Disfavored Precedent as a Challenge for the Lower Federal Courts

Lower Federal Courts Face Recurring Uncertainty When the U.S. Supreme Court Weakens but Does Not Overrule Precedent. We Offer a Decisional Authority Model to Guide Lower Courts in Handling Such Disfavored Decisions.

What should lower federal courts do when the U.S. Supreme Court chips away at or begins ignoring a precedent without overruling it? Consider, for example, the Court’s treatment of its 1971 decision in Bivens v. Six Unknown Named Agents of Federal Bureau of Narcotics, which allowed private plaintiffs to bring damages claims against federal officials for constitutional violations. Although the Court has never overruled Bivens, it has rejected every Bivens claim it has considered since 1980 and has emphasized that the remedy is “disfavored.”

There are many other examples. For decades, various justices disparaged the 1971 Lemon test for Establishment Clause claims, yet the Court did not abrogate it until 2022. Likewise, the Court for years ignored the Chevron doctrine, despite deciding a number of cases involving agency interpretations of statutes, but did not overrule it until 2024. In recent years, the Court has also cast doubt on precedents concerning presidential removal of executive officials, congressional delegations of authority to the executive branch, and qualified immunity.

We describe this phenomenon as “disfavoring” precedent. To treat a precedent as “disfavored” means to give it less weight than what would have seemed appropriate at the time that it was issued. When the Supreme Court disfavors but does not overrule a precedent, what happens in the lower federal courts, which do almost all of the federal judicial business? How should they treat such precedents?

In a recent article,³¹ we explored how the Court may disfavor its precedents, and we discussed how we believe lower courts should respond. Our suggestions draw on case studies, which gathered data — both quantitative and qualitative — on federal courts of appeals decisions involving the Lemon test, Bivens actions, and Chevron deference. We highlight key findings here.

Legal Framework

The Need for a New Model of Vertical Precedent

“Vertical precedent” refers to the obligation of a lower court to give effect to a higher court’s decision. Among the existing accounts of vertical precedent, there seem to be two major perspectives: the prediction model and the authority model. We argue that neither model, at least as currently formulated, provides a normatively attractive vision of vertical precedent.

Under the prediction model, a lower federal court should make a “prediction” about how the current Supreme Court might rule in a case and act accordingly. But such an effort has the potential to undermine the rule of law. The prediction model “promotes a conception of law as the sum of the views of the particular judges who happen to sit on the high court at any time, rather than a conception of law as an impersonal, ideal whole.”³²

The prediction model also raises practical concerns. There is a danger, for example, that lower court predictions could turn out to be wrong, in part because it may be difficult to predict how much weight the Supreme Court will give to its own precedents. Another challenge is the passage of time. Given the slow pace of litigation, the composition of the Supreme Court may change before a case reaches the Court. Accordingly, the lower court’s prediction may be based on the “wrong” set of justices.

A prediction model may be particularly problematic for the modern Supreme Court. Since 1988, the Court has had broad discretionary certiorari jurisdiction, such that it hears only a small fraction of the hundreds of thousands of cases handled annually by the lower federal courts. So the Court simply cannot police on a case-by-case basis “incorrect” lower court predictions. That may help explain why the Court, beginning in 1989, started admonishing lower courts not to predict future Supreme Court rulings — for example, by engaging in anticipatory overruling. Instead, the Court has insisted, lower courts should follow an authority model.

The authority model avoids the normative challenges of the prediction model because it treats existing Supreme Court precedent as governing law. But the Court has been notably unclear about what the authority model means, suggesting at times that lower courts should apply the original meaning of a Supreme Court decision and ignore subsequent developments.

Yet such a vision of judicial decision-making is artificially narrow. Supreme Court doctrine consists not simply of one case, such as Lemon, Bivens, or Chevron, but of the entire body of precedent that applies, extends, or distinguishes an earlier ruling. Nor is it likely that the Supreme Court really means for lower courts to zero in on a single past case without regard for subsequent changes to the doctrine. The scope and effect of a precedent are not captured merely in the text of the original decision; the meaning of a precedent is affected by other legal materials, including other precedents. Accordingly, lower federal courts properly take into account some subsequent developments that signal that a precedent is in disfavor.

But which developments count? The existing case law and commentary do not tell lower courts which Supreme Court declarations properly signal that a precedent is on thin ice. We aim to provide more clarity.

Disfavoring Signals and the Decisional Authority Model

To treat a precedent as “disfavored” means to give it less weight than what would have seemed appropriate at the time that it was issued. To give this concept more practical content, we offer a taxonomy of five signals that the Court may use to disfavor its precedent. We then offer our “decisional authority model” to provide needed guidance on which signals matter.

First, the current justices may make disparaging statements about a precedent in either majority or other opinions. Many justices attacked the Lemon test in this way, describing it as untethered from history and unworkable in practice. Justice Antonin Scalia colorfully referred to the test as akin to “some ghoul in a late-night horror movie that repeatedly . . . stalks our Establishment Clause jurisprudence.”³³

Second, the Court may apply a precedent narrowly. Since 1980, the Court has narrowed its Bivens decision by rejecting many subsequent efforts to sue federal officials for constitutional violations.

Third, the Supreme Court may simply ignore a precedent — by failing to rely on it in cases where it would seem by its terms to apply. For years, the Court gave such silent treatment to the Chevron doctrine. As the Court observed in its recent decision overturning Chevron, it had “not deferred to an agency interpretation under Chevron since 2016.”³⁴

Fourth, the Court may issue decisions in related areas of law that seem to undermine a precedent. Some commentators asserted that the Supreme Court’s application of the major questions doctrine to reject federal agency action had cast further doubt on the continuing vitality of Chevron deference.

Fifth, the Supreme Court may shift away from the methodology that underlies a precedent. Several justices have challenged Bivens on this ground, arguing that, because the Court today is less inclined to imply private rights of action into federal statutes, it should be equally unwilling to infer private causes of action to enforce the Constitution.

Which of these signals count? Under our decisional authority model, lower federal courts should rely only on the first and second categories, and only to the extent that the Supreme Court has criticized, distinguished, or narrowed its precedent in binding decisions. By contrast, lower courts should not rely at all on the third, fourth, or fifth categories — the failure to cite a precedent, changes in related areas of law, or methodological shifts. Supreme Court silence, as well as changes in methodology or other areas of law, provide uncertain and unreliable information about the legal status of a past precedent. Nor should lower courts rely on the first category, to the extent that the disparaging statements come only from plurality, concurring, or dissenting opinions.

Our decisional authority model — with its focus on binding decisions, rather than separate opinions or silence — comports with core features of our federal judicial system. Our system rests on a broad assumption that decisions of the Supreme Court should be supported by a majority of the sitting justices. Yet if lower federal courts were to treat separate opinions as effectively undermining a past precedent, that would give a single justice or a subset of the Court considerable power over the content of federal law. Indeed, that is particularly true in the modern judiciary, where the Supreme Court can review only a fraction of lower court cases, and thus would have trouble policing such lower court conduct. Our decisional authority model is designed to ensure that the Court as a whole may decide the fate of its precedent.

Case Studies

We examined how the lower courts reacted to disfavored Supreme Court precedent in three different issue areas. For each case study, we began with a summary of how the Court signaled its disfavor and then discussed the lower court response. Notably, we found that the doctrines studied here — Lemon, Bivens, and Chevron — had considerable staying power in the lower courts, even after they had lost favor in the Supreme Court. But we also identified some confusion and disagreement among lower court judges as to how to make sense of disfavoring signals from the Court.³⁵

Lemon

1. Disfavoring at the Supreme Court

The 1971 test in Lemon v. Kurtzman instructed courts to examine whether a government action had a secular purpose; whether the effect of the government action was to favor religion; and whether the government had somehow become too entangled with religion.³⁶ Justice Sandra Day O’Connor later offered a gloss on this test, advising courts that, in determining the purpose or effect of government action, they should focus on endorsement — that is, whether a reasonable observer would understand the action as favoring adherents over nonadherents.³⁷ Although the Court relied heavily on the Lemon and endorsement approach in subsequent decisions, the approach was long subject to significant and sustained criticism.

For several decades, the Court vacillated between applying the Lemon test and ignoring it. Some observers assumed that Lemon was “dead” after the Court in its 1992 decision in Lee v. Weisman applied a coercion test, rather than Lemon, to invalidate a state-sponsored high school graduation prayer.³⁸ Yet the Court turned back to Lemon just one year later in examining a religious group’s use of public school property.³⁹ To further complicate the picture, on the same day in 2005, a Court majority relied on Lemon to hold invalid Ten Commandments displays at two county courthouses,⁴⁰ while a plurality declined to use Lemon in upholding a different Ten Commandments display on state capitol grounds.⁴¹

The Court’s 2019 decision in American Legion v. American Humanist Ass’n suggested continuing disagreement about Lemon. Justice Samuel Alito’s plurality opinion (joined by Chief Justice John Roberts and Justices Stephen Breyer and Brett Kavanaugh) declined to apply Lemon in upholding the display of a cross on public land, indicating that a historical test rather than Lemon should guide (at least) the evaluation of “longstanding monuments, symbols, and practices.”⁴² Justice Neil Gorsuch (joined by Justice Clarence Thomas) concurred in the judgment, arguing primarily that the plaintiffs lacked standing to challenge the cross, but further offering the view that “Lemon was a misadventure.” Separately concurring, Justice Thomas announced that he “would take the logical next step and overrule the Lemon test in all contexts.” Meanwhile, Justice Elena Kagan — who provided the fifth vote to uphold the cross — declined to join the parts of Justice Alito’s opinion that rejected Lemon, stating that the “test’s focus on purposes and effects is crucial in evaluating government action in this sphere.”

This leads us to the Supreme Court’s 2022 decision in Kennedy v. Bremerton School District. The Court there rejected a lower court ruling that a school district would violate the Establishment Clause if it permitted a public high school football coach to pray (sometimes, alongside student players) on the 50-yard line. Justice Gorsuch’s opinion announced that the court of appeals “erred” by relying on the Lemon and endorsement approach, explaining that “the ‘shortcomings’ associated with [Lemon‘s] ‘ambitiou[s],’ abstract, and ahistorical approach to the Establishment Clause became so ‘apparent’ that this Court long ago abandoned Lemon and its endorsement test offshoot.”⁴³

2. Lower Court Response

We set out to determine how lower courts responded to the Court’s vacillating approach to Lemon. Did the lower courts treat Lemon as “abandoned” or as good law? We examined a random set of federal court of appeals decisions from 1971 to 2023. We found that, until the Supreme Court’s decision in Kennedy, the lower federal courts generally treated the Lemon test as good law.

Our qualitative review of the cases confirmed this finding. Although the courts of appeals recognized that the Lemon test was subject to criticism, they viewed it as governing law. A 2009 Tenth Circuit case involving a Ten Commandments display is illustrative. The Court stated that, although “certain Supreme Court Justices have harshly criticized” Lemon, “[t]he Lemon test clings to life because the Supreme Court, in the series of splintered Establishment Clause cases since Lemon, has never explicitly overruled the case. While the Supreme Court may be free to ignore Lemon, this court is not.”⁴⁴

We also found, however, notable confusion among some courts of appeals as to how to approach Lemon, following the splintered 2019 decision in American Legion. Several courts of appeals “count[ed] heads” in American Legion — adding up the votes of the plurality and concurrences — to conclude that “Lemon [was] dead” in the context of public displays or even in all cases.⁴⁵ Meanwhile, writing in 2021, Seventh Circuit Judge David Hamilton criticized this practice of “add[ing] together” the votes from different opinions in American Legion, asserting that “[p]erhaps the broader demise of Lemon is finally imminent, but Lemon has survived scholars’ predictions of its demise for decades.”⁴⁶

Bivens

1. Disfavoring at the Supreme Court

The Supreme Court’s 1971 decision in Bivens announced that private individuals could bring damages claims against federal officials for constitutional violations.⁴⁷ Bivens itself involved a Fourth Amendment claim arising out of a search by drug enforcement officials. In 1979, Davis v. Passman permitted a Bivens claim by a former congressional staffer alleging sex discrimination under the Fifth Amendment’s Due Process Clause.⁴⁸ One year later, Carlson v. Green allowed an Eighth Amendment claim for failure to provide adequate medical treatment to a federal prisoner.⁴⁹ The Court’s initial decisions indicated that any Bivens claim should be permitted, unless Congress provided an alternative remedy that was designed to be as effective as Bivens, or there were “special factors” that counseled against allowing a Bivens claim.

Yet soon thereafter, the tide turned against Bivens. The Court repeatedly rejected Bivens claims, finding them barred by special factors or alternative remedies, even those not as effective as Bivens. Then, beginning in 2001, a few justices mounted a frontal attack on the entire concept of a Bivens remedy. In a series of concurring opinions, Justices Scalia and Thomas urged the Court to “limit Bivens and its two follow-on cases (Davis and Carlson) to the precise circumstances that they involved.”⁵⁰ That approach, Justices Scalia and Thomas argued, would be more consistent with the Court’s refusal to imply causes of action in federal statutes.

In its 2017 decision in Ziglar v. Abbasi, the Court made some important changes to the Bivens doctrine. First, it decided that a court must determine if a case presents a “new context” — that is, differs “in a meaningful way from previous Bivens cases decided by this Court” (although the Court added that “trivial” differences would not be sufficient to create a new context). Second, if a case does present a new context, a court must consider whether there are any “special factors” that counsel against allowing a Bivens claim (a category that now includes any alternative remedies).⁵¹

In 2020, writing separately in Hernández v. Mesa, Justice Thomas (joined by Justice Gorsuch) urged the Court to go further and overrule Bivens: “It is time to correct this Court’s error and abandon the doctrine altogether.”⁵² Yet in its 2022 decision in Egbert v. Boule, the Court declined to overrule Bivens — interestingly, in an opinion also written by Justice Thomas — and instead further narrowed Bivens by rejecting the Fourth and First Amendment claims in that case.⁵³ Meanwhile, Justice Gorsuch, now writing alone, renewed the call to expressly overrule Bivens.

2. Lower Court Response

How have the lower federal courts responded to the Supreme Court’s growing hostility toward Bivens? We examined courts of appeals decisions from 1971 to 2023. We found three different eras of activity with respect to Bivens.

During the first era, which lasted from the 1970s until the mid-1990s, the federal courts of appeals both heard fewer Bivens appeals and accepted a greater proportion of Bivens claims than they would in later eras. Notably, the courts of appeals accepted claims arising under not only the Fourth, Fifth, and Eighth amendments but also other constitutional provisions, such as the First Amendment, and in factual scenarios that differed dramatically from Bivens, Davis, and Carlson.

The second era, which comprises the 20-year period from the mid-1990s until around 2017, reflects a good deal more variation and uncertainty surrounding Bivens. During this period, the Supreme Court repeatedly rejected Bivens claims but did not revise the Bivens framework. Meanwhile, the number of court of appeals decisions discussing Bivens increased dramatically. And although the courts of appeals continued to grant relief to Bivens claimants in a number of cases, the percentage of successful Bivens claims dropped substantially. It is also interesting to observe the grounds upon which Bivens claimants lost: When courts of appeals rejected Bivens claims, they did not do so simply by applying the Bivens test. Instead, courts dismissed many claims on the merits — without analyzing whether a Bivens action was cognizable — or on procedural grounds, such as a failure to exhaust administrative remedies or filing beyond the statute of limitations. Indeed, it seemed during this era, perhaps because of both the volume of cases and the uncertainty surrounding the Court’s approach to Bivens, that the lower courts often preferred to avoid addressing a Bivens question directly.

The third era began with the Supreme Court’s 2017 decision in Ziglar v. Abbasi, which offered a new framework for Bivens claims and helped to calm much of the uncertainty of the second era. During this third era, the volume of Bivens litigation has decreased dramatically, as has the number of successful Bivens claims. Moreover, when the courts of appeals reject Bivens claims, they generally do so by applying the Ziglar framework (new context, special factors) rather than on separate grounds. But notably — and even after the Hernández and Egbert decisions in 2020 and 2022 underscored the disfavored status of Bivens — some Bivens plaintiffs still manage to succeed.

Chevron

1. Disfavoring at the Supreme Court

In its 1984 decision, Chevron U.S.A. Inc. v. Natural Resources Defense Council Inc., the Supreme Court set forth a two-step analysis for judicial review of an administrative agency’s interpretation of a statute that it administers. First, if Congress had spoken directly to the question presented, the court was to apply the “unambiguously expressed intent of Congress.” But, if the statute was silent or ambiguous, the court should accept the agency’s “permissible construction of the statute.”⁵⁴

This decision became the basis for the Chevron deference doctrine. For a long time, Chevron was not controversial on the Supreme Court. But some fundamental disagreements emerged in 2013, as members of the Court openly questioned Chevron. The Court also soon began applying a “major questions doctrine” to inform how delegations of power to administrative agencies were to be interpreted, which limited the space for Chevron deference. Appearing to sense the Court’s lack of enthusiasm for the doctrine, counsel before the Supreme Court began de-emphasizing it. In 2023, the Court finally granted certiorari to consider whether to retain the doctrine, and it formally overturned Chevron in June 2024.⁵⁵

2. Lower Court Response

In examining the lower federal courts’ response to Chevron, we focused on recent years, when the doctrine appeared to be most in disfavor. We coded a random sample of federal courts of appeals decisions from 2013 to 2023. We found that the Chevron doctrine continued to be an active doctrine in the lower courts even when it seemed to observers to have lost life in the Supreme Court. To the extent that these courts took note of the controversy over Chevron, they typically responded like the Third Circuit did in a 2018 decision, where the court stated that “[t]he Chevron doctrine of deference to federal agencies is open to question, but it is the law.”⁵⁶

Our data provide some reason to think, however, that courts in the last few years began to be affected by the disfavoring signals concerning Chevron. Decisions discussing Chevron started sharply falling at that point, and there was also a noticeable drop in decisions applying the doctrine. It appears that the government was invoking the doctrine less often, perhaps because it knew that the doctrine was on thin ice in the Supreme Court. And our qualitative review suggests that at least some courts were becoming more resistant to the doctrine. An example is the Fifth Circuit’s 2023 en banc decision in Cargill v. Garland, which involved the meaning of “machinegun” in a criminal gun control statute.⁵⁷ Although the government did not invoke Chevron, the court of appeals made a point of explaining in detail why Chevron was inapplicable. In this decision and others, appellate courts or concurring or dissenting judges referenced the skepticism of particular Supreme Court justices toward the doctrine.

Disfavoring and Vertical Precedent

Under our decisional authority model, the Supreme Court may properly signal disfavor only through binding decisions that criticize, distinguish, or otherwise narrow an earlier decision. Supreme Court silence, changes in related areas, and methodological shifts should not be sufficient to signal that a precedent is in disfavor. This model should help guide lower courts in dealing with disfavored precedent going forward. We also offer thoughts on the tradeoffs between disfavoring and overruling precedent.

Lessons for Lower Courts

In some respects, the scheme of vertical precedent is working reasonably well. Our empirical analysis suggests that lower courts generally do not aim to “predict” whether the Supreme Court will overrule a decision. For the most part, lower courts continued to apply the Lemon and Chevron tests, even after they had lost favor with some justices, until those decisions were formally overruled in 2022 and 2024.

But we did find some important disagreement among lower court judges, particularly in recent years, about which Supreme Court signals should be considered. Some lower court judges seemed to be willing to look beyond binding decisions to declarations of a subset of the Court. In the wake of American Legion in 2019, several courts of appeals opted to “count heads” — adding up votes in the plurality opinion and two concurrences — to declare that Lemon was “dead.” In a 2024 opinion, a court of appeals judge relied heavily on Justice Gorsuch’s concurrence in Egbert in proclaiming that “Bivens claims are no longer judicially cognizable.”⁵⁸ And some appellate judges were influenced by the skepticism of Chevron expressed by individual justices.

We argue that lower courts should not enable a subset of the Court, in separate opinions, to change the law. Indeed, the Lemon case study underscores the trouble with that approach. Supreme Court precedent in this area was quite mixed. Although some justices disparaged Lemon for decades, a majority of the Court continued to apply the test from 1971 through at least 2005.

Likewise, the Supreme Court’s failure to cite a precedent tells lower federal courts very little about the continuing vitality of that precedent. The Court for lengthy periods of time ignored the Lemon test, leading scholars to speculate that the test had been abandoned. But then the Court would revive and rely on the Lemon test. The lower federal courts thus had good reason to continue to treat the test as governing law. The story of Lemon helps explain why the lower courts likewise reasonably opted to continue applying Chevron, despite the Supreme Court’s silence on the doctrine. The Court’s failure to cite a decision, even for considerable periods of time, does not necessarily signal that the precedent has been interred or abandoned.

Some other signals are also unreliable. Although the Supreme Court itself may take separate doctrinal changes into account in determining whether to adhere to a precedent, lower federal courts should not rely on such “tea leaves.” Any such reliance would be close to a prediction model and create a substantial risk of error. Nor should methodological shifts cause a precedent to be disfavored. The Court is never perfectly consistent in methodology across issue areas, so the mere lack of such consistency is not a reliable signal. There is also substantial debate about whether and to what extent the obligation of stare decisis, including vertical stare decisis, applies to methodology, and that is another reason for caution in the lower courts.

Tradeoffs in Disfavoring Versus Overturning

When the Supreme Court disfavors — without overturning — precedent, it should expect some disuniformity in the lower courts about the proper application of the precedent. This disuniformity, moreover, is likely to grow over time if the Court continues to leave the viability of a precedent in doubt. Such disuniformity has both benefits and costs.

Disagreement among the lower courts may be beneficial by providing valuable information to the Court about how to develop a doctrine and whether to retain it. In fact, sometimes the Supreme Court may allow the lower courts to experiment with a doctrine so that the Court may learn from the different approaches. The second era of Bivens from the mid-1990s to 2017 can be seen as an era of experimentation. The lower court decisions during that era provided lots of information — demonstrating how courts could use the existing doctrine both to deny most claims and to grant Bivens claims beyond the Fourth, Fifth, and Eighth amendment contexts of Bivens, Davis, and Carlson. The lower courts’ efforts to use other off-ramps to dispose of Bivens cases likely also signaled to the Court the confusion and uncertainty surrounding its existing doctrine. This lower court experimentation likely informed the Court’s decision in Ziglar to modify the test for Bivens claims.

Such disuniformity, however, can also undermine the predictable application of federal law, lead to unnecessary litigation, and strain judicial legitimacy. The Court offered more guidance to the lower courts when it modified the Bivens test in its 2017 Ziglar decision than it had through case-by-case rejections of Bivens claims in previous decades. Yet Ziglar and subsequent decisions have still left room for variation among the lower courts. For example, the Fifth Circuit in a 2021 decision suggested that any factual distinction between Bivens, Davis, or Carlson and a new case — including the difference between a home and a parking lot — was enough to create a “new context,”⁵⁹ while the Seventh Circuit in a 2023 decision described such factual distinctions — such as between a home and a hotel — as “trivial.”⁶⁰

The key point here is that there are tradeoffs between merely disfavoring a precedent and overturning it, and the Court should consider these tradeoffs in deciding what to do with precedents that it no longer supports. Sometimes, failing to take the formal step of overruling allows for helpful experimentation and feedback from the lower courts or properly accounts for institutional differences in the lower courts. At other times, it simply generates confusion and needless litigation. There is also an issue of accountability to be considered: Given that the Court was responsible for generating the initial precedent, it should arguably bear some responsibility for determining its path forward, rather than continuing to send mixed signals to the lower courts.

---

CURTIS BRADLEY is the Allen M. Singer Distinguished Service Professor at the University of Chicago Law School. He teaches and writes about the federal judiciary, separation of powers issues, and U.S. foreign relations law.

TARA LEIGH GROVE is the Vinson & Elkins Chair in Law at the University of Texas School of Law. In 2021, Grove served on the Presidential Commission on the Supreme Court, a bipartisan commission that examined proposals for Court reform.

---

1. This article originally appeared as Curtis A. Bradley & Tara Leigh Grove, Disfavored Supreme Court Precedent in the Lower Federal Courts, 111 Va. L. Rev. 1353 (2025). Readers should refer to the original publication for further citations and discussion.
2. Michael C. Dorf, Prediction and the Rule of Law, 42 UCLA L. Rev. 651, 655 (1995).
3. Lamb's Chapel v. Ctr. Moriches Union Free Sch. Dist., 508 U.S. 384, 398 (1993) (Scalia, J., concurring in the judgment).
4. Loper Bright Enters. v. Raimondo, 144 S. Ct. 2244, 2269 (2024).
5. For those interested in our methodology: To assemble a list of court of appeals decisions for Lemon, we ran the following search in Westlaw in "Federal Courts of Appeals": adv: SY("establishment clause") & TE(lemon). There were 366 cases in the results. We placed them in an Excel file in reverse chronological order. We coded every fifth case from 1971 to 2023. For Bivens, we ran the following search in Westlaw in "Federal Courts of Appeals": adv: SY(Bivens). There were 1,610 cases. We placed them in an Excel file in reverse chronological order. We coded every tenth case from 1971 to 2014 and, because we were especially interested in recent trends, we coded every fifth case from 2015 to 2023. For Chevron, we coded federal appellate decisions from 2013 to 2023. We ran the following search in Westlaw in "Federal Courts of Appeals": adv: SY(defer!) & TE(chevron). Narrowing the results to cases decided beginning in 2013 produced a sample size of 288 cases. We placed them in an Excel file in reverse chronological order and coded every fifth case.
6. Lemon v. Kurtzman, 403 U.S. 602, 612–13 (1971).
7. See Lynch v. Donnelly, 465 U.S. 668, 687–88 (1984) (O'Connor, J., concurring) (noting that government endorsement of religion "sends a message to nonadherents that they are outsiders, not full members of the political community, and an accompanying message to adherents that they are insiders, favored members of the political community").
8. See Lee v. Weisman, 505 U.S. 577, 587–88, 594 (1992).
9. See Lamb's Chapel, 508 U.S. at 395–97.
10. McCreary Cnty. v. ACLU, 545 U.S. 844, 859–66 (2005).
11. See Van Orden v. Perry, 545 U.S. 677, 681, 685–86, 692 (2005) (plurality opinion).
12. Am. Legion v. Am. Humanist Ass'n, 139 S. Ct. 2067, 2074, 2080–85, 2090 (2019) (plurality opinion).
13. Kennedy v. Bremerton Sch. Dist., 142 S. Ct. 2407, 2427–28 (2022) (second alteration in original) (quoting Am. Legion, 139 S. Ct. at 2079–81).
14. Green v. Haskell Cnty. Bd. of Comm'rs, 568 F.3d 784, 796–98, 797 n.8 (10th Cir. 2009) (internal quotation marks omitted).
15. See, e.g., Kondrat'yev v. City of Pensacola, 949 F.3d 1319, 1326–27 (11th Cir. 2020).
16. Woodring v. Jackson Cnty., 986 F.3d 979, 1002, 1003 n.4 (7th Cir. 2021) (Hamilton, J., dissenting).
17. Bivens v. Six Unknown Named Agents of Fed. Bureau of Narcotics, 403 U.S. 388, 389, 395–97 (1971).
18. Davis v. Passman, 442 U.S. 228, 229–31, 248–49 (1979).
19. Carlson v. Green, 446 U.S. 14, 18–20, 23 (1980).
20. Correctional Services Corp. v. Malesko, 534 U.S. 61, at 63–65, 75 (Scalia, J., joined by Thomas, J., concurring).
21. See Ziglar v. Abbasi, 582 U.S. 120, 136–37, 139–40 (2017).
22. Hernández v. Mesa, 140 S. Ct. 735, 752–53 (2020) (Thomas, J., joined by Gorsuch, J., concurring).
23. Egbert v. Boule, 142 S. Ct. 1793, 1804–07 (2022).
24. Chevron U.S.A. Inc. v. Nat. Res. Def. Council Inc., 467 U.S. 837, 842–43 & n.11 (1984).
25. Loper Bright, 144 S. Ct. at 2257, 2272–73.
26. S.E.R.L v. Att'y Gen., 894 F.3d 535, 554–55 (3d Cir. 2018) (citation omitted).
27. Cargill v. Garland, 57 F.4th 447, 465–69 (5th Cir. 2023) (en banc) (explaining why Chevron does not apply).
28. Mohamed v. Jones, 100 F.4th 1214, 1235, 1237–38, 1244 (10th Cir. 2024) (Tymkovich, J., dissenting).
29. Byrd v. Lamb, 990 F.3d 879, 880–82 (5th Cir. 2021) (per curiam).
30. Snowden v. Henning, 72 F.4th 237, 239–40, 247 (7th Cir. 2023).
31. This article originally appeared as Curtis A. Bradley & Tara Leigh Grove, Disfavored Supreme Court Precedent in the Lower Federal Courts, 111 Va. L. Rev. 1353 (2025). Readers should refer to the original publication for further citations and discussion.
32. Michael C. Dorf, Prediction and the Rule of Law, 42 UCLA L. Rev. 651, 655 (1995).
33. Lamb’s Chapel v. Ctr. Moriches Union Free Sch. Dist., 508 U.S. 384, 398 (1993) (Scalia, J., concurring in the judgment).
34. Loper Bright Enters. v. Raimondo, 144 S. Ct. 2244, 2269 (2024).
35. For those interested in our methodology: To assemble a list of court of appeals decisions for Lemon, we ran the following search in Westlaw in “Federal Courts of Appeals”: adv: SY(“establishment clause”) & TE(lemon). There were 366 cases in the results. We placed them in an Excel file in reverse chronological order. We coded every fifth case from 1971 to 2023. For Bivens, we ran the following search in Westlaw in “Federal Courts of Appeals”: adv: SY(Bivens). There were 1,610 cases. We placed them in an Excel file in reverse chronological order. We coded every tenth case from 1971 to 2014 and, because we were especially interested in recent trends, we coded every fifth case from 2015 to 2023. For Chevron, we coded federal appellate decisions from 2013 to 2023. We ran the following search in Westlaw in “Federal Courts of Appeals”: adv: SY(defer!) & TE(chevron). Narrowing the results to cases decided beginning in 2013 produced a sample size of 288 cases. We placed them in an Excel file in reverse chronological order and coded every fifth case.
36. Lemon v. Kurtzman, 403 U.S. 602, 612–13 (1971).
37. See Lynch v. Donnelly, 465 U.S. 668, 687–88 (1984) (O’Connor, J., concurring) (noting that government endorsement of religion “sends a message to nonadherents that they are outsiders, not full members of the political community, and an accompanying message to adherents that they are insiders, favored members of the political community”).
38. See Lee v. Weisman, 505 U.S. 577, 587–88, 594 (1992).
39. See Lamb’s Chapel, 508 U.S. at 395–97.
40. McCreary Cnty. v. ACLU, 545 U.S. 844, 859–66 (2005).
41. See Van Orden v. Perry, 545 U.S. 677, 681, 685–86, 692 (2005) (plurality opinion).
42. Am. Legion v. Am. Humanist Ass’n, 139 S. Ct. 2067, 2074, 2080–85, 2090 (2019) (plurality opinion).
43. Kennedy v. Bremerton Sch. Dist., 142 S. Ct. 2407, 2427–28 (2022) (second alteration in original) (quoting Am. Legion, 139 S. Ct. at 2079–81).
44. Green v. Haskell Cnty. Bd. of Comm’rs, 568 F.3d 784, 796–98, 797 n.8 (10th Cir. 2009) (internal quotation marks omitted).
45. See, e.g., Kondrat’yev v. City of Pensacola, 949 F.3d 1319, 1326–27 (11th Cir. 2020).
46. Woodring v. Jackson Cnty., 986 F.3d 979, 1002, 1003 n.4 (7th Cir. 2021) (Hamilton, J., dissenting).
47. Bivens v. Six Unknown Named Agents of Fed. Bureau of Narcotics, 403 U.S. 388, 389, 395–97 (1971).
48. Davis v. Passman, 442 U.S. 228, 229–31, 248–49 (1979).
49. Carlson v. Green, 446 U.S. 14, 18–20, 23 (1980).
50. Correctional Services Corp. v. Malesko, 534 U.S. 61, at 63–65, 75 (Scalia, J., joined by Thomas, J., concurring).
51. See Ziglar v. Abbasi, 582 U.S. 120, 136–37, 139–40 (2017).
52. Hernández v. Mesa, 140 S. Ct. 735, 752–53 (2020) (Thomas, J., joined by Gorsuch, J., concurring).
53. Egbert v. Boule, 142 S. Ct. 1793, 1804–07 (2022).
54. Chevron U.S.A. Inc. v. Nat. Res. Def. Council Inc., 467 U.S. 837, 842–43 & n.11 (1984).
55. Loper Bright, 144 S. Ct. at 2257, 2272–73.
56. S.E.R.L v. Att’y Gen., 894 F.3d 535, 554–55 (3d Cir. 2018) (citation omitted).
57. Cargill v. Garland, 57 F.4th 447, 465–69 (5th Cir. 2023) (en banc) (explaining why Chevron does not apply).
58. Mohamed v. Jones, 100 F.4th 1214, 1235, 1237–38, 1244 (10th Cir. 2024) (Tymkovich, J., dissenting).
59. Byrd v. Lamb, 990 F.3d 879, 880–82 (5th Cir. 2021) (per curiam).
60. Snowden v. Henning, 72 F.4th 237, 239–40, 247 (7th Cir. 2023).