70 Am. U. L. Rev. 387 (2020).

* Bridget C.E. Dooling is a research professor at The George Washington University Regulatory Studies Center. The author thanks Steve Balla, Jerry Ellig, Rachel Potter, Philip Wallach, Brian Feinstein, Matt Lawrence, Laura Stanley, Mark Febrizio, Camille Chambers, and the participants at the Oversight and Accountability Symposium hosted by American University Washington College of Law in November 2020, particularly Eloise Pasachoff and Jeff Lubbers, for helpful comments on earlier drafts. She also thanks the Brookings Institute for its financial support of a related essay published in its Series on Regulatory Process and Perspective in March 2020.

Into the Void: The GAO’s Role in the Regulatory State

The study of Congress largely focuses on its members, committees, and leaders.1See infra note 11. Meanwhile, the study of federal agencies tends to focus on those that fall within the executive branch.2See infra note 12. But many agencies and institutions exist to support Congress, and administrative law and political science scholarship has largely overlooked their internal operations. Far more than ministerial, some of these agencies have the ability to shape policy in very meaningful ways.3See, e.g., About MedPAC, MedPAC, http://www.medpac.gov/-about-medpac- [https://perma.cc/4LXK-MRJZ] (explaining that the Medicare Payment Advisory Commission “advise[s] . . . Congress on issues affecting the Medicare program” and “analyz[es] access to care, quality of care, and other issues affecting Medicare”); see also infra notes 15–17 and accompanying text. They also have their own institutional interests distinct from the legislators they serve.4See infra Section IV.A (describing congressional agencies’ incentives and norms that guide the agencies’ behavior). As instruments of Congress, though, these agencies are generally exempt from the Administrative Procedure Act and a host of other “good government” statutes that govern agency conduct.5See infra note 23 and accompanying text (noting that congressional agencies are not subject to the APA or other statutes that use the APA’s definition of “agency”). This does not mean that they operate in a lawless fashion. Instead, other statutes and internal norms guide their conduct.6See infra notes 96–101 and accompanying text (describing congressional agencies’ emphasis on the non-partisan, objective nature of their work, and agencies’ actions to buffer themselves from the political environment). This Article suggests what we might learn from closer study of these understudied agencies.

One example of such an agency is the U.S. Government Accountability Office (GAO). Known mostly for its role as an auditor, the GAO has expanded its activities over time.7See infra notes 28–30 and accompanying text’’’ (discussing the variety of functions the GAO now performs, including auditing, issuing legal opinions, and providing policy analyses). As part of this expansion, the GAO built on its prior practice of offering legal opinions and decisions on appropriations matters to become a referee in an increasingly important area of the administrative state: issuing legal opinions on which agency actions are “rules” under the Congressional Review Act85 U.S.C. §§ 801–08 (2018). (CRA).9See infra notes 62–69 (describing how the GAO stepped into the role of providing legal opinions regarding whether agency action constitutes a “rule”). The significance of these opinions, which are not binding as a matter of law, has grown as legislators use them strategically in regulatory politics.10See infra Section IV.B (discussing the range of motives legislators have for seeking CRA legal opinions, from wanting clarity about how to apply the CRA to signaling political disapproval). The full outcome of the Senate races in the 2020 election, which was not yet available when this Article went to press, will inform near-term use of the CRA to disapprove rules. Whether CRA disapproval is a potent tool in 2021 or not, we can expect legislators to continue to request the GAO’s legal opinions in strategic ways.

This Article examines how Congress has used the GAO’s legal opinion function to expand the application of the CRA. Part I describes congressional agencies to shed light on the wide array of functions that they serve and to highlight one in particular: the GAO. Part II explains the CRA, the statute that allows Congress to disapprove certain rules using fast-track procedures. Part III describes the significance of the GAO’s CRA legal opinions. Part IV sketches a strikingly wide variety of policy implications of this phenomenon and suggests areas for future study.

I. Congressional Institutions and Agencies 

In political science and law, the study of Congress has largely focused on its members, committees, and leaders.11See, e.g., R. Douglas Arnold, Congress and the Bureaucracy: A Theory of Influence 3 (1979) (examining congressional influence over bureaucratic decisions regarding the geographic allocation of funding); Benjamin Ginsberg & Kathryn Wagner Hill, Congress: The First Branch 148 (2019) (offering an overview of congressional committee and personal staffs); Douglas L. Kriner & Eric Schickler, Investigating the President: Congressional Checks on Presidential Power 6 (2016) (describing congressional investigative oversight as a tool to “expose wrongdoing, force executive-branch officials to answer difficult questions, and raise the political costs of noncompliance”); The Oxford Handbook of Political Institutions (R.A.W. Rhodes, Sarah A. Binder & Bert A. Rockman eds., 2008) (providing an overview of the study of political institutions); Jack M. Beermann, Congressional Administration, 43 San Diego L. Rev. 61, 127–29 (2006) (providing an overview of congressional oversight institutions, including the GAO); Brian D. Feinstein, Avoiding Oversight: Legislator Preferences and Congressional Monitoring of the Administrative State, 8 J. L. Econ. & Pol’y 23, 24 (2011) (examining how members of Congress value and view oversight subcommittee service); Brian D. Feinstein, Congress in the Administrative State, 95 Wash. U. L. Rev. 1187, 1195 (2018) (exploring congressional oversight of the executive branch through committee and subcommittee oversight hearings); Kenneth Lowande & Rachel Augustine Potter, Congressional Oversight Revisited: Politics and Procedure in Agency Rulemaking, J. Pol. (forthcoming) (manuscript at 1) (analyzing congressional lawmakers’ use of procedural mechanisms to exert oversight of executive policymaking); Kenneth Lowande, Who Polices the Administrative State?, 112 Am. Pol. Sci. Rev. 874, 874 (2018) (presenting oversight relationships between individual legislators and executive agencies). The study of bureaucracy in federal government tends to focus on executive branch agencies.12See, e.g., Sean Gailmard & John W. Patty, Learning While Governing: Expertise and Accountability in the Executive Branch 3 (2013) (analyzing information development and transmission as a product of incentives facing executive branch bureaucrats); David E. Lewis, The Politics of Presidential Appointments: Political Control and Bureaucratic Performance 1–2, 9–10 (2008) (studying the Federal Emergency Management Agency and its response to Hurricane Katrina to demonstrate agency appointees’ influence on public policy); Rachel Augustine Potter, Bending the Rules: Procedural Politicking in the Bureaucracy 2 (2019) (exploring the role of unelected bureaucrats in the rulemaking process); William G. Resh, Rethinking the Administrative Presidency: Trust, Intellectual Capital, and Appointee-Careerist Relations in the George W. Bush Administration 2 (2015) (examining relationships between career executives and political appointees); James Q. Wilson, Bureaucracy: What Government Agencies Do and Why They Do It (1989); Terry M. Moe, The Politics of Bureaucratic Structure, in Can the Government Govern? 267, 268 (John E. Chubb & Paul E. Peterson eds., 1989) (describing the politics of agency structural design); Sean Gailmard & John W. Patty, Formal Models of Bureaucracy, 15 Ann. Rev. Pol. Sci. 353, 353 (2012) (reviewing themes in research applying formal modeling to bureaucratic institutions). The same is true for administrative law, which has mostly concerned itself with judicial review of agency actions subject to the Administrative Procedure Act.13See, e.g., Bennett v. Spear, 520 U.S. 154 (1997); Auer v. Robbins, 519 U.S. 452 (1997); Chevron, U.S.A., Inc. v. NRDC, Inc., 467 U.S. 837 (1984); Vt. Yankee Nuclear Power Corp. v. Nat. Res. Def. Council, Inc., 435 U.S. 519 (1978). The literature, vast and provocative as it is, nevertheless mostly overlooks the agencies and institutions that support Congress.14This is changing. See, e.g., Jesse M. Cross, Legislative History in the Modern Congress, 57 Harv. J. on Legis. 91, 99–101 (2020) (detailing the nonpartisan offices within Congress and describing how labor is divided in both partisan and nonpartisan congressional offices). More research is on the way. See, e.g., Jesse M. Cross & Abbe R. Gluck, The Congressional Bureaucracy 1–11 (George Mason Univ. Antonin Scalia Law Sch., CSAS Working Paper No. 20–23), https://administrativestate.gmu.edu/wp-content/uploads/sites/29/2020/10/Cross-Gluck-The-Congressional-Bureaucracy.pdf [https://perma.cc/5AND-KT6H] (introducing the concept of “congressional bureaucracy” by discussing the function and purpose of nine nonpartisan institutions internal to Congress); Cody A. Drolc, Watchdog or Partisan Pawn? Agenda Setting and Government Accountability Office Oversight (working paper) (on file with author) (analyzing partisan influences on the GAO).

Congressional institutions and agencies—which are separate from any individual legislator’s staff—provide a host of services to members of Congress, and, perhaps surprisingly, directly to the public.15See infra note 17 and accompanying text. These agencies and institutions employ several thousand people.16Ginsberg & Hill, supra note 11, at 148 (estimating that the legislative branch employs almost 20,000 people on Capitol Hill). Table 1 provides a glimpse of the breadth of institutions and agencies that reside in the legislative branch.

Table 1: Congressional Institutions and Agencies17Office of Mgmt. & Budget, A Budget for America’s Future Appendix 13–48 (2020), https://www.whitehouse.gov/wp-content/uploads/2020/02/appendix_fy21.pdf [https://perma.cc/A469-57YS]. This table adopts the groupings used in the Legislative Appendix of the source document. One exception is that, while the budget for the U.S. Tax Court is reflected in the legislative branch’s budget, id. at 38–39, the court is excluded from this table because of Freytag v. Commissioner, 501 U.S. 868 (1991). In that case, the Supreme Court determined that the Tax Court is “independent of the Executive and Legislative Branches.” Id. at 891; see Harold Dubroff & Brant J. Hellwig, The United States Tax Court: An Historical Analysis 265 (2d ed. 2014).


Office of the Vice President

Offices of the President Pro Tempore and Pro Tempore Emeritus

Leadership Offices (e.g., Offices of the Majority and Minority

Leaders, Whips, Conferences, Secretaries)

Committees (e.g., Appropriations, Policy)

Office of the Chaplain

Office of the Parliamentarian

Office of the Secretary

Office of the Sergeant at Arms and Doorkeepers

Office of the Legislative Counsel of the Senate

Office of Senate Legal Counsel

Caucus on International Narcotics Control

House of Representatives

Leadership Offices (e.g., Offices of the Speaker, Floor Leaders,


Standing Committees, Special and Select

Committee on Appropriations

Office of the Clerk (including Chaplain and Historian)

Office of the Sergeant at Arms

Office of the Chief Administrative Officer

Office of Diversity and Inclusion

Office of the Whistleblower Ombudsman

Office of the Inspector General

Office of General Counsel

Office of the Parliamentarian

Office of the Law Revision Counsel of the House

Office of the Legislative Counsel of the House

Office of Interparliamentary Affairs

Joint Committees

Joint Congressional Committee on Inaugural Ceremonies of 2021

Joint Economic Committee

Joint Committee on Taxation

Office of the Attending Physician

Office of Congressional Accessibility Services

Capitol Police
Office of Congressional Workplace Rights
Congressional Budget Office
Architect of the Capitol
Capitol Visitor Center
Botanic Garden

Library of Congress

Office of the Librarian

Library Services

Law Library

Office of the Inspector General

Office of the Chief Information Officer

Office of the Chief Operating Officer

Copyright Office

Congressional Research Service

National Library Service for the Blind and Print Disabled

Duplication Services

Fedlink Program and Federal Research Program

Government Publishing Office
Government Accountability Office
Medicare Payment Advisory Commission
Medicaid and CHIP Payment and Access Commission
U.S.-China Economic and Security Review Commission
Commission on International Religious Freedom
Other Boards, Commissions, and Trust Funds


It is difficult to characterize the functions of these agencies and institutions as a cohesive group. Their missions range broadly to include research services, legal advice, and audits and investigations on behalf of Congress; marvelous libraries and exquisite gardens that are open to the public; publication services for the entire federal government; law enforcement; and the day-to-day functioning of the buildings and grounds used by the legislature. The variety is noteworthy in itself, and particularly in comparison to the typical conception of Congress as a collection of voting members and their immediate staff.

Some of these agencies make decisions that directly affect the public, making them akin to federal executive branch agencies in terms of their functions. The Copyright Office manages the registration of copyrights.18Office of Mgmt. & Budget, supra note 17, at 34. The U.S. Capitol Police is a law enforcement agency with jurisdiction over the grounds of the U.S. Capitol, other congressional buildings, and those who enter them, as well as other areas.19USCP Fast Facts, U.S. Capitol Police, https://www.uscp.gov/media-center/uscp-fast-facts [https://perma.cc/S6KT-VGQV]. The GAO adjudicates federal procurement disputes for would-be contractors known as “bid protests.”20See Bid Protests, Appropriations Law, & Other Legal Work: FAQs, U.S. Gov’t Accountability Office, https://www.gao.gov/legal/bid-protests/faqs [https://‌perma.cc/76V8-N5ZZ]. The advisory commission on Medicare and the advisory commission on Medicaid and the Children’s Health Insurance Program (CHIP) convene experts drawn from the public to inform policy choices on these massive social programs.21See Office of Mgmt. & Budget, supra note 17, at 39–40 (discussing the responsibilities of the Medicare Payment Advisory Commission and the Medicaid and CHIP Payment and Access Commission, including providing analysis and making recommendations to Congress, the U.S. Department of Health and Human Services, and the states on issues affecting the programs). However, congressional agencies are not subject to the Administrative Procedure Act22Pub. L. No. 79-404, 60 Stat. 237 (1946). (APA) and therefore the other cross-government, “good government” statutes that use the APA’s definition of “agency” to determine applicability23In the Administrative Procedure Act, “the Congress” is excluded from the Act’s definition of “agency.” 5 U.S.C. § 551(1); accord 4 C.F.R. § 81.1(a) (2019) (“GAO is not subject to the Freedom of Information Act.”). The GAO established its own set of public disclosure rules, which “follows the spirit of [FOIA].” Id. While congressional agencies are arguably distinct from “the Congress,” if that means the entity that is authorized to make laws under the Constitution, they are undeniably part of the legislative branch. See Bowsher v. Synar, 478 U.S. 714, 731 (1986) (finding that the Comptroller General of the United States, the head of the GAO, is part of the legislative branch). like the Freedom of Information Act (FOIA)245 U.S.C. § 552 (2018). and the Federal Advisory Committee Act.25Pub. L. No. 92-463, 86 Stat. 770 (1972).

While many of the agencies and institutions noted above deserve additional study, the balance of this Article focuses on one of them: the GAO. The GAO employed over 3,000 people and had a budget of $638 million in fiscal year 2019.26See Office of Mgmt. & Budget, supra note 17, at 37–38. It began as the General Accounting Office in 1921 and Congress renamed it the Government Accountability Office in 2004.27See Office of Mgmt. & Budget, supra note 17, at 37–38. This name change reflected the GAO’s broader mission, which expanded, over time, from the government’s “fiscal watchdog” to the “congressional watchdog.”28 While the GAO’s mission has grown, the agency remains best known for its audit and evaluation work.29See Nancy Kingsbury, The Government Accountability Office and Congressional Uses of Federal Statistics, 631 Annals Am. Acad. Pol. & Soc. Sci. 43, 43–44 (2010) (explaining that the GAO’s audit and evaluation work helps Congress oversee federal programs and operations); Simon D. Norton & L. Murphy Smith, Contrast and Foundation of the Public Oversight Roles of the U.S. Government Accountability Office and the U.K. National Audit Office, 68 Pub. Admin. Rev. 921, 925 (2008) (highlighting the GAO’s structural and qualitative oversight of federal agencies); Allen Schick, Congress and the “Details” of Administration, 36 Pub. Admin. Rev. 516, 521–22 (1976) (explaining the GAO’s expansion into evaluations). It has a host of other functions, however, including “policy analyses, as well as [the issuance of] legal opinions, bid protest adjudications, and investigations.”30Kingsbury, supra note 29, at 44. It is this legal opinion function, particularly legal opinions that the GAO dispenses for purposes of the CRA, which this Article examines.31The GAO website explains that it “issues decisions regarding federal bid protests, appropriations law, and other legal matters” including “reviewing how federal agencies comply with the Congressional Review Act and the Federal Vacancies Reform Act.” Bid Protests, Appropriations Law, and Other Legal Work, U.S. Gov’t Accountability Office, https://www.gao.gov/legal [https://perma.cc/7HLN-QG73].

II. The Congressional Review Act

The CRA is an unusual statute that gives Congress fast-track procedures to disapprove, within a prescribed window of time after its issuance, a rule issued by a federal agency.325 U.S.C. § 802 (2018).

Congress, of course, has other ways to roll back an agency’s rule, either by legislating to overturn the agency’s rule or by restricting the agency’s ability to implement the rule using an appropriations rider,33See Todd Garvey & Daniel J. Sheffner, Cong. Res. Serv., R45442, Congress’s Authority to Influence and Control Executive Branch Agencies 13–14 (2018) (noting that Congress employs appropriations riders as a tool to prohibit or condition the use of funds to guide agency action). See generally Jason A. MacDonald, Congressional Power over Executive Branch Policy Making: Limitations on Bureaucratic Regulations, 1989–2009, 43 Presidential Stud. Q. 523 (2013) (arguing that Congress uses limitation riders in appropriations bills to prohibit specific agency action, particularly when Congress disagrees with the executive branch, and predicting that limitation riders are removed once Congress and the executive branch’s goals align); Jason A. MacDonald, Limitation Riders and Congressional Influence over Bureaucratic Policy Decisions, 104 Am. Pol. Sci. Rev. 766 (2010) (explaining how congressional majorities strategically use limitation riders in appropriations bills to influence policies in ways the majorities would otherwise be unable to). but the CRA offers a different pathway that has greater appeal at certain times.34See Beermann, supra note 11, at 83–88 (“Substantively, the CRA is unnecessary because Congress always had the power to legislatively override agency rules.”). The CRA is a sort of book-end to the initial delegation of rulemaking or other legal authority, giving Congress a streamlined opportunity to disapprove a rule once issued by the agency.35See Daniel Cohen & Peter L. Strauss, Congressional Review of Agency Regulations, 49 Admin. L. Rev. 95, 102 (1997) (describing one of the goals of the CRA as shifting responsibility to Congress to review agency rules). The scope of what qualifies as a “rule,” then, is pivotal. The CRA does not direct any particular agency to adjudicate that question,36See 5 U.S.C. §§ 801–08. and, as will be discussed below, the GAO has stepped into the role upon request from legislators.

Under the CRA, a member of the House or Senate may introduce a resolution for disapproval during a period of time spanning sixty days of continuous session, beginning the day Congress received the rule.37Maeve P. Carey & Christopher M. Davis, Cong. Res. Serv., IF10023, In Focus: The Congressional Review Act (2018). This is a relatively short window of time in which Congress may act under the CRA, therefore it functions somewhat like a statute of limitations. Within that window, the statute allows Congress to expedite the resolution through Congress—avoiding, for example, filibuster in the Senate.38§ 802 (containing the CRA’s disapproval procedures); see also Maeve P. Carey & Christopher M. Davis, Cong. Res. Serv., IF10023, In Focus: The Congressional Review Act (2018) (describing procedures for a joint resolution of disapproval under the CRA); Maeve P. Carey & Christopher M. Davis, Cong. Res. Serv., R43992, The Congressional Review Act (CRA): Frequently Asked Questions 14 (2020) (noting that the CRA contains “fast track” procedures for both committee and floor consideration of a CRA disapproval resolution in the Senate, though it does not contain fast track procedures for a disapproval resolution in the House). Once the resolution has passed in both chambers, it is presented to the President for veto or signature.39Maeve P. Carey & Christopher M. Davis, Cong. Res. Serv., R43992, The Congressional Review Act (CRA): Frequently Asked Questions 13 (2020) (noting that a veto is almost certain given the President’s own administration made the rule, thus creating a “de facto supermajority requirement”).

This presentment process40Presentment cures the legislative veto defect that was at issue in Immigration and Naturalization Service v. Chadha, 462 U.S. 919 (1983). Id. at 956–57 (declaring unconstitutional a statute that allowed one house of Congress to override executive action without the constitutionally-mandated bicameralism and presentment processes); see Paul J. Larkin, Jr., Reawakening the Congressional Review Act, 41 Harv. J.L. & Pub. Pol’y 187, 197 (2018) (explaining that “[t]he CRA falls between the quick-acting legislative veto and the deliberative process that Congress ordinarily uses to enact legislation”). means that, as a practical matter, the CRA is only likely to be used to disapprove a rule in certain moments linked to a new presidential term. That is, disapproval of a rule under the CRA is most feasible immediately following an election where the White House changes parties and the incoming President is politically aligned with both chambers of Congress.41See Note, The Mysteries of the Congressional Review Act, 122 Harv. L. Rev. 2162, 2169 (2009) (illustrating the importance of the CRA during transition years when the presidential administration changes parties by detailing how the Bush Administration used the Act). Otherwise, a President would likely veto any congressional resolution to disapprove a rule from his own administration or that of a prior President from his party.42Maeve P. Carey & Christopher M. Davis, Cong. Res. Serv., R43992, The Congressional Review Act (CRA): Frequently Asked Questions 1 (2020). This assumes general political alignment between an outgoing President of one party and an incoming President of the same party. It is conceivable, though, that on any particular regulatory issue, an incoming President might disagree enough to disapprove a rule of a prior President from his same party. This might not hold for independent agencies, which would not necessarily be expected to be in complete alignment with the President at any given time. Indeed, Congress passed and President Trump signed two disapproval resolutions for Consumer Financial Protection Bureau (CFPB) actions, even though Democrats controlled the House while Republicans controlled the Senate. S.J. Res. 57, 115th Cong. (2018) (enacted); H.R.J. Res. 111, 115th Cong. (2018) (enacted). At the time, the Supreme Court had not yet ruled in Seila Law v. Consumer Financial Protection Bureau, 140 S. Ct. 2183 (2020); while the agency’s status now is somewhat ambiguous, at the time of the CRA disapproval, the agency was understood to be independent. See id. at 2207–08 (holding that the structure of the CFPB violated the separation of powers principle and that its Director must be removable by the President at will, therefore the CFPB cannot exist as an entirely independent agency). In addition, both chambers of Congress must pass the resolution before it may go to the President, so if either chamber is out of political alignment with the White House, the disapproval is unlikely to make it to the President.43The exception here, of course, is if members cross the aisle to vote for disapproval of a rule. This happened recently in response to the U.S. Department of Education’s borrower defense rule. Student Assistance General Provisions, Federal Family Education Loan Program, and William D. Ford Federal Direct Loan Program, 84 Fed. Reg. 49,788 (Sept. 23, 2019) (to be codified at 34 C.F.R. pts. 668, 682, 685). In that case, a CRA resolution of disapproval passed out of the Democrat-controlled House and the Republican-controlled Senate, only to be vetoed by the Republican President. H.R.J. Res. 76, 116th Congress (2019); S.J. Res. 56, 116th Congress (2019).

Table 2 shows the party affiliation of the President alongside those of the majorities in the House and the Senate for each presidential term beginning with the 1997 inauguration. The year refers to the presidential term beginning January 20 of that year. Because a President is unlikely to disapprove actions of his own administration, midterm changes in Congress are not displayed below.

Table 2. Political Alignment Conducive to CRA Disapproval44Party Division, U.S. Senate, https://www.senate.gov/history/partydiv.htm [https://perma.cc/QLS3-VFSL]; Party Divisions of the House of Representatives, 1789 to Present, U.S. House of Representatives: History, Art & Archives, https://‌history.house.gov/Institution/Party-Divisions/Party-Divisions/ [https://perma.cc/7883-NWWJ].

Presidential Inauguration Year President House Majority Senate Majority
1997 Democrat Republican Republican
2001 Republican Republican Republican45Republicans had the majority from Jan. 20, 2001 until June 6, 2001. The Senate seats were split evenly among the parties from Jan. 20, 2001 until June 6, 2001, providing Republican Vice President Richard Cheney the deciding vote and the Republicans the majority.
2005 Republican Republican Republican
2009 Democrat Democrat Democrat
2013 Democrat Republican Democrat
2017 Republican Republican Republican
202146At publication time, the Senate majority is unknown because two Georgia runoff elections will be held in January 2021. Becky Bohrer, Republican Dan Sullivan Re-Elected in Alaska Senate Race, AP News (Nov. 11, 2020), https://apnews.com/article/
Democrat Democrat Unknown


Since the enactment of the CRA, the White House, Senate, and House have politically aligned in a manner conducive to CRA disapproval resolutions in three instances: the beginnings of President George W. Bush’s term in 2001, President Barack H. Obama’s term in 2009, and President Donald J. Trump’s term in 2017.47See supra note 44 and accompanying text. These are shown in bold in Table 2. At those times, the party affiliation of the newly elected President changed and was aligned with that of the new Congress starting that same year.48See supra note 44 and accompanying text.

As such, while there was some CRA disapproval activity following the CRA’s enactment in 1996, including introductions of resolutions and votes in one or both chambers,49Sam Batkins, Congress Strikes Back: The Institutionalization of the Congressional Review Act, 45 Mitchell Hamline L. Rev. 351, 362–64 (2019). the first CRA disapproval did not become law until the early days of the George W. Bush administration in 2001. That disapproval was for a rule issued late in the Clinton Administration that set workplace ergonomic standards.50Pub. L. No. 107-5, 115 Stat. 7 (2001) (disapproving a Department of Labor rule regarding ergonomics); see also Batkins, supra note 49, at 366–67 (discussing the legislative history of the ergonomics disapproval).

Democrats did not take advantage of the next window of CRA disapproval that occurred at the beginning of President Obama’s term. As Bethany A. Davis Noll and Richard L. Revesz explain, “[R]ather than resort to using the Congressional Review Act, the Obama administration focused its early attention on filling cabinet and subcabinet positions and used regular rulemaking procedures, rather than the Congressional Review Act, to overturn at least some of the targeted rules.”51Bethany A. Davis Noll & Richard L. Revesz, Regulation in Transition, 104 Minn. L. Rev. 1, 18–19 (2019); see also Adam M. Finkel & Jason W. Sullivan, A Cost-Benefit Interpretation of the “Substantially Similar” Hurdle in the Congressional Review Act: Can OSHA Ever Utter the E-Word (Ergonomics) Again?, 63 Admin. L. Rev. 707, 728–29 (2011) (explaining that after considering CRA disapprovals, “in the end, the Obama Administration used executive procedures” instead of the CRA disapproval process).

Another factor that may have depressed Democrats’ use of the CRA to disapprove rules is that, once a rule is disapproved, an agency may not issue another rule that is in “substantially the same form” unless Congress affirmatively permits an agency to do so in subsequent legislation.525 U.S.C. § 801(b)(2) (2018). This language is undefined in the CRA and the issue has not yet been litigated, though scholars and others have proposed various methods of applying the standard.53See, e.g., Michael J. Cole, Interpreting the Congressional Review Act: Why the Courts Should Assert Judicial Review, Narrowly Construe “Substantially the Same,” and Decline to Defer to Agencies Under Chevron, 70 Admin. L. Rev. 53, 83–88, 91–100 (2018) (arguing that courts should interpret the CRA’s “substantially the same” requirement through a cost-benefit analysis and that Chevron deference should not be applied); Finkel & Sullivan, supra note 51, at 734–37 (examining seven possible interpretations of “substantially the same”); Stephen Santulli, Use of the Congressional Review Act at the Start of the Trump Administration: A Study of Two Vetoes, 86 Geo. Wash. L. Rev. 1373, 1389–91 (2018) (hypothesizing that courts will apply an “anticircumvention” framework to interpreting the “substantially the same” prohibition to ensure that the new regulation is not intended to thwart the purpose of CRA disapprovals). Meanwhile, the legal uncertainty around how a court would apply this prohibition places a chill on downstream regulatory activity,54Finkel & Sullivan, supra note 51, at 730–31 (noting that, following Congress’s joint resolution disapproving a Department of Labor ergonomics rule, the Secretary of Labor commented that “[s]he did not want to ‘expend valuable—and limited—resources on a new effort’ if another regulation would be invalidated as substantially similar”).[/mfn ] and therefore may deter regulation-minded Administrations from using the CRA to disapprove a rule in the first place. Regulation is an important tool for modern presidents because it offers the opportunity for unilateral action. Therefore, it is not surprising that a President might prefer not to limit his regulatory discretion unnecessarily by walling off “substantially the same” actions to those disapproved.

In all, since its enactment in 1996, Congress used the CRA to disapprove a rule only once prior to President Trump’s inauguration.54Brad Plumer, Republicans Will Try a Little-Used Tactic to Kill Five Obama Regulations This Week, Vox, (Feb. 1, 2017, 10:50 AM) https://www.vox.com/energy-and-environment/2017/2/1/14459630/congressional-review-act-obama-regs. From that point on, however, CRA disapprovals have sharply increased, resulting in sixteen disapproved rules.55CRA Tracker, GW Regulatory Studies Center, https://‌regulatorystudies.columbian.gwu.edu/sites/g/files/zaxdzs3306/f/downloads/CRA/GW%20Reg%20Studies_CRATracker_03.12.20.pdf [https://perma.cc/2JVF-GMA5]. President Trump’s campaign promise of deregulation partially explains the increase; the Trump Administration has used CRA disapprovals as one of several strategies to achieve deregulation.56Keith B. Belton & John D. Graham, Trump’s Deregulation Record: Is It Working?, 71 Admin. L. Rev. 803, 814–15 (2019) (commenting on the Trump Administration’s efforts towards deregulation); Noll & Revesz, supra note 51, at 2–3 (noting that Trump had made “aggressive use” of the CRA).

A more subtle explanation lies in a procedural aspect of the CRA in which the GAO plays an important role. Congress is only likely to use the CRA to disapprove a rule when the political stars align, as discussed above. A critical question is how far back, during those times of alignment, Congress can reach using the CRA. The conventional wisdom—to the extent that there is such a thing about a statute as arcane as the CRA—is that a rule is vulnerable to disapproval for a set and relatively short period of time that is determined using a technical but fairly straightforward calculation, and which begins shortly after the agency issues the rule.57See Maeve P. Carey & Christopher M. Davis, Cong. Res. Serv., IF10023, In Focus: The Congressional Review Act (2018) (explaining the calculation). As discussed below, Congress has used the GAO’s legal opinions to expose rules to vulnerability even years after they were issued.

The full results of the 2020 election were not yet available when this Article went to print. President Trump’s defeat by Vice President Joseph R. Biden satisfies a key condition for use of CRA disapproval resolutions: a change in party affiliation in the White House. If Democrats obtain a majority the Senate—having retained their majority in the House—then the CRA might well be used to disapprove rules again. This would depend on Democrats’ willingness to risk courts’ application of the “substantially the same” test, which could inhibit future rules, but such willingness is plausible at least for a subset of the Trump Administration’s regulations. If the Democrats do not obtain a majority in the Senate, bipartisan cooperation would be required to disapprove a presumably even smaller set of Trump Administration rules. Using legal opinions from the GAO, Congress might be able to disapprove certain rules issued at any point in the Trump Administration.

III. The GAO’s CRA Legal Opinions

As noted above, the GAO plays a role in Congressional oversight of rules issued by federal agencies. This Part provides background on the GAO’s two activities related to the CRA: reporting on major rules and providing legal opinions. The first is written into the law, while the second is something that the GAO does at legislators’ requests but arguably also of its own accord. First, the CRA requires the GAO to report on whether major agency rules have complied with the CRA’s procedural requirements.585 U.S.C. § 801(a)(2)(A) (2018). These assessments, which the GAO sends to relevant congressional committees, tend to be straightforward recitations of what the agencies did or did not do in making their rules.59See, e.g., Letter from Shirley A. Jones, Managing Assoc. Gen. Counsel, Gov’t Accountability Office, to the chairs and ranking members of the U.S. Senate Comm. on Small Bus. & Entrepreneurship and the U.S. House of Representatives Comm. on Small Bus. (June 26, 2020), https://www.gao.gov/assets/710/708776.pdf [https://‌perma.cc/L2N5-E9LC] (providing an assessment of the Small Business Administration and Department of the Treasury’s compliance with the procedural steps required by the CRA). The GAO churns these out with regularity.60Congressional Review Act: Reports on Major Rules, U.S. Gov’t Accountability Office, https://www.gao.gov/legal/other-legal-work/congressional-review-act#reports [https://perma.cc/JC5J-8WKY]. What Congress does with these reports upon receipt is not well-documented, but the reports may sometimes lead to oversight activity.

Second, members of Congress have, from time to time, asked the GAO to weigh in on whether an action is a “rule” under the CRA.61Maeve P. Carey & Christopher M. Davis, Cong. Res. Serv., R43992, The Congressional Review Act (CRA): Frequently Asked Questions 11–12 (2020). If the GAO issues an affirmative opinion, that opinion can be used by legislators to start the clock on the CRA window within which legislators may introduce a resolution of disapproval.62Congressional Research Service analysts Maeve Carey and Christopher Davis explain the mechanics: “To avail themselves of the CRA’s disapproval mechanism following such an opinion, Senators have published the GAO opinion in the Congressional Record. It appears that, in these cases, the Senate has considered the date of publication of the GAO opinion in the Congressional Record to be the beginning of the periods for congressional review. Normally, when agencies submit their rules to Congress under the CRA, a record of each rule’s receipt is published in the Congressional Record. The publication of the GAO opinion in the Congressional Record fulfills this same purpose: notifying Congress that a rule is now available for review under the CRA.” Id. at 12 (footnotes omitted).

Congress can only use the CRA’s fast-track procedures described above to disapprove “rules,” while actions that are not “rules” are outside of Congress’s power to disapprove under the CRA.63See 5 U.S.C. § 804(3). The definition of a “rule” is technical and encompasses more than those rules promulgated using notice-and-comment rulemaking under the APA.64See generally Valerie C. Brannon & Maeve P. Carey, Cong. Res. Serv., R45248, The Congressional Review Act: Determining Which “Rules” Must Be Submitted to Congress 1 (2019) (noting some agency actions that do not fall within the APA’s notice-and-comment requirements may still be considered “rules” under the CRA and therefore subject to the CRA’s disapproval procedures). The definition, in short, hinges more on what a document does than on what it is called, so it is sometimes unclear whether an agency action is a “rule” or not.65Maeve P. Carey & Christopher M. Davis, Cong. Res. Serv., R43992, The Congressional Review Act (CRA): Frequently Asked Questions 7 (2020) (noting that whether the CRA covers a particular agency action may depend on the specific nature of the action and its effect). Perhaps partly due to this ambiguity, agencies sometimes fail to meet their legal duty under the CRA to send all of their rules to Congress.66See Curtis W. Copeland, Hundreds of Recent Final Rules Are Technically Unlawful, Reg. Rev. (Sept. 15, 2014), https://www.theregreview.org/2014/09/15/15-copeland-final-rules-unlawful [https://perma.cc/B3RW-FW2S].

The CRA does not direct the GAO to resolve this ambiguity by giving legal opinions, nor does it vest any other particular body with the express authority to determine what counts as a “rule” under the CRA. Instead, and as explained more below, legislators requested the GAO’s legal opinions, and the GAO obliged, stepping into the void and setting in motion a new responsibility for the GAO that would expand over time.

Given the GAO’s other legal work to prepare legal opinions and decisions on federal bid protests, appropriations law, and other matters, legislators made a reasonable choice.67The GAO’s website describes its legal work. About GAO: Legal Decisions, U.S. Gov’t Accountability Office, https://www.gao.gov/about/what-gao-does/legal-decisions [https://perma.cc/9DD6-JJ24]. This is particularly the case because of the GAO’s reputation as a reasonably independent, expert fact finder and arbiter. The GAO’s fulfillment of this function, though, does not imbue its opinions with legal weight. As analysts at the Congressional Research Service have noted, “[s]tanding alone, a GAO opinion deciding whether an agency action is a ‘rule’ covered by the CRA does not have legal effect.”68Valerie C. Brannon & Maeve P. Carey, Cong. Res. Serv., R45248, The Congressional Review Act: Determining Which “Rules” Must Be Submitted to Congress 24 (2019). This is somewhat akin to the GAO’s other work. “[L]ike Inspector General offices, the GAO has no direct authority to make anything happen. Its reports may or may not result in the changes it recommends. Its effect depends on its persuasive authority, which may vary significantly depending on the political and institutional context.” Eloise Pasachoff, Federal Grant Rules and Realities in the Intergovernmental Administrative State: Compliance, Performance, and Politics, 37 Yale J. on Reg. 573, 598 (2020).

Legislators have sought the GAO’s legal opinions under the CRA since shortly after it became law, suggesting that legislators attribute weight to those opinions.69See infra Chart 1 And, another congressional institution gave the GAO’s CRA legal opinions a boost in terms of their practical weight. The Senate Parliamentarian “has considered the date of publication of the GAO opinion in the Congressional Record to be the beginning of the periods for congressional review.”70Maeve P. Carey & Christopher M. Davis, Cong. Res. Serv., R43992, The Congressional Review Act (CRA): Frequently Asked Questions 12 (2020). This means that legislators can use GAO legal opinions to make a rule vulnerable to disapproval under the CRA.71Of course, this would only work if the agency failed to submit its rule to Congress as required by the CRA. As the GAO explained in its legal opinion about an Internal Revenue Service (IRS) notice, the IRS decision to submit its notice to the GAO constituted an agency determination that the notice was a rule, thereby “obviat[ing]” the need for the GAO to opine. Letter from Thomas H. Armstrong, Gen. Counsel, U.S. Gov’t Accountability Office, to Charles E. Schumer, Senator, U.S. Senate, and Ron Wyden, Senator, U.S. Senate Comm. (Sept. 15, 2020), https://www.gao.gov/assets‌/710/709404.pdf [https://perma.cc/HQZ2-DUA8].

There was a time when the executive branch was cool to the idea of the GAO offering its legal opinion on this issue. Testifying before a joint hearing of Congress in 1997, Sally Katzen, then Administrator of the Office of Information of Regulatory Affairs, explained that “it is the agency promulgating the regulation that has the responsibility for determining whether a particular issuance is a ‘rule’ under the [CRA].”72Tongass Land Management: Joint Hearings Before the S. Comm. on Energy & Nat. Res. and H. Comm. on Res., 105th Cong. 19 (1997) (statement of Sally Katzen, Administrator, Office of Information and Regulatory Affairs, Office of Management and Budget). Arguments over which branch may define key terms are not unique to the CRA context and are probably—in the words of James Madison—indicators of ambition counteracting ambition.73See The Federalist No. 51 at 264 (James Madison) (Ian Shapiro ed., 2009) (explaining that a national government premised upon a system of check and balances must resist “encroachment[]” on a department’s power with “commensurate” action). They do not, however, settle the matter.

At that same hearing—which took place one week after the GAO issued one of its first CRA legal opinions finding that the United States Forest Service Tongass National Forest Land and Resource Management Plan was a “rule” under the CRA74Letter from Robert P. Murphy, Gen. Counsel, U.S. Gen. Accounting Office, to Ted Stevens, Chairman, U.S. Senate Comm. on Appropriations, Frank H. Murkowski, Chairman, U.S. Senate Comm. on Energy & Nat. Res., and Don Young, Chairman, U.S. House of Representatives Comm. on Res. (July 3, 1997), https://www.gao.gov‌/assets/170/168383.pdf [https://perma.cc/34ZW-85HF].—Robert Murphy, the GAO’s General Counsel, said:

I cannot say that [the] GAO has a special role under the statute for making that determination. The decision, the opinion, that we issued last week on the question was done in our role as adviser to the Congress in response to the request of three chairmen of congressional committees.75Tongass Land Management, supra note 63, at 20 (statement of Robert P. Murphy, General Counsel, General Accounting Office).

This suggests that the GAO found itself obliged to opine upon being asked. Since that hearing, legislators have continued to request the GAO’s legal opinion, and the GAO has continued to give it.76See infra Chart 1.

More recently, the Office of Management and Budget’s (OMB) General Counsel dismissed the weight of the GAO’s legal opinions, writing that “[w]hen an agency of the Legislative Branch interprets a law differently than the Executive Branch, the Executive Branch is not bound by its views.”77Memorandum from Mark Paoletta, Gen. Counsel, Office of Mgmt. & Budget, to Agency Gen. Counsels (Nov. 5, 2019), https://www.whitehouse.gov/wp-content/uploads/2019/11/Memo-to-Agencies-on-A-11.pdf [https://perma.cc/ZKE6-L2EF]. The timing of this memorandum suggests it is responsive to issues outside of the CRA. Shortly after this memo was issued, the GAO issued two determinations critical of the OMB’s actions. The first rebuked the Office of Information and Regulatory Affairs for working on certain rules during the partial shutdown from December 22, 2018 through January 25, 2019 in violation of the Antideficiency Act. Letter from Thomas H. Armstrong, Gen. Counsel, Gov’t Accountability Office, to Robert C. Scott, Chairman, U.S. House of Representatives Comm. on Educ. & Labor, and Andy Levin, Vice Chairman, U.S. House of Representatives, Comm. on Educ. & Labor (Dec. 19, 2019), https://www.gao.gov/assets/710/703496.pdf [https://perma.cc/65XR-H54W]. The second found that OMB violated the Impoundment Control Act when it withheld from obligation certain funds for security assistance to Ukraine. U.S. Gov’t Accountability Office, B-331564, Decision in the Matter of Office of Management and Budget—Withholding of Ukraine Security Assistance (Jan. 16, 2020). Both determinations note that the OMB was aware of the GAO’s inquiries into these matters. And yet, when President Trump signed disapproval resolutions into law that Congress introduced based on the GAO legal opinions that started the clock on congressional reviews, he arguably acquiesced to the GAO’s role in this process.78E.g., S.J. Res. 57, 115th Cong. (2018) (enacted) (disapproving a 2013 CFPB bulletin from 2013); Consumer Fin. Prot. Bureau, CFPB Bull. No. 2013-02, Indirect Auto Lending and Compliance with the Equal Credit Opportunity Act (2013). Of course, it’s also possible that the President was relying on analysis from attorneys in the executive branch, independent of the GAO’s legal opinion. While signing a disapproval resolution under the CRA is not the same as accepting that the GAO’s legal opinions are binding upon the executive, it may make it harder, rhetorically, for future presidents to push back on a disapproval resolution by arguing that it is improper to use a GAO legal opinion to start the CRA clock. Ultimately, however, the Constitution gives Congress authority over its own procedures.79U.S. Const. art. I, § 5, cl. 2. If Congress wants to rely on the GAO’s legal opinions, that choice probably cannot be successfully challenged. Perhaps more importantly, it may also make it harder for the executive branch to shrug off GAO legal opinions in other domains such as the Antideficiency Act.

A. The GAO’s Record of CRA Legal Opinions

The GAO received its first request for a CRA legal opinion just five months after President Clinton signed the CRA into law in March 1996.80Letter from Robert P. Murphy, Gen. Counsel, U.S. Gen. Accounting Office, to Larry E. Craig, Chairman, U.S. Senate Subcomm. on Forests and Pub. Land Mgmt. of the Comm. on Energy & Res. (Sept. 16, 1996) (noting the date of the Senator’s request was August 27, 1996), https://www.gao.gov/assets/370/365493.pdf [https://perma.cc/QN5B-S8H9]. Senator Larry Craig wrote to the GAO asking for its opinion about whether a memorandum from the Secretary of Agriculture was a “rule” under the CRA.81Id. at 1. The GAO answered yes, reasoning that the memorandum was a “statement of general . . . applicability and future effect”—a term of art used in the definition of a rule—and not subject to any exemptions.82Id. at 6–10 (alteration in original).

Since that first inquiry in 1996, the GAO has offered twenty-eight more legal opinions about whether an agency action was a “rule” in response to inquiries from members of Congress. This is a relatively small number of inquiries given the thousands of actions that likely qualify as rules every year. Across these twenty-nine opinions, the GAO opined that the action at issue was a “rule” sixteen times (fifty-five percent of the time), giving reasons for its opinions in detailed letters that are readily available on the GAO’s website.83Congressional Review Act: Legal Opinions, U.S. Gov’t Accountability Office, https://www.gao.gov/legal/other-legal-work/congressional-review-act#legal_opinions [https://perma.cc/2A6Y-GN2M]. The GAO sometimes opines about multiple actions in one letter. For purposes of counting legal opinions, this Article counts the GAO’s legal opinion about each action as a separate legal opinion, even if it was communicated in a single letter with other legal opinions.

This fairly even split might be because most of the members’ inquiries have been about agency letters, memos, fact sheets, and other documents that can broadly be described as guidance rather than the “rules” that come out of the notice-and-comment rulemaking process under the APA.84One exception was when the GAO was asked about the status of a notice of proposed rulemaking, which is the document that goes out for public comment under the notice-and-comment procedures of the APA; the GAO responded that the proposed rule was not a “rule” under the CRA because it was not yet final. Letter from Susan A. Poling, Gen. Counsel, U.S. Gov’t Accountability Office, to Harry Reid, Majority Leader, U.S. Senate, Mitch McConnell, Minority Leader, U.S. Senate, Barbara Boxer, Chairman, U.S. Senate Comm. on Env’t & Pub. Works, and Thomas Carper, Chairman, U.S. Senate Comm. on Homeland Sec. & Governmental Affairs (May 29, 2014), https://www.gao.gov‌/assets/670/663690.pdf [https://perma.cc/GE8V-HMT5]. In these close cases, the ambiguity of a “rule” is especially present. GAO legal opinions, to date, provide a decisive answer in the face of such ambiguity, unless the GAO cites a procedural reason not to opine.85See Letter from Thomas H. Armstrong to members of Congress, supra note 72.

B. Increased Frequency of Opinions

The frequency of the GAO’s legal opinions offers some insight into how Congress uses its ability to request opinions. The GAO issued between zero and two legal opinions in response to inquiries from members of Congress every year until the frequency of inquiries increased significantly in 2017, as shown in Chart 1.

Chart 1: CRA Legal Opinions 1996–202086U.S. Gov’t Accountability Office, supra note 84 (last visited Dec. 17, 2020).

As noted above, prior to 2017, Congress disapproved only one rule under the CRA. The increase in legal opinions in 2017 and beyond is consistent with increased use of the CRA to disapprove rules in the last few years. It also coincides with a period of advocacy arguing that legislators could use the CRA to reach back farther in time than was previously understood to be possible.87See, e.g., Kimberley A. Strassel, Opinion, A GOP Regulatory Game Changer, Wall St. J. (Jan. 27, 2017), https://www.wsj.com/articles/a-gop-regulatory-game-changer-1485478085 (stating that Todd Gaziano appeared before members of Congress to explain that the CRA “grants them far greater powers”); Jonathan Wood & Todd Gaziano, Three Cheers for the Congressional Review Act, Nat’l Rev. (June 29, 2017, 8:00 AM), https://www.nationalreview.com/2017/06/congressional-review-act-finally-some-accountability-washington (stating that the review period has yet to begin for many rules that agencies have not submitted to Congress); Red Tape Rollback, www.redtaperollback.com [https://perma.cc/NG9L-UQKR] (arguing that the CRA “requires agencies to submit every rule to Congress”).

C. Increased Time Between Agency Actions and Legislator Requests

Another way to consider significance of the GAO’s legal opinions is to look at when they are requested. Table 3 shows the number of days between when an agency action was issued and when a legislator requested the GAO’s legal opinion. This span of time is important because Congress limited the reach of the CRA to a set period of time—and the longer the rule is vulnerable, the more it strains the idea that the CRA provides Congress with limited time to review the rule.

Table 3: Days Between Agency Action and Legislator Request88This excludes two actions where the request date is unknown. Data on file with author.

Minimum Maximum Average
1996–2016 8 205 89
2017–2020 9 2944 842


Table 3 shows a significant increase in the maximum and average number of days. Legal opinions before 2017 were requested within the timeframe for the conventional understanding of the CRA window: an average of 89 calendar days. Starting in 2017, the number of days elapsed between the action and the opinion grew almost tenfold to an average of 842 days. This increase was driven by requests for opinions up to 2944 days after an agency action.

IV. Implications

The legal opinions that the GAO issues to help Congress administer the CRA provide a view into Congressional oversight of rulemaking, but they also have implications far beyond regulatory policy. This Part offers several possibilities for future study along three main dimensions. First, it is plain that the congressional agencies, as a group, deserve much more attention than they have received in the academic literature to date. The GAO’s choices in how and whether to issue these opinions provide material for a rich case study in agency incentives and bureaucratic decision-making. Second, a diverse set of factors could motivate legislators to request a legal opinion from the GAO and could help scholars understand legislator decision-making and this political moment better. Third, with respect to the CRA itself, the use of the GAO’s legal opinions to make certain executive actions vulnerable to disapproval deserves a closer look to assess whether this trend aligns with the compromise that the CRA text reflects.

A. Congressional Agencies in the Political Environment

The biggest question provoked by the GAO’s CRA legal opinions has nothing to do with the CRA. Congressional agencies are numerous, and they operate within a very political environment. Research suggests that, at least on the investigation side, the GAO is not a passive recipient of Congressional inquiries. To what extent do congressional agencies have their own incentives, set their own agendas and priorities, and otherwise make their own choices? What guides and governs the actions of congressional agencies? The APA does not apply to these agencies, and therefore a host of other “good government” statutes do not apply to them either—but that does not mean that these agencies operate in a lawless manner. Instead, their behavior suggests that a complex set of incentives and norms evolve to guide their actions where statutes do not expressly reach. This might be an effective replacement for the APA in congressional agencies, given their unique missions, or it might not.89Along these lines, it could be useful to explore whether the normative justifications for the APA and other administrative law statutes are relevant for congressional agencies. In rulemaking, for example, an agency action can bind members of the public, while the GAO’s CRA legal opinions are used to shape congressional procedure. However, to the extent that congressional agencies take actions that are more like agency rulemaking, the normative justifications may be present. The answer could inform our understanding of the importance of the APA in general; a natural experiment of sorts.

With respect to the GAO’s incentives, Professor Anne Joseph O’Connell’s dissertation cracked open an enticing avenue for further study.90Anne Margaret Joseph, Political Appointees and Auditors of Politics: Essays on Oversight of the American Bureaucracy (May 2002) (unpublished Ph.D. dissertation, Harvard University) (on file with ProQuest). Two of her three essays focus on the GAO.91Id. at iii–iv. Both consider how the GAO’s institutional incentives differ from those of the legislators requesting investigations from the GAO.92See id. at 4–9. The gap between these incentives can lead to strategic behavior both on the part of the legislator—for example, requesting investigations to advance the legislator’s substantive knowledge or perhaps “to build their chances for reelection”93Id. at 187.—and with respect to the GAO.94See id. at 97–99, 178, 182–84. The GAO, like all federal agencies, has limited resources and must prioritize which investigations to conduct and in what manner. How does the GAO prioritize within its CRA legal opinion work, its other legal opinion work, and other, non-opinion-related work?

With respect to norms, congressional agencies emphasize the non-partisan, objective nature of their work.95The websites of the GAO, the Congressional Budget Office (CBO), and the Congressional Research Service offer ready examples. “[GAO] is an independent, nonpartisan agency that works for Congress.” About GAO, U.S. Gov’t Accountability Office, https://www.gao.gov/about [https://perma.cc/AL5K-L24Y]. “CBO’s analysis is objective, impartial, and nonpartisan.” 10 Things to Know about CBO, Cong. Budget Office, https://www.cbo.gov/about/10-things-to-know [https://perma.cc/JUT7-GR5V]. “CRS analysis is confidential, authoritative, objective and nonpartisan.” About CRS: Values, Cong. Res. Serv., https://www.loc.gov/crsinfo/about/values.html [https://perma.cc/GDG8-N3HJ]; see also Ginsberg & Hill, supra note 11, at 144–48. See generally Philip G. Joyce, The Congressional Budget Office: Honest Numbers, Power, and Policymaking 7 (2011) (describing efforts to solidify the reputation of the Congressional Budget Office as credible and non-partisan). By statute, the Congressional Budget Office must be led by a Director appointed “without regard to political affiliation and solely on the basis of his fitness to perform his duties” and staffed by personnel selected on those same criteria.962 U.S.C. § 601(a)(2), (b) (2018). The Comptroller General may serve for a fifteen-year term, much longer than that of a two-term president.9731 U.S.C. §  703(b) (2018).

Provisions like this are a source of legitimacy for these agencies, though they do not completely insulate the agencies from political attack.98E.g., Rebecca Shabad, Report: Elmendorf out as CBO director, Hill (Dec. 22, 2014, 6:03 PM), https://thehill.com/policy/finance/227920-elmendorf-out-at-cbo-report-says [https://perma.cc/35T7-B7YU] (asserting that Republican leaders wanted to replace the CBO head with a candidate from their own party); Erik Wasson, House Budget Chairman Backs Embattled CBO Head, Hill (Jan. 6, 2011, 11:57 AM) (highlighting the fierce political debates surrounding the head of the CBO), https://thehill.com‌/policy/finance/136311-ryan-conrad-stand-up-for-cbo-head [https://perma.cc/DQ7R-34XV]; see Ginsberg & Hill, supra note 11, at 145–47. Quite recently, in response to a GAO report on the government’s coronavirus response, White House advisor Peter Navarro said, “It is well-known that the GAO is neither non-partisan or [sic] independent but simply a useful idiot for the Democrat Party and Big Labor.”99Alexander Nazaryan, White House Dismisses Criticism of Persistent Coronavirus Supply Chain Problems as Work of ‘Useful Idiot,’ Yahoo! News (Sept. 22, 2020), https://news.yahoo.com/white-house-dismisses-criticism-of-persistent-coronavirus-supply-chain-problems-as-work-of-useful-idiot-190104765.html [https://perma.cc/86S5-8LMD]. These periodic attacks create incentives for congressional agencies to seek to buffer themselves from the political environment in which they operate. While some of these norms might be invisible, like internal staff norms transmitted through the workplace culture,100See Kevin R. Kosar, Why I Quit the Congressional Research Service, Wash. Monthly (Jan./Feb. 2015), https://washingtonmonthly.com/magazine/janfeb-2015/why-i-quit-the-congressional-research-service [https://perma.cc/NQ2T-A35B] (describing a cultural shift among CRS analysts as their work shifted from conducting proactive research to producing reactive responses, staffing diminished, and workloads increased). the GAO’s legal opinions offer an observable set of actions that one can observe and theorize.

For example, the GAO has elected to make its legal opinions public, and not just those that opine that an action is a rule.101“Once a decision or opinion is issued to the requestor, GAO will publicly release the decision or opinion and post a copy on our Web site.” U.S. Gov’t Accountability Office, GAO-06-1064SP, Procedures and Practices for Legal Decisions and Opinions 9 (2006). This surely serves a purpose to the public as a whole. By making the legal opinions public, an agency will have a better sense of which actions might be “rules,” the public can better understand what to expect in terms of oversight, and legislators and their staffs have access to the opinions to inform themselves. But this choice may serve the GAO itself, as well, particularly in combination with how it writes the legal opinions. The GAO has elected to compose the opinions with a wealth of information about the requestor, the matter under consideration, and the detailed legal reasoning that informed its opinion.102See e.g., Letter from Susan A. Poling to members of Congress, supra note 85 (responding to proposed rules submitted by the Environmental Protection Agency). This reason-giving, somewhat akin to the reason-giving that an agency provides in its proposed and final rules103See, e.g., Jerry L. Mashaw, Reasoned Administration: The European Union, the United States, and the Project of Democratic Governance, 76 Geo. Wash. L. Rev. 99 (2007) (arguing that reason-giving in administration is requisite for democratic legitimacy). or that a judge provides in his opinions,104See, e.g., Robin J. Effron, Reason Giving and Rule Making in Procedural Law, 65 Ala. L. Rev. 683 (2014) (noting that judges are administrators of the procedural rules that govern cases and arguing that rule-makers should employ the technique of reason-giving to regulate select procedural devices). could similarly be a way to bolster the legitimacy of the GAO’s actions. One empirical question is whether, as the political importance of the GAO legal opinions has increased, the sophistication of the legal opinions has increased with it. The GAO also documents, in its legal opinions, that it contacted the agency that issued the action to obtain that agency’s views.105See generally U.S. Gov’t Accountability Office, supra note 84 (compiling the legal opinions of the GAO). This looks to be the result of an earlier bone of contention between the GAO and the executive branch.106Writing about GAO investigations, Professor O’Connell wrote that “[u]ntil several years ago, members of Congress sprung [GAO] reports on administrative agency officials in hearings, creating ill will. The GAO’s current policy is to seek comments from a targeted agency and include them in its reports.” Joseph, supra note 91, at 184 (footnote omitted). An interesting contrast is the Congressional Research Service, which tends to work on a confidential basis responding to legislator inquiries, but which has been criticized for not making its reports public. Tajha Chappellet-Lanier, Transparency Advocates Push Congress on Public Access to CRS Reports, FedScoop (Feb. 4, 2020), https://www.fedscoop.com/transparency-advocates-push-congress-public-access-crs-reports [https://perma.cc/H2XD-XBL7]. How else do agencies in the executive and legislative branches play off of each other?

There is even another layer of norms: the GAO’s willingness to offer legal opinions in the CRA context. Presumably, the GAO could have declined to fulfill this role; the statute does not require it.107See supra note 36 and accompanying text. Further, when the GAO stepped into the void, it might not have anticipated that the significance of its CRA legal opinions would increase over time. As the GAO considers its own role, might the use of its opinions to reach back years, beyond the period immediately following a rule’s issuance, combined with the politically-charged domain of regulatory policy, encourage it to exit the role to protect its reputation and credibility? Or is the role so institutionalized within the GAO that this is unthinkable?

Turning to the other congressional agencies, some of the lesser-known institutions of Congress shape policy in meaningful ways. For some, that is their direct mission,108One example is the Medicare Payment Advisory Commission (MedPAC): “The Commission’s statutory mandate is quite broad: In addition to advising the Congress on payments to private health plans participating in Medicare and providers in Medicare’s traditional fee-for-service program, MedPAC is also tasked with analyzing access to care, quality of care, and other issues affecting Medicare.” MedPAC, supra note 3. and some exist to “provide the legislative branch with information and a capacity for action.”109Ginsberg & Hill, supra note 11, at 144. Further, other individuals in non-policy roles serve alongside policymakers in such a way as to influence policy in more subtle ways.110See, e.g., History of the Library of Congress, Library of Cong., https://www.loc.gov‌/about/history-of-the-library [https://perma.cc/HAN9-7KZX] (discussing various institution-building efforts of Librarians of Congress). We can accept as a given that congressional agencies have their own institutional interests, distinct from the legislators that they serve.111See, e.g., Daniel Hawkings, Capitol Police: A Spending Force, Roll Call (Apr. 7, 2016, 6:00 AM), https://rollcall.com/2016/04/07/capitol-police-a-spending-force [https://perma.cc/24N7-5EEW] (reporting on the exponential growth of the U.S. Capitol Police with limited congressional oversight and budgetary restraint). Are there tensions, though, between the actions and processes of these lesser-studied agencies and the expert, non-partisan roles we ask them to play, and which they defend for themselves?

Beyond the dynamics operating within any given congressional agency, how do the various congressional agencies interact with each other? In the CRA legal opinion context, the Senate Parliamentarian greatly increased the relevance of GAO legal opinions by considering their publication in the Congressional Record as the start of the CRA review period for Congress.112See supra note 71 and accompanying text. This was a pragmatic workaround to the problem of an agency failing to give Congress notice of a rule. But this decision has allowed agency actions to be vulnerable to CRA disapproval years after they were issued, which has significant political consequences. Are there other instances of the combined efforts of congressional agencies coming together to lead to significant political effects? What can we learn from them?

All of these considerations dovetail with the emerging literature on internal administrative law.113See, e.g., Gillian E. Metzger & Kevin M. Stack, Internal Administrative Law, 115 Mich. L. Rev. 1239 (2017) (offering a conceptual and historical account of internal administrative agency norms and structures). That literature has not yet crossed over into the study of congressional agencies. Perhaps it should. One characteristic of internal administrative law is that it “lack[s] the element of enforcement through independent courts.”114Id. at 1245. Might the Parliamentarian’s choice to permit publication of the GAO’s legal opinions to start the CRA review period offer an example of how internal administrative law can be enforced, or at least supported, by another institutional actor? And where else might we observe that type of interaction in the bureaucracy?

In terms of practical applications, the foregoing is especially relevant, given current interest in the relationship between congressional agencies and congressional capacity to govern.115Ginsberg & Hill, supra note 11, at 144–45; Brian D. Feinstein, Congressional Government Rebooted: Randomized Committee Assignments and Legislative Capacity, 7 Harv. L. & Pol’y Rev. 139 (2013); Kosar, supra note 101. See generally Congress Overwhelmed: The Decline in Congressional Capacity and Prospects for Reform (Timothy M. LaPira, Lee Drutman & Kevin R. Kosar eds., forthcoming 2020). In January 2019, the House established the Select Committee on the Modernization of Congress to make recommendations about the future of Congress.116H.R. Res. 6, 116th Cong. (2019) (enacted). The Committee’s work is ongoing, but many of the recommendations have included reforms for congressional agencies and institutions.117See Recommendations, Select Comm. on the Modernization of Cong., https://modernizecongress.house.gov/recommendations [https://perma.cc/K264-9YVJ] (listing various proposed changes to make the House more effective, efficient, and transparent). For example, the Committee has proposed new life for the Office of Technology Assessment, a congressional agency that existed from 1972 until the late 1990s.118Id.; Jory Heckman, House Modernization Committee Recommends Bringing Back Office of Technology Assessment, Fed. News Network (July 26, 2019, 12:24 PM), https://federalnewsnetwork.com/congress/2019/07/house-modernization-committee-recommends-bringing-back-office-of-technology-assessment [https://perma.cc/BW8P-AFV8]; see also Ginsberg & Hill, supra note 11, at 144–45 (describing the initial purpose of the Office of Technology Assessment and the current concern that Congress lacks “basic technological literacy”). Apart from these recent developments, institutional reform is often a topic in Congress. For example, in 2000, legislators came close to piloting a new regulatory review office inside the GAO to advise on regulations promulgated by the executive branch agencies.119Truth in Regulating Act of 2000, Pub. L. No. 106-312, 114 Stat. 1248. The pilot was never funded. William Yeatman, The Case for Congressional Regulatory Review, CATO Inst. 10 (Apr. 14, 2020), https://www.cato.org/sites/cato.org/files/2020-04/PA888_edit.pdf [https://perma.cc/8J9B-KGA3]. With recent efforts towards institutional reform in Congress, closer study of congressional agency function and dysfunction could provide important insights to legislators and advocates. How these agencies are structured likely informs how well they can serve Congress and the public.

B. Legislator Requests of CRA Legal Opinions

As Professor O’Connell observed, legislators may have different reasons to request an investigation from the GAO.120Joseph, supra note 91, at 187. So, too, with respect to CRA legal opinions. Reasons might include genuine uncertainty, boundary-testing, clock-starting, signaling, and protection.

First, legislators might be genuinely uncertain about how to apply the CRA’s broad definition of a rule to the multitudes of different policy documents that agencies issue. As the GAO explained in its first legal opinion on the subject:

Many agency rules are not described as such. They may be referred to as a guideline, direction, directive, instruction, clarification, manual section, policy, etc. While how an agency describes a document may be considered in determining whether the document is a rule under the APA, the courts primarily consider the substantive effect of the document.121Id. at 6.

The GAO then explained that whether a document is a rule “requires an examination of what it is intended to accomplish.”122Id. at 6–7. If Congress wants to base its decisions on a careful, fact-intensive analysis, it is reasonable that members of Congress would want an expert fact-finder like the GAO to weigh in.

Second, and relatedly, legislators might also be interested in testing boundaries of the definition of a rule. If the legislator is concerned about a policy document that generates disagreement as to whether it is a rule, being able to rely on an expert opinion from the GAO could help the legislator blunt criticism that he is stretching the law by trying to introduce a resolution of disapproval.123One example of this might be when the GAO was asked about the CRA’s applicability to proposed rules. The GAO opined that a proposed rule is not a rule for CRA purposes. See supra note 85.

Third, legislators could seek the GAO’s legal opinion solely as a procedure to start the CRA review period, even when they do not have a substantive need for an opinion. In this way, Congress can use a GAO legal opinion to make rules vulnerable to disapproval under the CRA. The descriptive statistics above, particularly the increase in the number of elapsed days between the issuance of an agency action and an associated legislator request, shows that legislators are doing this with increased frequency since 2017.

Fourth, legislators might also request GAO legal opinions to send political signals. A legislator might want to send signals to the agency that issued the document in question, as a shot across the bow to indicate displeasure with the policy; to constituents and other public stakeholders, as a way to show action;124See generally David R. Mayhew, Congress: The Electoral Connection (1974) (discussing legislators’ use of signaling for reelection purposes). or to other legislators or political actors as part of larger debates and negotiations.

Fifth, when the GAO issues a legal opinion that an agency action is a rule, legislators sometimes, but not always, follow that opinion with the introduction of a disapproval resolution. In October 2019, for example, the GAO opined that a letter from the Board of Governors of the Federal Reserve was a rule under the CRA.125Letter from Thomas H. Armstrong, Gen. Counsel, U.S. Gov’t Accountability Office, to Thom Tillis, Senator, U.S. Senate (Oct. 22, 2019), https://www.gao.gov/assets‌/710/702190.pdf [https://perma.cc/TDP7-BCC6]. No disapproval resolution has been introduced.126See Congress.gov, https://www.congress.gov/search?q={%22congress%22:%‌22116%22,%22source%22:%22legislation%22,%22search%22:%22congressId:116%20AND%20billStatus:\%22Introduced\%22%22}&searchResultViewType=expanded&pageSort=dateOfIntroduction:desc [https://perma.cc/TYB7-FY6Q]. Earlier in 2019, however, the GAO issued a legal opinion that a guidance document issued jointly by the Departments of Health and Human Services and Treasury was a rule under the CRA.127Letter from Thomas H. Armstrong, Gen. Counsel, U.S. Gov’t Accountability Office, to Ron Wyden, Ranking Member, U.S. Senate Comm. on Fin., and Frank Pallone, Jr., Chairman, U.S. House of Representatives Comm. on Energy and Commerce (July 15, 2019), https://www.gao.gov/assets/710/700266.pdf [https://perma.cc/8KAQ-XKYL]. Just over two weeks later, a disapproval resolution was introduced in both the House and the Senate.128S.J. Res. 52, 116th Cong. (2019); H.R.J. Res. 74, 116th Cong. (2019). Neither resolution was signed into law.129See Public Laws, Congress.gov, https://www.congress.gov/public-laws/116th-congress. This recent episode suggests that a request for a legal opinion does not necessarily indicate that a legislator intends to introduce a disapproval resolution. Why else might legislators request legal opinions, then? Might it be to protect the action by letting the agency know that it should submit the action to Congress to start the CRA review period, or by developing a record that the action is not a rule subject to the CRA’s disapproval procedures? Here the most interesting possibility is not about the GAO’s role but the way in which requests for legal opinions shed light on the politics of the time.

C. Effects of GAO Legal Opinions on CRA Vulnerability

Lastly, and specific to the CRA, is it a problem that GAO legal opinions are used to make rules vulnerable to disapproval under the CRA?

On the one hand, yes. Congress limited the applicability of the CRA’s fast-track procedures.130See supra note 32 and accompanying text. Using a legal opinion to make these rules vulnerable much later could thrust large numbers of rules into doubt. This includes those rules that were not sent to Congress, whether published in the Federal Register or issued elsewhere. The tradeoffs are somewhat analogous to those presented by statutes of limitations, most notably reliance interests.131See Charles C. Callahan, Statutes of Limitation—Background, 16 Ohio St. L.J. 130 (1955) (summarizing the history of statutes of limitation in the United States); Tyler T. Ochoa & Andrew J. Wistrich, The Puzzling Purposes of Statutes of Limitation, 28 Pac. L.J. 453, 467–68 (1997) (noting the effects of statutes of limitations on the reliance interests of nonparties). With a statute of limitations, valid claims trade off against the ability for parties to move forward. In the rulemaking context, members of the public may have invested resources or made other decisions in reliance upon the contents of an agency action that the GAO later opines is a rule and is therefore suddenly vulnerable. But the legal opinion, in conjunction with its use to start the CRA review period, makes CRA disapproval more accessible to legislators—by allowing them to reach back further in time—than was perhaps intended by the CRA.

On the other hand, no. If an agency fails to fulfill its legal obligation to notify Congress of its rule, that essentially deprives Congress of its ability to exercise oversight. This is especially the case for a rule that the agency did not publish in the Federal Register, but which might solely reside on an agency’s website, for example. While it might be reasonable to hold legislators responsible for knowing what the agencies publish in the Federal Register, expecting legislators to know the entire contents of every agency’s website is not reasonable. Also, there’s no evidence at present that legislators want to use the CRA to disapprove such an overwhelming number of rules such that the results would be significantly destabilizing.

The ultimate question is, if an agency failed to classify its action as a rule under the CRA, what is the remedy? Agencies are expected to apply a somewhat ambiguous definition of “rule” to their actions and notify Congress if they issue rules. But is this ambiguity—which was, of course, crafted by Congress—enough to deny Congress the oversight tool it afforded itself in the CRA? This question is worth additional consideration, perhaps borrowing from statutes of limitations and other analogous doctrines to shed light on the dilemma. With statutes of limitations, for example, the clock can start again in certain situations. So, too, with the CRA?


A subtle but burgeoning phenomenon, the GAO’s issuance of legal opinions for purposes of the CRA, has expanded the way that legislators can use the CRA to disapprove rules. The CRA is an increasingly important oversight tool in an era of expanded presidential unilateralism, and the GAO’s role in shaping the CRA’s scope poses several important questions. These questions span the proper role and functioning of congressional agencies, the political dynamics that drive legislator interaction with these agencies, and the tradeoffs between oversight using the CRA and reliance interests. While academic studies of Congress and federal agencies have largely overlooked congressional agencies, this brief analysis of the GAO’s legal opinions shows that a closer consideration of congressional agencies like the GAO could improve our understanding of a host of dynamics at play in the federal government.

Share this post