws, though he was acquitted).424 Concurring, Jackson reiterated his concern that legislative history was practically not usable by most lawyers and that its use tended “to involve the Court in political controversies which are quite proper in the enactment of a bill but should have no place in its interpretation.”425

The second of Jackson’s critiques from the bench came in United States v. Public Utilities Commission of California (1953), on the line between federal and state authority in electricity regulation. Both the federal government and the California agency briefed legislative history from the 1930s, but Reed’s majority opinion ended up relying on legislative history from the 1920s that appeared in neither brief. Jackson was especially disturbed by the obscurity of this older legislative history, which the California lawyers said they had been unable to obtain at any libraries near their San Francisco offices, nor even through inter-library loan requests to Harvard and the Library of Congress. The lawyers had finally accessed the material “only four days before [oral] argument, in Washington at the Library of [the U.S. Supreme] Court.”426 It seemed that, while the federal lawyers might easily have accessed the material on which Reed was relying, the California lawyers could not have. “The practice of the Federal Government relying on inaccessible law has heretofore been condemned,” declared Jackson, recalling the days before the Federal Register was established, when the executive branch had issued orders without publishing them.427

Jackson’s critique drew diverse reactions. Horsky thought he might be onto something: “To the extent that [Jackson’s] criticism is warranted, the Washington lawyer, in Washington, at least, has an enormous advantage.”428 Finley sought to rebut Jackson by arguing, based on her own inquiries, that the materials in Public Utilities Commission were far more accessible than the California lawyers had said, although her answer in some ways proved Jackson’s larger point. The committee reports and floor debates, she noted, “were in several San Francisco libraries.”429 But she admitted that the hearings would have required loans both from UC-Berkeley and from the California State Library in Sacramento (ninety miles away). She also failed to note that the California lawyers had inquired at Berkeley, which (they said) had been “unable to supply” the material.430 Finley suspected the problem was that the California lawyers did not know how to identify which hearings and documents constituted the relevant legislative history, and she said the California librarians could hardly be blamed for not finding the material if the lawyers could not come up with the references.431 But of course, the fact that professional law librarians in California were incapable of identifying the relevant documents was itself significant. It underscored how opaque legislative history was to lawyers and librarians who (unlike Finley) were not in a professional position where it paid to monitor Congress on a regular basis.

An even more obtuse response to Jackson was the rebuttal by Hart and Sacks. Though The Legal Process was shrewd and circumspect about some aspects of legislative history, its treatment of the accessibility question was glib. Hart and Sacks simply cited the federal regulations for depository libraries (barely pausing to note that those regulations had key exceptions and might not be followed), and they concluded that some depository library in San Francisco “might reasonably have been expected to have the material” cited in Public Utilities Commission.432 This ignored the question of whether lawyers could use legislative history on a regular basis if doing so required assembling materials from multiple depository libraries. And of course it also ignored the problems of identifying the relevant documents and of wading through their unindexed vastness. We might attribute Hart and Sacks’s refusal to grapple with these issues to their own experience as lawyers, which was confined to the rare institutions that made legislative history easy to use: Hart’s only practice experience had been at DOJ and OPA in 1937-45,433 while Sacks’s only practice experience had been at Covington in 1950-52,434 by which time Finley had built much of her tremendous library. Hart and Sacks were insiders, and that status gave them a blinkered view of the problem. This, in turn, reflected the status of legislative history as an insider’s source.435

V. conclusion: then to now

Whether courts should rely upon legislative history is the most contested practical question in statutory interpretation. This Article’s documentation of the statist origin of that reliance draws our attention to a neglected aspect of the politics surrounding the question, which we typically think of as concerning judges and legislators, when in fact it may concern the executive branch just as much, or more. My hope is that we will ultimately reach a well-informed understanding of how legislative history and judicial interpretive method interact with executive-branch power in our own era. Reaching that understanding will require primary research on how our interpretive practices have evolved from the 1940s to the present. Here I note some early data and offer some hypotheses about that evolution.

It seems that, once legislative history became normal at the Supreme Court in the 1940s, its level of use did not shift dramatically in the 1950s or 1960s, but it did intensify greatly in the 1970s and early 1980s. Then came the reaction: usage dropped steeply from the mid-1980s through the 1990s and has remained low.436 Similarly, according to Bridgman’s study of the U.S. circuit courts, usage in those tribunals rose greatly in the 1970s and early 1980s, followed by a major decline.437

It would appear that federal agency litigation in Washington, D.C., has remained the driving force behind shifts in legislative history usage. According to Bridgman’s data, in 1950-69, the U.S. Supreme Court accounted for 47% of all citations to legislative history in that court and all U.S. circuit courts combined.438 In about 1970, the D.C. Circuit rapidly transformed into a de facto specialized court for agency litigation,439 and during 1970-85—the years of heaviest legislative history usage—the Supreme Court and D.C. Circuit (despite their low caseloads) together accounted for 44% of all citations to legislative history in those two courts and all other U.S. circuit courts combined.440 Further, as Bridgman shows, the reduction in use of legislative history from the mid-1980s onward was more pronounced in those two courts than anywhere else.441

Why did the use of legislative history intensify in 1970-85 and decline thereafter? And, given that these shifts were most pronounced in the Supreme Court and D.C. Circuit, what did the shifts have to do with the federal administrative state and with the D.C. establishment built up around it?

Begin with the intensified use of legislative history in 1970-85. I hypothesize that this reflected two developments that occurred circa 1970. First, a rising disenchantment with bureaucracy led the judiciary and especially the D.C. Circuit to become far more independent, willful, and aggressive in reviewing the actions of agencies than in the preceding generation.442 Second, the “game” of producing, influencing, and briefing congressional discourse came to include a larger and more diverse set of players, beyond the familiar agency personnel and the Covington-style firms connected with them. These new players included a new cadre of D.C. activists who opposed industry and fought for the environment, consumers, and the like;443 a larger, better-organized, and more confrontational corps of D.C. business lobbyists who emerged in response to the anti-industry activists;444 and a huge number of new professional congressional staffers, hired mainly by Democratic lawmakers to reduce Congress’s dependence on agencies at a time when Republicans usually held the presidency.445 As these new players forced their way into the game, the legislative record became more likely to include a wider diversity of statements about a statute’s meaning, beyond those of the agency, and the non-agency briefs in a case were more likely to delve into the legislative record and to put forward pieces of it that did not fit the agency view. The bureaucracy’s advantage in shaping legislative history and presenting it to courts diminished (thought it did not disappear).446 Judges, in their heightened scrutiny of agency action, took advantage of the greater diversity in congressional discourse (and in the briefing thereof) to find and marshal statements of meaning to challenge the agency view and impose their own. Judges had to cram more citations into their opinions to refute the understandings of the hitherto-dominant agencies. Even when agencies won, they now needed to provide the courts with more comprehensive histories to refute the readings urged by new players in the more diverse environment. This, too, resulted in more citation-heavy opinions.

What of the judicial reaction against legislative history that occurred from 1985 onward? I hypothesize that the intensified use of legislative history proved exhausting and overwhelming to judges. They became tired and cynical after a few years of trying to process and marshal the disparate pieces of an increasingly diverse and incoherent legislative record—a task judges found especially onerous now that they were less inclined to rely upon the bureaucracy as a guide to that record. But at the same time, judges did not want to relinquish the authority over the bureaucracy that they had asserted since 1970. Textualism was an attractive new means for retaining that authority. Narrowing the materials of decision to statutory text (and related sources like canons of construction and dictionaries) placed judges on a more equal footing with agencies than was the case with legislative history. Whereas legislative history was so voluminous and complex that judges could never approach the mastery that agency personnel achieved through their continuous interaction with Congress, statutory text was circumscribed enough that generalist judges could process it relatively easily—and could thus more easily justify defying the agency view. This was even more true of canons and dictionaries, which were as legible to generalist judges as to anyone else. Thus, textualism reasserted the omnicompetence of the judiciary and of generalist judging, despite the enormously complex and subject-specific processes of policy formulation that actually characterize the administrative state. Just as the shift toward legislative history in the 1940s was statist, I suspect that the turn against it has been anti-statist, in that it has been a means to maintain judicial power against the bureaucracy.447

Appendix I: Methodology

Data.All data produced for this project is available online at the Yale Law Journal’s website (

Why I Did Not Use the Supreme Court Database. In conducting this project, I did not use the Supreme Court Database (SCD) originated by Harold Spaeth. Though this database is frequently used in quantitative studies of the Supreme Court, it currently goes back only to 1946, so it was not usable for the period I wanted to study. Scholars are currently extending the database backward in time, but it is estimated that the data on 1930-46 will not be released until 2014.448 Since I began this project in mid-2011, waiting for the Spaeth Database would have required a two-and-a-half year delay (possibly more). I decided such a delay was not worth it, for two reasons. First, the SCD does not include any information about citations (even for the years that it does cover and will cover). Its only use for me would be that it distinguishes cases interpreting federal statutes from other cases. Thus, even if I had delayed starting my study until the SCD data was released for the relevant years, I would nonetheless have had to gather most of the data for my study independently. Second, I think the data I have gathered can be retrofitted to the SCD fairly easily, if any future scholars want to combine my data with the SCD. There may be a few cases where my categorizations (of cases as statutory versus not) diverge from those in the SCD, but I think my cite-counting method is explicit enough that counts could be done for any cases that are counted as statutory by the SCD and not by me.

Identifying Federal Statutory Cases.I defined federal statutory cases as cases decided by the Court with a full, signed opinion that involved the interpretation of a federal statute, including any federal constitutional challenge to a federal statute. One research assistant (RA) categorized cases (as statutory versus not) for the years 1910, 1915, 1920, 1925, 1930-41, and 1950. A second RA categorized cases for 1942-45. To check objectivity and reliability, I had the first RA take a 25% random sample of all cases decided by the Court with a full, signed opinion from 1942-45 (yielding 144 cases) and categorize them as statutory versus not. There was 93% agreement between the two RAs, and disagreements were about even in both directions.449 A third RA categorized the cases for 1900 and 1905. I personally categorized the first 60 cases in 1900 (by order in the U.S. Reports), and there was 93% agreement between the RA and me, and the disagreements were even in both directions.

Counting Citations in Federal Statutory Cases. RAs performed all counting of citations by hand, not by electronic means. I considered this necessary because citation formats in the early twentieth century were less formal and consistent than they have now become.450

There are many possible ways to define a “citation” to legislative history for purposes of counting. Let me explain the method I chose.

First, one must decide exactly which kinds of materials to count. On this point, I included what I think most lawyers would say are the conventional materials: House bills, Senate bills, House committee reports, Senate committee reports, House hearings, Senate hearings, House documents (in the serial set, including House Executive Documents), Senate documents (in the serial set, including Senate Executive Documents), all conference reports, the Congressional Record (since 1873), Congressional Globe (1833-73), Register of Debates (1824-33), and Annals of Congress (1789-1824), plus joint hearings. This includes all genres that Carro and Brann included and found in more than negligible amounts.451

Second, one must decide what sort of reference to any of the materials above counts as a “citation.” I had my research assistants count any reference to any of the materials above where a formal document title and/or document number was given, such as “S. Rept. No. 76, 72nd Cong.,” “41 Cong. Rec. 3401,” or “Hearings Before the House Ways and Means Committee on Estate Taxes (1935).”452 If a single document (such as “S. Rept. No. 76, 72nd Cong.”) was cited once and then cited again later in the opinion, that would count as two cites. If it were cited a third time, that would count as three, and so on. Similarly, the use of “id.” would count as a cite in itself.

For each reference, I had the research assistants count one cite for each individual page number or hyphenated set of page numbers. If the document was cited with no page number, I had them simply count the reference as one cite. My reason for counting page numbers is that, when a reference to a document gives several specific pages, this generally reflects greater research into the document and engagement with what it said. There is a big difference, in research time, between citing one page in a volume of the Congressional Record and citing six pages. Likewise, there is a big difference, in research time, between citing one page in an 800-page hearing and citing six pages.453

To give an example, say an opinion contains this sentence: “The committee chair emphasized this point repeatedly. Cong. Rec., 73rd Cong., 1st Sess., pp. 7897, 7912, 7914-15, 8131.” That would count as four citations, since four pages and/or hyphenated sets of pages are given.454

One RA counted the years 1920, 1925, 1932-41, and 1950. A second RA counted the years 1942-45. To check objectivity and reliability, I had the first RA count citations in a random sample of cases from 1942-45 (which sample totaled 101 cases, with 167 opinions). There were some deviations up and down between the two RAs in individual cases, but apparently these were random, with almost no bias. The first RA produced a total of 304 citations, the second 297 citations, a difference of less than 3%.455

A third RA counted the years 1900, 1905, 1910, 1915, 1930, and 1931. To check objectivity and reliability, I had a fourth RA do a duplicate count on 1931 (which had 111 cases, with 128 opinions). Again, there were some deviations up and down between the two RAs in individual cases, but apparently these were random, with almost no bias. The third RA produced a total of 111 citations for 1931, the fourth 116, a difference of less than 5%.456

Measuring the Proportion of Cases Citing Any Legislative History. The data on which cases contained any citation to legislative history were a by-product of counting the number of citations in each case.

As noted above, one RA counted the years 1920, 1925, 1932-41, and 1950. A second RA counted the years 1942-45. To check objectivity and reliability, I had the first RA count citations in a random sample of cases from 1942-45 (which sample totaled 101 cases, with 167 opinions). There was agreement in 100 cases (99%) as to whether the case contained at least one citation to legislative history or none.

As noted above, a third RA counted the years 1900, 1905, 1910, 1915, 1930, and 1931. To check objectivity and reliability, I had a fourth RA do a duplicate count on 1931 (which had 111 cases, with 128 opinions). There was agreement in 108 cases (97%) as to whether the case contained at least one citation to legislative history or none.

Calculating Citation/Opinion Ratios for Individual Justices. In calculating citation/opinion ratios for individual Justices, we confront the question of how to define “opinion.” In the first half of the twentieth century, it was common for a Justice to concur or dissent with only a very brief statement, sometimes only a sentence (perhaps written in the third person), giving no reasoning, cursory reasoning, or simply citing the reasoning of the lower court’s opinion. Deciding which of these short utterances to count as “opinions” is difficult. I therefore adopted the somewhat arbitrary but happily objective method of excluding all separate opinions that did not exceed one paragraph. I felt comfortable adopting such a method in this study because, for the whole period 1930-45, no opinion that did not exceed one paragraph cited any legislative history.

Locating Briefs for Cases. In locating briefs, we sought filings that went to the merits of the case and contained legal argument. This included the regular merits briefs, reply briefs, supplemental briefs, and appendices under separate cover, as well as amicus briefs. We did not count petitions for certiorari, briefs in opposition to certiorari, petitions for rehearing, or any filings that contained no legal argument (e.g., filings concerning purely factual matters). If a case was decided after reargument, then, for all purposes, we counted only the briefs on reargument, not the initial briefs.

The project required locating three distinctly defined (but in some ways overlapping) sets of briefs. First, it was necessary to locate briefs for all cases in the sample from 1940-45 used for the opinion/brief citation-matching exercise. Second, it was necessary to locate all federal briefs for all statutory cases in 1930-45. Third, it was necessary to locate all briefs (not only federal but also non-federal) for all statutory cases in 1938-41.

In locating all three of these sets, we used Gale Cengage Learning’s database, The Making of Modern Law: U.S. Supreme Court Records and Briefs, 1832-1978 (which is drawn, for the period in question, from the collection of the Association of the Bar of the City of New York). For all cases in all three sets, we also checked for the relevant kinds of briefs in the hard-copy collection housed at the Yale Law Library. According to an order of the Supreme Court, the ABCNY and Yale have been official repositories of Supreme Court briefs since June 1931, and they were already receiving briefs from the Clerk’s office before then.457

To gauge the completeness of the results of our search of these two repositories for the three sets of briefs, consider the following:

As to cases in the opinion/brief matching sample in 1940-45: We found a facially complete set of briefs (i.e., at least one full merits brief on each side in every docket-numbered dispute decided within the case) in 68 of 69 cases, and the one case that was facially missing any briefs contained only one legislative history citation. (I therefore excluded this case from the matching exercise.)

As to federal briefs in all statutory cases in 1930-45: We found federal briefs, or other evidence that a case was federally briefed, in well over one thousand cases. In 1,086 of these cases, the federal briefs that we found constituted a facially complete set. That is, we found at least one full federal merits brief corresponding to every docket-numbered dispute decided within the case that included a federal party. There were only 15 cases in which one or more federal briefs were facially missing.

As to federal and non-federal briefs in all statutory cases in 1938-41: We can divide our findings between federally briefed cases and non-federally-briefed cases. As to federally briefed cases: there were 267 of these. In them, we found a facially complete set of federal briefs in 263, which is 98.5%. And we found a facially complete set of non-federal briefs in 258, which is 96.7%. That is, we located at least one non-federal brief on each side in every docket-numbered dispute decided within the case (excluding the federal side of a dispute that included a federal party). As to non-federally-briefed cases: there were 91 of these. In them, we found facially complete sets of briefs in a disappointingly low proportion of cases (about 80%); I therefore say nothing about this category of briefs in the study.

Dividing the Population of Statutory Cases Citing Legislative History in 1940-45 into Strata and Drawing the Sample. I divided the population of statutory cases citing legislative history in 1940-45 into strata and drew a sample for purposes of opinion/brief citation matching. The process of dividing the population and taking the sample is recounted in the body of the paper, supra Section III.A.458

Matching Citations in Cases to Citations in Briefs.In attempting to match a case’s citations to the corresponding briefs, we first isolated the individual citations in the case and made a list of them. In isolating individual citations for purposes of the list, we followed the same method as for counting citations in cases. For example, say an opinion included this passage: “See House Rept. No. 131; Sen. Rept. No. 76, pp. 13, 27; 43 Cong. Rec. 2713-16, 2942-45.” This would be broken down into five citations, and for each one of those citations, we would search for “matches” in all the briefs. The breakdown would look like this:

·        House Rept. No. 131

·        Sen. Rept. No. 76, p. 13

·        Sen. Rept. No 76, p. 27

·        43 Cong. Rec. 2713-16

·        43 Cong. Rec. 2942-45

If the opinion referenced any document, giving no page number, and the brief referenced that same document, giving a page number or not, then we counted a “match.”

If the opinion referenced a document and did give a page number, the process of counting matches was somewhat more complicated. For committee reports, conference reports, and bills, we counted a match whenever the brief cited the same document, regardless of whether the brief gave a page number, and regardless of what page numbers the brief gave. The rationale for this is that committee reports, conference reports, and bills were typically short and relatively easy to use, since they usually had detailed sub-headings. However, in the case of hearings, House Documents (in serial set), Senate Documents (in serial set), and floor debates (Congressional Record, etc.), we counted a match only when the opinion cited a certain page of a document and the brief cited a page of the same document that was within five pages of the page cited in the opinion. The rationale for this is that hearings, House Documents, Senate Documents, and floor debates are generally long and often desultory, so that a brief, in citing such material, is useful to the Justice and his clerk only if it points to specific passages they can use. (Similarly, if the opinion cited a hyphenated set of pages in a document, we would count a match if the brief cited a page of the same document that was within five pages of any page in the opinion’s hyphenated set of pages.)

In looking for citations in briefs, we looked at every page of the brief. We did not simply rely upon the brief’s Table of Authorities, which (we found) was often unreliable.

For the work of matching, the cases were divided about equally between two RAs. To check objectivity and reliability, I had the two RAs duplicate each other’s work for four sets of cases, corresponding to the four strata of the population:

(A) “Heavy” federally briefed cases (15 cases, all included in sample). The two RAs duplicated each other’s work for 13 of the 15 cases.459 In the duplicated cases, I was able to compare the percentages reported by the two RAs for the four categories of citations: those matching federal and non-federal briefs; those matching federal brief(s) only; those matching non-federal brief(s) only; and those matching no briefs. For these four categories, the differences between the first and second RAs were very small: 1.0, -1.8, -0.8, and 1.5 percentage points, respectively.

(B) “Heavy” non-federally-briefed cases (1 case, included in sample). The two RAs duplicated each other’s work for this case. I was able to compare the percentages reported by them for the two categories of citations: those matching a brief and those not. For the two categories, the differences were small: 0.4 and -0.4 percentage points, respectively.

(C) “Regular” federally briefed cases (214 cases, of which we drew a 20% sample amounting to 43 cases). I had the two RAs duplicate each other’s work for a random subset of this part of the sample (which was 15 cases out of 43, or 34%).460 For the four categories listed for stratum (A) above, the differences were very small: -1.8, -1.2, 2.5, and 0.4 percentage points, respectively.

(D) “Regular” non-federally-briefed cases (44 cases, of which we drew a 20% sample amounting to 9 cases). I had the two RAs duplicate each other’s work for a random subset of this part of the sample (which was 4 cases out of 9, or 44%).461 For the two categories listed for stratum (B) above, the differences were very small: 1.5 and -1.5 percentage points, respectively.

Counting Citations in Briefs. When counting citations in briefs, we followed the same method as for counting citations in cases. Within any given brief, we always counted all legislative history citations, whether they appeared in the body of the filing or the appendix.

One RA counted federal briefs for the years 1930-40 and 1945 and non-federal briefs for the years 1938-40. A second RA counted federal briefs for the years 1941-44 and non-federal briefs for the year 1941. To check objectivity and reliability, I had the second RA perform a duplicate count for the federal briefs in 1938, which had 92 statutory cases.462 As with the citation-counting for the cases, there were some divergences up and down in individual cases, but these were apparently random, with almost no bias. For 1938, the first RA identified 67 cases with federal briefs and counted 286 citations in those briefs; the second RA identified 68 cases with federal briefs and counted 295 citations in those briefs. (Sixty-six of the cases were identified by both RAs as having federal briefs.) The two RAs thus produced citation/brief ratios of 4.27 and 4.34. These ratios are within 2% of each other.463

Matching Citations in Supreme Court Cases to Citations in Lower-Court Cases. For every case, we sought matches with the corresponding lower-court opinion. We looked only to the opinion of the court immediately below. For example, if the case was decided by a U.S. district court and then by a U.S. circuit court, we looked only at the latter’s opinion. We looked at lower-court opinions only if they were published in a reporter. (We did not attempt matches with, say, a never-reported opinion that was printed in the transcript of record.) Across the whole sample (68 cases), all but three had corresponding lower-court opinions for us to process.

For the work of matching, the cases were divided about equally between two RAs. To check objectivity and reliability, I had them duplicate each other’s work for four sets of cases, corresponding to the four strata of the population:

(A) “Heavy” federally briefed cases (15 cases, all included in sample). The two RAs duplicated each other’s work for 13 of the 15 cases.464 In the duplicated cases, I was able to compare the percentages reported by the two RAs for matches with lower-court opinions. One RA got 6.1% and the other 5.3%, a very small difference of 0.8 percentage points.

(B) “Heavy” non-federally-briefed cases (1 case, included in sample). The two RAs duplicated each other’s work for this case. Both found zero matches with the lower-court opinion.

(C) “Regular” federally briefed cases (214 cases, of which we drew a 20% sample amounting to 43 cases). I had the two RAs duplicate each other’s work for a random subset of this part of the sample (which was 15 cases out of 43, or 34%).465 Both found zero matches across all cases.

(D) “Regular” non-federally-briefed cases (44 cases, of which we drew a 20% sample amounting to 9 cases). I had the two RAs duplicate each other’s work for a random subset of this part of the sample (which was 4 cases out of 9, or 44%).466 One RA got 10.6% and the other 10.9%, a very small difference of -0.3 percentage points.

Appendix Ii: Lobbying Connections Of Lawyers And Litigants On Non-federal Briefs Heavily Citing Legislative History, 1938-41

¶ 1. Brief for Petitioners, Elec. Bond & Share Co. v. SEC, 303 U.S. 419 (1938) (No. 636). 37 cites. This case concerned the Public Utility Holding Company Act of 1935. Brief-signing attorney John F. MacLane had represented the Committee of Public Utility Executives in Senate hearings on the bill. Public Utility Holding Company Act: Hearings on S. 1725 Before the S. Comm. on Interstate Commerce, 74th Cong. 287-88 (1935) (statement of John F. MacLane).

¶ 2. Brief for Appellees, United States v. Klamath & Moadoc Tribes, 304 U.S. 119 (1938) (No. 707). 72 cites. This case was a claim by Klamath Indians for a taking by the federal government, particularly concerning a treaty of 1864, a statute of 1906 under which the taking occurred, a statute of 1920 providing jurisdiction to hear the suit, and a statute of 1936 amending the jurisdictional statute. Of the legislative history in the brief (pp. 3-4, 14-15, 34, 41-42, 57-59, 66-67, 74, 78-79, 84, 112-14, 131, 139), more than half the citations pertained to deliberations specifically concerning the Klamath tribe in the 70th Congress (1927-29) or later. By the time of that Congress, brief-signing attorney Daniel B. Henderson had become a lobbyist for the Klamath tribe on their claims against the federal government and had begun testifying before Congress on the tribe’s behalf. Indians of the Klamath Reservation, Oreg.: Hearing on S. 3006 Before the S. Comm. on Indian Affairs, 70th Cong. 4-5 (1929) (statement of Daniel B. Henderson); see also Interior Department Appropriation Bill, 1929: Hearings on H.R. 9136 Before a Subcomm. of the S. Comm. on Appropriations, 70th Cong. 130 (1928) (same).

¶ 3. Brief for Respondent, Helvering v. Nat’l Grocery Co., 304 U.S. 282 (1938) (No. 723). 47 cites. No evidence of lobbying.

¶ 4. Brief for Respondent, Kessler v. Strecker, 307 U.S. 22 (1939) (No. 330). 38 cites. This case concerned immigration statutes from 1900 through the 1930s, on which the brief’s legislative history focused. I have found no evidence that any of the attorneys lobbied on those statutes. However, brief-signing attorney Carol Weiss King had worked as a full-time left-wing activist (including on immigration issues) since about 1920. See Ann Fagan Ginger, Carol Weiss King: Human Rights Lawyer, 1895-1952 (1993). She had written a brief for the ACLU on deportation law that was inserted in a congressional hearing transcript in 1926. Deportation of Alien Criminals, Gunmen, Narcotic Dealers, Defectives, Etc.: Hearings Before the H. Comm. on Immigration and Naturalization, 69th Cong. 77-86 (1926). King would testify before Congress repeatedly in the years shortly after submitting the Strecker brief. Ginger, supra, at 308-09, 432-33, 502.

¶ 5. Brief for Pure Milk Ass’n [Appellee], United States v. Borden Co., 308 U.S. 188 (1939) (No. 397). 48 cites. This case concerned (apart from the labor issues discussed supra notes 313-314 and accompanying text) the Capper-Volstead Act of 1922 (CVA), particularly the exemption from the antitrust laws that it granted to farmer cooperatives. The brief’s discussion of legislative history pertained to CVA. One of the interest groups to lobby for CVA had been the National Milk Producers Federation (NMPF), represented by its Secretary, Charles W. Holman. Authorizing Association of Producers of Agricultural Products: Hearings on H.R. 2373 Before a Subcomm. of the S. Comm. on the Judiciary, 67th Cong. 169 (1921) (statement of Charles W. Holman, Secretary, National Milk Producers Federation). Holman remained ensconced as Secretary of NMPF through the 1940s, continuously lobbying Congress on cooperative legislation. E.g., Agricultural Marketing Agreements: Hearings on S. 3426 Before a Subcomm. of the S. Comm. on Agriculture and Forestry, 76th Cong. 129 (1940) (same). Meanwhile, one of the litigants in Borden Co. was Don N. Geyer, manager of the Chicago Pure Milk Association and a major participant in NMPF, through which he surely knew Holman. From the late 1920s through the 1930s, Geyer gave papers at conferences organized by Holman. E.g., Am. Inst. of Cooperation, American Cooperation: A Collection of Papers and Discussions Comprising the Fifth Summer Session of the American Institute of Cooperation 7, 105 (1930). As of 1935, Geyer served on the ten-member NMPF executive committee. Oleomargarine: Hearing on H.R. 5586 and H.R. 5587 Before a Special Subcomm. of the H. Comm. on Agriculture, 74th Cong. 146 (1935).

¶ 6. Brief for Robert G. Fitchie et al. [Appellees], United States v. Borden Co., 308 U.S. 188 (1939) (No. 397). 64 cites. See supra note 313 and accompanying text for more details.

¶ 7. Brief on Behalf of Leslie G. Goudie [Appellee], United States v. Borden Co., 308 U.S. 188 (1939) (No. 397). 35 cites. See supra note 314 and accompanying text for more details.

¶ 8. Brief for the Borden Company et al. [Appellees], United States v. Borden Co., 308 U.S. 188 (1939) (No. 397). 33 cites. Perhaps the attorneys briefed such material in response to the initiative of other parties in the case (see ¶¶ 6, 7), including the federal government.

¶ 9. Brief for Respondent Mrs. Julia Caroline Sponenbarger, United States v. Sponenbarger, 308 U.S. 256 (1939) (No. 72). 422 cites. This was a test case for Fifth Amendment and statutory claims to compensation by numerous landowners for alleged losses from flood-control projects arising from the Mississippi Flood Control Act of 1928. The legislative history in the landowners’ unbelievably long and repetitive brief consisted mainly of two kinds of citations: floor debates on the 1928 Act (which appeared at pages 105-18 and in Appendix B, at pages 303-49, and accounted for about three-fourths of the legislative history citations in the brief) and committee hearings in 1934-38 (which were scattered throughout and accounted for about one-fourth of the legislative history citations). Brief-signing attorney Lamar Williamson had become the chief advocate for the claimants by 1934 and had begun testifying at congressional hearings on their behalf at that time, pursuing the goal of compensation through the alternative means of congressional lobbying and litigation simultaneously. Flood Control on the West Side of the Mississippi River in Arkansas, and on the South Side of the Arkansas River: Hearings on H.R. 8146 and H.R. 8048 Before the H. Comm. on Flood Control, 73d Cong. 9 (1934) (statement of Lamar Williamson); see also Flood Control in the Mississippi Valley: Hearings Before the H. Comm. on Flood Control, 74th Cong. 592, 602, 611 (1935) (same). Also, as of 1935, Williamson “worked closely with” U.S. Senator John Overton of Louisiana on related compensation legislation, advising him on drafting. Martin Reuss, Designing the Bayous: The Control of Water in the Atchafalaya Basin, 1800-1995, at 183 (2004). I have found no evidence that Williamson or his co-counsel lobbied on the 1928 Act; their research on that phase of the congressional process may have been purely retrospective.

¶ 10. Appendices to Brief for Respondent, FCC v. Sanders Bros. Radio Station, 309 U.S. 470 (1940) (No. 499). 236 cites. This case concerned the Radio Act of 1927 and amendments and other legislative activity relating to it, up to and including the Communications Act of 1934. Brief-signing attorney Louis G. Caldwell had represented a leading commercial broadcaster in the first lawsuit ever to win recognition of a common-law property right in radio frequencies, in autumn 1926. The “radio industry applauded [the decision] instantly,” and broadcasters across the country began filing similar suits. Thomas W. Hazlett, The Rationality of U.S. Regulation of the Broadcast Spectrum, 33 J.L. & Econ. 133, 149-52 (1990) (discussing the decision); see also Radio Board Chief Counsel Elected Amherst Trustee, Wash. Post, July 1, 1928, at A5 (noting Caldwell’s role). Within a few weeks, Senator C.C. Dill, later an author of the Radio Act, had the decision reprinted in the Congressional Record. 68 Cong. Rec. 215-19 (1926). The decision “ignited legislative activity in Washington,” Hazlett, supra, at 151-52, 160, and was one of two key cases that “created an incentive for [federal] radio broadcasting legislation,” Gerald W. Brock, The Second Information Revolution 38 (2003). Congress passed the Radio Act in 1927, “[w]ith the support of the broadcast industry.” Id. at 39. As of 1928, Caldwell was considered “one of the leading commercial broadcasting attorneys in the nation.” In July of that year, he was hired by the new Federal Radio Commission as its first general counsel, partly because he had privately developed an impressive plan for implementing the Commission’s (extremely pro-broadcaster) policies. Robert W. McChesney, Free Speech and Democracy! Louis G. Caldwell, the American Bar Association and the Debate over the Free Speech Implications of Broadcast Regulation, 1928-1938, 35 Am. J. Legal Hist. 351, 356-58 (1991). Returning to the private sector in late 1928, Caldwell immediately became the first chairman of the ABA’s new Standing Committee on Radio Law (the main source of legislative proposals in the field going forward) and served as the principal lobbyist for the broadcasters (hired by the “entire industry” to represent it at congressional hearings in 1931). Id. at 358, 365-66, 370, 374. This constitutes direct evidence that Caldwell lobbied on amendments to the Radio Act from 1928 onward and, in my view, strong circumstantial evidence that he lobbied on the 1927 Radio Act itself.

¶ 11. Brief for Respondent, United States v. City & Cnty. of San Francisco, 310 U.S. 16 (1940) (No. 587). 68 cites. This case concerned the Raker Act of 1913, which authorized San Francisco to use the Hetch Hetchy valley for hydropower, under certain conditions. Brief-signing attorney Robert M. Searls was San Francisco’s “special counsel for the Hetch Hetchy project and a key representative of the city during the legislative process” that led to the Raker Act. John Warfield Simpson, Dam!: Water, Power, Politics, and Preservation in Hetch Hetchy and Yosemite National Park 245 (2005); see also M.M. O’Shaughnessy, Hetch Hetchy: Its Origin and History 52 (1934).

¶ 12. Brief for Respondents, Perkins v. Lukens Steel Co., 310 U.S. 113 (1940) (No. 593). 295 cites. This case concerned steel companies’ challenge to the DOL’s interpretation of the Walsh-Healey Act of 1936, which imposed wage and hour regulations on all firms that had supply contracts with the federal government. Brief-signing attorney O. Max Gardner (the former governor of North Carolina and a personal friend of President Roosevelt) was perhaps the single most powerful and widely connected lobbyist in Washington; from 1933 onward, he established and ran a whole firm that systematically monitored congressional committee proceedings and floor debates on a daily basis. Joseph L. Morrison, Governor O. Max Gardner: A Power in North Carolina and New Deal Washington 142, 189-90 (1971). Gardner’s main client was the textile industry, and he was also highly engaged in labor issues. Id. at 142, 149-50, 172, 179-80, 190. The textile industry had more workers subject to the Walsh-Healey Act than any other industry. Edward F. Denison, The Influence of the Walsh-Healey Public Contracts Act Upon Labor Conditions, 49 J. Pol. Econ. 225, 231 tbl.1 (1941). I think this is extremely strong circumstantial evidence that Gardner must have been involved in the lobbying process that led to the Act.

¶ 13. Brief for Respondent, United States v. George S. Bush & Co., 310 U.S. 371 (1940) (No. 613). 29 cites. This case concerned the tariff. The brief’s discussion of legislative history, id. at 41-48, covered a tariff bill passed in 1922, another passed in 1929-30, and a third rejected in 1933. Marion De Vries, formerly a judge on the U.S. Court of Customs Appeals, was from 1922 to 1939 a leading tariff lobbyist, representing fourteen interest groups in the landmark revision of 1929-30. E.E. Schattschneider, Politics, Pressure, and the Tariff: A Study of Free Private Enterprise in Pressure Politics, as Shown in the 1929-1930 Revision of the Tariff 197-98 (1963); see also Biographical Directory of Federal Judges, Federal Judicial Center, (last visited Sept. 25, 2013) (entry of Marion De Vries). Brief-signing attorney Frank L. Lawrence corresponded with De Vries in 1928, reporting to him on the willingness of about a dozen importing and manufacturing companies to pay for securing a certain tariff policy; Lawrence and De Vries ultimately divided fees with each other. 3 Lobby Investigation: Hearings Before a Subcomm. of the S. Comm. on the Judiciary, 71st Cong. 2776, 2790-91 (1930) (statement of Marion De Vries).

¶ 14. Brief on Behalf of Congress of Industrial Organizations [as] Amicus Curiae, Apex Hosiery Co. v. Leader, 310 U.S. 469 (1940) (No. 638). 74 cites. See supra note 316 and accompanying text for more details.

¶ 15. Brief for Respondents, Apex Hosiery Co. v. Leader, 310 U.S. 469 (1940) (No. 638). 64 cites. See supra note 317 and accompanying text for more details.

¶ 16. Petitioner’s Brief in Reply to Briefs of Respondents and Amici Curiae, Apex Hosiery Co. v. Leader, 310 U.S. 469 (1940) (No. 638). 25 cites. See supra note 318 and accompanying text for more details.

¶ 17. Brief for Appellees, United States v. Am. Trucking Ass’ns, 310 U.S. 534 (1940) (No. 713). 32 cites. This case concerned the interaction between the Motor Carrier Act of 1935 and the Fair Labor Standards Act of 1938. The legislative history in the brief covered both statutes, plus post-1938 developments. Id. at 25-38, 53-54. Brief-signing attorney J. Ninian Beall had represented the American Trucking Associations at the 1937 joint House and Senate hearings on the FLSA. Fair Labor Standards Act of 1937: Joint Hearings on S. 2475 and H.R. 7200 Before the S. Comm. on Educ. & Labor and the H. Comm. on Labor, Pt. 2, 75th Cong. 743 (1937) (statement of J. Ninian Beall, General Counsel, American Trucking Associations). Past representatives of the Associations had played a similar role with respect to the Motor Carrier Act. E.g., To Amend the Interstate Commerce Act: Hearings on S. 1629, S. 1632, and S. 1635 Before the S. Comm. on Interstate Commerce, Pt. 1, 74th Cong. 305 (1935) (statement of William G. Fitzpatrick, American Trucking Associations).

¶ 18. Brief for Respondent, United States v. Stewart, 311 U.S. 60 (1940) (No. 13). 39 cites. This case concerned the Federal Farm Loan Act of 1916, on which the brief’s legislative history focused. I have found no evidence that Stewart or any of the attorneys lobbied for that Act. However, brief-signing attorney Ernest L. Wilkinson was experienced at congressional lobbying more generally, most famously for Indian tribes, frequently testifying at committee hearings and drafting bills and committee reports since the mid-1930s. Woodruff J. Deem & Glenn V. Bird, Ernest L. Wilkinson: Indian Advocate and University President 129, 155-56, 196-201 (1982). His clients also included trade associations. Id. at 100.

¶ 19. Brief for Respondent, FCC v. Columbia Broad. Sys. of Cal., 311 U.S. 132 (1940) (No. 39). 32 cites. This case concerned the Communications Act of 1934. Brief-signing attorney Duke M. Patrick had written an analysis of the bill that was inserted in the transcript of the Senate hearings by the chairman of the Legislative Committee of the National Association of Broadcasters. Federal Communications Commission: Hearings on S. 2910 Before the S. Comm. on Interstate Commerce, 73d Cong. 68-69 (1934). Sixteen months earlier, Patrick had been General Counsel of the FCC’s predecessor agency and had testified before Congress on related issues. See Fees for Radio Licenses: Hearing on S. 5201 Before a Subcomm. of the S. Comm. on Interstate Commerce, 72d Cong. 20 (1933) (statement of Duke M. Patrick, General Counsel, Federal Radio Commission); To Amend the Radio Act of 1927: Hearings on H.R. 7716 Before the S. Comm. on Interstate Commerce, Pt. 2, 72d Cong. 35 (1932) (statement of Duke M. Patrick, Assistant General Counsel, Federal Radio Commission).

¶ 20. Brief for Appellees on Reargument, United States v. N. Pac. Ry. Co., 311 U.S. 317 (1940) (No. 3). 25 cites. This case concerned the vast federal land grants to the Northern Pacific Railway Company. The legislative history in the brief, id. at 15, 44-45, 69-80, was about the acts pertaining to these grants, mainly a Joint Resolution of 1870, which was specific to the Northern Pacific, and for which it obviously must have lobbied. Brief-signing attorney Lorenzo B. daPonte was general counsel to the Northern Pacific and would have had access to the company’s records and institutional memory.

¶ 21. Brief for Petitioner, Benitez Sampayo v. Bank of Nova Scotia, 313 U.S. 270 (1941) (No. 90). 48 cites. This case dealt with several statutory provisions related to farm bankruptcies. The brief’s discussion of legislative history focused overwhelmingly on congressional activity on bankruptcy from 1937 through 1940. Id. at 35-63. Brief-signing attorney Elmer McClain had testified in favor of farmer-debtors before the House Subcommittee on Bankruptcy (chaired by Walter Chandler, of Chandler Act fame) in December 1937. Farm Mortgage Moratorium: Hearings on S. 2215 and H.R. 6452 Before Special Subcomm. on Bankruptcy of the S. Comm. on the Judiciary, 75th Cong. 53 (1938) (statement of Elmer McClain). Also, McClain since at least 1937 had been connected to U.S. Senator William Lemke, a champion of farmer-debtors, having served as his co-counsel in defending the constitutionality of Lemke’s signature legislation, the Frazier-Lemke Farm Mortgage Moratorium Act of 1935. Wright v. Vinton Branch of Mountain Trust Bank of Roanoke, 300 U.S. 440, 442 (1937).

¶ 22. Brief for Respondent, Benitez Sampayo v. Bank of Nova Scotia, 313 U.S. 270 (1941) (No. 90). 52 cites. I have found no evidence of lobbying. Perhaps the attorneys briefed such material in response to the initiative of an opposing party, represented by a lawyer-lobbyist. See ¶ 21.

¶ 23. Brief for Respondents, Brooks v. Dewar, 313 U.S. 354 (1941) (No. 718). 44 cites. This case concerned the Taylor Grazing Act of 1934. I have found no evidence that any of the attorneys lobbied on this Act. But brief-signing attorney Milton B. Badt was experienced in testifying to Congress on Indian affairs. See Survey of Conditions of the Indians in the United States: Hearings Before a Subcomm. of the S. Comm. on Indian Affairs, Pt. 28, 72d Cong. 14841 (1934) (statement of Milton B. Badt). Also, as of 1940, he was being contacted by U.S. Senator Patrick McCarron regarding congressional investigation of federal grazing administration. Margaret Patricia McCarron, Patrick Anthony McCarron, Part II, Nev. Hist. Soc’y Q., Spring 1969, at 3, 9-10.

¶ 24. Brief for the Petitioner, Textile Mill Sec. Corp. v. Comm’r, 314 U.S. 326 (1941) (No. 34). 49 cites. Lobbying was the very subject of this case. The litigant itself (Textile Mills Securities Corporation) lobbied for passage of the Settlement of War Claims Act of 1928, and one of the issues in the case was whether the company could deduct its expenses for that lobbying, which arguably turned on whether the Act was “favor legislation” or “debt legislation.” The legislative history appearing in the company’s brief pertained mainly to that Act and its nature. Id. at 53-76.

¶ 25. Respondent’s Brief, United States v. Santa Fe Pac. R.R. Co., 314 U.S. 339 (1941) (No. 23). 58 cites. In this case, the federal government prosecuted a land claim against the Santa Fe Pacific Railroad on behalf of the Walapai Indians. The large majority of legislative history citations in the Railroad’s brief were to an unusual genre of legislative history: a Senate Document titled The Walapai Papers, S. Doc. No. 74-273 (1936). This was a huge cache of historical records on Walapai land rights from federal agency files that had been assembled and printed by Congress in 1936 at the behest of the Railroad’s patron, U.S. Senator Carl Hayden of Arizona, as part of his deliberate effort to help the Railroad’s case. Hayden enthusiastically distributed the Papers to the Railroad’s attorneys even before they were published. Christian W. McMillen, Making Indian Law: The Hualapai Land Case and the Birth of Ethnohistory 139-40 (2007).