A recent essay at the Harvard Law Review blog on the Congressional Review Act (CRA) demonstrates an earnest interest in the arcane corner of public policy where administrative law, statutory interpretation, and legislative procedure all pile up. Further, by offering a constructive reform it does something most of the peanut gallery (myself a season ticketholder) frequently avoid.
It also betrays common misunderstandings of the CRA.
Briefly, that essay argues that the CRA, having undergone a recent surge in use of its regulatory disapproval mechanism, is asymmetric in effect and results entirely in regulatory “destruction.” The author then proposes a modification to remedy this one-way ratchet.
However there are discrepancies meriting scrutiny, some directionally defensible but inexact and others manifestly fallacious. Not the least of which are several citations which not only do not support the claims therein, but in some cases actually argue counter to them. More broadly it provides a good opportunity to clarify some misunderstandings with respect to the CRA which often go unchallenged.
Stand Up and be Counted
First, the author points out that the CRA, by “waiv[ing] the Senate filibuster,” effectively allows passage by only “fifty votes in the Senate,” but elsewhere by variously “a simple majority,” “fifty-one votes,” and “just over fifty percent.” All permissible descriptions perhaps but the imprecision undermines a major premise for the author’s reform proposal.
Formally the law affords the resolution privileged status in the Senate making a motion to proceed non-debatable, allowing passage with support of a simple majority of a quorum present and voting (the Constitutional passage threshold for all duly passed laws, Art. I, Sec. 5). This is why resolutions have passed with as few as 49 votes (see H.J. Res 131 (2026); H.J. Res 37 (2017)).1
Second, they characterize use of the CRA’s disapproval mechanism as having “exploded in 2025” following “only a handful of successful CRA resolutions in the statute’s first thirty years” (1996-2025). However one chooses to characterize the recent increase, explosion/handful is not how most would describe the expansion from 16 resolutions (2017) to 22 (2025). This even excludes an additional three in 2021 for a total of 19 in nine years. What’s more, these counts are all included in the very The Regulatory Review article linked as the basis for this claim and repeatedly elsewhere in the essay (Jones and Revesz (2026)).2
The Forgotten Biden Years
Substantively, it’s worth focusing on the omitted 2021 examples. Notwithstanding the intractability of categorizing any given administrative action as entirely regulatory or deregulatory (more below), these three are described by the Biden Administration’s own Statements of Administration Policy as undoing the prior Administration’s deregulatory actions.3 That “Congress can only use the statute’s fast-track procedure to invalidate regulations” (emphasis in original) is true, but claiming it “is solely deregulatory” as they mean it does not follow. While conceding that just over seven percent of all successful resolutions running in the regulatory direction is compelling evidence the CRA strongly tends to be deregulatory, it is not “inherently” so.
This is important because the entire premise of the subsequent proposal is a need to resolve a structural “asymmetry [to] help Congress reassert its power over the administrative state.”
A Rule is a Rule is a Rule
This whole issue is further convoluted in that it’s not always obviously the case that a regulation is exclusively regulatory or deregulatory. Rules can run into the hundreds of pages, with a mix of various restrictions, allowances, prohibitions, waivers, reporting, and so on. Admittedly I’m being a bit precious here. We can stipulate there is a nontrivial universe of regulations that on net do fall into one category or the other.
But there’s a common and incorrect assumption about the procedural scaffolding of regulations.
Without loss of generality - limiting ourselves to so-called legislative rules promulgated subject to the Administrative Procedure Act (APA) under 5 U.S.C. § 553 - the APA does not distinguish between regulations and so-called deregulations—between newly promulgated and repealed, those that impose restrictions and those removing them.
The notice-and-comment requirements – and the same standards of judicial reviewability – apply equally whether the EPA is tightening automobile emission standards or loosening them, ceteris paribus.
While empirically the CRA has historically provided a deregulatory mechanism, that is orthogonal to the essay’s argument that the CRA is a “one-way ratchet” and the law’s “substantially same form” language (5 U.S.C. § 801(b)(2)) can only function to abrade agency regulatory authority.
There Oughta Be a Law
On this I concur:
If Congress wishes to restore ‘congressional accountability’ over agency rulemaking, that would reasonably include Congress affirming that it wants an agency to take an action.
Notwithstanding pettifogging over the meaning of “action,” as I emphasized in my Congressional testimony on the subject of Congress’s prerogative in light of Loper Bright, “’That is not to say that Congress cannot or does not confer discretionary authority on agencies. Congress may do so…and it often has.’” (Loper Bright Enters. v. Raimondo, 603 U.S. ___, slip op. at 26 (2024) ). In other words, Congress indeed ought to affirmatively direct “an agency to take action,” which it not only is fully permitted to but required to do so, without CRA ever entering into it—reform or no.
The essay’s core normative desire resides in the very same neighborhood as my own. Directions there, however, are not as they argue nor does Congress suffer a lack of capable tools. And as was laid bare above, the tool in question does not in fact work in one direction as claimed.
But Congressional accountability is constrained by capacity (among other things). This is an institutional deficit and not a procedural defect. To wit, House committee staff has declined forty percent since 1994. Modifying a fast-track lever affixed to regulations it already has authority over does not resolve the asymmetry the author identifies. It would only further burden an under-resourced institution, and therefore a solution misapplied to the actual problem.
1 The binding theoretical minimum is a majority of a quorum meaning a simple majority of 51 Senators, thus 26 aye votes.
2 Not essential to the essay’s argument, but the author provides a historical context unsupported by their own linked source. I’ve added years and context within this quoted passage in bold to demonstrate why the sequencing is incoherent:
“[After] INS v. Chadha (1983), the legislative branch sought a new means of exercising control over its delegations to agencies. As distrust in agency bureaucracy grew over the next decade (the linked article specifically argues the shift originated in the 60s), Congress settled on a new framework through which to ‘redress the balance’ between the legislative and executive branches by ‘provid[ing] a formal Congressional review process of regulations used by Federal agencies.’ Thus, the CRA was born (1996).”
3 See Statements of Administration Policy: S.J. Res 13 (2021), S.J. Res 14 (2021), and S.J. Res 15 (2021).