Article Correction for the 51-49 Rule of Constitutional and Statutory Interpretation; Loper Bright and Interpreting Out Ambiguity (11/9/26; 11/10/26)

I write this blog to add nuance, dialing back a key claim I made in my article Loper Bright Flip Flops on Chevron Deference: A Tax Lawyer's Perspective, 79 Tax Law. 323 (2026), here. In broad strokes, I questioned whether Loper Bright’s claims about the state of the law at the enactment of the APA in 1946, noting particularly the role of a famous tax case, Dobson v. Commissioner, 320 U.S. 489 (1943) that Loper Bright ignores and the statutory text interpreted in Dobson. I do not change that discussion. More importantly for this blog, I questioned Loper Bright’s claims that

(i) Chevron required courts to defer to an agency interpretation over an interpretation that they determined to be the best interpretation; and

(ii) through rigorous interpretation, courts can interpret out interpretive ambiguity (aka interpretive equipoise) where a court could not determine with sufficient confidence which of two competing interpretations was the best interpretation.

I do not change (i) in any respect.

I add considerable nuance to (ii).

Focusing on (ii), I posited that courts honest with themselves would sometimes encounter interpretive issues that would not permit a best interpretation with sufficient confidence and, in that state which I posited as a range, the court would either (a) need a default tool to decide the issue (such as Chevron provided prior to Loper Bright and other interpretive tools (such as lenity) in other contexts) or (b) use some other nonprincipled method (such as flipping coins or going with the court’s policy preferences) to break the tie. (I called that process “law-finding” like fact-finding.) I said (pp. 338-339) using a more likely than not standard for legal interpretation (remember that standard for law-finding in the penalty for tax shelters?):

More likely than not is often conceptualized as at least 51%–49%, with 50%–50% being a state of interpretive equipoise where one cannot determine probability one way or the other. Even that acknowledges a range of equipoise from 49.0001% to 50.9999%. In any event, how can anyone make a meaningful distinction between a 51% interpretation and a 49% interpretation? If (as I think) that cannot be done in any meaningful way, then legal interpretive equipoise involves a range of reasonable interpretations, say of 55% and 45% or 60% and 40% or even some larger spread.n58 If that is right, then a lawyer’s opinion or a court opinion will have to treat the range as being a state of interpretive equipoise where the “best” and “not best” interpretations cannot be determined.
   n58 See Kavanaugh, supra note 48. See also Ryan Doefler, How Clear Is Clear?, 198 VA. L. REV. 651, 682 (2023) (on the motivated reasoning affecting interpretive outcomes stating: 
In an ordinary statutory case, a judge could, without embarrassing herself, mischaracterize a merely likely interpretation as “unlikely.” And because it is an ordinary case— that is, a case in which the evidentiary threshold is more likely than not—that would be enough to alter the outcome. Here, again, the question is why think that judges would have an easier time exaggerating or understating the likelihood of various readings if the epistemic threshold were, say, 65-35, as opposed to 51-49? In any statutory case, there must be some line past which the moving party prevails. And wherever that line happens to be, the risk of willfulness or motivated reasoning will be greater the closer a case is to that line.

See also my article pp. 345-349 where I say that Loper Bright’s claim of no ambiguity/equipoise is a legal fiction much in the way that Chevron’s claim of interpretation from ambiguity permitted a presumption of interpretive authority.

I subsequently reviewed and commented in a blog on an article where, in comments to my blog entry, the author and I discussed my range of ambiguity/equipoise argument. Comments on Highly Recommended Article Extending Skidmore "Deference" Approved in Loper Bright (Federal Tax Procedure Blog 3/7/26; 3/8/26), here. The author, Mitchell Zaic, argued (highly summarized) that, under Loper Bright, equipoise would be encountered only at 50%-50% level of confidence, meaning (again highly summarized) that 51%-49% confidence level requires the judge to apply the 51% interpretation. Mr. Zaic disagreed with my thesis that a judge cannot make a meaningful determination that a 51% confidence level is better than the 49% interpretation. (Of course, even that argument says there is a range of ambiguity between 51% and 49%; but as a practical matter most would say that anything in that constricted range was 50-50.)

I have recently read two articles that develop the “51-49 rule” in constitutional interpretation, an exercise that is not materially different from statutory interpretation. Aaron Tang, Second-Order Constitutional Theory, 93 U. Chi. L. Rev. 1447 (2026), here; and John O. McGinnis & Michael B. Rappaport, The Constitution's Own Interpretive Method: Advancing the Original Methods Approach, 102 Notre Dame L. Rev. (forthcoming 2027) (SSRN here).

I start with Professor Tang’s article. Professor Tang categorizes constitutional interpretation as involving first-order process, meaning applying the interpreter’s preferred constitutional interpretation method (originalism, pragmatism, even textualism, etc.), and then if first-order process does not resolve ambiguity, applying a second-order process to resolve ambiguity and decide the case. Professor Tang devotes (pp. 1467-1478) a significant portion of the article to listing and discussing “Some Common Second-Order Theories” used by the Supreme Court to resolve first-order uncertainty. This uncertainty after first-order process (the method of interpretation) basically parallels my argument of ambiguity or equipoise. Second-order processes resolve the ambiguity or equipoise in the way Chevron deference did; indeed in this regard, one of the second order processes Professor Tang discusses (pp. 1471-1474) is “Legislative Deference.”

First up on Professor Tang’s list of Common Second-Order Theories is “The 51-49 rule” (pp. 1467-1471). Professor Tang describes it:

The most common second-order approach to resolving legal uncertainty is so familiar as to be unstated: rule in favor of whichever answer has more evidence to support it than any other. Professor Gary Lawson has written the most important and comprehensive descriptive account of this default rule, which he calls the “best-available-alternative” standard: In difficult cases, courts typically adopt the constitutional interpretation that is better than all other options, no matter how close or uncertain the evidence.n109
   n109 LAWSON, EVIDENCE OF THE LAW, supra note 40, at 202–03; see also Gary Lawson, No History, No Certainty, No Legitimacy . . . No Problem: Originalism and the Limits of Legal Theory, 64 FLA. L. REV. 1551, 1561–63 (2012); Gary Lawson, Proving the Law, 86 NW. U. L. REV. 859, 890–91 (1992).

The process as described by Professors Lawson and Tang involves law-finding like fact-finding where the party having the affirmative of the legal interpretation (i.e., must convince the court of the legal interpretation) to prevail. I will further address burden of proof in law-finding in this context at the end of this blog.

Professor Tang says that the label “51-49 rule” was not Lawson’s but the label of “fellow originalists McGinnis and Rappaport” the authors of the second article I cite and link above, who describe the rule as requiring the interpreter to choose “the better interpretation of a provision, even if it is only slightly better,” with Professor Tang elaborating in footnote 111 on p. 1468:

This 51–49 rule, they [Professor McGinnis & Rappaport] argue, is “a key to reducing uncertainty” about the Constitution— and they argue that it is called for by the Constitution itself. Id. I return to the second claim below. See infra Part II.B.3. I note now, though, that the 51–49 rule does not “reduc[e] uncertainty” about the Constitution in a strict sense; it is not, in other words, a first-order tool that would allow a legal thinker to become more confident that a given constitutional interpretation is the correct one. Instead, the rule is better thought of as a tool for deciding a constitutional question in the face of uncertainty when people could reasonably reach different conclusions as to which outcome possesses 51% of the relevant evidence.

Professor Tang says (pp. 1468-1469) that the 51-49 “naming convention * * * efficiently captures the rule’s distinctive feature: When the relevant body of evidence points in competing directions, choose the outcome that has more evidence to support it, however marginal the difference.”

So how is this relevant to statutory interpretation which is the domain of Loper Bright? Professor Tang says (p. 1471 n. 124) , correctly I think,

Or consider the Court’s recent embrace of what is effectively the 51–49 rule in the related context of statutory interpretation, where it rejected the longstanding second-order rule of Chevron deference to agencies and instead directed judges to “use every tool at their disposal to determine the best reading of [an ambiguous] statute.”n124
   n124 Loper Bright Enters. v. Raimondo, 144 S. Ct. 2244, 2266 (2024) (emphasis added). Indeed, this Article’s distinction between first- and second-order theories of constitutional adjudication could apply with similar conceptual force to the development of a fully worked-out approach to statutory interpretation.

My key point about the 51-49 rule is that, from my perspective, the 51-49 rule for  marginal differences unresolved by first-order processes does not permit principled resolution of the interpretive issue. Yet it appears that mainstream analysis may be more consistent with Mr. Zaic’s interpretation of the process required by Loper Bright. However, I stick by my conclusion that this genre of marginal differences do not permit principled resolution of the interpretive issue because the judge finding the interpretations at 51-49 cannot articulate a reason that 51 prevails over 49; the judge will not say that he flipped a coin or use some other arbitrary process (including his own policy preferences) to decide the case.

Professor McGinnis and Rappoport’s new article (cited and linked above) is similar, describing (p. 108) the 51-49 rule as instructing “interpreters to select the better interpretation, even if it is only marginally better than the alternative reading. Under this framework, instances of true indeterminacy, such as those involving irreducibly vague terms, are rare; when they do arise, they concern borderline issues of secondary importance.”

I said I would return to the use of burden of proof concepts with respect to statutory interpretation, a law-finding exercise rather than a fact-finding exercise. In his concluding discussion of Some Common Second-Order Theories, Professor Tang says (pp. 1477-1478):

To summarize, each of the leading second-order theories entails two elements: a burden of proof and a decision rule for what to do if that burden is not satisfied. Sometimes the burden of proof is articulated in qualitative terms—such as the Thayerian requirement of proof of a constitutional violation that is “clear” or “beyond a reasonable doubt.” The 51–49 rule, by contrast, is quantitative. Sometimes the requisite burden of proof is unspoken; sometimes it is the subject of open contestation. Importantly, what these differing proof thresholds reflect is competing values and individual attitudes, not some objective empirical truth in the world about when first-order uncertainty is too great to tolerate.

Added 11/9/26 7:30pm: While on burden of proof, I return to Dobson which I mentioned above. Dobson was a 1943 unanimous opinion written by Justice Robert Jackson (who had formerly been Chief Counsel of IRS, DOJ Tax Division AAG, Solicitor General, and Attorney General). Dobson interpreted the standard of review for Tax Court legal interpretations to require deference because (i) the Tax Court was statutorily an agency to which standard agency deference applied and (ii) the statutory standard of review permitted reversal only if the Tax Court interpretation was "not in accordance with law." Just focusing on the latter text, the Tax Court interpretation had to be found affirmatively "not in accordance with law," which means that the Tax Court interpretation prevailed unless it was not in accordance with law because an opposing interpretation was best. Doubts as to the opposing interpretation are resolved in favor of the Tax Court interpretation.

Not surprisingly that statutory text found its way into APA § 706(2)(A) which permits a court to set aside an agency interpretation only if it is found "not in accordance with law." Applying the Dobson interpretation of that statutory text, if the agency interpretation could not be found affirmatively to be not in accordance with law, it could not be set aside. Does anyone thinking about this and the trajectory from Dobson, Skidmore (which also affirmed deference as well as the deference light for which it is used now), to APA § 706(2)(A) think that deference should not apply. I know. The Supreme Court said so on a totally inadequate development of the law in Loper Bright. Which brings up another great Justice Jackson aphorism about the Supreme Court: “We are not final because we are infallible, but we are infallible only because we are final." Brown v. Allen, 344 U.S. 443, 540 (1953) (Jackson, J., concurring). So, the Supreme Court's Loper Bright is infallible at least for the time being until the present Court realizes and confesses its error (not likely unless Skidmore is expanded to cover the ground deference formerly occupied) or a more prudent majority appears on the Court.

Added 11/10/26 3:30 pm: I have just revised the Working Draft of my book to make the following additions (in red face). The following will be inserted at Practitioner Edition p. 91; Student Edition 58. The red font indicates additions to the texts.

          1. What does a court do when, after it applies Loper Bright’s exhortations to determine the “best” interpretation, the court still finds statutory ambiguity (interpretive equipoise)? Ambiguity (interpretive equipoise) existed before Loper Bright and exists after Loper Bright. Remember the lesson of lenity–ambiguity (interpretive equipoise) requires a default rule to resolve the case. The court must have a default rule or decide by some arbitrary device–like flipping a coin.[1] I think the debate, when sharply focused, is not whether there can be ambiguity space for Loper Bright’s command to reach the best interpretation but how constricted that space, which I call a range, is. Using percentage levels of interpretive confidence in agency interpretation and an opposing interpretation, 51% for either interpretation might permit Loper Bright best interpretation, meaning the interpretations at 49% is not the Loper Bright best interpretation (this is referred to as a “51-49 rule”);[2] but note that 51-49 still has an ambiguity range from 50.999 to 49.001. (For those, like me, who think the range is wider where an interpreter cannot with intellectual honesty distinguish between 55% and 45% interpretations (or 60% and 40%) is the range of ambiguity where some default rule of decision is required and failing such a default, arbitrary decision is required; however, this may be semantics since a court unable to distinguish one or the other interpretation on a best-not best scale would treat the split as 50-50 (regardless of the range of ambiguity) unless it is willing to go with an arbitrary result in the absence of a default rule.[3])

[1] ChatGPT Plus says that this is an anomaly in Loper Bright. See ChatGPT Plus on Loper Bright's Superficial Claim of Court's Ability to Determine a Single Best Reading in State of Ambiguity (Federal Tax Procedure Blog 9/24/26) (discussing the ChatGPT responses which conclude that irreducible ambiguity “may be the central theoretical vulnerability in Loper Bright when one takes its ‘single, best meaning’ language literally”).

[2] This 51-49 rule was originally articulated as rule of constitutional interpretation but conceptually applies for statutory interpretation under Loper Bright. Aaron Tang, Second-Order Constitutional Theory, 93 U. Chi. L. Rev. 1447, 1471 n. 124 (2026) (citing Loper Bright and saying that the discussion of first- and second-order theories for constitutional interpretation (including the 51-49 rule) “could apply with similar conceptual force to the development of a fully worked-out approach to statutory interpretation.”

[3] See Article Correction for the 51-49 Rule of Constitutional and Statutory Interpretation; Loper Bright and Interpreting Out Ambiguity (Federal Tax Procedure Blog 11/9/26) (adding nuance to an earlier claim I made in an article about ranges for interpretive ambiguity/equipoise).