abstract. Conventional wisdom often portrays Congress’s role as legislative. This view is incomplete. While Congress engages in general oversight and factfinding through hearings, reports, and the like to advance legislation, members of Congress also dedicate enormous amounts of time and energy to advocating on behalf of individual constituents in administrative proceedings. This is not legislating; it is the congressional bully pulpit, directed at the trenches of administrative adjudication. Described by Professor Jerry L. Mashaw as “mysterious,” this activity throws into question textbook accounts of the modern separation of powers and the administrative state.

Our Feature is the first to examine systematically this form of congressional intervention in administrative adjudication. We make four contributions. First, we discuss the doctrinal landscape around congressional interventions and explain what administrative and constitutional law misses about congressional control by focusing nearly exclusively on judicial review as the predominant constraint on agencies. Second, using a unique and comprehensive dataset of over two million cases appealed to the Board of Veterans’ Appeals (BVA) from 2003 to 2017, we provide an empirical portrait showing the scope and importance of congressional inquiries into pending veterans’ appeals. Between 4-11% of cases advanced on the docket and resolved by BVA are subject to a congressional inquiry. Veterans are twice as likely to receive expedited treatment when a member of Congress inquires on their behalf. Third, we investigate the distributive consequences of this system, which deviates significantly from the neutral, rationality-based model envisioned by the Administrative Procedure Act. Fourth, we spell out the implications of our findings for contemporary separation-of-powers doctrine and administrative law. The model of Congress as a site for the resolution of individual grievances—often thought to be a bygone relic of the Constitution’s Petitions Clause—remains alive and well.

authors. Lindsey Gailmard is Postdoctoral Scholar, Regulation, Evaluation, and Governance Lab (RegLab), Stanford University. Daniel E. Ho is William Benjamin Scott and Luna M. Scott Professor of Law, Stanford Law School; Professor of Political Science and Professor of Computer Science (by courtesy), Stanford University; Senior Fellow, Stanford Institute for Human-Centered Artificial Intelligence; Senior Fellow, Stanford Institute for Economic and Policy Research; Director, RegLab, Stanford University. Mark S. Krass is Nonresident Affiliate, RegLab, Stanford University. The views presented here are those of the authors alone and do not necessarily represent the position of any government agency. We thank David Ames, Maggie Blackhawk, Emily Bremer, Cody Drolc, Cassandra Handan-Nader, Anne McDonough, Austin Peters, James D. Ridgway, and participants at the RegLab meeting for thoughtful comments, as well as Derek Ouyang, Kit Rodolfa, and Joyce Tagal for help.

Introduction

Oversight is central to the legitimacy of administrative agencies.1 But for individual claimants before agency adjudicators, judicial review is often thought to be the only external check on agency action.2 That understanding is wrong. Members of Congress play a role in the adjudication of individual cases before administrative agencies through what we call “congressional intervention”: the inquiries that members of Congress submit to agencies advocating for their constituents to get a favorable or faster decision on a pending claim, or requesting a status update on one of these claims, of which there are thousands each year. Members of Congress collectively invest millions of dollars and thousands of hours into lobbying agencies on their constituents’ behalf. We ask whether all that effort matters—and, given the distributive, democratic, and due-process consequences of congressional intervention, whether it should.

These questions are important. Administrative adjudications affect the fundamentals of life for the millions of people who file asylum claims, rely on federal-assistance programs like Social Security Disability Insurance to make ends meet, or otherwise find themselves interacting with a federal agency. In fiscal year 2023, the Executive Office for Immigration Review completed 526,382 immigration-court cases,3 and the Social Security Administration (SSA) conducted 246,399 hearings for disability benefits before administrative law judges (ALJs).4 In high-volume adjudication settings like these, delay is pervasive.5 If congressional outreach influences how agencies treat otherwise-similar claims, then members’ favored constituents might have a leg up in obtaining timely or favorable decisions. This means members may secure electoral advantage—courtesy of taxpayers—that degrades democratic competition, while agencies may open backdoor channels for currying favor with Congress. At a time when the vision of a civil service committed to procedural neutrality is in question,6 this is the ideal time to interrogate the failures of agency adjudication—and Congress’s highly personalistic, and even clientelist, role in correcting them after the fact.7

Congressional intervention in agency adjudication is conventionally seen as tremendously beneficial. Claimants may need all the help they can get to navigate complex bureaucratic processes. Members of Congress can be ideal conduits for information between constituents and agencies. After all, they are in the business of cultivating strong relationships with their constituents and might have stronger links to vulnerable populations in their districts than federal agencies do.8 And because congressional offices have come to specialize in advocating for constituents before agencies, they may be able to package a claimant’s narrative in a way that is easier for the agency to digest—for instance, by foregrounding the facts that are material to the agency’s decision.9

What’s more, many legislators view interceding with agencies and providing their constituents with a voice in the bureaucracy as central to their job descriptions.10 Members of Congress make tens of thousands of inquiries to agencies every year.11 The collective effort expended on this task is enormous: one estimate is that some 20-30% of congressional office budgets can go toward constituent engagement,12 and veterans’ claims are commonly understood to comprise a significant part of that
work.13 While politicians rarely interface with agency officials themselves absent a compelling personal stake—they usually leave that to professional “caseworkers” in their offices14—legislators do manage their casework operations and intervene when problems filter up to them.15 In other words, the rise of congressional interventions reveals an implicit understanding of what it means to represent constituents in government, with legislators spending their scarce energy and resources lobbying the executive branch on behalf of individual claimants, possibly at the expense of making law. Is that effort paying off—and should it?

The doctrinal and scholarly consensus is that congressional interventions in individual cases are rare and rarely change outcomes.16 That view flows from the very foundation of the Administrative Procedure Act (APA). The APA was drafted with an eye to agencies’ newfound powers to issue rules with the force of law, enforce those rules, and resolve claims through in-house adjudication. To counterbalance those powers, the APA required, among other things, that agency action be based only on the rational consideration of evidence collected on an open agency record—and it backed that requirement with judicial review.17 The APA’s requirement that agencies act only in response to their rational view of the evidence implies that legislative pressure cannot determine the outcome of any individual case. Deciding a case on the basis of political pressure would mean relying on an irrelevant factor outside the record.18

Because the APA treats courts as the main guarantors of agency rationality, it is no wonder that administrative-law scholars have fixated on judicial review while largely ignoring the role that members of Congress play in adjudication. Occasionally, scholars have acknowledged congressional interest in constituents’ cases. As far back as 1983, Professor Jerry L. Mashaw noted the “mysterious” role played by congressional intervention in benefits adjudication.19 Mashaw was skeptical that such interventions mattered, suggesting that members of Congress were mostly engaged in marketing when they “claim[ed] [to] have been effective in their intervention.”20 If interventions did have an impact, Mashaw thought it would mostly be to waste agencies’ time writing letters back to Congress instead of resolving cases.21 Other administrative-law luminaries, like Professors Thomas W. Merrill and Robert A. Kagan, have noted in passing the role of such congressional inquiries as external checks on agency behavior.22 Professor Jack M. Beermann discussed constituency casework in greater depth as one mechanism of “congressional administration” among many, including Congress’s appropriations power and other formal legislative powers.23 The Administrative Conference of the United States (ACUS) recently recommended that agencies improve their procedures for handling congressional inquiries.24 Despite this recommendation,25 the supporting report stressed that congressional inquiries, while legitimate, by construction have a limited role in the adjudication of cases.26 To our knowledge, no study has ever systematically assessed the law and empirics of such congressional interventions in agency adjudication.

We start to unravel the mystery of congressional intervention by focusing on the case of veterans’ appeals. In recent years, the Veterans Benefits Administration (VBA) has decided a record number of claims for disability compensation.27 Many of these decisions are subsequently appealed to the Board of Veterans’ Appeals (Board or BVA). As the number of appeals to the Board has also steadily increased, overburdened Veterans Law Judges face pressure to issue hasty decisions to avoid adding to the backlog of cases.28

Unable—or perhaps unwilling—to give BVA the resources necessary to manage its growing backlog, Congress has supplemented its legislative efforts with more traditional forms of oversight. There have been angry speeches, as when Representative Morgan Luttrell, member of the House Committee on Veterans’ Affairs, said in a hearing that the “Board leadership’s primary concern is issuing a high number of decisions on veterans’ claims on appeal” at the expense of “ensur[ing] that these decisions are correct and fair for each and every veteran.”29 There have been committee inquiries, like Senator Jon Ossoff’s October 2023 letter to the Veterans Affairs Secretary demanding explanations for decision delays and condemning them as “betray[ing] the sacred compact we make with those who wear the uniform and undermin[ing] faith in our institutions.”30And there have been reports, like the damning assessment issued by the Government Accountability Office in 2023 criticizing BVA’s Quality Review Program after a study conducted by one of us documented the manipulation of that program to manufacture high accuracy rates.31 We leave these traditional tools of oversight to one side.

We focus instead on members’ responses to individual requests for help dealing with pending adjudicatory matters before administrative agencies—what others may call “casework” or “constituency service”32 and what BVA attorneys refer to as “congressional inquiries.”33 We use the term “congressional intervention” to distinguish these actions from congressional involvement in nonadjudicatory proceedings (e.g., submitting a comment in a rulemaking). More specifically, we focus on the requests of individual constituents with cases pending before an administrative agency that involve evidentiary hearings.34 Of course, members of Congress are also frequently recruited to help with “larger-scale matters,” like local governments’ grant applications or federal enforcement actions targeted at significant businesses.35 And they routinely pursue constituents’ particular interests through legislative efforts—that is, through pork-barrel politics. We set aside these grander kinds of congressional favors and concentrate only on interventions in routine, individual adjudications.

Legislators intervene in thousands of appeals to BVA on behalf of their constituents every year. We examine the effect of these inquiries by examining a procedural back door into the veterans’ appeals process created by Congress’s own short-term fix in 1994: the opportunity to request that BVA decide a particular veteran’s case more quickly due to extenuating circumstances.36 Under the 1994 Board of Veterans’ Appeals Administrative Procedures Improvement Act, a case could be “advanced on motion for earlier consideration and determination” provided “the case involve[ed] interpretation of the law of general application affecting other claims or for other sufficient cause shown.”37 We ask whether congressional intervention succeeds in pressuring BVA to exercise its congressionally created discretion to decide cases more quickly in favor of legislators’ favored constituents.

The example of BVA shows that congressional inquiries do matter in practice. They matter to the agency, which must handle thousands of incoming requests each year; to litigants, whose cases may be profoundly affected by them; and to the fairness of the system as a whole, conditioning timely justice on access to a channel of assistance that comes laden with many distributive imperfections. Given scarce adjudicatory resources, if decisions are expedited for some, costs of increased delay are left for others to pay.

We make four specific contributions to advance our argument. First, we highlight how administrative law’s fixation on judicial review as the predominant constraint on administrative actions ignores important mechanisms of congressional control. And we interrogate the constitutional and administrative-law basis for this type of informal congressional oversight.

Second, we empirically document and assess congressional intervention, examining when members of Congress intervene and whether such inquiries have any effect on agency adjudication. We leverage a unique dataset of all appeals to BVA between 2003 and 2017 obtained through Freedom of Information Act requests. Our data cover nearly eighty thousand congressional inquiries on behalf of claimants into the status of a case and documents whether BVA exercised its procedural discretion to expedite. We provide new insight into the impact of congressional contact on procedural discretion. And we assess the extent to which congressional resources and attention augment BVA’s capacity to address veterans’ claims by providing information relevant to adjudication, or whether they simply provide a mechanism for members of Congress to claim credit for the actions and services of the bureaucracy. And we explore the relationship between congressional interventions and legislative behavior, including bill sponsorship and voting records.

Third, we assess how BVA’s reliance on inquiries to inform advancement decisions affects fairness. Over our study period, a substantial number of cases advanced on the docket involved congressional intervention; in 2017, the last year of our data, there were 1,720. Depending on the year, those cases constituted between 4-11% of advanced cases that BVA resolved. Moreover, for claimants who did not document one of the criteria for expedited treatment (e.g., financial hardship, terminal illness, advanced age), congressional intervention on their behalf doubled their likelihood of being advanced on the docket relative to similar cases without congressional intervention. The quantity of cases advanced with inquiries highlights their potential to displace adjudication of those claimants who have not sought congressional assistance or whose legislators lack a developed framework for making requests to the Department of Veterans Affairs (VA). We investigate the correlation between congressional intervention and demographic characteristics by combining data on inquiries with district-level census data. There is suggestive evidence that congressional intervention not only reduces decision wait times but also increases the likelihood of a favorable disposition.

Finally, we argue that these empirical findings yield important lessons for practice and theory. As a practical matter, our results suggest a serious gap in the legal framework governing congressional interventions and casework. While some agencies have formulated explicit rules on handling congressional contacts,38 others—like BVA—have no explicit policy on how to reflect congressional inquiries in the administrative record. Agencies should fill that gap. They should work to learn as much as possible from inquiries, treating each one as a sign of potential systematic problems and a tool for proactively identifying process improvements. For example, if congressional inquiries indicate that veterans eligible for expedited processing because of their age are not receiving it, then BVA ought to view that as a signal to improve automated flagging of claims that meet advanced-age criteria. Agencies should also indicate inquiries on the public docket. Our results suggest that the lack of a transparent policy could be especially problematic in mass-adjudication agencies like BVA where adjudication resources are scarce. Relying on congressional inquiries creates a basic due-process concern that well-represented litigants may be in a better position to exploit unwritten levers like congressional interventions to avoid delay at the expense of veterans who lack these tools. Agencies ought to think carefully about how to make access to their attention fair. Indeed, given the statutory and constitutional frailty of such congressional interventions and the serious inequities they introduce, our results support improving safeguards on such contacts.

Our results also show the limits of a formalist view of the separation of powers. Judges tend to emphasize that Congress may only legitimately exercise legislative power—which it can do only as a body.39 In that world, individual members of Congress have no constitutionally recognized “power” of their own.40 The formalist approach has a certain appeal. The involvement of individual members of Congress in the execution or interpretation of laws could breed unfairness, for instance.

At least as a descriptive matter, our results show that the formalist view is incomplete. Individual members of Congress are key to finding resolutions for constituents aggrieved by executive action. They do this mostly by unsticking bureaucratic processes plagued by delay. This problem-solving function is not, in the main, formally codified in statutes. But members of Congress see it as a core part of their jobs as representatives. So does the public. In short, an accurate picture of how mass-benefit programs are administered by the federal government must include members of Congress.

Our objection to a restrictive formalist view of the separation of powers turns out to be connected to a rich history. An emerging scholarship has sought to uncover the role Congress once played in resolving retail-level injustices, especially in the context of mass-benefit programs. As Professor Maggie Blackhawk has shown, from the Founding through the late 1940s, the right to petition Congress provided a mechanism by which the politically powerless could seek congressional assistance, and petitions were resolved in ways that “took many forms, not all of them clearly delineated as adjudicative, legislative, or executive.”41 At midcentury, Congress mostly replaced petitioning with executive bureaucracies fit for purpose. The establishment of the Court of Claims and the Pensions Bureau, and the implementation of the APA and the Legislative Reorganization Act of 1946, all but destroyed petitions as an avenue for Congress to right individual wrongs.42

But a form of petitioning survives. Our work suggests that the practice of legislators acting on the basis of constituent complaints to ensure the proper execution of laws lives on today in the less glamorous incarnation known as casework. In effectively representing constituents before administrative agencies, members of Congress draw on an alternative vision of congressional representation that goes beyond roll-call voting or bill sponsorship—beyond a purely legislative role.

The normative implications of this revelation are complex. On the one hand, as political scientists have long understood, congressional interventions allow legislators to signal their responsiveness to constituent needs, improving their political standing.43 In addition to amplifying the voices and concerns of their constituents, legislators provide constituents with a voice in otherwise-faceless bureaucratic processes, helping to legitimize and humanize them, and a lever of power for constituents who otherwise lack it in Washington. Their participation in administration may even help to constitutionalize the administrative state by ensuring congressional control after a statute is passed, even when the statute contains vast delegations to the executive branch.44 On the other hand, the inequities introduced by congressional intervention might undermine the fairness and rationality values embedded in administrative law. Despite internal ethics rules forbidding such behavior, legislators may intervene to represent interests that may be politically problematic to represent publicly,45 may intervene more on behalf of constituents belonging to certain demographic groups,46 or—least surprising of all—may simply structure their interventions to maximize their own political benefits. In this way, congressional interventions are a window into the tradeoff between expertise and democracy that pervades administrative law.47

This Feature is divided into six Parts. Part I lays out the doctrinal landscape. Administrative law as a field has fixated nearly exclusively on judicial review, with little attention paid to informal mechanisms of congressional control. We show that congressional intervention in agency adjudication lies in significant tension with the APA and a formalistic conception of the separation of powers. Part II synthesizes the theory and scholarship around congressional intervention. Members of Congress may want to boost their own popularity by helping constituents navigate the complicated federal bureaucracy. But if intervention does not in fact affect the process, as is commonly believed, congressional interventions may simply allow legislators to claim credit for an award of benefits even without exerting any real influence. Both explanations pose puzzles for the separation of powers. And an alternative theory that frames congressional inquiries as a means of gathering information to be used in legislation has never been empirically tested.

Part III articulates why veterans’ benefits adjudication provides an important setting in which to unravel the mystery of congressional intervention. We first provide the legal and institutional context for veterans’ adjudication at BVA, reviewing the origins of BVA’s caseload crisis and explaining how BVA tracks and manages congressional intervention. Prior studies of congressional influence have not focused on more formal types of adjudication that require evidentiary hearings, and we know relatively little about the prevalence, distribution, and effects of interventions in these settings. Because our dataset was used by BVA itself to administer cases, it enables a comprehensive assessment based on what was known to the agency at the time of decision.

Part IV presents empirical evidence from our novel dataset of all two million veterans’ appeals at BVA from 2003 to 2017, including nearly eighty thousand congressional inquiries. First, we show that congressional intervention is prevalent and still increasing, with thousands of inquiries per year. Second, we show that members of Congress take highly divergent approaches to constituency service, even adjusting for the veteran population. Broadly, congressional inquiries favor relatively older, male veterans who have attorneys. Finally, we study the impact of interventions on procedural outcomes (e.g., whether a case is “advanced on the docket” to be decided earlier48) and substantive outcomes. BVA regulations allow a case to be expedited for any of the following three reasons: (1) the veteran is “seriously ill”; (2) the veteran is experiencing “severe financial hardship”; (3) or for “other sufficient cause,”49 which includes, but is not limited to, “administrative error resulting in a significant delay,” “administrative necessity,” or the veteran being “75 or more years of age.”50 Adjudicators appear to possess substantial discretion when deciding whether a veteran has serious illness or financial hardship. As one congressional staffer opined, financial-hardship criteria are “up to who is reviewing.”51 We show that a congressional inquiry makes advancement twice as likely on average and translates into a 158-day reduction in wait times, and that such advancements are likely caused not merely by any marginal information the legislator provides to the agency. We also present evidence that congressional interventions may increase the odds of a favorable disposition on the merits.

Part V discusses the extent to which our empirical results may or may not generalize to other adjudicatory settings and discusses empirical limitations. While our results may not generalize to all forms of congressional intervention and we are unable to observe congressional screening of constituent requests, we offer institutional reasons to think that the selection of meritorious claims does not drive our findings.

Part VI draws out the implications for law and policy. We focus both on the lessons for the separation of powers and on potential policy interventions. We consider a range of interventions, including (1) transparency, namely requiring that congressional inquiries be disclosed or be entered into the formal record; (2) guidance, namely providing more precision around when a case may be advanced for undocumented reasons; (3) internal process improvements, namely requiring agencies to treat congressional inquiries as performance indicators for systematic improvements of case review; and (4) restrictions on congressional inquiries.

The Feature thus aims to show that while Congress expressly provided for the judicial review of agency action through the APA’s requirement of reasoned decision-making, that is not its only strategy for shaping the outcomes of administrative processes. Nor does Congress influence the executive branch only when acting as a collective body to pass legislation. Individual members of Congress shape the resolution of vast numbers of veterans’ appeals using the congressional bully pulpit—members’ power to cajole and persuade administrators to adopt the members’ priorities as their own—in a manner that is arguably at odds with administrative-law and constitutional theory. The case of veterans’ appeals demonstrates that constitutional and administrative law must grapple with the proper scope not only of judicialized administration, but of legislative administration as well.