Implicit Rejection of Massachusetts v. EPA: The Prominence of the Major Questions Doctrine in Checks on EPA Power – Frances Williamson

Posted by on Aug 15, 2022 in Per Curiam

Download PDF

Implicit Rejection of Massachusetts v. EPA: The Prominence of the Major Questions Doctrine in Checks on EPA Power

Frances Williamson*

Over the past four decades, many landmark administrative law cases have involved challenges to environmental agency action.[1] Environmental regulation exemplifies the tensions within the administrative state: unelected agencies tackle highly technical problems with pervasive nationwide effects that impact industries and individuals alike. The Court’s recent decision in West Virginia v. Environmental Protection Agency is one such example.[2] In West Virginia, the Court considered whether the EPA’s broader conception of its authority to regulate “outside-the-fence”[3] is within the power granted by the Clean Air Act (CAA). The 6-3 opinion, authored by Chief Justice Roberts, takes aim at broad assertions of agency power in the absence of clear congressional intent through the major questions doctrine (MQD), a tool of statutory interpretation typically used as an appendage of Chevron analysis that the Court now relies upon  as a standalone principle.[4] The prominence of the MQD in West Virginia signals the Court’s growing skepticism of broad agency power and its willingness to use the doctrine in statutory interpretation. In addition, the majority and concurrence in West Virginia suggest that the current Court would have decided Massachusetts v. EPA,[5] another landmark environmental case, differently.[6]

I. Background

Section 111(a) of the CAA grants the EPA the power to regulate power plants as stationary sources under a standard called the “best system of emission reduction” (BSER).[7] Under this framework, the EPA applies a system of reduction, determined “best” by the EPA Administrator, to emissions from new and existing plants.[8] In its 2015 Clean Power Plan Rule (CPP), the EPA set the BSER for existing coal and gas power plants in a way that “caused a shift toward wind, solar, and natural gas.”[9] This approach provided the EPA with significant flexibility in the demands it placed on power plants—“in translating the BSER into an operational emissions limit, EPA could choose whether to require anything from a little generation shifting to a great deal.”[10] The EPA then settled on a program that forced the States to implement generation-shifting practices.[11]

Shortly after the 2016 presidential election, the Trump Administration repealed the rule and replaced the CPP with the Affordable Clean Energy (ACE) Rule.[12] The ACE Rule was challenged in the D.C. Court of Appeals in 2020, which determined that, contrary to what the Trump EPA asserted in its repeal of the CPP, the EPA possessed the authority to force generation shifting.[13] Shortly after this decision, and the 2020 election, President Biden’s EPA requested a stay of the court’s mandate so that the CPP did not “immediately go back into effect.”[14] However, several states challenged the EPA’s authority to regulate emissions under the CPP, culminating in West Virginia.

II. Opinion

In West Virginia, the Court considered whether the EPA had the authority to “restructure the Nation’s overall mix of electricity generation. . . [as] the ‘best system of emission reduction’ within the meaning of Section 111.”[15] Writing for the Court, Chief Justice Roberts asserted that this is a “major questions case”[16] because it involved an agency “asserting highly consequential power beyond what Congress could reasonably be understood to have granted.”[17] The majority asserted that the EPA constructed a regulatory framework devoid of clear statutory roots; basing its “newfound power in the vague language” of §111(d) of the CAA, the EPA concluded that this vague language “allowed it to adopt a regulatory program that Congress had conspicuously and repeatedly declined to enact itself.”[18] The Court stated that the EPA had never used a systematic approach like generation shifting that looked beyond individual source emissions reduction or technological standards, and Congress would not employ the “previously little-used backwater of Section 111(d)” to confer such authority on the EPA.[19] Essentially, the EPA’s interpretation of Section 111(d) gave the EPA unprecedented authority to demand “much greater reductions in emissions based on a very different kind of policy judgment: that it would be ‘best’ if coal made up a much smaller share of national electricity generation.”[20] The majority also described how the Court could not ignore the fact that Congress “considered and rejected” such a course of action “multiple times.”[21] To the question of “whether the ‘best system of emission reduction’ identified by EPA in the [CPP] was within the authority granted to the Agency in Section 111(d),”[22] the Court employed the MQD and answered with a resounding “no.”

III. Concurrence and Dissent

Justice Gorsuch’s concurrence further bolstered the history behind the MQD and traced its roots to the separation of powers doctrine.[23]  Justice Gorsuch first outlined four questions that identify whether agency action is based on a valid, clear statement by Congress.[24]  Ultimately, he determined that the EPA’s assertion of power in the form of the CPP fit each category, and despite the dissent’s criticism, Justice Gorsuch concluded that the Court’s decision returned power to the “people’s elected representative[s]”;[25] agency power could not be used to circumvent constitutional design.

In her dissent, Justice Kagan criticized the majority and concurrence, claiming that the majority flouted “judicial modesty” in an attempt to constrain the EPA and alleging that the concurrence failed to rigorously analyze the statutory text.[26] Justice Kagan stated that Congress used “capacious” language[27]  to permit the EPA to address “new and big” problems.[28] Specifically targeting the majority’s use of the MQD, Justice Kagan argued that the MQD was inappropriate in the instant case because there was no “misfit” between agency action and statute; she emphasized that the CPP “fits perfectly” within the EPA’s wheelhouse. [29] Justice Kagan concluded that simply answering “major issues of public policy” did not expose an agency to judicial scrutiny.

IV. Implications for Massachusetts v. EPA and the Administrative State

As the nation’s approach to climate change continues to be litigated, environmental law will become increasingly salient. The impact of the majority’s use of the MQD in West Virginia is not limited to power plant regulation; this case signaled a shift in the Court’s view of agency power. In the past, the MQD appeared only as an interpretative device under the Chevron doctrine,[30] where judges used it in the “step one” analysis to determine whether a “statute [was] silent or ambiguous.”[31] In West Virginia, the doctrine appeared without mention of Chevron.

The MQD’s prominence in West Virginia implicitly rejected the logic used in past administrative and environmental decisions, notably Massachusetts v. EPA (2007).[32] In Massachusetts, the Court rejected the EPA’s argument that it lacked the statutory authority to regulate greenhouses gases on a principle similar to the MQD; Congress would have spoken clearly if it intended the EPA to regulate greenhouse gases because of their political and economic significance.[33] The EPA relied on the Court’s decision in Brown & Williamson,[34] and argued that “imposing emission limitations on greenhouse gases would have even greater economic and political repercussions than regulating tobacco.”[35] The Court, however, rejected these arguments and determined that the “EPA’s reliance” on Brown & Williamson was “misplaced.”[36] The majority stated that the analogy failed because 1) while it was “unlikely that Congress meant to ban tobacco products,” it was plausible Congress meant to allow the EPA to “only regulate emissions;”[37] and 2) in Brown & Williamson, the Court “pointed to an unbroken series of congressional enactments that made sense only if adopted ‘against the backdrop of the FDA’s consistent and repeated statements that it lacked authority under the FDCA to regulate tobacco,’”[38] while the EPA had “not identified any congressional action that conflict[ed] with the regulation of greenhouse gases from new motor vehicles.”[39] For these reasons, the Court rejected the EPA’s reliance on Brown & Williamson and declared that the EPA had the authority to regulate greenhouse gases.

The result of West Virginia runs counter to the 2007 decision in Massachusetts, suggesting that today’s Court would decide Massachusetts differently. In West Virginia, the Court rejected its reasoning in Massachusetts; it construed EPA regulation of emissions as a major question and interpreted congressional silence as a limit rather than an authorization. First, the West Virginia Court departed from Massachusetts by holding that the EPA’s CPP posed a major question because it did not “only regulate emissions” but reconfigured the national landscape of power generation. Emphasizing that the EPA could demand “much greater reductions in emissions based on a very different kind of policy judgment” than that which Congress authorized, the Court diverged from its earlier reasoning in Massachusetts where it had rejected reliance on Brown & Williamson.[40] In Massachusetts, the Court wrote that there was nothing “counterintuitive to the notion that EPA can curtail the emission of substances” that were harming the climate, so a comparison with Brown & Williamson was unsuitable.[41] The Court highlighted that because there was no “mismatch” between regulating pollutants and emissions, unlike the FDA and tobacco, the EPA had not overreached its statutory authority.[42] The MQD as articulated by the West Virginia Court, however, did not rely on statutory “mismatch”;[43] despite Justice Kagan’s urging, the MQD weighed principles such as the separation of powers against agency interpretation.[44] While the Court of 2007 saw nothing incongruous between forcing the EPA to regulate an entirely new category of pollutants and Congressional ambiguity on the matter, the 2022 Court would view Congressional silence as a red flag; extreme authorizations of agency power require extremely clear language from Congress. Therefore, if faced with the same question the Court faced in Massachusetts, today’s Court would curb the EPA’s power in the absence of clear congressional authorization.

Second, in West Virginia, the Court departs from its reasoning in Massachusetts by interpreting congressional silence as a limit, rather than a blessing, on regulation. The West Virginia Court pointed to “an unbroken series of congressional enactments that made sense only if adopted ‘against the backdrop’” of the EPA’s inability to regulate beyond individual sources under Section 111(d) of the CAA;[45] the Court concluded that the petitioners had “identified . . . congressional action that conflicts with”[46] the CPP. The Court stated that the EPA did not succeed in its claim that the “vague language” of §111(d) of the CAA “allowed it to adopt a regulatory program that Congress had conspicuously and repeatedly declined to enact itself.”[47] The majority highlighted Congress’s reluctance to impose the kind of regulatory scheme outlined in the CPP as well as the absence of “outside-the-fence” regulation of power plants.[48] The Court did not ignore that Congress “considered and rejected” such a course of action “multiple times” and that a systematic approach like generation-shifting remains unprecedented under the CAA.[49] In Massachusetts, on the other hand, the Court seemed content to state that “the EPA ha[d] not identified any congressional action that conflict[ed] in any way with the regulation of greenhouse gases from new motor vehicles,” another reason why the petitioners should not have relied on Brown & Williamson, where Congress denied an agency the ability to regulate tobacco.[50] The Massachusetts Court saw the EPA’s failure to “disavow[] the authority to regulate greenhouse gases” as an open door through which additional regulatory power could enter.[51] The West Virginia Court took the opposite stance and recognized the similarities between the EPA’s actions and those of the FDA in Brown & Williamson; the West Virginia Court concluded that Congress’s failure to avow the agency’s authority to regulate never opened the door in the first place. The West Virginia Court would view the major policy implications and congressional silence in Massachusetts as a resounding “no” to regulation, not a “yes.”

The petitioners in West Virginia raised similar points as the petitioners in Massachusetts, but this time, the Court accepted the arguments and placed the Brown & Williamson decision at the forefront of a “series of significant cases” that address the “particular and recurring problem”[52] of agency aggrandizement of power. In doing so, the Court embraced Brown & Williamson as a limit on EPA power. While the Court did not name the MQD in its analysis in Massachusetts, the EPA employed the principles underlying the doctrine—the absence of explicit Congressional authorization and the unique political history of climate change.

V. Implications of Justice Gorsuch’s Concurrence for Massachusetts v. EPA and the Administrative State

Justice Gorsuch’s proposed analysis of the MQD differed from Chief Justice Roberts’s two-step analysis by forcing a more rigorous examination of the potential ambiguity in the statutory text. In West Virginia, Chief Justice Roberts performed (in the words of Justice Kagan) a two-step MQD analysis, first asking if the agency action qualifies as an “extraordinary case” and then asking where the agency finds its “clear congressional authorization” for its action.[53] It was unclear which method the Court adopted in Massachusetts; Justice Kagan remained convinced that past Courts only used a “common sense” approach that analyzed “the fit between the power claimed, the agency claiming it, and the broader statutory design.”[54] Despite Justice Kagan’s assertion, Chief Justice Roberts’s two-step approach mapped onto the Court’s analysis in Massachusetts; the difference is that the Massachusetts Court stopped at step one: the question of whether the issue of greenhouse gas regulation was an “extraordinary case.” Because there was no “mismatch” between the EPA’s authority and the agency action at-issue, the Massachusetts Court did not view the question as “extraordinary” and stopped at the first step of Chief Justice Roberts’s major questions analysis.[55] Although the Court addressed congressional authorization, it did so only by pointing to the absence of congressional disapproval. The Massachusetts Court did not rigorously analyze the second element (whether Congress expressly authorized the agency action) after they determined that the issue was not extraordinary.

Justice Gorsuch proposed four questions that courts must ask when performing major questions analysis:

  • Is the language in the legislative provision “oblique” or obscure?[56]
  • Does the “age and focus of the statute the agency invokes” align with the “problem the agency seeks to address”?[57]
  • Does the desired interpretation match the agency’s “past interpretations of the relevant statute”?[58]
  • Is there is a “mismatch” between the “agency’s challenged action and its congressionally assigned mission and expertise”?[59]

Justice Gorsuch inverted the framework used by Chief Justice Roberts and the Massachusetts Court by grounding the question of congressional authority in the first step, not the second. Instead of simply asking whether there is a congressional authorization after the determination of “extraordinariness,” Justice Gorsuch made the inquiry the threshold question—if the language was “oblique,” then the Court must become skeptical of whether there was legitimate congressional delegation. The second, third, and fourth questions of Justice Gorsuch’s framework tackled the traditional first question—answering whether the agency action was “extraordinary” by analyzing its history, past interpretations, and alignment with the agency’s purpose.

If the Court analyzed Massachusetts under Justice Gorsuch’s framework, it would likely reach a different result than that reached in 2007.

First, under Justice Gorsuch’s framework the Massachusetts Court would probably consider the language used in the legislative provision at issue in Massachusetts as “oblique”;[60] indeed, in its 2007 opinion the Massachusetts Court acknowledged that the CAA provided a “capacious definition of ‘air pollutant.’”[61] The use of the word “capacious” indicated that the 2007 Court knew that Congress did not speak with sharp clarity when defining “air pollutant” in the CAA. Therefore, analyzing Massachusetts under the first factor of Justice Gorsuch’s framework, the Court would likely view the CAA’s language as “oblique.”

Second, the Court would probably reason that the “age and focus of the statute the agency invokes” does not align with the “problem the agency seeks to address.”[62] The Massachusetts Court stated that “the Congresses that drafted §202(a)(1) might not have appreciated the possibility [of] global warming,” and thus Congress had made an “intentional effort to confer the flexibility necessary to forestall such obsolescence.”[63] Justice Gorsuch’s third factor, however, would likely highlight such a lack of foreseeability as evidence that the age of the statute does not align with the agency’s attempted use of the statute. In West Virginia, Justice Gorsuch used the example of the forty-year gap between the adoption of OSHA provisions and the COVID-19 pandemic as illustrative of a statute that does not align with the problem it seeks to address.[64] Similarly, in Massachusetts, the provision of the CAA the Court cited was first passed in 1977 and amended in 1990, decades before the climate change issue addressed in the case arose. Therefore, the age of the statute did not align with the problem the EPA sought to address, failing Justice Gorsuch’s second factor.

Third, under Justice Gorsuch’s framework the Court would probably conclude that the EPA’s desired statutory interpretation in Massachusetts did not match the agency’s “past interpretations.”[65] After debating this point, the Massachusetts Court concluded that the “EPA had never disavowed the authority to regulate greenhouse gases, and in 1998 it in fact affirmed that it had such authority.”[66] The EPA’s one-time 1998 affirmation of this regulatory authority, however, contrasted starkly with the EPA’s interpretation that it definitely lacked the authority to do so. Thus, the Massachusetts EPA seemed to fail Justice Gorsuch’s third factor of matching agency “past interpretations.”[67]

Fourth, under Justice Gorsuch’s framework, the Court would probably conclude that there was a “mismatch” between the Massachusetts EPA’s “challenged action and its congressionally assigned mission and expertise.”[68] The 2007 Court determined that the CAA’s definition of “air pollutant” “embrace[d] all airborne compounds of whatever stripe,” a determination that granted the EPA broad authority to regulate emissions across almost any industry.[69] While Justice Kagan would likely agree with the Massachusetts Court that there was no “mismatch” between EPA authority and the regulation of emissions from (essentially) the entire automotive industry,[70] today’s Court disagreed: the majority saw the economic and industrial significance of generation-shifting as outside the EPA’s wheelhouse—they lacked the expertise necessary to redefine how the entire nation receives energy.[71] As in Massachusetts, the EPA in West Virginia was tasked with regulation that involved policymaking tools they traditionally did not use.[72] Under Justice Gorsuch’s fourth factor, today’s Court would see this break from tradition as a sign that the EPA acted in a way that did not match its statutory authority.

Thus, Justice Gorsuch’s MQD framework strongly suggests that today’s Court would rule differently in Massachusetts. Justice Gorsuch’s framework also indicates that today’s Court is focused on the more demanding threshold question of congressional ambiguity rather than the abstract idea of agency-statute “mismatch.” Justice Gorsuch’s concurrence revealed the intensified rigor of the major questions analysis, and it signaled that pervasive assertions of agency authority would be met with pervasive skepticism.


West Virginia “announce[d] the arrival” of the major questions doctrine as a standalone doctrine in power plant regulation, environmental law, and administrative law.[73] Advocates should look to this case as foreshadowing the Court’s approach to non-delegation and agency authority. Some may say this opinion came as no surprise; some of the Justices have expressed skepticism about expansive delegation in the past.[74] However, the birth of the MQD as an independent tool of statutory interpretation does not simply mark a shift in environmental jurisprudence. Instead, it undermines regulatory precedent by chipping away at the logic of Massachusetts, provides a new tactic for opponents of the administrative state, and signals that today’s Court is willing to tackle the problem that broad delegation poses to foundational constitutional principles.

* Harvard Law School Class of 2023.

[1] See Chevron U.S.A., Inc. v. Nat. Res. Def. Council, 467 U.S. 837 (1984); Massachusetts v. EPA, 549 U. S. 497 (2007); Utility Air Regulatory Group v. EPA, 573 U.S. 302 (2014).

[2] West Virginia v. EPA, 142 S.Ct. 2587 (2022).

[3] Transcript of Oral Argument at 12, West Virginia v. EPA, 142 S.Ct. 2587 (2022) (No. 20-1530). The phrase “outside the fence” essentially references regulation of power plant emissions that goes beyond the plant itself—i.e., the regulation does not impact the infrastructure within the fence-line of the single power plant.

[4] The MQD has appeared, in different forms, in recent cases before the Court. In the per curiam opinion Alabama Association of Realtors v. Department of Health and Human Services, the Court based its decision, in part, on the principle that Congress did not clearly authorize the Centers for Disease Control and Prevention (CDC) to exercise such significant, broad power. 141 S.Ct. 2485 (2021) (per curiam). The Court wrote “[w]e expect Congress to speak clearly when authorizing an agency to exercise powers of ‘vast economic and political significance.’” Alabama Ass’n of Realtors, 141 S.Ct. 2485, 2489 (2021) (per curiam) (citing Utility Air Regulatory Group, 573 U.S. at 324 (2014) (quoting FDA v. Brown & Williamson Tobacco Corp., 529 U. S. 120, 160 (2000))) (internal quotations omitted). Even if the statute had been ambiguous (the Court determined it was not), because Congress did not explicitly grant the CDC the power it claimed, the CDC could not exercise the “breathtaking amount of authority” to regulate landlord-tenant relationships across the nation. Id. Although not named, the MQD accompanied the Court’s statutory analysis and bolstered the conclusion that the CDC exceeded its authority.

Within six months, the Court issued another per curiam opinion in January 2022, National Federation of Independent Business v. Department of Labor, Occupational Safety & Health Administration, where it determined that the Occupational Health and Safety Administration (OSHA) exceeded its statutory authority in issuing a vaccine mandate for all who work for an employer with 100 or more employees. 141 S.Ct. 661 (2022) (per curiam). The opinion only referenced the MQD. Id. at 667. However, Justice Gorsuch’s concurrence named the MQD as the animating principle behind the decision and distinguished the major questions doctrine from the closely related nondelegation doctrine. Id. at 669 (“The nondelegation doctrine ensures democratic accountability by preventing Congress from intentionally delegating its legislative powers to unelected officials.”). Justice Gorsuch connected the MQD with the separation of powers, writing that the federal government must “properly invoke a constitutionally enumerated source of authority to regulate in this area or any other” and “act consistently with the Constitution’s separation of powers.” Id. at 667.

[5] 549 U.S. 497 (2007).

[6] Chief Justice Roberts, Justices Thomas, Alito, and Breyer were all on the Court at the time Massachusetts v. EPA was decided. Chief Justice Roberts authored a dissenting opinion that argued that the plaintiffs lacked standing; he was joined by Justices Thomas, Alito, and Scalia.

[7] 42 U.S.C. §7411(a)(1) (“The term ‘standard of performance’ means a standard for emissions of air pollutants which reflects the degree of emission limitation achievable through the application of the best system of emission reduction which (taking into account the cost of achieving such reduction and any nonair quality health and environmental impact and energy requirements) the Administrator determines has been adequately demonstrated.”).

[8] See 42 U.S.C. §7411(b)(1); 42 U. S. C. §7411(b)(1).

[9] West Virginia, 142 S.Ct. at 2601.

[10] Id.

[11] Id. at 2604.

[12] 84 Fed. Reg. 32,520 (July 8, 2019) (since vacated). The Court in West Virginia describes that the CPP was repealed on the grounds that it exceeded its statutory authority. 142 S.Ct. at 2604. The subsequent ACE Rule determined that the BSER only required minor, source-specific facility upgrades. Id.

[13] West Virginia, 142 S.Ct. at 2605. The MQD was cited by the Trump Administration EPA as a reason for the repeal of the CPP. 84 Fed. Reg. 32,520 at 32,529 (July 8, 2019).

[14] West Virginia, 142 S.Ct.. at 2606 (“The court’s decision, handed down on January 19, 2021, was quickly followed by another change in Presidential administrations. One month later, EPA moved the Court of Appeals to partially stay the issuance of its mandate as it pertained to the Clean Power Plan.”). The EPA chose to stay the reimplementation of the CPP because they were reconsidering promulgating a new §111(d) of the CAA. Id. The “EPA believed that such a result would not make sense while it was in the process of considering whether to promulgate a new Section 111(d) rule.” Id.

[15] Id. at 2606. The Court also addressed the issue of standing. The Court first held that the states had standing because they were injured by the Court of Appeals’ judgment which vacated the ACE Rule and seemed to reinstate the CPP: “to the extent the Clean Power Plan harms the States, the D.C. Circuit’s judgment inflicts the same injury.” Id. The Court did not find the government’s mootness argument persuasive—the case was not mooted just because the EPA voluntarily decided not to enforce the CPP, partially because the EPA actively defended the “legality of [the CPP’s] approach.” Id. at 2607.

[16] Chief Justice Roberts defended the major questions “thread” from attack by the dissent, claiming that cases like FDA v. Brown & Williamson, 529 U.S. 120 (2000) (a case comparison the Court rejected in Massachusetts v. EPA, 549 U.S. 497 (2007)), Utility Air Regulatory Group v. EPA, 573 U.S. 302 (2014), and the recent case NFIB v. OSHA, 142 S.Ct. 661 (2022), support the relevance and validity of major questions analysis in the space of agency authority.

[17] West Virginia, 142 S.Ct. at 2609.

[18] Id. at 2610.

[19] Id. at 2613.

[20] Id. at 2612.

[21] Id. at 2614 (citing Brown & Williamson, 529 U.S. at 144).

[22] Id. at 2616.

[23] Id. at 2616 (2022) (Gorsuch, J., concurring). Justice Gorsuch specifically emphasized the importance of the major questions doctrine with the rise of the administrative state in the 1970s, writing that the “constitutional lines at stake here are surely no less important than those this Court has long held sufficient to justify parallel clear-statement rules.” Id. at 2620.

[24] Id. at 2622.

[25] Id. at 2624.

[26] West Virginia, 142 S.Ct. at 2633 (Kagan, J., dissenting); see also id. at 2641 n.8 (“Nowhere will you find the concurrence ask: What does the phrase ‘best system of emission reduction’ mean? §7411(a)(1). So much for ‘begin[ning], as we must, with a careful examination of the statutory text.’”) (citing Henson v. Santander Consumer USA Inc., 137 S.Ct. 1718 (2017)).

[27] Id. at 2632.

[28] Id. at 2628.

[29] Id. at 2633.

[30] The majority does not employ the Chevron doctrine, or mention the case, in the opinion. See West Virginia, 142 S.Ct. 2599.

[31] Chevron, 467 U.S. at 843. This language may conjure images of other notable doctrines and axioms of interpretation—notably, the nondelegation doctrine and the assertion that Congress does not “hide elephants in mouseholes.” Whitman v. Am. Trucking Ass’ns, Inc., 531 U.S. 457, 468 (2001).

[32] 549 U.S. 497. Only Justice Kagan referenced the case in her dissent; the majority did not mention the case, even in reference.

[33] 549 U.S. at 512.

[34] 529 U.S. at 120.

[35] Massachusetts, 549 U.S. at 512.

[36] Id. at 530.

[37] Id. at 531.

[38] Id. (internal citation omitted).

[39] Id.

[40] West Virginia, 142 S.Ct. at 2612.

[41] Massachusetts, 549 U.S. at 531.

[42] West Virginia, 142 S.Ct. at 2624 (Gorsuch, J., concurring).

[43] Id.

[44] West Virginia, 142 S.Ct. at 2637 (Kagan, J., dissenting) (“It is EPA (that’s the Environmental Protection Agency, in case the majority forgot) acting to address the greatest environmental challenge of our time.”).

[45] Massachusetts, 127 S.Ct. at 1461 (internal citation omitted).

[46] Id.

[47] West Virginia, 142 S.Ct. at 2595.

[48] Transcript of Oral Argument at 12, West Virginia v. EPA, 142 S.Ct. 2587 (2022) (No. 20-1530).

[49] West Virginia, 142 S.Ct. at 2614 (internal citation omitted).

[50] Massachusetts, 127 S.Ct. at 1461.

[51] Id.

[52] West Virginia, 142 S.Ct. at 2609.

[53] Id. at 2608–10 (internal quotations omitted).

[54] West Virginia, 142 S.Ct. at 2634 (Kagan, J., dissenting).

[55] The Massachusetts Court held that “there [was] nothing counterintuitive to the notion that EPA [could] curtail the emission of substances that [were] putting the global climate out of kilter,” Massachusetts, 127 S.Ct. at 1461, and then succinctly determined that “[t]here [was] no reason . . . to accept EPA’s invitation to read ambiguity into a clear statute.” Id.

[56] West Virginia, 142 S.Ct. at 2620 (Gorsuch, J., concurring).

[57] Id. at 2623. (Gorsuch, J., concurring).

[58] Id.

[59] Id.

[60] West Virginia, 142 S.Ct. at 2620 (Gorsuch, J., concurring).

[61] Massachusetts, 127 S.Ct. at 1462.

[62] West Virginia, 142 S.Ct. at 2623 (Gorsuch, J., concurring).

[63] Massachusetts, 127 S.Ct. at 1462.

[64] West Virginia, 142 S.Ct. at 2623 (Gorsuch, J., concurring).

[65] Id.

[66] Massachusetts, 127 S.Ct. at 1461.

[67] West Virginia, 142 S.Ct. at 2623 (Gorsuch, J., concurring).

[68] Id. at 2623 (Gorsuch, J., concurring).

[69] Massachusetts, 127 S.Ct. at 1443. In addition, the 2007 Court disregarded the EPA’s argument that its actions impeded on authority traditionally reserved for the Department of Transportation by writing that the obligations of the two agencies “may overlap, but there is no reason to think [they] cannot both administer their obligations and yet avoid inconsistency.” Id. at 1462.

[70] West Virginia, 142 S.Ct. at 2637 (Kagan, J., dissenting) (“the ‘how’ of generation shifting creates no mis-match with EPA’s expertise.”).

[71] Justice Gorsuch referenced the majority’s phrasing: “requires technical and policy expertise not traditionally needed in [the] EPA’s regulatory development.” Id. at 2624 (Gorsuch, J., concurring).

[72] Id.

[73] 142 S.Ct. at 2633 (2022) (Kagan, J., dissenting).

[74] See, e.g., Gundy v. United States, 139 S.Ct. 2116 (2019) (Gorsuch, J., dissenting, joined by Roberts, C.J., and Thomas, J.).

Read More »

Berger v. North Carolina State Conference of the NAACP: A Victory for Federalism and State Autonomy – David Thompson

Posted by on Aug 11, 2022 in Per Curiam

Download PDF

Berger v. North Carolina State Conference of the NAACP: A Victory for Federalism and State Autonomy

David Thompson[1]

In a term defined by landmark opinions and culture war fodder, an eight to one Supreme Court opinion concerning a state legislature’s right to intervene may seem unremarkable. Indeed, in comparison to Dobbs,[2] Bruen,[3] and West Virginia,[4] the opinion in Berger[5] received little attention. The commentariat did not, this time, clutch their pearls and breathlessly wail about the demise of democracy. But for those who care about foundational principles of federalism, the Berger opinion is far from picayune or uninteresting. It is worthy of careful study and consideration.

The case stems from the decision of the people of North Carolina to require “[v]oters offering to vote in person” to “present photographic identification.”[6] North Carolinians voted to add this language to the state constitution in November 2018, additionally providing that “[t]he General Assembly shall enact general laws governing the requirements of such photographic identification, which may include exceptions.”[7] Acting in accordance with this charge, the General Assembly approved Senate Bill 824 (S.B. 824), effectuating the constitutional demand for a voter ID law.[8]

A pitched political battle ensued. The North Carolina Governor, Roy Cooper, a Democrat elected in 2016, vetoed the bill. The Republican-controlled General Assembly rebuffed him, overriding the veto and allowing the law to take effect on December 19, 2018. Not wasting any time, the North Carolina Chapter of the National Association for the Advancement of Colored People (NAACP) sued the very next day. The NAACP named the Governor and the members of the State Board of Elections (the Board) as defendants and alleged that S.B. 824 violated the Federal Constitution.[9]

The Attorney General, a former Democratic State Senator with a public history of opposing voter ID legislation and undermining a defense of the prior law, stepped in claiming an intent to defend the Board. The Board was, in turn, populated with appointees of the Governor. As this suit unfolded in the midst of a chaotic 2020 election cycle, the Board understandably had an overwhelming interest in seeking clarity about whether it should apply the state’s voter ID law. Its stated interest was to lend stability and clarity to the Board’s task of election administration rather than to vigorously defend the law itself.[10]

Seeking to defend their handiwork, the Speaker of the State House of Representatives and the President pro tempore of the State Senate (the legislative leaders) moved to intervene. Their primary interest was simple: they wanted to win. Despite this divergence of interests among state officials, applying a presumption that the Attorney General and the Board adequately represented the legislative leaders’ interests, the District Court denied the legislative leaders’ motion to intervene. As a result, the state’s most important interest—vindicating its laws—went without representation.[11] All this, despite North Carolina law explicitly permitting the two legislative houses to participate in the State’s defense.[12]

For those needing a refresher in civil procedure, the relevant language can be found in Federal Rule of Civil Procedure 24(a)(2). The district court “must permit anyone to intervene who” can show “an interest relating to . . . the subject of the action,” and who “is so situated that disposing of the action may as a practical matter impair or impede the movant’s ability to protect its interest, unless existing parties adequately represent that interest.”[13] The question of adequate representation in this context had split the Circuit Courts. In Berger, the District Court and, later, the Fourth Circuit, applied a strong presumption that the Attorney General adequately represented the State’s interests, thus keeping the legislative leaders out of the case. The Seventh Circuit applied an even stronger presumption of adequate representation when faced with a claim of intervention of right by state officials under similar circumstances, embracing a presumption that was only surmountable in instances of “gross negligence or bad faith.”[14] Elsewhere, the Sixth Circuit applies no such presumption—instead requiring only a minimal showing of inadequacy for a state official to intervene in circumstances similar to those here.[15]

The District Court and the Fourth Circuit erred grievously in finding that the Attorney General, on behalf of the Board, adequately represented the legislative leaders’ interests. These courts operated on the assumption that the Attorney General represented the State’s interest and thus the interests that the legislative leaders sought to vindicate. They assumed, wrongly, that the Board and the legislative leaders’ interests overlapped “fully.”[16] But these two groups  represented two distinct state interests: an interest in election administration and an interest in defending the law of the state, respectively.

North Carolina law recognizes that a tension often exists between legislative interests and administrative interests, the latter of which inherently rests with the Executive Branch. To ameliorate this tension, the State grants the legislative leaders “standing to intervene on behalf of the General Assembly as a party in any judicial proceeding challenging a North Carolina statute or provision of the North Carolina Constitution.”[17] The lower courts were wrong to ignore this directive of North Carolina law and assume that the Attorney General provided adequate representation of the legislative interest. North Carolina explicitly contemplates that the State’s interests may be best represented by multiple branches of government. To reject this understanding evinces a limited respect for the virtues of federalism and system of government with co-equal branches attending to different state interests.

Fortunately, the Supreme Court stood as a bulwark against this infringement on North Carolina’s ability to choose how it will be represented in court. The Court, in an 18-page opinion penned by Justice Gorsuch, agreed that the legislature “had claimed an interest” in the lawsuit “that may be practically impaired or impeded without their participation.”[18] The majority held that “when a State chooses to allocate authority among different officials who do not answer to one another, different interests and perspectives, all important to the administration of state government, may emerge.”[19] The Attorney General may, in circumstances like that here, represent an interest that diverges in some vital way from that of the legislature. As such, the Court recognized that, were it not allowed intervention of right, the legislature’s interest in defending its law could be seriously impaired.

After dispensing with this question, the Court had only to determine that the legislature was not adequately represented in the action. Such a finding flowed easily from the holding that the legislators and the Attorney General may seek to vindicate competing interests. The Court noted that “a State’s chosen representatives should be greeted in federal court with respect, not adverse presumptions.”[20] The lower courts had improperly relied on a presumption that the Board adequately represented the legislative leaders’ interests—and had found that the legislative leaders could not overcome this presumption. Setting aside this “adverse presumption,” the Court held that, “when a duly authorized state agent seeks to intervene to defend a state law,” the “presumption of adequate representation is inappropriate.”[21]

This ruling, though unlikely to be the subject of any campaign ads, confirms a basic tenet of our federal constitutional system: “[w]ithin wide constitutional bounds, States are free to structure themselves as they wish.”[22] Where state law, as here, clearly contemplates separate representation of the state’s competing interests, the courts would “do much violence to our system of cooperative federalism” were they to override the state’s selection of representatives.[23] The Court instead chose to give great deference to the choices a state has made in structuring its government. Here, each different state “agent” is expected to represent distinct interests that the State has: defense of the law and administration of the law. The Court, evincing the appropriate respect for federalism, agreed that the duly authorized agents should have their day in court.

Berger was not the only time last term when the Court championed this principle of federalism in a case involving intervention by state officials. Earlier in the term, in Cameron v. EMW Women’s Surgical Center,[24] the Court reversed a Sixth Circuit ruling that barred the Kentucky Attorney General from intervening in defense of an abortion law that the state Health Secretary had declined to defend on appeal. In the ruling, the Court acknowledged that a State may “empower multiple officials to defend its sovereign interests in federal court.”[25] The Court further noted that a state “clearly has a legitimate interest in the continued enforceability of its own statutes . . . and a State’s opportunity to defend its laws in federal court should not be lightly cut off.”[26] Finding that the lower court’s decision to keep the Attorney General from intervening improperly hampered the State’s ability to defend its interests in the manner which it had designed for itself, the Court permitted the Attorney General to intervene

Viewed in this light, Berger and Cameron are birds of a feather. In both cases, the Supreme Court effectuated a desire to respect the governmental structure the respective states have chosen for themselves. These decisions represent not just an important victory for federalism, but also an opportunity to have more views aired in federal litigation on key constitutional questions. In an age of divided government and political polarization, these rulings have significant implications. Where state law permits it, we may see an increase in the number of Rule 24(a)(2) motions on the part of legislatures – which, in many cases, have their own perspective on how best to vindicate vital state interests in litigation.

[1] Managing Partner, Cooper & Kirk, PLLC. Lead counsel for the Petitioner in Berger.

[2] Dobbs v. Jackson Women’s Health Org., 142 S.Ct. 2228 (2022).

[3] New York State Rifle & Pistol Ass’n, Inc. v. Bruen, 142 S.Ct. 2111 (2022).

[4] W. Virginia v. Env’t Prot. Agency, 142 S.Ct. 2587 (2022).

[5] Berger v. N. Carolina State Conf. of the NAACP, 142 S.Ct. 2191 (2022).

[6] N.C. Const. art. VI., § 2(4).

[7] Id.; Berger, 142 S.Ct. at 2197.

[8] Berger, 142 S.Ct. at 2197.

[9] Id. at 2197–98.

[10] Id. at 5, 16.

[11] Id. at 5–8.

[12] N.C. Gen. Stat. Ann. § 1-72.2 (2021).

[13] Fed. R. Civ. P. 24(a)(2) (emphasis added).

[14] Planned Parenthood of Wisconsin, Inc. v. Kaul, 942 F.3d 793, 799 (7th Cir. 2019).

[15] N.E. Ohio Coal. for Homeless v. Blackwell, 467 F.3d 999, 1007–08 (6th Cir. 2006); The Sixth Circuit is arguably closest to Supreme Court precedent in this area, as Trbovich v. Mine Workers established a minimal burden and promised intervention even where interests may overlap, but not fully. See Trbovich v. United Mine Workers of Am., 404 U.S. 528 (1972).

[16] Berger, 142 S.Ct. at 2197–98.

[17] N.C. Gen. Stat. Ann. § 1-72.2 (2021).

[18] Berger, 142 S.Ct at 2201.

[19] Id.

[20] Id. at 2205.

[21] Id. at 2204.

[22] Id. at 2197.

[23] Id. at 2214.

[24] Cameron v. EMW Women’s Surgical Ctr., P.S.C., 142 S.Ct. 1002 (2022).

[25] Id. at 1004.

[26] Id. at 1011.

Read More »

Kennedy v. Bremerton School District: The Final Demise of Lemon and the Future of the Establishment Clause – Daniel L. Chen

Posted by on Aug 8, 2022 in Per Curiam

Download PDF

Kennedy v. Bremerton School District: The Final Demise of Lemon and the Future of the Establishment Clause

Daniel L. Chen[1]

Nearly three decades ago, Justice Scalia famously lamented that the much-maligned test from Lemon v. Kurtzman[2] remained binding precedent: “Like some ghoul in a late-night horror movie that repeatedly sits up in its grave and shuffles abroad, after being repeatedly killed and buried, Lemon stalks our Establishment Clause jurisprudence.”[3] This past June, in Kennedy v. Bremerton School District, the Supreme Court finally interred Lemon once and for all, declaring that “this Court long ago abandoned Lemon and its endorsement test offshoot.”[4] Though the precise time of death is indeterminate, all nine members of the Court now agree that Lemon no longer governs.[5]

In place of Lemon’s “ambitious, abstract, and ahistorical approach,” the Court returned to “original meaning and history,” concluding that “the Establishment Clause must be interpreted by reference to historical practices and understandings.”[6] This nuanced historical approach not only offers the best way forward for resolving Establishment Clause controversies, but will also prove largely consistent with existing Supreme Court precedent.

I.   The Lemon Test and the Departure from History

Modern Establishment Clause jurisprudence began in 1947 with Everson v. Board of Education of Ewing, when the Supreme Court for the first time incorporated the Establishment Clause against the States.[7] From that time onwards, the Court looked primarily to historical practice to guide its Establishment Clause analyses. In Everson, although the majority and dissent disagreed about what precisely constituted a religious establishment, both sides agreed that history served as the touchstone for their inquiries.[8] This historical method dominated the Court’s Establishment Clause jurisprudence for decades.[9]

In 1971, in Lemon v. Kurtzman, the Court departed from this historical inquiry.[10] The case concerned an Establishment Clause challenge to Pennsylvania’s and Rhode Island’s statutes providing aid to nonpublic schools.[11] The Pennsylvania statute provided financial reimbursements to private schools for secular educational services including teachers’ salaries, textbooks, and educational materials.[12] The Rhode Island statute supplemented the salaries of teachers of secular subjects in private schools provided the teacher did not inculcate religion in his or her classes.[13] The Court held, in an 8-1 decision, that both laws violated the Establishment Clause because private parochial schools were able to obtain funding on equal footing with private secular schools.[14]

The Court began its analysis by stating “[t]he language of the Religion Clauses of the First Amendment is at best opaque,” and that it could “only dimly perceive the lines of demarcation in this extraordinarily sensitive area of constitutional law.”[15] But instead of even attempting the type of textual or historical inquiry long demanded by its precedents, the Court assessed “cumulative criteria developed by the Court” and “gleaned” a novel three-part test to govern all Establishment Clause cases.[16] The so-called Lemon test prohibited any government action that (1) lacks a secular purpose, (2) has the primary effect of advancing or inhibiting religion, or (3) excessively entangles the government in religion.[17] The Court held that Pennsylvania’s and Rhode Island’s statutes ran afoul of the third prong and invalidated both statutes.[18]

Before long, the Lemon test became the subject of withering criticism by commentators and jurists alike.[19] The test was inherently malleable: what was meant by “secular purpose,” what was the baseline to compare whether religion was advanced or inhibited, and what kinds of “entanglement” were permitted or prohibited? This baked-in indeterminacy led to unpredictability in results and conflicting (or nonexistent) guidance to lower courts. And the criticism came across the ideological spectrum. For example, Professor Jesse Choper summarized the Court’s early applications of Lemon thusly:[20]

[A] provision for therapeutic and diagnostic health services to parochial school pupils by public employees is invalid if provided in the parochial school,[21] but not if offered at a neutral site, even if in a mobile unit adjacent to the parochial school.[22] Reimbursement to parochial schools for the expense of administering teacher-prepared tests required by state law is invalid,[23] but the state may reimburse parochial schools for the expense of administering state-prepared tests.[24] The state may lend school textbooks to parochial school pupils because, the Court has explained, the books can be checked in advance for religious content and are “self-policing”;[25] but the state may not lend other seemingly self-policing instructional items such as tape recorders and maps.[26] The state may pay the cost of bus transportation to parochial schools,[27] which the Court has ruled are “permeated” with religion; but the state is forbidden to pay for field trip transportation visits “to governmental, industrial, cultural, and scientific centers designed to enrich the secular studies of students.[28]

The Court itself acknowledged this embarrassing jurisprudence but attempted to justify its decisions as “sacrific[ing] clarity and predictability for flexibility.”[29] But to many, this was but “a euphemism . . . for . . . the absence of any principled rationale.”[30] Indeed, over time, the Court was forced to overrule many of its early decisions based on Lemon.[31]

Given this state of affairs, the Court attempted to clarify its Establishment Clause jurisprudence by modifying Lemon’s “effects” prong, beginning with Justice O’Connor 1984 concurrence in Lynch v. Donnelly.[32] This new “endorsement” test asked whether a “reasonable observer” would consider the challenged government action to endorse religion. But this test fared no better.

Utilizing the endorsement test, the Court held, for instance, that a single crèche, surrounded by a “fence-and-floral frame,” bearing a plaque stating a private organization donated the display, and located in the “most public” part of a county courthouse was unconstitutional.[33] But what about a crèche located in the “heart of the shopping district” that displayed a banner reading “SEASONS GREETINGS” along with miniature and life-sized figures of Jesus, Mary, Joseph, angels, shepherds, kings, and animals—all surrounded by a Santa Claus house, reindeer, candy-striped poles, a Christmas tree, carolers, a clown, an elephant, a teddy bear, and hundreds of colored lights?[34] Constitutional, of course.[35]

Just like Lemon, the endorsement test relied on “little more than intuition and a tape measure,”[36] “unguided examination of marginalia,” and an “Establishment Clause geometry of crooked lines and wavering shapes.”[37] This was a jurisprudence in which “a judge [could] do little but announce his gestalt.”[38]

Given these shortcomings, the Court began departing from Lemon and the endorsement test, “repeatedly emphasiz[ing] [its] unwillingness to be confined to any single test or criterion in this sensitive area,”[39] and that Lemon’s three elements were “no more than helpful signposts.”[40] In fact, the Court has not applied Lemon in Establishment Clause cases in almost two decades, and in recent years, it has come back to focus on history.

II. The Return to History

The push to refocus on history first occurred in Marsh v. Chambers.[41] There, the Court upheld the practice of opening a state legislature session with a prayer by a chaplain paid with public funds, explaining that such practices were “deeply embedded in the history and tradition of this country,” such that “[f]rom colonial times through the founding of the Republic and ever since, the practice of legislative prayer has coexisted with the principles of disestablishment and religious freedom.”[42] Interestingly, Marsh was decided only twelve years after Lemon, but the Court did not mention Lemon, leading Justice Brennan to state in dissent that the Court was merely “carving out an exception to the Establishment Clause, rather than reshaping” it.[43]

Town of Greece v. Galloway came three decades later.[44] The decision expressly relied on Marsh to conclude that a municipality’s decision to open its monthly board meetings with a prayer did not violate the Establishment Clause.[45] To begin, the Court rejected the notion that it “carv[ed] out an exception” in Marsh and instead held that “Marsh stands for the proposition that it is not necessary to define the precise boundary of the Establishment Clause where history shows that the specific practice is permitted.”[46] “Any test the Court adopts must acknowledge a practice that was accepted by the Framers and has withstood the critical scrutiny of time and political change.”[47] In short, “the Establishment Clause must be interpreted by reference to historical practices and understandings.”[48]

Most recently, in American Legion v. American Humanist Association, a plurality of the Court explained that “the Lemon court ambitiously attempted to find a grand unified theory of the Establishment Clause,” but later cases had “taken a more modest approach that focuses on the particular issue at hand and looks to history for guidance.”[49] Notably, six members of the Court agreed Lemon did not govern, and the dissent never once invoked Lemon to justify its reasoning.[50]

III. Kennedy v. Bremerton School District

The Court’s decision in Kennedy thus enters the scene at a time when the Supreme Court had effectively overruled Lemon, yet lower courts had repeatedly failed to heed that instruction. The lower court opinion in Kennedy was a prime example of this.[51]

Joseph Kennedy, a high school football coach at Bremerton High School, made it his practice to quietly pray and give thanks at the conclusion of football games.[52] After shaking hands with players and coaches, Kennedy would take a knee at the 50-yard line and give a brief, quiet prayer.[53] Sometimes, Kennedy prayed on his own; other times, players would voluntarily join him; still other times, opposing players would join.[54] Separately, Kennedy would give motivational speeches with religious imagery and pray in the locker room with his players.[55]

The District eventually learned about Kennedy’s locker-room prayers and religious speeches and asked him to cease those practices.[56] Kennedy complied with the District’s request, but also felt pressure to abandon his own private practice of quiet, on-field post-game prayers.[57] Kennedy asked the District to allow him to continue this private religious expression, but the District refused.[58] Although it noted that Kennedy had complied with its previous request, it forbade him from engaging in any “overt actions that could appear to a reasonable observer to endorse prayer.”[59] Kennedy refused to cease his practices, and the District placed him on administrative leave.[60]

Kennedy sued under the Free Exercise Clause and the Free Speech Clause.[61] Kennedy sought a preliminary injunction but lost in the lower courts. The Supreme Court denied certiorari, but four Justices explained that their votes were based on the preliminary posture of the case and that the denial of certiorari should not be interpreted as agreement with the lower courts’ reasoning.[62]

The case went back down, and the Ninth Circuit again ruled against Kennedy.[63] First, the panel again rejected his Free Speech claim because it found that his expression qualified as government speech since it occurred on the field during his time as a government employee.[64] The Ninth Circuit also noted that even if Kennedy’s practices were private speech, the District had an “adequate justification” for its disciplinary measures: an “objective observer” would conclude that the District had “endorsed” his religious activity by refusing to censor it, thereby violating the Establishment Clause.[65] As to Kennedy’s Free Exercise claim, the District conceded that it targeted Kennedy specifically because his conduct was religious.[66] Nonetheless, the Ninth Circuit, applying the endorsement test, upheld the District’s actions.[67] It concluded that the District had satisfied strict scrutiny because had it failed to discipline Kennedy, the District would have violated the Establishment Clause.[68]

The Ninth Circuit denied a petition for rehearing en banc, but eleven judges dissented. Notably, Judge Nelson explained that “the Supreme Court ha[d] effectively killed Lemon,” so the panel’s reliance on that decision was misguided.[69]

The Supreme Court began its analysis with the Free Exercise Clause.[70] The majority explained that because the District restricted Kennedy’s activities “because of their religious character,” its actions were by definition not neutral.[71] Nor were the District’s actions generally applicable.[72] The District created a “bespoke requirement” that only applied to Kennedy’s religious exercise and then pretextually claimed that he had failed to supervise students after games, even though other coaching staff were not required to do so.[73]

What’s important here is that even amidst ongoing discussions about the state of Free Exercise law, how to analyze neutrality and general applicability,[74] and whether Employment Division v. Smith[75] should be overruled,[76] all members of the Court—including those in dissent—agreed that the District’s actions did not fall under Smith’s lenient standard.[77] Indeed, the District had to concede that its policies were not neutral or generally applicable.[78]

The Court then addressed Kennedy’s Free Speech claim.[79] The Court analyzed Kennedy’s situation under the Pickering-Garcetti framework, which sets a different, more lenient standard for restricting the speech rights of government employees.[80] It concluded that Kennedy’s prayers were his own private speech because they were not “‘ordinarily within the scope’ of his duties as a coach.”[81] Moreover, Kennedy would pray at times when other coaches were permitted to attend to personal matters, including checking sports scores on their phones and greeting friends in the stands.[82]

What’s interesting about the Court’s Free Speech analysis is that the Court was not as unified as on the Free Exercise analysis. To begin, Justice Kavanaugh did not join the Court’s opinion as to the Free Speech Clause at all.[83] This is significant: the Court’s Free Speech analysis explains in a footnote that because the prayer was private speech and could not be credited to the District, the Court did “not decide whether the Free Exercise Clause may sometimes demand a different analysis” under the Pickering-Garcetti framework.[84] Justice Kavanaugh’s refusal to join this part of the opinion raises the question whether he would subject Free Exercise claims to Pickering-Garcetti at all—which could mean that he believes religious expression is entitled to greater protection. This would align with his earlier stated views that government actions violating “the bedrock principle of religious equality” are unconstitutional and wholly distinguishable from cases “where the government itself is engaging in religious speech.”[85]

Justice Thomas’s concurrence is similar. Justice Thomas reiterated that the Court’s decision does not decide whether or how government employees’ Free Exercise rights may be different from those belonging to the general public.[86] But in so doing, he cited to a concurrence from Justice Scalia in Borough of Duryea, Pa. v. Guarnieri, which cautioned against importing a doctrine from the Free Speech Clause into the Petition Clause.[87] Justice Scalia’s concurrence states, and Justice Thomas quotes, that any limitations on a constitutional provision must be justified by the provision’s “history” and “tradition.”[88] And to top it off, Justice Thomas noted that “the Court has never before applied Pickering balancing to a claim brought under the Free Exercise Clause,” strongly indicating that he would not apply Pickering in such cases.[89]

Justice Alito’s one-paragraph concurrence is of a piece, clarifying his view of the free speech issue. Justice Alito stated that the speech at issue was “unlike that in any of our prior cases” and agreed that the Court did not decide what standard applied under the Free Speech Clause, instead holding that the District’s actions could not be justified by any standard.[90] That may indicate, consistent with Justice Kavanaugh and Justice Thomas, that religious exercise issues may be categorically different from free speech issues. And this may be the case because unlike free speech—where almost anything goes—religious exercise must be both religious and sincere,[91] meaning the types of actions protected by the Free Exercise Clause are far more limited than those protected by the Free Speech Clause.

This brings us to the Court’s Establishment Clause holding, arguably the most important part of the case. Because the Court found that the Free Exercise Clause and the Free Speech Clause protected Kennedy’s religious expression, the Court had to assess whether the Establishment Clause provided any justification for the District’s actions. The Court held it did not because there was “only the ‘mere shadow’ of a conflict” based on “a misconstruction of the Establishment Clause.”[92]

That “misconstruction” was the District’s and the Ninth Circuit’s reliance on Lemon and the endorsement test.[93] The Court reiterated that “the ‘shortcomings’ associated” with Lemon’s “ambitious, abstract, and ahistorical approach to the Establishment Clause became so apparent that this Court long ago abandoned Lemon and its endorsement test offshoot.”[94] Citing Town of Greece and American Legion, the Court stated that Lemon and the endorsement test had been supplanted by a test based on “historical practices and understandings.”[95]

The Court, however, did not explain precisely how the historical analysis cashes out. To be sure, the Court held that “a historically sensitive understanding of the Establishment Clause” must take “coercion” into account because “coercion . . . was among the foremost hallmarks of religious establishments the framers sought to prohibit when they adopted the First Amendment.”[96] And it concluded that on the facts here, the District had not demonstrated that Kennedy’s private, post-game prayers coerced any students into praying.[97] Indeed, the Court took pains to explain that “[t]he exercise in question involves . . . giving ‘thanks through prayer’ briefly and by himself ‘on the playing field’ at the conclusion of each game he coaches” and “does not involve leading prayers with the team or before any other captive audience.”[98]

But at first glance, the Kennedy opinion itself offers little more about the “hallmarks of religious establishments,” leading the dissent to claim that the majority’s “test offers essentially no guidance for school administrators.”[99] A closer examination of the opinion, however, strongly suggests the path forward for future Establishment Clause cases.

IV. So what is an Establishment of Religion?

In a section explaining that the Establishment Clause prohibits government coercion of religious exercise, Kennedy states that “[n]o doubt, too, coercion along these lines was among the foremost hallmarks of religious establishments the framers sought to prohibit when they adopted the First Amendment.”[100] Then, curiously, the opinion includes a footnote at the end of that sentence that includes four notable citations.[101] That footnote, footnote 5, is a cipher for interpreting how the Court interprets the Establishment Clause by reference to history and tradition.

The first citation is to a specific portion of Justice Scalia’s dissent in Lee v. Weisman, where he explains that “one of the hallmarks of historical establishments of religion was coercion of religious orthodoxy and of financial support by force of law and threat of penalty.”[102] Another citation concerns James Madison’s statements during the ratification debates, where he explained that Establishment Clause prohibited Congress from “establish[ing] a religion to which they would compel others to conform.”[103]

The remaining two sources are where things get really interesting. Those sources are Justice Gorsuch’s concurrence in Shurtleff v. City of Boston,[104] a case decided earlier this Term, and well-known scholarship authored by Professor Michael McConnell,[105] perhaps the leading law and religion scholar in the country.

In Shurtleff, the City of Boston created a public forum by permitting private groups to raise their own flags at City Hall.[106] Boston permitted all types of speakers to host their events and raise flags, never rejecting a single request until a religious group sought to raise a flag that included religious imagery.[107] Boston refused access to the religious group, asserting that permitting the group’s speech would endorse religion.[108] The Court rejected that argument, holding that Boston could not exclude speech based on the speech’s religious viewpoint.[109]

Though the majority opinion did not mention Lemon, Justice Gorsuch’s Shurtleff concurrence explained that Lemon was the root of the problem but had long been overruled, and in its place, courts must consult history.[110] Importantly, his concurrence stated that “our constitutional history contains some helpful hallmarks that localities and lower courts can rely on.”[111]

Citing to Professor McConnell’s scholarship and adopting that position in whole, Justice Gorsuch concluded that historical establishments “often bore certain other telling traits”: (1) “the government exerted control over the doctrine and personnel of the established church,” (2) “the government mandated attendance in the established church and punished people for failing to participate,” (3) “the government punished dissenting churches and individuals for their religious exercise,” (4) “the government restricted political participation by dissenters,” (5) “the government provided financial support for the established church, often in a way that preferred the established denomination over other churches,” and (6) “the government used the established church to carry out certain civil functions, often by giving the established church a monopoly over a specific function.”

Indeed, we know that the Kennedy opinion adopts these six hallmarks as the touchstone for future Establishment Clause challenges because it cites not just broadly to Justice Gorsuch’s concurrence, but specifically to the very pages containing this analysis.[113] And if there were any remaining doubt, footnote 5 refers specifically to Professor McConnell’s original scholarship laying out these six categories.[114] Thus, by incorporating the Shurtleff concurrence and Professor McConnell’s work, Kennedy makes clear that government conduct violates the Establishment Clause only when that conduct exhibits these historical characteristics of a religious establishment.[115]

Some commentators have claimed that any historical approach is lacking and would allow for religious indoctrination by public schools.[116] But the Kennedy opinion itself refutes this premise, explaining that the case concerned only Kennedy’s private prayers, not his locker-room sermons.[117] Indeed, the opinion strongly indicated that such cases would come out differently due to concerns about “a captive audience” and “compell[ing] attendance and participation in a religious exercise.”[118]

This makes sense under the historical approach. Professor McConnell has explained that “[t]he historical approach is consistent with the vast majority of the Court’s existing precedent, and indeed provides a better explanation for most of the cases.”[119] This includes Torcaso v. Watkins[120] because with test oaths, the government restricts political participation by dissenters.[121] It includes Larkin v. Grendel’s Den[122] because granting churches veto power over liquor licenses assigns civil authority to religious groups.[123] And it also includes Engel v. Vitale[124] because in school prayer, the government controls religious doctrine by composing an official prayer.[125]

It is important to note, then, that the historical approach will not be as disruptive as some claim.[126] And in its favor, the historical approach provides objectivity and predictability to the Establishment Clause analysis. Rather than “assume a baseline of complete secularism in government affairs,” which “is ahistoric, produces hostility toward religion, and impoverishes public culture,” “[a] more objective baseline consists of the body of historical practices that have been widely accepted throughout the nation’s history and are consistent with the historical meaning of the Establishment Clause.”[127]

To be sure, important questions remain unanswered. Most notably, is sharing a single characteristic of a historical religious establishment enough to render government conduct unconstitutional? The examples from Torcaso, Grendel’s Den, and Engel suggest that at least in some circumstances, yes, a single hallmark is enough. But that leads to additional questions: does it depend on the specific historical establishment? Are some hallmarks more important than others? And if more than a single hallmark is necessary, should a “sliding scale” approach apply whereby stronger showings on some hallmarks make up for weaker showings on others?

Kennedy doesn’t answer these questions, but that isn’t unusual. Often, when the Court announces a new rule of constitutional law, it provides a general principle that requires future elaboration. For example, even in the religious liberty context, lawyers need only look back ten years to Hosanna-Tabor Lutheran Church v. EEOC,[128] where the Supreme Court recognized the existence of the ministerial exception. There, the Court determined that the plaintiff’s formal title, the substance reflected in that title, the plaintiff’s own use of that title, and the important religious functions performed by the plaintiff all weighed in favor of concluding that she was a minister.[129]

Though the Court looked at these four factors, it did not fully explain whether all four were necessary or how they might relate to each other.[130] But it didn’t have to, as the case was easily resolved because the plaintiff fulfilled all four of them.[131] Over time, however, the lower courts applied and refined the factors and concluded that the fourth factor—the important religious functions performed by an employee—was the most important.[132] And before long, the Supreme Court confirmed this by holding that “the significance of th[e] factors” in Hosanna-Tabor “did not mean that they must be met—or even that they are necessarily important—in all other cases.”[133] Instead, “[w]hat matters, at bottom, is what an employee does.”[134]

Kennedy utilizes this same approach. This does create a measure of ambiguity, but it can also be considered “a commendable example of judicial minimalism” whereby “the Court decides this case, and states a general principle, but does not try to work out all its implications in advance, in the abstract.”[135] Instead, Establishment Clause jurisprudence will be decided in future cases with concrete facts, and it will require additional legal scholarship to further elaborate the contours of historical religious establishments.

[1] Daniel Chen is counsel at the Becket Fund for Religious Liberty. Becket filed an amicus brief in the case discussed in this article, but the views expressed here do not necessarily reflect the views of Becket or its clients. The author thanks his colleagues Lori Windham, Becky Ricketts, Eric Rassbach, and Hari Asuri for their thoughts and contributions. John Heo and Ari Spitzer provided excellent editorial assistance. Any errors are his own.

[2] 403 U.S. 602 (1971).

[3] Lamb’s Chapel v. Ctr. Moriches Union Free Sch. Dist., 508 U.S. 384, 398 (1993) (Scalia, J., concurring).

[4] 142 S. Ct. 2407, 2427 (2022); see also id. at 2449 (Sotomayor, J., dissenting) (“The Court now goes much further, overruling Lemon entirely and in all contexts.”).

[5] See id.

[6] Id. at 2428 (quoting Town of Greece, N.Y. v. Galloway, 572 U.S. 565, 576 (2014)).

[7] 330 U.S. 1 (1947).

[8] Compare id. at 9–15 (surveying history of religious establishments in England and the colonies) with id. at 33 (Rutledge, J., dissenting) (“No provision of the Constitution is more closely tied to or given content by its generating history than the religious clause of the First Amendment.”).

[9] See, e.g., McGowan v. Maryland, 366 U.S. 420, 437–40 (1961) (examining “the place of Sunday Closing Laws in the First Amendment’s history”); Torcaso v. Watkins, 367 U.S. 488, 490 (1961) (invalidating religious test oaths because they were one of the elements of “the formal or practical” religious establishments that “many of the early colonists left Europe and came here hoping to” escape); Sch. Dist. of Abington Twp. v. Schempp, 374 U.S. 203, 294 (1963) (Brennan, J., concurring) (“[T]he line we must draw between the permissible and the impermissible is one which accords with history and faithfully reflects the understanding of the Founding Fathers”); Walz v. Tax Comm’n of City of N.Y., 397 U.S. 664, 680 (1970) (upholding tax exemptions for churches based on “more than a century of our history and uninterrupted practice”); see also Shurtleff v. City of Boston, 142 S. Ct. 1583, 1606 n.6 (2022) (Gorsuch, J., concurring) (collecting these cases).

[10] 403 U.S. 602 (1971).

[11] Id. at 607–610.

[12] Id.

[13] Id.

[14] Id. at 625.

[15] Id. at 612.

[16] Id.

[17] Id.

[18] Id. at 614–22.

[19] See, e.g., Mark V. Tushnet, Reflections on the Role of Purpose in the Jurisprudence of the Religion Clauses, 27 Wm. & Mary L. Rev. 997, 1004 (1986) (explaining that cases involving “‘deeply ingrained practices’” as “not readily susceptible to analysis under the ordinary Lemon approach”); Douglas Laycock, Towards a General Theory of the Religion Clauses: The Case of Church Labor Relations and the Right to Church Autonomy, 81 Colum. L. Rev. 1373, 1380–88 (1981) (criticizing the “unstructured expansiveness of the entanglement notion” and the potential that certain constructions of the effects prong may result in “the establishment clause threaten[ing] to swallow the free exercise clause”); Jesse H. Choper, The Religion Clauses of the First Amendment: Reconciling the Conflict, 41 U. Pitt. L. Rev. 673, 680–81 (1980); Lamb’s Chapel v. Ctr. Moriches Union Free School Dist., 508 U.S. 384, 398–99 (Scalia, J., concurring) (“[As of 1993], no fewer than five of the currently sitting Justices have” called for Lemon to be overruled, “and a sixth has joined an opinion doing so.” (collecting cases)).

[20] Choper, supra note 19, at 680–81.

[21] Meek v. Pittenger, 421 U.S. 349 (1975).

[22] Wolman v. Walter, 433 U.S. 229 (1977).

[23] Levitt v. Comm. for Pub. Educ., 413 U.S. 472 (1973).

[24] Comm. for Pub. Educ. v. Regan, 444 U.S. 646 (1980).

[25] Board of Educ. v. Allen, 392 U.S. 236 (1968).

[26] Meek, 421 U.S. 349.

[27] Everson v. Board of Educ., 330 U.S. 1 (1947).

[28] Wolman v. Walter, 433 U.S. 229, 252 (1977).

[29] Regan, 444. U.S. at 662.

[30] Choper, supra note 19, at 681.

[31] Michael W. McConnell, No More (Old) Symbol Cases, 2019 Cato Sup. Ct. Rev. 91, 104 (2019) (citing Agostini v. Felton, 521 U.S. 203 (1997) (overruling Sch. Dist. v. Ball, 473 U.S. 373 (1985) and Aguilar v. Felton, 473 U.S. 402 (1985)); Mitchell v. Helms, 530 U.S. 793 (2000) (plurality) (overruling Wolman v. Walter, 433 U.S. 229 (1977), and Meek v. Pittenger, 421 U.S. 349 (1975)).

[32] 465 U.S. 668, 688 (1984) (O’Connor, J., concurring); see, e.g., County of Allegheny v. ACLU, Greater Pittsburgh Chapter, 492 U.S. 573, 593 (1989); Allegheny, 492 U.S. at 630 (O’Connor, J., concurring in part and concurring in judgment).

[33] Allegheny, 492 U.S. at 579–81.

[34] Lynch, 465 U.S. at 671.

[35] Id.

[36] Allegheny, 492 U.S. at 675–76 (1989) (Kennedy, J., concurring in judgment in part and dissenting in part).

[37] Lamb’s Chapel v. Ctr. Moriches Union Free Sch. Dist., 508 U.S. 384, 399 (1993) (Scalia, J., concurring).

[38] Am. Jewish Congress v. City of Chicago, 827 F.2d 120, 129 (7th Cir. 1987) (Easterbrook, J., dissenting); see also Utah Highway Patrol Ass’n v. Am. Atheists, Inc., 132 S. Ct. 13, 14–15, 17, 19, 21–22 & n.3 (2011) (Thomas, J., dissenting from denial of certiorari) (“Establishment Clause jurisprudence [is] in shambles,” “nebulous,” “erratic,” “no principled basis,” “Establishment Clause purgatory,” “impenetrable,” “ad hoc patchwork,” “limbo,” “incapable of consistent application,” and a “mess.”).

[39] Lynch v. Donnelly, 465 U.S. 668, 679 (1984).

[40] Van Order v. Perry, 545 U.S. 677, 686 (2005) (quoting Hunt v. McNair, 413 U.S. 734, 741 (1973)).

[41] 463 U.S. 783 (1983).

[42] Id. at 786.

[43] Id. at 796 (Brennan, J., dissenting).

[44] 572 U.S. 565 (2014).

[45] Id. at 575.

[46] Id. at 575, 577.

[47] 572 U.S. at 577.

[48] Id. at 576 (emphasis added).

[49] Am. Legion v. Am. Humanist Ass’n, 139 S. Ct. 2067, 2087 (2019).

[50] 139 S. Ct. at 2103–13 (Ginsburg, J., dissenting).

[51] See Kennedy v. Bremerton Sch. Dist., 991 F.3d 1004 (9th Cir. 2021).

[52] Kennedy v. Bremerton Sch. Dist., 142 S. Ct. 2407, 2416 (2022).

[53] Id.

[54] Id.

[55] Id. The dissent asserted that these facts should have played a part in the majority’s reasoning. But at the time of its decision, the District only justified its actions based on Kennedy’s private post-game prayers. As the Court explained, “[g]overnment ‘justification[s]’ for interfering with First Amendment rights ‘must be genuine, not hypothesized or invented post hoc in response to litigation.’” Id. at 2432 (quoting United States v. Virginia, 518 U.S. 515, 533 (1996)).

[56] Id. at 2416.

[57] Id. at 2417.

[58] Id.

[59] Id. at 2418.

[60] Id. at 2418–19.

[61] Id. at 2419.

[62] 139 S. Ct. 643 (2019).

[63] 991 F.3d 1004 (9th Cir. 2021).

[64] Id. at 1014–16.

[65] Id. at 1018–19.

[66] Id. at 1020.

[67] Id. at 1020–21.

[68] Id.

[69] 4 F.4th 910, 945–46 (9th Cir. 2021) (Nelson, J., dissenting from the denial of rehearing en banc).

[70] 142 S. Ct.  at 2421.

[71] Id. at 2422.

[72] Id.

[73] Id. at 2423.

[74] See, e.g., Roman Catholic Diocese of Brooklyn v. Cuomo, 141 S. Ct. 63 (2020); Tandon v. Newsom, 141 S. Ct. 1294 (2021).

[75] 494 U.S. 872 (1990).

[76] Fulton v. City of Philadelphia, 141 S. Ct. 1868, 1883 (Alito, J., concurring).

[77] See Kennedy, 142 S. Ct. at 2426 n.3; id. at 2446 (Sotomayor, J., dissenting).

[78] Id. at 2422.

[79] Id. at 2423.

[80] Id.

[81] Id. at 2424.

[82] Id. at 2425.

[83] Id. at 2411.

[84] Id. at 2425 n.2.

[85] Morris Cty Bd. of Chosen Freeholders v. Freedom from Religion Found., 139 S. Ct. 909 (2019) (Kavanaugh, J., respecting the denial of certiorari); Shurtleff v. City of Boston, 142 S. Ct. 1583, 1594–95 (2022) (Kavanaugh, J., concurring).

[86] Kennedy, 142 S. Ct. at 2433 (Thomas, J., concurring).

[87] Id. (citing Borough of Duryea, Pa. v. Guarnieri, 564 U.S. 379. 405–06 (2011) (Scalia, J., concurring in part)).

[88] Kennedy, 142 S. Ct. at 2433 (Thomas, J., concurring).

[89] Id.

[90] Id. at 2433–34 (Alito, J., concurring).

[91] See, e.g., Wisconsin v. Yoder, 406 U.S. 205, 235 (1972) (“The Amish in this case have convincingly demonstrated the sincerity of their religious beliefs[.]”); id. at 216 (belief that is “philosophical and personal rather than religious . . . does not rise to the demands of the Religion Clauses.”).

[92] Kennedy, 142 S. Ct. at 2432.

[93] Id.

[94] Id. at 2427.

[95] Id. at 2428.

[96] Id. at 2429.

[97] Id.

[98] Id. at 2416; see also id. at 2432 n.7.

[99] Id. at 2450 (Sotomayor, J., dissenting).

[100] Id. at 2429.

[101] Id. at 2429 n.5.

[102] 505 U.S. 577, 640–42 (1992) (Scalia, J., concurring) (emphasis in original).

[103] 1 Annals of Cong. 730–31 (1789).

[104] 142 S. Ct. 1583, 1609 (Gorsuch, J., concurring).

[105] Michael W. McConnell, Establishment and Disestablishment at the Founding, Part I: Establishment of Religion, 44 Wm. & Mary L. Rev. 2105 (2003).

[106] Shurtleff v. City of Boston, 142 S. Ct. 1583 (2022).

[107] Id. at 1588.

[108] Id.

[109] Id. at 1593.

[110] Id. at 1604 (Gorsuch, J., concurring).

[111] Id. at 1609.

[112] Id. (citing McConnell, supra note 105, at 2131–81).

[113] Kennedy, 142 S. Ct. at 2429 n.5 (citing Shurtleff, 142 S. Ct. at 1609–10 (Gorsuch, J., concurring).

[114] Id. at 2429 n.5.

[115] Shurtleff, 142 S. Ct. at 1609–10 (Gorsuch, J., concurring); McConnell, supra note 105 at 2115–30 (explaining different hallmarks of establishment present in various degrees among the colonies).

[116] See, e.g., Mark Joseph Stern, Supreme Court Lets Public Schools Coerce Students Into Practicing Christianity, Slate (June 27, 2022, 4:19 PM),

[117] Kennedy, 142 S. Ct. at 2422.

[118] Id. at 2431–32 (cleaned up).

[119] Michael W. McConnell, The Supreme Court And The Cross, Hoover Institution (Mar. 1, 2019),

[120] 367 U.S. 488 (1961).

[121] McConnell, supra note 119.

[122] 459 U.S. 116 (1982).

[123] McConnell, supra note 119.

[124] 370 U.S. 421 (1962).

[125] McConnell, supra note 119. See also Shurtleff, 142 S. Ct. at 1609 (Gorsuch, J., concurring).

[126] Indeed, Lemon has already been overruled in its original context of assessing whether religious organizations can be included in public-benefit programs. The Court now determines whether the government program grants benefits based on “neutral, secular criteria” and whether there exists a “historic and substantial” tradition against including religious organizations. Zelman v. Simmons-Harris, 536 U.S. 639, 653 (2002); Espinoza v. Montana Dep’t of Revenue, 140 S. Ct. 2246, 2258 (2020).

[127] McConnell, supra note 119.

[128] 565 U.S. 171 (2012).

[129] Id. at 192.

[130] Id. at 190 (“We are reluctant, however, to adopt a rigid formula for deciding when an employee qualifies as a minister.”).

[131] Id. at 190 (“It is enough for us to conclude, in this our first case involving the ministerial exception, that the exception covers Perich, given all the circumstances of her employment.”).

[132] See, e.g., Temple Emanuel of Newton v. Mass. Comm’n Against Discrimination, 463 Mass. 472, 486, 975 N.E.2d 433, 443 (2012) (“Therefore, the ministerial exception applies to the school’s employment decision regardless whether a religious teacher is called a minister or holds any title of clergy.”); Cannata v. Cath. Diocese of Austin, 700 F.3d 169, 177 (5th Cir. 2012) (“Application of the exception, however, does not depend on a finding that Cannata satisfies the same considerations that motivated the Court to find that Perich was a minister within the meaning of the exception. Rather, it is enough to note that there is no genuine dispute that Cannata played an integral role in the celebration of Mass and that by playing the piano during services, Cannata furthered the mission of the church and helped convey its message to the congregants.”); Conlon v. InterVarsity Christian Fellowship, 777 F.3d 829, 835 (6th Cir. 2015) (“[W]e . . . hold that where both factors—formal title and religious function—are present, the ministerial exception clearly applies.”); Fratello v. Archdiocese of N.Y., 863 F.3d 190, 205 (2d Cir. 2017) (“Where, as here, the four considerations are relevant in a particular case, ‘courts should focus’ primarily ‘on the function[s] performed by persons who work for religious bodies.’” (citation omitted)); Grussgott v. Milwaukee Jewish Day Sch., Inc., 882 F.3d 655, 658 (7th Cir. 2018) (“[O]ther courts of appeals have explained that the same four considerations need not be present in every case involving the exception.”).

[133] Our Lady of Guadalupe Sch. v. Morrissey-Berru, 140 S. Ct. 2049, 2063, 207 L. Ed. 2d 870 (2020).

[134] Id. at 2064.

[135] Michael W. McConnell, Reflections on Hosanna-Tabor, 35 Harv. J. L. & Pub. Pol’y 821, 835 (2012).


Read More »

Carson v. Makin and the Relativity of Religious Neutrality – Lael Weinberger

Posted by on Aug 4, 2022 in Per Curiam

Download PDF

Carson v. Makin and the Relativity of Religious Neutrality

Lael Weinberger[1]

For years, religious education has produced controversies about law and religion. On the one hand, government support for or endorsement of religious education was the paradigm case for Establishment Clause violations in the mid-twentieth century.[2] On the other hand, government discrimination against religious believers on the basis of their faith has long been a paradigm case for a Free Exercise violation.[3] The Supreme Court has taken many cases in which these two principles are in apparent tension. Throughout this long line of cases, the idea of “neutrality” appears repeatedly. In the latest decision in this line, Carson v. Makin,[4] the majority and the dissent both claim to defend the true tradition of neutrality. But what each thinks neutrality means is quite different. The disagreement reveals the limits of neutrality as a guiding principle for Religion Clause jurisprudence.

I.               The Case

Over the years, school funding programs have created many opportunities to test the limits of the Religion Clauses. Carson v. Makin presents some of those familiar issues with a twist: the necessity of a school choice was, in a sense, forced upon parents. The case arose in Maine. A large percentage of the state is rural and sparsely inhabited. This creates challenges for the state’s “school administrative units.”[5] Less than half of these “school administrative units” have a secondary school (high school) of their own.[6]

To help parents navigate the challenge of not having a local high school, the state of Maine offers tuition funds to parents to cover the expenses of enrolling their children in private high schools of the parents’ choice.[7]

But Maine’s high school funding program excluded “sectarian” schools from eligibility under the program.[8] Parents who wanted to enroll their high-school-age children in religious high schools couldn’t receive any of the state funding—funding that would have been provided freely had the students been enrolled in nonreligious schools.

The plaintiffs in Carson v. Makin were two couples, both parents of high school students, who wanted to enroll their children in religious schools.[9] They sued the state, arguing that Maine was engaged in religious discrimination by refusing to make the tuition awards available to them simply on the ground that their children were attending religious schools.[10]

Presented this way, the case appeared to be an easy one. In 2017, the Supreme Court held in Trinity Lutheran that the Free Exercise Clause prohibited the state of Missouri from discriminating against religious schools when it made funding grants to improve school playgrounds available to private schools generally.[11] And in 2020, the Supreme Court held in Espinoza that the state of Montana’s prohibition on using scholarship funds to pay for education at religious schools violated Free Exercise Clause.[12] Surely Maine’s prohibition on using the tuition funds for religious schools of the parents’ own choosing would also violate the Free Exercise Clause.

But the First Circuit thought otherwise. It looked to Locke v. Davey, a 2004 case in which the Supreme Court held that the government could exclude students “pursuing a degree in devotional theology” from an otherwise generally-available state scholarship program.[13] There is potential for “tension” between the Religion Clauses, the Court opined, and there is also “play in the joints” in which states could address the tension in different ways.[14] It was fine—and not a Free Exercise problem—for a state to avoid an Establishment Clause problem by excluding this kind of exclusion of religious programs.[15] The First Circuit thought that Locke provided the right framework for evaluating the Maine program. It noted with approval that the Maine legislature had rested its policy “on its ‘interest in maintaining a religiously neutral public education system in which religious preference is not a factor.’”[16] It distinguished Trinity Lutheran and Espinoza on the ground that those cases prohibited discrimination on the basis of religious status but permitted limits on funds going to religious uses.[17] The religious school in Trinity Lutheran couldn’t be excluded from funding because it was religious, but the government could limit funding from supporting training in ministry, as in Locke. The First Circuit reasoned that the Maine program was more like Locke.[18]

Yet for the six-justice majority at the Supreme Court, the right frame of reference was Trinity Lutheran and Espinoza. According to Chief Justice Roberts, writing for the majority, Maine was not being neutral when it singled out religion as the characteristic that would disqualify a school from being eligible to receive public funds.[19] Writing for the dissent, Justice Breyer took the opposite tack with the same concept, neutrality. The state, he argued, would be best able to maintain a stance of neutrality, of being above or beyond the fray of religious sectarianism, if it is permitted to exclude religious educational institutions from public funding programs.[20]

Of course, each side had to do a bit of finagling the precedent, since neither urged any overruling. The majority managed to preserve Locke by insisting that it wasn’t an exclusion from a generally applicable grant program on the basis of religion, but was just a permissible decision by the state not to fund certain vocations. The dissent tried to reconcile its analysis with prior nondiscrimination cases by insisting that the Maine program was different in that it guaranteed a public education, which was effectively the same as guaranteeing a secular education.

II.             The Relativity of Neutrality

For decades, “neutrality” has been one of the recurring touchstones of Religion Clause jurisprudence and scholarship.[21] “In the relationship between man and religion, the State is firmly committed to a position of neutrality,” the Supreme Court proclaimed in 1963.[22] But neutrality has been a protean concept, having been used both as a justification for strict separation and as a critique of religion-excluding policies.

Neutrality’s protean character is on display again in Carson. The rhetoric of neutrality is powerful. Courts are deeply enmeshed in a tradition in which law is “neutral” in the sense of being impartial, unbiased, evenhanded, and fair. Courts are not supposed to prejudge issues. More broadly, a kind of evenhandedness is a fundamental component of rule of law norms. Laws that favor special interests are automatically suspect in the eyes of the public (even if the level of scrutiny to which they are constitutionally subjected has varied over time!). General applicability is a norm of lawmaking Constitution prohibits laws that prospectively target individuals rather than being generally applicable (e.g., bills of attainder).

In short, no position in American law benefits from being characterized as non-neutral. The position that seems to have the better claim on neutrality effectively seizes the rhetorical high ground.[23]

Both scholars and judges have argued (or at least assumed) that neutrality has special purchase in the domain of the Religion Clauses. The story told here is simple but compelling: governments have created social discord whenever they take sides in religious disputes. They have enforced conformity with a dominant religion and suppressed and persecuted dissent. In doing so, they have made religion into a point of political division, a motivation for political contestation and even unrest. The Religion Clauses of the First Amendment prohibit government establishment and guarantee free exercise. In light of this assumed background to the Religion Clauses, the argument goes, the framers had decided to take religion off the table of politics, thereby sparing the American polity from the divisiveness of religion as a flashpoint of controversy.[24]

But while this narrative is rhetorically compelling, it is simplistic. For one thing, it is not at all clear that the major concern among the founders was avoiding (say) something like the European wars of religion of the sixteenth and seventeenth centuries. And even if it was, it is not at all clear how to flesh out a guarantee of government neutrality in concrete terms.

Neutrality, it turns out, is a relative term rather than an absolute term, at least when it comes to religion. Consider ideas that circulated in the early republic. Supreme Court Justice Joseph Story famously expounded the First Amendment as requiring government neutrality only among Christian sects: “The real object of the amendment was, not to countenance, much less to advance Mahometanism, or Judaism, or infidelity, by prostrating Christianity; but to exclude all rivalry among Christian sects, and to prevent any national ecclesiastical establishment, which should give to an [sic] hierarchy the exclusive patronage of the national government.”[25]

Maybe Story was too narrow in his outlook. Maybe it wasn’t quite right to limit the matter to Christian sects. Maybe theistic sects would have been the better frame of reference. Many state constitutions required officeholders to affirm the existence of a divine being, but only one—South Carolina—explicitly established Christianity and required officeholders to specifically affirm the validity of the Christian Bible.[26]

One group widely presumed to not have much in the way of religious liberty protection was atheists, who were subject to various disabilities because they were presumed incapable of taking oaths. The rationale was that oath-taking was necessary for certain civic responsibilities, such as serving on a jury; but an oath was only credible if the oath taker actually believed in some system of divine retribution for violations of oaths.[27]

For many in the early republic, any religious liberty guarantee of neutrality was at best a neutrality among theistic believers.

By the mid-twentieth century, limiting religious liberty to theists was of course out of the question. This was the time, of course, when the Supreme Court incorporated the First Amendment against the states and began developing a jurisprudence of the Religion Clauses in earnest. Now the idea gained currency that the Religion Clauses prohibited the government from favoring, not only one religion over another, but religion over irreligion.[28] Neutrality now went a step further.

The problem that emerged over the twentieth century was that religion, like politics, gradually became visible everywhere. The personal became political. It also became religious, in this sense: religion and religious commitments have a way of coloring virtually every area of life if one looks hard enough.[29] Americans in the founding period or early republic were not likely to look very hard. They lived in a relatively homogeneous religious setting and were able to take a lot for granted—that their society was mostly not just theistic, but protestant; that public religious expressions were commonplace, among other things. But as America’s religious landscape became more diverse, less and less could be taken for granted.

For instance, most Americans at the turn of the nineteenth century were theists and found it unsurprising to have prayer in school or exhortations to religious duties in the curriculum. Noah Webster’s widely used spelling book, for instance, contained admonitions in its reading lessons such as:

  • “Rest in the Lord, and mind his word.”[30]
  • “No man may put off the law of God: My joy is in his law all the day.”[31]
  • “When they are at church they will sit, kneel, or stand still; and when they are at home, will read some good book, that God may bless them.”[32]

Of course there were religious minorities in America even in the early nineteenth century. A number of Americans did not attend church, and some theologically orthodox Christians may have objected on theological grounds to the moralistic message that (for instance) God will bless those who have proper posture at church. But the number of Americans who were offended by these exhortations was relatively small.

That number of dissenters would be larger by the mid twentieth century. Jewish Americans were still numerically a small minority, but they had become enough a part of the society’s consciousness that the term “Judeo-Christian” gained currency, and sociologist Will Herberg argued for a tripartite coalition of American religion consisting of (as the title of his influential book put it) Protestant – Catholic – Jew.[33]  The idea of public schools instructing children that church was normative became open to question in a way that it would not have been a generation earlier.

But one of the oddities that often was overlooked in the mid twentieth century was that religious choices besides traditional religion could still be religious choices. Instead, the assumption was that one could have a category of religion for one subset of issues in life, and another category of secularity that is separate and other from religion. Yet as the courts have recognized in some contexts, absence of traditional religious commitments is itself a religious decision.[34] Atheism (for instance) is a religious choice. Conversely, many traditional religions make ethical and devotional demands on adherents throughout the day and affect adherents’ views on a wide range of topics. Decisions to not pray publicly in the course of a day, or to not evaluate a public policy issue in light of a religiously-infused ethical analysis, are religiously-freighted decisions.

Religion can also affect conduct in areas less directly connected to religious devotions. In the late eighteenth century, for instance, how one defined marriage was not generally seen as an activity deeply shaped by religious commitments. In the early twenty-first century, by contrast, one might look back at the assumptions about marriage in the eighteenth century and find them deeply shaped by the broadly Christian religious culture of the time.

The more one recognizes religious diversity, and the breadth of decisions one can make about religious commitments and observances, the more one must recognize the impossibility of absolute neutrality. Indeed, some theologians and philosophers have argued that religion is virtually omnipresent in its influence and effect on the life, outlook, and conduct of individuals.[35] The more one watches for it, the more one can find religious relevance in decisions across the range of human activity—and, yes, of government activity.

The government inevitably will make decisions affecting religion when it engages in such an all-encompassing project as education.[36] Secular education is not really neutral as between religion and irreligion. To many religious adherents, it is an education that sides with irreligion insofar as it assumes that one can go through daily life without reference to religion.

III.           Locating Relative Neutrality in School Funding Policy

Justice Breyer’s dissent in Carson takes the common, and understandable, position that a secular education is a religiously neutral education. He believes that by funding religious education, the government violates its neutrality. But this reflects a simplistic interpretation of what constitutes religion. A secular education is religiously neutral only if one knows that a secular education does not take a position on any issue upon which a religious adherent might disagree. But we know that there are people of faith who believe that it is religiously deficient to educate children without reference to God; to go through a day without prayer or acknowledgment of God by the educators tasked with developing young minds; or to teach about the natural world without expressing gratitude to the creator. Secular education is not neutral to these observers.[37]

Chief Justice Roberts’ majority opinion can also be faulted for lacking self-awareness about the limits of neutrality. Insisting that equal funding for religious educational institutions is the path to neutrality is not being neutral toward those religiously committed to not supporting traditional religion in any form. The majority’s conclusion is also bound to be non-neutral in its effects.[38] If tax dollars for scholarship grants are available only to schools in which children are enrolled, this will naturally select for religious groups that have a certain minimum number of adherents in a given community. There are more Christian schools in America than Jewish schools for the simple reason that there are more adherents to Christianity. And in many communities, there are not sufficient numbers of Jews to make it feasible to start a school. Minority religions will need critical mass to be able to take advantage of these kinds of equal-access policies. The result will be that the distribution of these funds will be “lumpy” rather than an even distribution across religious perspectives. (One way to reduce the lumpiness would be to make grants available for home education. But I suspect few judges or legal scholars would be interested in extrapolating the religious neutrality position into an argument for home education funding.)

Neither the majority nor the dissent in Carson is neutral in an absolute sense. Such absolute neutrality is impossible to achieve. Religious neutrality is a myth.

Yet if the courts are to use the language of neutrality—and for now they give no indication of wishing to abandon it—then the majority’s position is the easiest to defend in the terms I have argued for here—that of a relative neutrality. The law of religious freedom is neutral not in an absolute sense but in a relative sense—relative to a particular set of background conditions. Here, the state has already established a secular education. The state is not neutral as between religion and irreligion in education; it already has a default setting in favor of non-religious education. But when the state gives parents grant money to fund education for their children when a public school is unavailable, the relatively religiously-neutral method of choosing who can and cannot get that money is to let the parents decide. The government’s thumb on the scales in favor of a particular religious outlook is then made less strong, less intrusive. The Court’s solution in Carson is to require the state to adopt a relatively religiously-neutral decision mechanism for allocating government funding for education. That approach can be critiqued for its effects or its lack of commitment to a secular education as such. But these are not critiques from a stable and objective point; they are themselves religiously non-neutral critiques. Allowing parental choice as the baseline condition, though, provides a kind of limited and relative neutrality.[39]


Neutrality is a temptingly simple way of talking about the relationship between the state and religion in America. But it only takes a bit of poking at this paradigm to realize that it is far from stable. For starters, religion has wide-ranging implications, and it is virtually impossible for the government to avoid taking a position on education that is not offensive to one or another religious perspective. In Carson, both majority and dissent tried to stake a claim to neutrality. But neither tried to define and delineate the meaning of religious neutrality in a rigorous way. If neutrality is to be defensible at all, it must be in a limited and relative form. And in that sense, the Carson majority has articulated a sensible, and relatively neutral, path for states funding education to follow when navigating the relationship between the Religion Clauses: let the parents choose. It may not be absolutely neutral, but that is unremarkable. Nothing is.



[1] Olin-Searle-Smith Fellow and Lecturer, Harvard Law School. Thanks to Charlie Capps, Owen Gallogly, Zalman Rothschild, Peter Salib, and Sarah Weinberger for feedback that made this paper better.

[2] See, e.g., Engel v. Vitale, 370 U.S. 421 (1962) (holding that school-sponsored prayer in public schools violated the First Amendment); School Dist. Of Abington v. Schempp, 374 U.S. 203 (1963) (holding unconstitutional school-sponsored Bible reading before class).

[3] See, e.g., Church of the Lukumi Babalu Aye, Inc. v. Hialeah, 508 U.S. 520 (1993).

[4] 142 S.Ct. 1987 (2022).

[5] Carson, 142 S.Ct. at 1993.

[6] Id.

[7] Id.

[8] Id. at 1994.

[9] Id.

[10] Id. at 1996.

[11] Trinity Lutheran Church of Columbia, Inc. v. Comer, 137 S. Ct. 2012 (2017).

[12] Espinoza v. Montana Dep’t of Revenue, 140 S. Ct. 2246 (2020).

[13] Locke v. Davey, 540 U.S. 712, 715 (2004).

[14] Id. at 718–19.

[15] Id. at 724.

[16] Carson ex rel. O.C. v. Makin, 979 F.3d 21, 42 (1st Cir. 2020).

[17] Id. at 33–35.

[18] Id. at 44–46.

[19] Carson ex rel. O.C. v. Makin, 142 S. Ct. 1987, 1998 (2022).

[20] Id. at 2010 (Breyer, J., dissenting).

[21] For samples from an extensive literature and caselaw, see, e.g., John Witte, Jr., Religion and the American Constitutional Experiment 289 (2005); Thomas B. Colby, A Constitutional Hierarchy of Religions? Justice Scalia, the Ten Commandments, and the Future of the Establishment Clause, 100 NW. U. L. Rev. 1097, 1099–1101 (2006); R. George Wright, Can We Make Sense of “Neutrality” in the Religion Clause Cases?: Seven Rescue Attempts and A Viable Alternative, 65 SMU L. Rev. 877, 882 (2012); Walz v. Tax Comm’n of City of New York, 397 U.S. 664, 669–70 (1970); Epperson v. Arkansas, 393 U.S. 97, 104 (1968) (summarizing the First Amendment as requiring “governmental neutrality between religion and religion, and between religion and nonreligion”); Wallace v. Jaffree, 472 U.S. 38, 60 (1985) (“the government must pursue a course of complete neutrality toward religion”).

[22] Sch. Dist. of Abington Twp., Pa. v. Schempp, 374 U.S. 203, 226 (1963).

[23] For another critical account of neutrality in the Religion Clauses, see generally Frederick Mark Gedicks, The Rhetoric of Church and State: A Critical Analysis of Religion Clause Jurisprudence (1995).

[24] See, e.g., Carson v. Makin, 142 S. Ct. 1987, 2004–05 (2022) (Breyer, J., dissenting); Espinoza v. Montana Dep’t of Revenue, 140 S. Ct. 2246, 2284 (2020) (Breyer, J., dissenting).

[25] 3 Joseph Story, Commentaries on the Constitution 728 (§ 1871) (1833).

[26] South Carolina Const. Art. XXXVIII (1778). See also Michael W. McConnell, The Origins and Historical Understanding of Free Exercise of Religion, 103 Harv. L. Rev. 1409, 1436 (1990); Vincent Phillip Muñoz, The Original Meaning of the Establishment Clause and the Impossibility of Its Incorporation, 8 U. Pa. J. Const. L. 585, 639 (2006); Carl H. Esbeck, Dissent and Disestablishment: The Church-State Settlement in the Early American Republic, 2004 B.Y.U. L. Rev. 1385, 1491–94 (2004).

[27] See Jud Campbell, Testimonial Exclusions and Religious Freedom in Early America, 37 Law & Hist. Rev. 431 (2019).

[28] See Epperson v. Arkansas, 393 U.S. 97, 104 (1968).

[29] See Roy A. Clouser, The Myth of Religious Neutrality (2d ed. 2005); David S. Caudill, Law and Belief: Critical Legal Studies and Philosophy of the Law-Idea, in Christian Perspectives on Legal Thought 109 (Michael W. McConnell, Robert F. Cochran, Jr., & Angela C. Carmella eds., 2001); Lael D. Weinberger, Religion Undefined: Competing Frameworks for Understanding ‘Religion’ in the Establishment Clause, 86 U. Det. Mercy L. Rev. 735 (2009).

[30] Noah Webster, The American Spelling Book 43 (rev. ed. 1820).

[31] Id.

[32] Id. at 46.

[33] Will Herberg, Protestant—Catholic—Jew: An Essay in American Religious Sociology (1955).

[34] See, e.g., Wallace v. Jaffee, 472 U.S. 38, 38 (1985) (explaining that “the individual freedom of conscience protected by the First Amendment embraces the right to select any religious faith or none at all”); United States v. Seeger, 380 U.S. 163, 166 (1965) (finding that a “belief that is sincere and meaningful [and] occupies a place in the life of its possessor parallel to that filled by the orthodox belief in God” counts as sufficiently religious for purposes of conscientious objector status); Welsh v. United States398 U.S. 333, 343–44 (1970) (beliefs of comparable “strength” to “more traditional religious convictions” could be sufficiently religious for conscientious objector status even where the individual specifically said the beliefs were not religious)

[35] See, e.g., Weinberger, supra note 29, at 744–47 (collecting sources).

[36] See, e.g., Rockne M. McCarthy & James W. Skillen, Disestablishment a Second Time: Genuine Pluralism for American Schools (1982); David G. Leitch, Note, The Myth of Religious Neutrality by Separation in Education, 71 Va. L. Rev. 127 (1985).

[37] See also Gabriël A. Moens, The Menace of Neutrality in Religion, 2004 B.Y.U. L. Rev. 535, 536 (2004).

[38] See also Alan E. Brownstein, Interpreting the Religion Clauses in Terms of Liberty, Equality, and Free Speech Values-A Critical Analysis of “Neutrality Theory” and Charitable Choice, 13 Notre Dame J.L. Ethics & Pub. Pol’y 243, 254 (1999).

[39] See also Michael W. McConnell, Neutrality Under the Religion Clauses, 81 Nw. U. L. Rev. 146, 149 (1986) (arguing that the best interpretation of neutrality is one in which a regime “leave[s] decisions about religious practice to the independent judgment of the people”); Carl H. Esbeck, A Constitutional Case for Governmental Cooperation with Faith-Based Social Service Providers, 46 Emory L.J. 1 (1997).

Read More »

Antitrust and Modern U.S. Labor Markets: An Economics Perspective – Diana Furchtgott-Roth

Posted by on Aug 1, 2022 in Per Curiam

Download PDF

Antitrust and Modern U.S. Labor Markets: An Economics Perspective

Diana Furchtgott-Roth[1]

Among the most high-profile initiatives of the Federal Trade Commission (FTC) during the Biden Administration has been a focus on using antitrust law to address issues relating to economic inequality, specifically the role that corporate mergers purport to play in undermining labor market competition, and in turn harming workers.[2] Proponents of the FTC’s current approach appeal to academic analysis in support of their argument.[3]  We submit that this focus is misguided, and based on flawed assumptions with respect to both the state of the labor market and the purported growth in economic inequality.  Rather, we argue, when analyzed correctly, the data regarding workplace flexibility, labor market concentration, and so-called “income inequality” show that, if the FTC continues down this regulatory path, the workers the agency claims to protect will suffer the greatest harm.  The application of antitrust law to the labor market is unprecedented and, perhaps more importantly, antithetical to the well-being of workers.  For the reasons we explain below, it should be rejected.

In his latest book, How Antitrust Failed Workers,[4] University of Chicago Professor Eric Posner argues that firms exert wage-setting power over workers through labor market power[5] and employment concentration,[6] known in economic terms as monopsony. According to Professor Posner, workers have no choice but to accept the wages and terms offered by firms that have monopsony employment power, and firms use this power to keep down wages—a practice known as wage suppression.[7] Professor Posner claims that it is the job of the antitrust authorities, the Department of Justice and the Federal Trade Commission to oversee this abuse of employment monopsony power, just as they have the authority to review antitrust authority to review abuse of product and service markets.[8] Professor Posner, who has been hired as Counsel to the Assistant Attorney General[9] at the Antitrust Division of the U.S. Department of Justice to work in this area, has published other articles on the topic, most recently “Antitrust and Labor Markets: A Reply to Richard Epstein,” in the latest volume of the New York University Journal of Law and Liberty.[10]

However, data from the Bureau of Labor Statistics of the U.S. Department of Labor show that exercise of monopsony power is generally not occurring in today’s 21st century economy, nor has it been a characteristic of labor markets over the past half century.[11] In 2022, at the time of this writing, there are over 11 million unfilled jobs,[12] and employers are raising wages and improving benefits to find workers.[13] In addition to normal turnover over the past several decades, which has been substantial and greater than other countries, the “Great Resignation” has seen workers leave in substantial numbers for jobs with better pay and more favorable working conditions.[14]  In addition to better pay, 45 percent of workers surveyed by the Pew Research Center moved to new jobs to have more flexibility to work full-time or part-time from home.[15] Rather than spend a career with one employer, workers have an average of 12 different jobs before the age of 54 according to the data from the National Longitudinal Survey.[16]

Professor Posner states that “the anticompetitive behavior of employers causes significant harm to social welfare,”[17] although he admits that “the frequency with which collusion takes place in labor markets is an open question.”[18] It is quite possible that employer power was a feature of some geographically isolated labor markets in the 19th century, when some small towns were characterized by one employer or a small number of employers, and workers in these remote towns faced substantial costs to move and instead generally spent much or all of their careers with one employer.[19]

Professor Posner’s concerns about labor concentration and monopsony have spread beyond the realm of academia to the Executive Branch and Congress.

President Biden directly addressed monopsony in his Executive Order 14036,[20] issued July 9, 2021, which referred to “the harmful effects of monopoly and monopsony—especially as these issues arise in labor markets.”[21]

On January 18, 2022, the Federal Trade Commission and the Antitrust Division of the Department of Justice asked for public comments[22] on illegal merger enforcement, including on “labor market effects of mergers.” (The comment period was extended until April 21, 2022, from March 21, 2022.[23]) The U.S. Department of the Treasury issued a report on March 7, 2022, entitled “The State of Labor Market Competition.”[24]

On April 6, 2022, the U.S. House of Representatives Select Committee on Economic Disparity and Fairness in Growth held a hearing entitled “(Im)balance of Power: How Market Concentration Affects Worker Compensation and Consumer Prices.” [25]

In July, 2022, the National Bureau of Economic Research released a paper by economists Antonio Falato, Hyunseob Kim, and Till M. Von Wachter. It concludes that “increases in the concentration of more powerful institutional shareholders reduce employment and labor income, because the objectives and interests of the shareholders more likely conflict with those of the workers.”[26]

On July 19, 2022, Chair Lina Khan of the Federal Trade Commission and General Counsel Jennifer Abruzzo of the National Labor Relations Board signed a Memorandum of Understanding pledging to cooperate in areas of mutual interest. The memorandum states, “Issues of common regulatory interest include labor market developments relating to the gig economy and other alternative work arrangements; claims and disclosures about earnings and costs associated with gig and other work; the imposition of one-sided and restrictive contract provisions, such as noncompete and nondisclosure provisions; the extent and impact of labor market concentration; the impact of algorithmic decision- making on workers; the ability of workers to act collectively; and the classification and treatment of workers.”[27]

Recommendations to relieve so-called monopsony power include reducing occupational licensing; strengthening antitrust laws with regard to labor markets; raising the minimum wage; expanding the power of unionized labor; increasing funding for the National Labor Relations Board; broadening the Affordable Care Act; and reducing the numbers of independent contractors.[28] The burden on new entrants to the labor force of excessive occupational licensing requirements for professions such as flower-arranging, tree-cutting, and hair braiding has been well-documented,[29] and most people agree that these licensing requirements should be reduced. The other recommendations, which have been advocated before as cures for supposed inequality, wage disparity, and other societal problems,[30] show that labor monopsony is a solution in search of a problem. These recommendations would reduce flexibility for workers rather than expanding it, as will be explained below.

In this paper, we first provide data from the U.S. Department of Labor to show that America’s labor markets are flexible, with high rates of employee turnover, [31] high rates of firm expansion and contraction;[32] regular movement of workers between states;[33] and an average of 12 jobs held over an individual’s career.[34]  Second, we argue that the concept of labor market concentration is not relevant in a modern economy, and product market concentration generally is not linked to labor market concentration. Third, since inequality is given as a rationale for FTC investigations of product market concentration, we provide data from academic studies demonstrating that inequality has not been increasing and that Americans’ income has in fact been rising over time.[35] FTC investigations are a solution to an invented labor monopsony problem. Fourth, we analyze recommendations made by Professor Posner and others and show that these would not add to workers’ welfare. Finally, we suggest that American workers not only would not benefit from a newly invented antitrust application to labor markets, but would be injured by Professor Posner’s solutions.

I.               Job Turnover in America’s Labor Force

America’s labor markets are the most flexible in the world. Organization for Economic Co-operation and Development data show that the United States has the highest employee turnover and shortest job tenure of developed countries.[36] Table 2 shows that over the past decade the median job tenure in the United States has been about 4 years, but for most other developed countries median tenure is 5 to 10 years.[37] American workers also find jobs faster than those in other countries. (See Table 1). In the United States in 2021, the latest year available, 23.1 percent of unemployed workers could not find work within 12 months.[38] This compares with 28.4 percent for the UK, 29.5 per cent in France, 32.6 percent in Germany and 58 percent in Italy. [39] In the European Union as a whole, the percentage of the unemployed out of work for longer than a year was 36.2 percent, and for the OECD as a whole it was 28.4 percent.[40]

In the United States in 2021, the latest full year available, with a labor force of 161 million, there were 69 million separations (of which 48 million were voluntary quits) and 76 million hires.[41] (See Table 3.) This churning labor market has long been a feature of the U.S. labor force and is not simply due to post-pandemic supply chain shortages.[42] The same results hold on an annual basis back to 2001.[43] For instance, in 2017, before the pandemic, with a labor force of 160 million, there were 63 million separations (of which 38 million were quits) and 66 million hires.[44] (See Figure 3.)The latest data from the Job Openings and Labor Turnover survey,[45] for May 2022, show over 11 million job openings in that month alone. There were 6 million separations in May, of which 4.3 million voluntary quits. The number of hires totaled 6.5 million hires. The number of layoffs and discharges, another subset of separations, was 1.4 million. (See Table 4 and Figure 1.) Every month millions of people separate from their jobs, and every month millions are hired. These data refute the description of America’s labor markets as concentrated.

Data show that the American labor force is in a constant state of turnover, with people leaving one job and finding another on a regular basis.[46] The Bureau of Labor Statistics publishes monthly data on labor turnover that include number of job openings; number of separations (people who have left a place of employment for any reason); number of quits (people who have voluntarily left their job); and number of hires (additions to employers’ payrolls).[47]

As well as collecting numbers of separations, quits, layoffs, and hires, the Bureau of Labor Statistics calculates these as a share of the workforce by dividing the number of separations, quits, layoffs, and hires by employment and multiplying that quotient by 100. If labor market concentration were harming workers, one would expect to see the lowest rates of movement—quits and hires—in low-wage jobs. Lower-wage workers are supposedly trapped in their occupations, and need government action to help them move up.[48] However, the Bureau of Labor Statistics data show that separations and hiring rates for low-wage workers are higher than average.[49] Low-wage occupations, such as leisure and hospitality and retail trade, showed quit rates of 5.5 percent and 4.0 respectively, compared to an economy-wide average quit rate of 2.8 percent.[50] Similarly, these industries showed separation rates of 6.7 percent and 5.0 percent respectively, compared to an average of 3.9 percent.[51] Hiring rates were 7.3 percent and 5.1 percent respectively, compared to an average hiring rate of 4.3 percent.[52] (See Table 5 and Figure 2.)

Some people may quit jobs twice in a year and be hired twice.[53] So these annual numbers do not necessarily translate into a share of the workforce and must be interpreted carefully.  To look at it another way, in 2017 a group equivalent to 26 percent of the workforce quit their jobs, and this figure rose to 33 percent in 2021.[54] (See Table 7.) In 2017 the hiring rate was 45 percent, and in 2021 it was 52 percent. [55]

The Labor Department data show differences between regions. [56] Although incomes in the South are lower, the South shows more turnover—contrary to the thesis that low-income workers are trapped in their jobs. Quits were highest in the South, at 28.6 percent in 2017 and 37 percent in 2021, and smallest in the Northeast, at 20 percent in 2017 and 25 percent in 2021.[57] Hiring rates, defined as the number of hires during the year as a share of annual average employment, were also high.[58] (See Table 8.)

Massive churning has existed since data collection began, as can be seen from rates of hires, separations, and quits. For instance, between 2001 and 2021, the hiring rate ranged from 38 percent of the workforce to 52 percent.[59] It was under 40 percent in only four years, from 2009 to 2012,[60] and it was above 50 percent in 2020 and 2021.[61] The quit rate ranged from 16 percent to 33 percent, and it was below 20 percent from 2009 to 2012, during the recession and its aftermath.[62] (See Table 9.)

At any point over the past 20 years, there have been millions of job openings. (See Table 10.)[63] Currently there are over 11 million job openings due to post-pandemic labor shortages.[64] Excluding 2021, job openings range from a low of 2.5 million per month in a recession year to 7.2 million in an expansionary year.[65] Many of these jobs get filled, and others open up.

Data from the U.S. Census Bureau show that Americans are willing to leave their home state in search of better jobs, better climate, or lower taxes. [66] Americans are not trapped in one state at the mercy of one employer. From July 1, 2020, to July 1, 2021, the populations of Florida, Texas, and Arizona saw the largest gains, and the populations of California, New York, and Illinois saw the largest numeric losses[67]. In terms of share of population, Idaho, Utah, and Montana showed the greatest percentage growth, and the District of Columbia, New York, Illinois, and Hawaii showed the greatest percentage decline.[68] (See Table 11.)

Businesses are also in a constant state of turnover. To measure this, the Bureau of Labor Statistics publishes a quarterly series entitled Business Employment Dynamics.[69] The latest data show that in the fourth quarter of 2021 total job gains from new and expanding companies were 9.6 million, and total job losses from firms that closed or were shrinking were 6.7 million.[70] The net job gain was 2.9 million jobs.[71] These data offer a window into the churning among American businesses. (See Table 12.)

In order to show the number of jobs people have over their lifetimes, the Bureau of Labor Statistics publishes information from the National Longitudinal Survey of Youth 1979,[72] a group of 10,000 people who have been interviewed regularly since 1979. The latest data show that people born between 1957 and 1964 held an average of 12 jobs between the ages of 18 to 54. This does not fit Professor Posner’s thesis that people are trapped by the employer in one job. (See Table 13.)

Professor Posner suggests that employers strengthen labor monopsonies by taking a “paternalistic interest in the well-being of workers and their families.”[73] Elements include company picnics, snacks at the workplace, gyms, and subsidized cafeterias.[74] These features of the workplace are not evidence of monopsony, but proof that the labor market is so competitive that in order to retain workers employers have to provide adequate compensation packages, including more money or more benefits. In either case, employees are free to leave—and the perks Professor Posner cites are only evidence of that freedom.

Some companies deliberately offer generous benefits to save employees’ time and encourage them to spend more time in the office. For example, Palantir, which conducts sophisticated analytical work for private and corporate global clients, provides many services above its generous wages to employees in its Georgetown office in Washington, D.C.[75] Employees can bring their dogs to work and take their dogs to meetings in the conference rooms.[76] The office has shower facilities, nap rooms, a whiskey room for an employee whiskey tasting club, a free laundry service, and three free meals daily.[77] The object is to retain highly-skilled and trained workers. But this does not mean that Palantir is a monopsony. In fact, it suggests the opposite: Palantir must compete vigorously for its workers, to the point of offering benefits well beyond the ubiquitous coffee machine and workday company picnic.

II.             Product Market and Labor Market Concentration Are Unrelated

Professor Posner wants the Federal Trade Commission and the Department of Justice to investigate labor concentration in the same way that these agencies investigate product market concentration.[78] However, the concept of labor concentration is not relevant in a modern economy, outside some highly specialized high-income professions. From an economic perspective, one speaks of a concentrated market as one in which a small number of entities are available to provide a good or service.[79] For instance, the Department of Justice and groups of state attorneys general are investigating Google for monopolizing search and online advertising[80]  as well as Facebook for suppressing competition from online rivals.[81] The Department of Justice is also investigating Amazon for suppressing competition in e-commerce[82], and Apple is being sued by Epic in U.S. District Court for monopolizing app purchases.[83] But the workers in these firms cannot be described as concentrated as they are not all similar—and they are not all employed in technology. High-skilled employees include computer programmers, program managers, app developers, and software developers and engineers. Others are employed in human resources, marketing, public relations, and federal affairs. Lower skilled workers have jobs as cleaners, security guards, and cafeteria workers.

Professor Posner cites non-compete agreements between Apple and Google, but people in the Bay Area move frequently between leading tech firms, including Facebook, Apple, and Google, because this is a faster way to get salary increases.[84] Moves are openly publicized on individuals’ LinkedIn pages, which show that people have worked for multiple companies. Employees go to start-ups if they want a chance to make millions, Facebook if they want more pay, and other firms if they want a better work-life balance. One famous case illustrates the irrelevance of non-compete agreements. In 2016 engineer Anthony Levandowski left Google for Uber to work on Uber’s self-driving cars—the same field that he worked on when at Google.[85] He was sued not for moving to Uber and breaking a non-compete agreement, but for taking 14,000 documents with him containing proprietary information about Google’s vehicles and salary structure.[86]  Employees in non-specialized fields such as human resources, marketing, security, and food preparation at firms with high product market concentrations are also free to move.

III.           Inequality Justifies FTC Action Against Supposed Labor Concentration

Professor Posner and others justify the need for FTC enforcement against labor concentration on the grounds that inequality has been increasing.[87] This concern is exaggerated for a number of reasons.

First, pre-tax, pre-transfer measures of income are not realistic measures of inequality because these measures do not reflect taxes paid and transfers received. Studies that use post-transfer, post-tax measures of income find no increases in inequality.[88] In 2019, the latest data available, the top one percent paid 39 percent of all individual income taxes and earned 20 percent of reported income. The top half of earners paid 97 percent and the bottom half of earners paid 3 percent of all federal individual income taxes. The bottom half of earners receive back a share of the 97 percent paid by the top half for programs including Medicaid, food stamps, the earned income tax credit, housing vouchers, and unemployment insurance.

Studies that account for the value of transfers, including those by University of Chicago professor Bruce Meyer, University of Notre Dame professor James Sullivan, and Cornell University professor Richard Burkhauser, come to different conclusions.[89]  They find that inequality of consumption—what people buy—has not increased.[90] Treasury Department economist Gerald Auten and Joint Committee on Taxation economist David Splinter reached similar results in a paper written in January, 2022.[91]  The Congressional Budget Office[92] has calculated that transfers have steadily reduced measures of inequality between 1979 and 2017.[93] (See Figure 6.)

Second, many measures do not account for the movement of women into the workforce during the 1970s and 1980s. It became increasingly common for mothers with children to work outside the home in the 1980s.[94] In the top fifth of the income distribution, households average two earners per family.[95] In the middle quintile, households have about one earner per household.[96] In the lowest fifth, there is one earner for every two households, with retirees and unemployed.[97]

Third, the size of households has changed since 1980.[98] Due to the increased prevalence of divorce and longer life expectancy, there are more households composed of one person or non-family households. These households tend to be in the lower quintile.[99]  This contributes to perceived inequality. On average, households in the bottom quintile have fewer than two members, and those in the top quintile have four members.[100] Some measures of inequality over the past forty years are picking up additional two-earner couples at the top of the distribution and additional numbers of singles at the bottom of the distribution.

Fourth, people move around the income distribution during their life cycle, and inequality captures natural life cycle changes. Some Americans may have negative income while at school, and then enter the workforce in their 20s.[101]  When they marry, their income could double.[102] They reach peak earnings in their 50s.[103] When they leave the paid workforce, they have lower incomes and live off accumulated assets.[104] Treasury Department data[105] show substantial movement even among the top 400 adjusted gross income earners. Of the 4,584 people who appeared in the top 400 U.S. taxpayers by adjusted gross income over the period 1992 to 2014, only 138 people appeared there for 10 or more years.[106] The top one percent are not static. Sometimes a one-time event, such as selling a company, selling a house, or a making a large capital gain can elevate one’s income bracket for only a short time.[107]

IV.           Adverse Consequences of Recommendations to Reduce Monopsony

Data presented above show that labor concentration does not exist in the United States. Recommendations to reduce supposed labor concentration include ending occupational licensing, changing labor laws to encourage unionization, and raising the minimum wage.[108]  These proposals—the left-of center’s policy agenda—will not reduce economic concentration or inequality. Rather, they will reduce economic growth, giving people fewer opportunities to change jobs and to move up the career ladder.

A.    Occupational Licensing

There is general agreement that state-based licensing requirements—such as mandatory professional certifications regulating practice in a wide variety of professions ranging from cutting hair to driving taxis —serve to reduce entry into particular professions and keep wages high.[109] Many people support reducing licensing requirements.[110] But existing workers support the current level of licensing requirements because the requirements keep wages high.[111] They harm outsiders trying to get in, not existing workers who have completed the certifications. By artificially reducing the labor supply, occupational licenses increase worker bargaining power. Unions function in much the same way: they create a legal monopoly in labor and give workers more leverage by restricting supply.  It is puzzling that Professor Posner and others support granting more power to union leaders while opposing licensing.

These licensing requirements, while harmful, are not an example of employers exploiting workers and preventing them from finding other jobs.  Employers have nothing to do with licensing, but states do.[112] States put in place licensing requirements for different professions, with professions differing by state.[113] As I wrote in Disinherited: How Washington Is Betraying America’s Young,[114] licensing requirements disproportionately hurt young workers, who cannot break into professions without substantial cost.[115] They also harm spouses in military families, who move frequently and cannot requalify before the next move.[116] Licensing requirements raise the price of labor to employers and the price of services to consumers.[117]

One example of how occupational licensing raises prices by creating a cartel is New York City’s medallion requirement for taxi operators.[118] These medallions, before Uber and Lyft entered the industry, used to be valued at millions of dollars.[119] The purchase of a medallion would guarantee a stream of revenue for a cab driver.[120] When ridesharing companies began to compete with New York taxi drivers, taxi drivers’ income (and the value of these medallions) plummeted.[121] Drivers who were still paying off their debts from purchasing their medallions were hit hard, and some committed suicide.[122] New York City restructured medallion holders’ loans at taxpayer expense.[123] As with other occupational licensing schemes, this had nothing to do with labor concentration.

B.     Encouraging Unionization

Professor Posner and others concerned about supposed labor concentration recommend making it easier for workers to unionize as a means of reducing the supposed concentration.[124] Only 6.1 percent of private-sector workers belong to labor unions,[125] perhaps because most other workers believe that the costs of union dues are not worth the benefits of being a union member. The largest share of unionized workers can be found in state and local government, where they can bargain over working conditions but not salaries.[126]

But unions are not promoters of the free market; they are cartels of labor. An authority no less than the U.S. Supreme Court, in a case involving the United Mine Workers in 1965, has said that it is the legitimate goal of any national labor union to eliminate competition over labor standards. [127] Unions were exempted from antitrust laws in 1914.[128] Opportunities for public school teachers, transit workers, police and firefighters, and nurses, are only offered where employees have to join the union as a condition of taking the job. Raises must be negotiated between union management and the employer. Exit from the job is limited because it takes twenty or twenty-five years for a worker’s retirement plan to fully vest, and if they leave beforehand, they lose their contributions and are left with no retirement assets outside Social Security. A frequently cited 2021 study by economists Elena Prager and Matt Schmitt, published in the American Economic Review, found that when large hospitals merged, nurses’ wage growth slowed .[129] The nursing industry is highly unionized, and unionization prevented wage growth.

C.    Increasing the Minimum Wage

Professor Posner and others recommend raising the minimum wage to make workers better off.[130] However, minimum wages discriminate against low-skill workers, because the value that these workers can provide is not worth a higher wage. Minimum wages were originally put into place in the early 20th century to exclude less-educated Black, immigrant, and female workers from jobs by raising wages above their earning power.[131] Then Harvard Law School professor Felix Frankfurter, in a 1916 brief to the Supreme Court supporting Oregon’s minimum wage, wrote, “[t]he state, therefore, may use means, like the present statute, of sorting the normal self-supporting workers from the unemployables and then deal with the latter appropriately as a special class….”[132]

Similarly, British minimum wage supporters Sidney and Beatrice Webb wrote in Industrial Democracy, “[t]here are races who, like the African negro, have no assignable minimum, but a very low maximum; they will work, that is, for indefinitely low wages….”[133] The Federal minimum wage, put in place by the Fair Labor Standards Act of 1938, led to lower, not higher, standards of living for women and minorities who entered domestic service and farming (professions exempted from the minimum wage), because they were excluded from other jobs as “unemployables.”[134]

Research by Nobel Prize winning University of California (Berkeley) economist David Card and the late Princeton University economics professor Alan Krueger claims to show that raising the minimum wage had no effect on employment.[135] This research is frequently cited by proponents of anti-trust enforcement of supposed labor concentration. The Card and Krueger studies are so often cited that it is worth addressing their results in more detail. Card and Krueger examined the effects of a 1992 increase in the minimum wage on fast-food restaurants in New Jersey and Pennsylvania and found that raising the minimum wage did not reduce employment. The studies, published in 1994[136] and 2000,[137] compared New Jersey with neighboring Pennsylvania, which did not raise the minimum wage.

The studies had numerous flaws. First, the authors did not include information on the share of workers who were employed at the minimum wage. If workers were paid more than the minimum wage, then an increase would not reduce employment. Second, the studies did not include county information such as income, unemployment, teen unemployment, labor force, labor-force-participation rates, and changes in state taxes and franchise fees. Third, the regression statistics explain little variance, and practically none of the coefficients are significant. Finally, the studies examine fast food restaurants, but exclude the hospitality industry, which also could have been affected. Other studies using a superior methodology have come to different conclusions. For instance, a 2014 study by University of California (San Diego) economists Jeffrey Clemens and Michael Wither[138] found that the 40% increase in the minimum wage between 2007 and 2010 reduced the employment-population ratio by 6.6 percentage points in states with a binding minimum wage.[139] Low-income workers were particularly harmed because they had fewer job opportunities.[140] Between 2007 and 2010, as the minimum wage rose in three installments from $5.15 an hour to $7.25 an hour,[141] the share of the population that was employed declined by 4 percentage points among adults aged 25 to 54 and by 8 percentage points among those aged 15 to 24.[142]

Since some states had minimum wages that were above the federal minimum wage, Clemens and Wither could compare changes in states where the federal minimum wage was binding to those states with higher minimum wages where it was not.[143] They found that some low skill workers who earned the old minimum wage were employed at the new minimum wage.[144] However, the higher federal minimum wage reduced employment. By the second year of the $7.25 minimum wage rate, employment of low-skill workers had declined by 6.6 percentage points, or 9% more in states with the binding federal minimum wage than in states with a higher state minimum wage. [145] University of California (Irvine) Professor David Neumark has also found negative effects of minimum wage increases among teens and low-skill workers.[146]

V.             Conclusion

Antitrust law does not, and should not, deal with labor markets in the United States. Antitrust law does not deal with labor markets because American workers compete with each other to be hired. Labor markets are among the most competitive markets in America.  The reason that antitrust law has not addressed labor markets is not because American lawyers have been too lazy or ignorant to miss a legal opportunity to promote the interest of clients. Rather, as Professor Posner has admitted, American lawyers and courts have generally not found that antitrust law applies to American labor markets.[147]

Those who see excessive labor concentration suggest that some independent contractors are misclassified, and should have the status of employees with full benefits.[148] Reducing the number of independent contractors would deprive workers of the benefits of flexible, part-time employment that enables them to monetize unused portions of their days and to combine work with other pursuits, such as studying and child-rearing. The pandemic, followed by the Great Resignation, has shown that many workers have market power and can bargain not only for higher wages, but also for flexible hours.[149]

Professor Posner and others are mistaken in asserting that American workers would benefit with a newly invented antitrust application to labor markets. The United States has a wide array of labor laws to protect American workers, at both the federal and state levels.  Antitrust is the wrong instrument to address workers’ problems.


Table 1

Long Term Unemployment Rate, Selected Countries, European Union, and OECD, 2021[150]


Country/Partnership Share of Unemployed Unable to Find Work Within 12 Months
United States 23.1
United Kingdom 28.4
France 29.5
Germany 32.6
Italy 58
OECD – Total 28.4
European Union 36.2



Table 2

Median Job Tenure – OECD Countries[151] (years)


Australia[152] Austria Belgium Canada[153] Colombia
2010 3 – 5[154] 5 – 10 5 – 10 5 – 10 5 – 10
2012 3 – 5 5 – 10 5 – 10 5 – 10 5 – 10
2014 3 – 5 5 – 10 5 – 10 5 – 10 5 – 10
2016 5 – 10 N/A N/A 5 – 10 5 – 10
2018 N/A[155] N/A N/A 5 – 10 5 – 10
2020 N/A N/A N/A 5 – 10 5 – 10


Costa Rica Czech Republic Denmark Estonia Finland
2010 3 – 5 5 – 10 3 – 5 1 – 3 5 – 10
2012 3 – 5 5 – 10 3 – 5 5 – 10 5 – 10
2014 .08 – .5 5 – 10 5 – 10 5 – 10 5 – 10
2016 3 – 5 N/A N/A  N/A N/A
2018 3 – 5 N/A  N/A N/A  N/A
2020 3 – 5 N/A  N/A N/A N/A


France Germany Greece Hungary Iceland
2010 5 – 10 5 – 10 5 – 10 5 – 10 3 – 5
2012 5 – 10 5 – 10 5 – 10 5 – 10 3 – 5
2014 5 – 10 5 – 10 5 – 10 5 – 10 3 – 5
2016 N/A N/A  N/A N/A N/A
2018 N/A N/A N/A N/A N/A
2020 N/A N/A  N/A N/A N/A




Ireland Italy Japan[156] Latvia Lithuania
2010 5 – 10 5 – 10 N/A 1 – 3 1 – 3
2012 5 – 10 5 – 10 5 – 10 3 – 5 3 – 5
2014 5 – 10 5 – 10 N/A 3 – 5 3 – 5
2016 N/A N/A N/A N/A N/A
2018 N/A N/A N/A N/A N/A
2020 N/A N/A  N/A N/A N/A


Luxembourg Netherlands New Zealand Norway Poland
2010 5 – 10 5 – 10 N/A 3 – 5 5 – 10
2012 5 – 10 5 – 10 N/A  3 – 5 5 – 10
2014 5 – 10 5 – 10 N/A 3 – 5 5 – 10
2016 N/A N/A N/A N/A N/A
2018 N/A N/A .5 – 1 N/A N/A
2020 N/A N/A .5 – 1 N/A N/A


Portugal Slovak Republic Slovenia Spain Sweden
2010  5 – 10 5 – 10 5 – 10 5 – 10 5 – 10
2012 5 – 10 5 – 10 5 – 10 5 – 10 5 – 10
2014 5 – 10 5 – 10 5 – 10 5 – 10 5 – 10
2016 N/A N/A N/A N/A N/A
2018 N/A N/A N/A N/A N/A
2020 N/A N/A N/A N/A N/A


Switzerland Turkey[157] United Kingdom USA[158] OECD
2010  5 – 10 N/A 5 – 10 4.4 3 – 5
2012 5 – 10 N/A 5 – 10 4.6 5 – 10
2014 5 – 10 N/A 5 – 10 4.6 5 – 10
2016 N/A N/A 5 – 10 4.2 5 – 10
2018 N/A N/A 5 – 10 4.2 5 – 10
2020 N/A N/A 3 – 5 4.1 5 – 10



Table 3

Total Separations, Quits, and Hires 2017 to 2022 – Millions[159]


Category   2017 2018 2019 2020 2021
Separations[160] 63 66 68 81 69
Quits[161] 38 40 42 36 48
Hires[162] 66 69 70 73 76



Table 4

Job Turnover in the U.S. Economy, May 2022[163]





Job Openings[164] 11.3
Hires 6.5




Layoffs and Discharges[165]


Other Separations[166] .32


Note: Detailed definitions and technical methodology available from source.



Table 5

Separations, Quits, and Hiring Rates for Low-Wage Occupations, May 2022[167]


Industry Separation Quits Hiring
(Percent) (Percent) (Percent)
Total 3.9 2.8 4.3
Leisure and Hospitality 6.7 5.5 7.3
Retail Trade 5.0 4.0 5.1


Note: For definitions of Separations, Quits, and Hires, see Table 1. Rates are computed by dividing the number by employment and multiplying that quotient by 100.



Table 6

Total Separations, Quits, and Hires, 2017[168]




Separations 63 million
Quits 38 million
Hires 66 million



Note: for definitions, see Table 3. Detailed definitions available from source.




Table 7

Share of Workforce Quitting Jobs, 2017 through 2021[169]


Quit Rate


2017 25.7
2018 27.1
2019 28.0
2020 25.2
2021 32.7


Note: for definitions, see Table 3. Detailed definitions available from source.


Table 8


Share of Workforce Quitting Jobs, 2017 through 2021, By Region[170]


Northeast        South               Midwest          West

(Percent)         (Percent)         (Percent)         (Percent)


2017                                         20.1                  28.6                  25.2                  26.1

2018                                         19.8                  30.2                  27.7                  27.4

2019                                         20.8                  31.2                  28.1                  28.4

2020                                         18.8                  28.8                  26.0                  23.7

2021                                         24.6                  37.1                  33.3                  31.4


Note: for definitions, see Table 3. Detailed definitions available from source.



Table 9

Job Turnover in the U.S. Economy, 2001-2021[171]


Year Hires[172] Separations[173] Quits[174]
2001 47.4 48.7 26.4
2002 44.5 44.9 23.4
2003 43.6 43.7 21.7
2004 45.9 44.4 23.4
2005 47.5 45.6 25.2
2006 47.5 45.9 26.1
2007 46.3 45.4 25.3
2008 41.1 43.6 22.2
2009 35.6 39.6 15.9
2010 38.1 37.3 17.1
2011 39.2 37.6 17.9
2012 39.8 38.1 18.6
2013 40.3 38.5 20.1
2014 42.4 40.3 21.9
2015 44.1 42.3 23.7
2016 44.2 42.6 24.9
2017 44.7 43.3 25.7
2018 46.0 44.5 27.1
2019 46.3 45.1 28.0
2020 51.1 56.8 25.2
2021 51.7 47.2 32.7


Table 10

Millions of Job Openings in U.S. Economy, 2001-2021[175]


Year   Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
2001 5.2 5.1 4.8 4.6 4.4 4.4 4.4 4.0 4.1 3.7 3.8 3.7
2002 3.7 3.4 3.6 3.5 3.5 3.4 3.4 3.5 3.3 3.5 3.5 3.2
2003 3.4 3.2 3.1 3.1 3.3 3.4 3.0 3.2 3.1 3.3 3.3 3.4
2004 3.4 3.5 3.5 3.5 3.7 3.3 3.8 3.5 3.8 3.9 3.5 4.1
2005 3.8 4.0 4.0 4.2 3.8 4.1 4.3 4.1 4.4 4.2 4.2 4.3
2006 4.4 4.3 4.7 4.8 4.5 4.6 4.4 4.7 4.7 4.6 4.6 4.6
2007 4.8 4.7 5.0 4.7 4.7 4.9 4.6 4.5 4.7 4.6 4.6 4.5
2008 4.6 4.3 4.2 4.0 4.2 3.8 3.7 3.7 3.2 3.4 3.2 3.1
2009 2.7 2.9 2.5 2.3 2.5 2.5 2.2 2.3 2.5 2.4 2.5 2.6
2010 2.8 2.7 2.7 3.2 3.0 2.8 3.1 3.0 2.9 3.2 3.2 3.1
2011 3.1 3.2 3.3 3.3 3.2 3.5 3.6 3.3 3.8 3.6 3.6 3.8
2012 3.9 3.6 4.0 3.8 3.8 3.9 3.7 3.8 3.9 3.8 3.9 4.0
2013 3.9 4.0 4.1 4.0 4.1 4.2 3.9 4.1 4.1 4.2 4.1 4.1
2014 4.1 4.4 4.4 4.6 4.7 5.0 4.8 5.3 4.9 5.0 4.8 5.1
2015 5.3 5.5 5.2 5.6 5.6 5.2 6.1 5.5 5.5 5.8 5.7 5.8
2016 6.0 5.8 6.1 5.8 5.8 5.7 6.0 5.7 5.9 5.6 6.0 6.0
2017 5.6 5.9 5.8 6.1 5.8 6.3 6.2 6.3 6.3 6.4 6.3 6.3
2018 6.6 6.6 6.8 6.9 7.0 7.2 7.2 7.2 7.3 7.4 7.6 7.4
2019 7.5 7.1 7.3 7.2 7.3 7.1 7.1 7.1 7.1 7.4 6.9 6.7
2020 7.2 7.0 5.9 4.7 5.4 6.0 6.6 6.3 6.5 6.9 6.8 6.9
2021 7.2 7.9 8.5 9.3 9.6 9.9 10.8 10.6 10.7 11.1 10.9 11.4


Table 11

2021 Population Estimates[176]


Numeric Growth
Rank Geographic Area Growth
1 Texas 310,288
2 Florida 211,196
3 Arizona 98,330
Percent Growth
Rank Geographic Area Growth
1 Idaho 2.9%
2 Utah 1.7%
3 Montana 1.7%
Numeric Decline
Rank Geographic Area Decline
1 New York -319,020
2 California -261,902
3 Illinois -113,776
Percent Decline
Rank Geographic Area Decline
1 District of Columbia -2.9%
2 New York -1.6%
3 Illinois -0.9%
4 Hawaii -0.7%


Table 12

Business Employment Dynamics – Fourth Quarter 2021[177]




Gross Job Gains 9.6
Gross Job Losses 6.7
Net Employment Gain 2.9


Table 13

Average Jobs Held by Age Group for Individuals Born between 1957-1964[178]


Age Group

(Years Old)


Avg Jobs Held

18-24 5.6
25-34[179] 4.5
35-44 2.9
45-54[180] 2.1
18-54[181] 12.4


Figure 1: Job Turnover in the US Economy, May 2022[182]





Note: Detailed definitions and technical methodology available from source.


Figure 2:  Separations, Quits and Hiring Rates for Low-Wage Occupations, May 2022[183]

Note: For definitions of Separations, Quits, and Hires, see Table 3. Rates are computed by dividing the number by employment and multiplying that quotient by 100.

Figure 3: Total Separations, Quits, and Hires, 2017[184]

Note: for definitions, see Table 3. Detailed definitions available from source.

Figure 4:   Share of Workforce Quitting Jobs, 2017-2021[185]

Note: for definitions, see Table 3. Detailed definitions available from source.

Figure 5: Share of Workforce Quitting Jobs, 2017 through 2021, By Region[186]

Note: for definitions, see Table 3. Detailed definitions available from source.

Figure 6[187]


[1] Adjunct Professor, George Washington University; former Chief Economist, U.S. Department of Labor. The author is grateful to Gustavo I. Alcantar of Columbia University for research assistance. All errors are her own. This paper, which received financial support from Littler’s Workplace Policy Institute, uses only publicly available data.

[2] Press Release, Federal Trade Commission, Federal Trade Commission and Justice Department Seek to Strengthen Enforcement Against Illegal Mergers (January 18, 2022),

[3] Impact of Corporate Power on Workers and Consumers: Hearing on the Imbalance of Power: How Market Concentration Affects Worker Compensation and Consumer Prices Before the H. Select Comm. on Economic Disparity and Fairness in Growth, 117th Cong. (2022) [hereinafter H. Select Comm.].

[4] Eric A. Posner, How Antitrust Failed Workers (Oxford University Press 2021).

[5] Id. at 1.

[6] Id. at 2.

[7] Id. at 1.

[8] Id. at 32–33.

[9] Chief of Staff and Senior Advisors, Antitrust Division, U.S. Department of Justice (May 25, 2022),

[10] Eric A. Posner, Antitrust And Labor Markets: A Reply To Richard Epstein, 15 N.Y.U. J.L. & Liberty 389 (2022). See also Suresh Naidu, Eric A. Posner & Glen Weyl, Antitrust Remedies for Labor Market Power, 132 Harv. L. Rev. 536 (2018); Ioana Marinescu & Eric Posner, Why Has Antitrust Law Failed Workers?,015 Corn. L. Rev. 1343 (2020).

[11] Bureau Of Labor Statistics, Job Openings And Labor Turnover – May 2022 (2022), [hereinafter May Job Openings and Labor Turnover Survey]; Bureau Of Labor Statistics, Number Of Jobs, Labor Market Experience, Marital Status And Health: Results From A National Longitudinal Survey (2021), [hereinafter National Longitudinal Survey].

[12] May Job Openings and Labor Turnover Survey, supra note 11.

[13] Bureau Of Labor Statistics, Employment Cost Index – June 2022 (2022),

[14] May Job Openings and Labor Turnover Survey, supra note 11.

[15] Pew Research Center, American Trends Panel Wave 103 February 2022 (2022),

[16] National Longitudinal Survey, supra note 11.

[17] Eric A. Posner, Antitrust And Labor Markets: A Reply To Richard Epstein, 15 N.Y.U. J.L. & Liberty 389, 389 (2022), 7622/15.2_Posner_Final.pdf.

[18] Id. at 398.

[19] Hardy Green, Company Towns in the United States, Oxford Research Encyclopedias (Mar. 28, 2018),

[20] Exec. Order No. 14036, 86 Fed. Reg. 36987 (July 9, 2021).

[21] See id. § 1.

[22] Press Release, Federal Trade Commission, Federal Trade Commission and Justice Department Seek to Strengthen Enforcement Against Illegal Mergers (January 18, 2022),

[23] Press Release, Federal Trade Commission, FTC and DOJ Extend Deadline for Public Comment on Ways to Strengthen Enforcement Against Illegal Mergers (March 15, 2022),

[24] U.S. Department Of The Treasury, The State Of Labor Market Competition (2022),

[25] H. Select Comm., supra note 3.

[26] Antonio Falato, Hyunseob Kim, & Till M. von Watcher, Shareholder Power and the Decline of Labor 31 (Nat’l Bureau of Econ. Rsch., Working Paper No. 30203, July 2022).

[27] Lina M. Khan & Jennifer A. Abruzzo, Memorandum to Understanding, Federal Trade Commission (July 19, 2022),

[28] U.S. Department Of The Treasury, The State Of Labor Market Competition (2022),; Eric A. Posner, How Antitrust Failed Workers (2021); Press Release, Federal Trade Commission, FTC and DOJ Extend Deadline for Public Comment on Ways to Strengthen Enforcement Against Illegal Mergers (March 15, 2022),

[29] Dick M. Carpenter, Lisa Knepper, Angela C. Erickson & John K. Ross, License to Work: A National Study of Burdens of Occupational Licensing, Institute For Justice (2012),

[30] For prior advocacy of recommendations, see the following. For health care, see Marquisha Jones & Jill Rosenthal, How Investing in Public Health Will Strengthen America’s Health, Center For American Progress (May 17, 2022),

For strengthening unions, see Aurelia Glass & David Madland, Unions can rebuild the middle class and narrow the racial wealth gap, Center For American Progress (Sep. 13, 2021),

For broadening union power, see Celine McNicholas, Margaret Poydock, & Lynn Poydock, How the PRO Act restores workers’ right to unionize, Economic Policy Institute (Feb. 4, 2021), For raising minimum wage, see Julia Cusick, STATEMENT: Raising the Contractor Minimum Wage and Phasing Out the Subminimum Wage for Disabled Contractors Will Improve Hundreds of Thousands of Lives, Center For American Progress (Nov. 22, 2021), For reducing numbers of independent contractors, see Karla Walter, Workers who risk their lives to run others’ errands should have basic benefits: Personal protection, a minimum wage, health insurance, and more, Center For American Progress (Apr. 2, 2020), For push for mandatory employer-provided parental leave, see Barbara Gault, Heidi Hartmann, Ariane Hegewisch, Jessica Milli & Lindsey Reichlin, Paid Parental Leave in the United States, Institute for Women’s Policy Research (2014),

[31] See May Job Openings and Labor Turnover Survey, supra note 11.

[32]  See Bureau Of Labor Statistics, Business Employment Dynamics – Fourth Quarter 2021 (Apr. 27, 2022), [hereinafter Employment Dynamics].

[33] See Press Release, United States Census Bureau, New Vintage 2021 Population Estimates Available for the Nation, States, and Puerto Rico (Dec. 21, 2021), [hereinafter Population Estimates].

[34] National Longitudinal Survey, supra note 11.

[35] See Diana Furchtgott-Roth, Introduction and Summary to Diana Furchtgott-Roth, United States Income, Wealth, Consumption, And Inequality 1–9 (2021); James Ewell, Kevin Corinth, & Richard V. Burkhauser, Income Growth and Its Distribution from Eisenhower to Obama: The Growing Importance of In-Kind Transfers (1959-2016), in Diana Furchtgott-Roth, United States Income, Wealth, Consumption, And Inequality 90–124 (2021); Gerald Auten & David Splinter, Top Income Shares and the Difficulties of Using Tax Data, in Diana Furchtgott-Roth, United States Income, Wealth, Consumption, And Inequality 125–152 (2021); Edward Conard, The Economics of Inequality in High-Wage Economies, in Diana Furchtgott-Roth, United States Income, Wealth, Consumption, And Inequality 262–286 (2021).

[36] OECD, Employment By Job Tenure Intervals – Persons,  (last visited May 23, 2022) (Select “Labour”, “Labour Force Statistics”, “Job tenure”, “Employment by job tenure intervals – persons”, “Employment by job tenure intervals – persons”).

[37] Id.

[38] Oecd, Long-Term Unemployment Rate, (last visited July 8, 2022).

[39] Id.

[40] Id.

[41] Bureau Of Labor Statistics, Job Openings And Labor Turnover – January 2022 (2022),, at 3 [hereinafter January Job Openings and Labor Turnover Survey].

[42] Id. at Tables 13–22.

[43] Bureau Of Labor Statistics, New Monthly Data Series on Job Openings and Labor Turnover Announced by BLS (2002),

[44] Id., at Tables 13, 15, & 17.

[45] May Job Openings and Labor Turnover Survey, supra note 11.

[46] May Job Openings and Labor Turnover Survey, supra note 11.

[47] Id.

[48] Posner, supra note 4 at 132, 135, 161–162.

[49] May Job Openings and Labor Turnover Survey, supra note 11, at Tables 2–4.

[50]Id., at Table 4.

[51]Id., at Table 3.

[52]Id., at Table 2.

[53]January Job Openings and Labor Turnover Survey, supra note 41, at § Job Openings and Labor Technical Note.

[54]Id., at Table 18.

[55]Id., at Table 14.

[56]Id. at Tables 1–22.

[57]Id. at Table 18.

[58]Id. at Table 14.

[59] Job Openings and Labor Turnover Survey, Bureau of Labor Statistics, (Select “Total nonfarm”, “Total US”, “All areas”, “Hires”, “All size classes”, “Rate”, and “Not seasonally adjusted”, then select “Get Data”, then adjust the “From:” input to “2001” and select “include annual averages”, then observe the data under “Annual” within the table).

[60] Id.

[61] Id.

[62] Job Openings and Labor Turnover Survey, Bureau of Labor Statistics, (Select “Total nonfarm”, “Total US”, “All areas”, “Quits”, “All size classes”, “Rate”, and “Not seasonally adjusted”, then select “Get Data”, then adjust the “From:” input to “2001” and select “include annual averages”, then observe the data under “Annual” within the table).

[63] Job Openings and Labor Turnover Survey, Bureau of Labor Statistics, (Select “Total nonfarm”, “Total US”, “All areas”, “Job openings”, “All size classes”, “Level – In Thousands”, and “Seasonally adjusted”, then select “Get Data”, then adjust the “From:” input to “2001”, then observe the data within the table).

[64] Id.

[65] Id.

[66] Population Estimates, supra note 33. These Americans are not solely those in higher income brackets, See Also Geographic Mobility by Selected Characteristics, U.S. Census Bureau (March 2022),,06,12,17,36,48&tid=ACSST5Y2020.S0701; Median Income In The Past 12 Months (In 2020 Inflation-Adjusted Dollars) By Geographical Mobility In The Past Year For Residence 1 Year Ago In The United States, U.S. Census Bureau (March 2022),,06,12,17,36,48&tid=ACSDT5Y2020.B07411.

[67]Id. at Tables 2 and 4.

[68]Id. at Tables 3 and 5.

[69] Employment Dynamics, supra note 32.

[70] Id.

[71] Id.

[72] National Longitudinal Survey, supra note 11.

[73] Posner, supra note 4, at 16.

[74] Id. at 17.

[75] Diana Furchtgott-Roth, Visit to Palantir Georgetown Office (August 2019).

[76] Id.

[77] Id.

[78] Posner, supra note 4 at 3.

[79] P.R.G. Layard & A.A. Walters, Microeconomic Theory 238 (1978).

[80] United States v. Google, LLC, No. 1:20-cv-03010 (D.D.C. filed October 20, 2020).

[81] Brief for United States as Amicus Curiae Supporting Plaintiffs, State of New York v. Facebook, Inc., No. 21-7078 (D.C. Cir. filed January 28, 2022).

[82] Letter from U.S. H. Comm. on the Judiciary to Merrick Garland, U.S. A.G. (March 9, 2022),–_amazon.pdf.

[83] Epic Games, Inc. v. Apple Inc., No. 4:20-cv-05640-YGR (N.D. Cal. Sept. 10, 2021). See also Epic Games, Inc. v. Apple Inc., Nos. 21-16506 & 21-16695 (9th Cir. filed Mar. 24, 2022).

[84] Conversation of Diana Furchtgott-Roth with Google employee (May 2022).

[85] Daniel Cooper, The creator of Google’s self-driving car now works for Uber, ENGADET (Aug. 18, 2016),

[86] United States v. Levandowski, No. 3:19-cr-00377 (N.D. Cal. Aug. 6, 2020).

[87] See Posner, supra note 4 at 1; U.S. Dep’t. Treasury, supra note 24 at ii; Kate Bahn, Testimony before H. Select Comm., supra note 3, page 6,

[88] Gerald Auten & David Splinter, supra note 35, 143. See also Ewell et al., supra note35, 113.

[89] Id. 

[90] Id.

[91] Gerald Auten & David Splinter, Income Inequality in the United States: Using Tax Data to Measure Long-Term Trends (J. Comm. on Taxation, February 18, 2022),

[92] Congressional Budget Office, The Distribution Of Household Income, 2017 (2020),

[93] Id.

[94] Bureau Of Labor Statistics, U.S. Dep’t of Labor, Women In The Workforce: A Databook Table 7 (2022), available at

[95] Bureau Of Labor Statistics, U.S. Dep’t of Labor, Consumer Expectations in 2020 Table 2 (2021), available at

[96] Id.

[97] Id.

[98] Table HH-1 of Historical Household Tables, U.S. Census Bureau (Nov. 22, 2021),; See also Table HH-4 of Historical Household Tables, U.S. CENSUS BUREAU (Nov. 22, 2021),; also Table H-9 of Historical Income Tables: Households, U.S. Census Bureau (Nov. 8, 2021),

[99] Table H-11 of Historical Income Tables: Households, U.S. Census Bureau (Nov. 8, 2021),; See also Table HH-4 of Historical Household Tables, U.S. Census Bureau (Nov. 22, 2021),

[100] Table H-11 of Historical Income Tables: Households, U.S. Census Bureau (Nov. 8, 2021),

[101] Adrienne L. Fernandes-Alcantara, Youth And The Labor Force: Background And Trends 23-24 (2018),

[102] Kathy Morris, Your Job Might Have Something To Do With Who You Marry, ZIPPIA (Jul. 19, 2020),

[103] Median Household Income in the United States in 2020, by age of householder, STATISTA (Oct. 8, 2021),

[104] Id. 

[105] The 400 Individual Income Tax Returns Reporting the Largest Adjusted Gross Incomes Each Year, 1992–2014, IRS (Jan. 4, 2016),

[106] Id.

[107] Topic No. 409 Capital Gains and Losses, IRS (May 19, 2022),

[108] Posner, supra note 4, at 127–28, 132.

[109] Carpenter et al., supra note 29; see also Posner, supra note 4.

[110] Carpenter et al., supra note 29.

[111]  Morris M. Kleiner & Alan B. Krueger, The Prevalence and Effects of Occupational Licensing, 48 B.J.I.R. 676, 685 (2010),; See also Kleiner et. al, Relaxing Occupational Licensing Requirements: Analyzing Wages and Prices for a Medical Service, 59 J.L.E. 261, 286-287 (2016),, Okech v. Thompson, where Tedy Okech challenged the State of Idaho for the right to braid hair without a cosmetology license. Okech v. Thompson, No.1:22-cv-106 (D. Idaho filed Mar. 8, 2022).; see also Harper v. City of Lincoln, where Cynthia Harper challenged the City of Lincoln, Nebraska for the right to sell cookies, cakes, breads, and jams from home. Harper v. City of Lincoln, No.CI 20-1706, 2020 WL 9596383, at *1 (D. Neb.  Oct. 26, 2020).

[112] Morris M. Kleiner & Kyoung Won Park, Battles Among Licensed Occupations: Analyzing Government Regulations on Labor Market Outcomes for Dentists and Hygienists 19-20 (Nat’l Bureau of Econ. Rsch., Working Paper No. 16560, Nov. 2010); In Louisiana, you need a license to be a retail or a wholesale florist. Louisiana Horticulture Commission, La. Dep’t. of Agric. and Forestry (May 2022),, []. In Idaho, you need a license to be a mortician, a shorthand reporter, or a residential facility care administrator. Licensing and Registration Search, Idaho Div. Occupational and Pro. Licensing (Mar. 22, 2022, 9:44 AM), [].

[113] Examples include funeral service directors, barbers, and auctioneers. See Suzanne Hultin, The National Occupational Licensing Database, Nat’l Conf. State Legis. (Mar. 1, 2022),, [].

[114] Diana Furchtgott-Roth & Jared Meyer, Disinherited: How Washington is Betraying America’s Young 8, 122 (2015).

[115] Id. at 83-–5.

[116] Military Spouse Employment: DOD Should Continue Assessing State Licensing Practices and Increase Awareness of Resources, U.S. Government Accountability Office (Jan. 27, 2021),, [].

[117] Morris M. Kleiner & Evgeny S. Vorotnikov, At What Cost: State and National Estimates of the Economic Costs of Occupational Licensing, Institute for Justice 1, 5 (2018),, [].

[118] Stewart Dompe & Adam C. Smith, Taxicab Cartels Restrict Entry into Market at the Expense of Consumers, Mercatus Center (Nov. 24, 2014),, [].

[119] Aaron Jacobs, Distressed Drivers: Solving the New York City Taxi Medallion Debt Crisis, COLUMBIA HUMAN RIGHTS LAW REVIEW ONLINE (Mar. 16, 2022),

[120] See Ydanis Rodriguez et al., Report of the Taxi Medallion Task Force, Taxicab Medallion Sale Prices Task Force, at 34 (January 31, 2020),, [].

[121] Jacobs, supra note 119, at 174.

[122] Jacobs, supra note 119, at 175.

[123] See Press Release, N.Y. Attorney General, Attorney General James’ Statement After NYC Agrees to Supplement Taxi Medallion Relief Program (Nov. 4, 2021),, [].

[124] Posner, supra note 4, at 132.

[125] Bureau Of Labor Statistics, USDL-22-0079, Union Members – 2021 (2022),, [].

[126] Id.

[127] See United Mine Workers of Am. v. Pennington, 381 U.S. 657, 662 (1965).

[128] 15 U.S.C. § 17.

[129] Elena Prager & Matt Schmitt, Employer Consolidation and Wages: Evidence from Hospitals, 111 Am. Econ. Rev. 397, 398 (2021),

[130] Posner, supra note 4, at 127; U.S. Dep’t. Treasury, supra note 24, at 52.

[131] Thomas C. Leonard, Illiberal Reformers: Race, Eugenics, and American Economics in the Progressive Era 140 (2016).

[132] Thomas Reed Powell, The Oregon Minimum-Wage Cases, 32 Pol. Sci. Q. 296, 310 (1917),

[133] Sidney Webb & Beatrice Webb, Industrial Democracy 698 (1920 ed. 1920).

[134] Leonard, supra note 131, at 129–40.

[135] David Card & Alan B. Krueger, Minimum Wages and Employment: A Case Study of the Fast-Food Industry in New Jersey and Pennsylvania, 84 Am. Econ. Rev. 772 (1994),; see also David Card & Alan B. Krueger, Minimum Wages and Employment: A Case Study of the Fast-Food Industry in New Jersey and Pennsylvania: Reply, 90 Am. Econ. Rev. 1397 (2000),

[136] David Card & Alan B. Krueger, Minimum Wages and Employment: A Case Study of the Fast-Food Industry in New Jersey and Pennsylvania, 84 Am. Econ. Rev. 772 (1994),

[137] David Card & Alan B. Krueger, Minimum Wages and Employment: A Case Study of the Fast-Food Industry in New Jersey and Pennsylvania: Reply, 90 Am. Econ. Rev. 1397 (2000),

[138] Jeffery Clemens & Michael Wither, The Minimum Wage and the Great Recession: Evidence of Effects on the Employment and Income Trajectories of Low-Skilled Workers, 170 J.  Pub. Econ. 53 (2019),

[139] Id. at 59, Table 2.

[140] Id. at 62, ¶1.

[141] U.S. Troop Readiness, Veterans’ Care, Katrina Recovery, and Iraq Accountability Appropriations Act, Pub. L. No. 110-28, 121 Stat. 112.

[142] Employment rate by age in the United States from 2000 to 2020, Statista (May 2, 2022),, [].

[143] Clemens & Wither, supra note 138, at 54, ¶8.

[144] Clemens & Wither, supra note 138, at 55, fig.2.

[145] Clemens & Wither, supra note 138, at 59, ¶6.

[146] See generally, David Neumark, J.M. Ian Salas & William Wascher, Revisiting the Minimum Wage-Employment Debate: Throwing Out the Baby with the Bathwater?, 67 ILR REVIEW, 608 (Supp. 2014),,salas,wascher-ilrr-14.pdf, []; see also David Neumark, J.M. Ian Salas & William Wascher, More on recent evidence on the effects of minimum wages in the United States, 3 IZA J. Lab. Pol’y, 1 (2014),, [].

[147] Posner, supra note 4, at 47, 48, 92, 3, 5.

[148] Posner, supra note 4, at 159-–60, 162, 163.

[149] May Job Openings and Labor Turnover Survey, supra note 11.

[150] OECD (2022), Long-term unemployment rate (indicator). doi: 10.1787/76471ad5-en, [] (last visited July 8, 2022).

[151] OECD, Employment by job tenure intervals, [], (last visited May 24, 2022) (Select “Labour”, “Labour Force Statistics”, “Job tenure”, “Employment by job tenure intervals – persons”, “Employment by job tenure intervals – persons”).

[152] For all countries excluding the U.S, the period associated with the year’s median is reported. As most of the data is skewed rightward per year, median estimation was performed. To determine skewness, the difference between the summation of persons with job tenure below five years and the summation of persons at or above five years was calculated. In addition to this indicator, the year’s median in relation to its mean was considered.

[153] Canada does not have observations for the period [0, 1) months.

[154] All rightward boundaries are non-inclusive.

[155] Indicates unbalanced data from the OECD for this time series.

[156] Japan does not have observations for the time periods [0, 1) months and [1, 6) months.

[157] Unbalanced data, possesses time periods different from other countries reported.

[158] Bureau of Labor Statistics 2020, Employee Tenure in 2020 (September 22, 2020 10:00 EST),                                                 [].

[159] January Job Openings and Labor Turnover Survey, supra note 41 at Tables 13, 15, & 17.

[160] Separations include all separations from payroll during the entire reference month.

[161] Quits include employees who left voluntarily with the exception of retirements of transfers to other locations during the reference month.

[162] Hires include all additions to the payroll during the entire reference month.

[163] May Job Openings and Labor Turnover Survey, supra note 11, at Tables 1–6.

[164] Job openings include all positions that are open on the last business day of the reference month.

[165] Layoffs and discharges include involuntary separations initiated by employer during the reference month.

[166] Other separations include retirements, deaths, transfers to other locations.

[167] May Job Openings and Labor Turnover Survey, supra note 11, at Tables 2–4.

[168] January Job Openings and Labor Turnover Survey, supra note 41 at Tables 13, 15, & 17.

[169] January Job Openings and Labor Turnover Survey, supra note 41 at Table 18.

[170] January Job Openings and Labor Turnover Survey, supra note 41 at Table 18.


[171] Bureau of Labor Statistics, Job Openings and Labor Turnover Survey, [] (last visited April 30, 2022). Refer to notes 59 & 62, replicate with Separations.

[172] Hires include all additions to the payroll during the entire reference month. The hires rate is computed by dividing the number of hires by employment and multiplying that quotient by 100.

[173] Separations include all separations from the payroll during the entire reference month. The separations rate is computed by dividing the number of separations by employment and multiplying that quotient by 100.

[174] Quits include employees who left voluntarily with the exception of retirements or transfers to other locations. The quits rate is computed by dividing the number of quits by employment and multiplying that quotient by 100.

[175] Job Openings and Labor Turnover Survey, supra note 171; refer to note 63.

[176] Population Estimates, supra note 33.

[177] Employment Dynamics, supra note 32.

[178] National Longitudinal Survey, supra note 11.

[179] Excludes individuals who turned age 18 before January 1, 1978

[180] Excludes individuals who had not yet turned age 55 when interviewed in 2018-19

[181] Jobs held in more than one age category were counted in each appropriate row, but only once in the total (last) row. Therefore, the overall average number of jobs is less than the sum of the number of jobs across age categories.

[182] May Job Openings and Labor Turnover Survey, supra note 11 at Tables 1–6.

[183] May Job Openings and Labor Turnover Survey, supra note 11 at Tables 2–4.

[184] January Job Openings and Labor Turnover, supra note 41 at Tables 13, 15, & 17.

[185] January Job Openings and Labor Turnover, supra note 41 at Table 18.

[186] Id.

[187] The Distribution of Household Income, supra note 92.

Read More »

Religious Autonomy in Carson v. Makin – Nick Reaves

Posted by on Jul 29, 2022 in Per Curiam

Download PDF

Religious Autonomy in Carson v. Makin

By Nick Reaves[1]

In Carson v. Makin, the Supreme Court confirmed that excluding only “sectarian” religious schools from its tuition aid program violated the “unremarkable” constitutional principle of religious neutrality—that one religion cannot be preferred to another.[2] But court watchers who view this case as the simple application of prior precedent may have missed one of its most important points: its embrace of church autonomy principles in the government funding context.

To help explain why Maine could not constitutionally pick and choose which religious schools to fund, the Supreme Court—for the first time—called on its line of religious autonomy precedent. The Court, citing Our Lady of Guadalupe,[3] explained that the line Maine sought to draw (excluding religious schools that “promote” faith and “inculcate” religious beliefs into the curriculum)[4] misunderstood the very purpose of religious schools. The Court then explained that even the process of “scrutinizing whether and how a religious school pursues its educational mission” unconstitutionally entangled the government in religious questions and could result in “denominational favoritism.”[5]

Carson is therefore far more than “Trinity Lutheran 3.0.”[6] First, it confirms that the status/use distinction discussed in prior government funding cases did not bear the weight that some had hoped. Second, it cabins Locke v. Davey’s[7] anti-establishment interests to the funding of “vocational religious degrees.”[8] And finally, it anchors the Court’s government funding cases in core Free Exercise doctrine, framing Maine’s actions as “exclud[ing] otherwise eligible schools on the basis of their religious exercise.”[9] Together, these doctrinal developments make clear that governments cannot use access to generally available funding as a wedge to interfere with the internal operations of religious organizations.

I.               Maine’s Tuition Aid Program

Maine, the most rural state in the Union,[10] provides families living in school districts without a public secondary school with funding to “pay the tuition . . . at the public school or the approved private school of the parent’s choice at which the student is accepted.”[11] The program imposes no geographic limits on schools the state will fund and, until 1981, parents could choose to send their children to any accredited religious or secular private school, with very few restrictions.[12] However, in 1981, Maine’s legislature limited the program to “nonsectarian” schools. While not defined by statute, the Maine Department of Education considers a religious school “sectarian” if it is “associated with a particular faith or belief system” and, in addition to teaching academic subjects, it “promotes the faith or belief system with which it is associated and/or presents the material taught through the lens of this faith.”[13] In 2018, two Maine families challenged this exclusion of “sectarian” schools, but both the district court and the First Circuit upheld the program under prior precedent. Then the Supreme Court granted certiorari.

At the Supreme Court, a win for the families seemed likely after just a few minutes of Maine’s oral argument. In what appeared to be a pre-planned hypothetical, Chief Justice Roberts asked Maine’s attorney how the state would treat two different religious schools under its tuition funding program. The first school, run by “Religion A,” “has a doctrine that they should provide service to their . . . neighbors . . . but there’s nothing in their . . . doctrine about propagating the faith,” so the school “look[s] just like a public school, but it’s owned by” a religious community.[14] The second school is run by “Religion B” and “its doctrine requires adherents to educate children in the faith.”[15] Religion B’s school, therefore, “is infused in every subject with their view of the faith.”[16] Responding to this hypothetical, Maine’s attorney confirmed what the Chief Justice surely already knew: that parents sending their children to the first school would receive tuition funding but parents at the second school would not, based on the religious differences between the two schools.[17]

Similarly problematic, in response to a question from Justice Barrett, Maine’s attorney acknowledged that when parents request funding to send their children to a school not already in the program, Maine’s Department of Education “does a little homework” to figure out whether it considers the school sectarian or non-sectarian.[18] In briefing, Maine explained that the “focus” of this inquiry “is on what the school teaches through its curriculum and related activities, and how the material is presented.”[19] And that “affiliation or association with a church or religious institution” was relevant, but not “dispositive.”[20] And at oral argument, Maine’s attorney added that sometimes these decisions can be made by a cursory review of “the school’s website . . . [o]r maybe . . . the student handbook.”[21]

Oral argument thus confirmed a fundamental flaw in Maine’s tuition assistance program. By excluding only “sectarian” schools, the program attempted to distinguish between religious schools based on Maine’s own assessment of whether and how a religious school “promotes a particular faith and presents academic material through the lens of that faith.”[22] And Maine had no objective criteria or measure for determining which religious schools crossed the line. Instead, the assessment was left to the discretion of Maine’s Department of Education—apparently sometimes based only on a quick skim of the school’s public facing materials.[23]

II.             Status vs. Use

Maine’s attempt to distinguish between permissible and impermissible types of religious education also revealed the impossibility of distinguishing between religious status and the religious use in generally available government aid programs.

Defending its program, Maine argued that it had an overriding anti-establishment interest in denying funding to schools that would put government money to a religious “use.”[24] According to the state, its exclusion of “sectarian” schools mapped directly onto the status/use distinction suggested by the Supreme Court’s holdings in Trinity Lutheran and Espinoza.[25] The First Circuit and Justice Breyer agreed. As the First Circuit saw it, Maine “does not bar schools from receiving funding simply based on their religious identity,” it instead excludes schools “based on the religious use that they would make of it in instructing chil­dren.”[26]

In both Espinoza and Trinity Lutheran, Chief Justice Roberts, writing for the majority, characterized the government aid program as one that discriminated first and foremost on religious status: religious schools need not apply.[27] But the Court, especially in Espinoza, was also careful to note that this did not implicitly sanction discrimination based on religious use.[28] The Court simply had no need to address the question.

Ignoring the Court’s warning in Espinoza, Maine nevertheless sought to test the constitutionality of religious use discrimination. When squarely presented, the Court had little trouble showing why Maine’s focus on religious use “misreads our precedents.”[29] The Court explained that neither Trinity Lutheran nor Espinoza “suggested that use-based discrimination is any less offensive to the Free Exercise Clause,” and relied on Maine’s own administration of its tuition aid program to “illustrate[ ] why.”[30] As the Court explained, Maine’s attempt to distinguish between religious schools based on “whether and how” they “pursue[ ] [their] educational mission” raised serious religious autonomy concerns.[31]

III.           Religious Autonomy

The principles of religious autonomy discussed in Carson are rooted in both Religion Clauses. As early as 1871, in Watson v. Jones, the Supreme Court held that civil courts must defer to religious bodies on “questions of discipline, or of faith, or ecclesiastical rule, custom, or law.”[32] Later, in Larson v. Valente, the Court held that “[t]he clearest command of the Establishment Clause is that one religious denomination cannot be officially preferred over another.”[33] The principles of religious autonomy therefore generally prevent the government from interfering in disputes over religious doctrine and in internal governance decisions “that affect[ ] the faith and mission” of religious institutions, including religious schools.[34] Keeping the government out of religious disputes therefore prevents both government entanglement in religious questions and religious favoritism.

One form of unconstitutional entanglement that impinges on religious autonomy is the attempt by governments to distinguish between “sectarian” and “non-sectarian” religious beliefs or organizations. In Town of Greece, the Supreme Court explained not only that such a distinction is inconsistent with our nation’s history and traditions, but also that drawing this distinction would force courts “to act as supervisors and censors of religious speech.”[35] And as Justice Thomas explained in his American Legion concurrence, such a distinction would result in “courts ‘trolling through religious beliefs’” and making “inevitably arbitrary decisions” regarding what is and is not “sectarian.”[36] This would obviously create serious religious autonomy concerns.

Protecting this religious autonomy is particularly important when it comes to religious education. In Hosanna-Tabor and Our Lady, the Supreme Court confirmed that religious schools must have exclusive control over who teaches the faith to protect their “independence in matters of faith and doctrine.”[37] As the Court reasoned, “a wayward” teacher could “contradict” the tenets of the faith and lead students “away from the faith.”[38] This would undermine a crucial component of religious education: transmitting the faith “to the next generation.”[39]

Against this constitutional backdrop, Maine’s tuition aid program raised significant religious autonomy concerns. By determining which schools were eligible for tuition funding based on what, how, and how much religion was infused into a school’s curriculum, Maine was conditioning benefits on a religious school’s conformity to the government’s preferred approach to religious education. Parents who needed the state’s tuition assistance were forced to choose schools with a certain religious perspective, and some religious schools may have even felt financial pressure to conform to Maine’s requirements to obtain (or maintain) eligibility. Perhaps even worse, Maine’s evaluation process was both essentially standardless and discretionary, further entangling the government in religious questions.

IV.           The Fiction of Value-Neutral Religious Education

To defend its program, Maine argued that it was not expressing hostility or opposition to the schools’ beliefs; it just wanted to prevent the “inculcation” or infusion of those beliefs into the curriculum and the school environment.[40] In other words, if religious schools could be religious without encouraging religion, they would be eligible for funding. But this argument fundamentally misunderstands religious education (and education more generally).

Maine’s argument assumes that non-sectarian schools (religious, private, or public) do not inculcate any values. But no type of education can be completely value neutral. As Justice Barrett explained at oral argument, “all schools, in making choices about curriculum and the formation of children, have to come from some belief system.”[41] “[I]n public schools, . . . the districts are” choosing “the kind of values that they want to inculcate in the students.”[42] Even Justice Breyer recognized as much. As he explained in his dissent, Maine’s public schools “seek first and foremost to provide a primarily civic education” and serve as “the primary vehicle for transmitting the values on which our society rests,” which he viewed as including “the preservation of a democratic system of government.”[43]

Education certainly includes the passing on of objective facts, but it also imparts ways of thinking and perspectives on history and current events. It inculcates civil (and sometimes religious) values, and even good (or bad) habits. So understood, Maine’s argument breaks down: because all schools impart values in one way or another, Maine is simply picking and choosing which values it deems appropriate and beneficial and which it deems, as explained further below, “fundamentally at odds with [the] values we [the State of Maine] hold dear.”[44]

Thus, while Maine may have sidestepped religious status discrimination, it walked right into the Court’s religious autonomy precedent by privileging some religious beliefs over others based on a discretionary evaluation of how religious schools pass on the faith.

V.             Carson’s Impact

Doctrinally, this decision confirms that state Blaine Amendments—laws often rooted in religious animus that exclude “sectarian” schools from public benefits—are well and truly dead. While Espinoza did much of the heavy lifting,[45] Carson’s unequivocal rejection of the status/use distinction ensures that states and lower courts can no longer rely on arguments about religious “use” to deny religious organizations equal access to generally available government funding programs.

In a similar vein, Carson confirmed that Locke v. Davey,[46] a 2004 case in which the Supreme Court upheld the State of Washington’s college scholarship program against a free exercise challenge,[47] “cannot be read beyond its narrow focus on vocational religious degrees” and cannot justify exclusion of “religious persons” based on “their anticipated religious use” of the government benefit.[48] Instead, the Court confirmed that Locke is justified only by the unique “historic and substantial state interest” against taxpayer funding for “church leaders.”[49] Outside of a theological seminary, it is hard to see how Locke has any life left.

Practically, while Carson may signal the end of the status/use distinction, it may also pave the way for the Supreme Court to wade into even deeper waters. On the same day the Court ruled against Maine, the state’s attorney general issued a press release calling the ruling “disturbing,” and associating the beliefs of the prevailing religious families and schools with “discrimination, intolerance, and bigotry.”[50] Tellingly, the attorney general revealed the value-laden motivations behind the state’s defense of its tuition aid program: he described the school’s religious beliefs as “inimical to a public education” and “fundamentally at odds with the values we hold dear.”[51] He then speculated that Maine may still be able to bar disfavored religious schools from its program because, he argued, they engage in “discriminatory practices” in violation of the Maine Human Rights Act’s “anti-discrimination provisions.”[52]

Unwilling to cede defeat, Maine appears poised to continue to exclude religious schools on a new theory: that conduct motivated by a school’s sincere religious beliefs (like a school’s decision to only hire coreligionists or only admit students who share the school’s religious beliefs) violates state antidiscrimination laws. This argument is not new. Variations on this theme are already working their way through the courts. In the religious student group context, for example, courts are confronting cases in which student clubs have been denied generally available benefits not because of their beliefs, but because their beliefs impose certain requirements on club leaders (like requiring leaders to conduct themselves in accordance with the club’s statement of faith).[53] Similar dynamics are at play in fights over the application of public accommodation laws and non-discrimination requirements for federal funding.[54]

But—much like the status/use distinction Carson rejected—this artificial attempt to distinguish between what religious communities believe and how those beliefs are concretely manifest cannot hold up in practice, for at least four reasons. First, Carson itself, by incorporating the principles of religious autonomy, confirmed that religious organizations must have the freedom to operate in accordance with their beliefs. Second, Fulton implicitly rejected the argument that religious beliefs could be separated from how a religious ministry puts those beliefs into practice—even in the context of government-contracted services.[55] Third, AOSI and Masterpiece forbade the targeted use of antidiscrimination provisions and gerrymandered programs to exclude religious applicants.[56] Finally, because Maine has exempted single-sex private schools (but not religious schools) from its nondiscrimination requirements, Tandon would require the state to justify this disparate treatment under strict scrutiny, which Maine would surely struggle to do.[57]

Maine may think it has “outmaneuver[ed]” the Supreme Court.[58] But Carson is further confirmation that the Court has already thought long and hard about these issues and is prepared to continue to protect religious autonomy and, with it, a healthy separation between church and state.[59]

[1] Nick Reaves is counsel at the Becket Fund for Religious Liberty. Becket filed an amicus brief in the case discussed in this article. But the views expressed here do not necessarily reflect the views of Becket or its clients. The author thanks his colleagues Mark Rienzi, Eric Rassbach, and Diana Verm. Any errors remain his own.

[2] Carson v. Makin, 142 S. Ct. 1987, 1997–2002 (2022).

[3] Our Lady of Guadalupe Sch. v. Morrissey-Berru, 140 S. Ct. 2049 (2020).

[4] Carson, 142 S. Ct. at 2002.

[5] Id. at 2001.

[6] Trinity Lutheran Church of Columbia, Inc. v. Comer, 137 S. Ct. 2012 (2017).

[7] Locke v. Davey, 540 U.S. 712 (2004).

[8] Carson, 142 S. Ct. at 2002.

[9] Id. (alteration in original) (emphasis added).

[10] Id. at 1993.

[11] Me. Revised Statute Annotated, Title 20-A, §5204(4) (Cum. Supp. 2021).

[12] Carson, 142 S. Ct. at 1994. Putting aside Maine’s requirement that schools be “non-sectarian,” the state’s only other requirement was that schools either be regionally accredited or satisfy “specified curricular requirements” such as using English as the language of instruction and offering a course in “Maine history.” Id. at 1993. For schools that are accredited, they need not also meet Maine’s curricular requirements. Id.

[13] Carson, 142 S. Ct. at 2007–08 (Breyer, J., dissenting) (alteration in original).

[14] Transcript of Oral Argument at 56, Carson v. Makin, 142 S. Ct. 1987 (2022) (No. 20-1088).

[15] Id. at 56–57.

[16] Id.

[17] Id. at 57.

[18] Id. at 90.

[19] Brief of Respondent at 6, Carson v. Makin, 142 S. Ct. 1987 (2022) (No. 20-1088).

[20] Id. at 5–6.

[21] Transcript of Oral Argument at 90, Carson v. Makin, 142 S. Ct. 1987 (2022) (No. 20-1088) (alteration in original).

[22] Carson v. Makin, 142 S. Ct. 1987, 2001 (2022).

[23] Maine argued both at oral argument and in its brief that many schools “self-identify as nonsectarian” and that it is “extremely rare” for Maine to be “forced to make a determination.” Brief of Respondent at 5, Carson v. Makin, 142 S. Ct. 1987 (2022) (No. 20-1088). Putting aside the selective (and potentially complaint-driven) enforcement problems with this system, there is also obviously no exemption in the Constitution for “just a little bit” of religious discrimination.

[24] Carson, 142 S. Ct. at 2006–07 (Breyer, J., dissenting); Brief of Respondent at 35–40, Carson v. Makin, 142 S. Ct. 1987 (2022) (No. 20-1088).

[25] Espinoza v. Mont. Dep’t of Revenue, 140 S. Ct. 2246 (2020).

[26] Carson v. Makin, 979 F.3d 21, 40 (1st Cir. 2020) (emphasis added); 142 S. Ct. at 2007 (Breyer, J., dissenting) (similar).

[27] Trinity Lutheran Church of Columbia, Inc. v. Comer, 137 S. Ct. 2012, 2024 (2017) (“The rule is simple: No churches need apply.”); Espinoza, 140 S. Ct. at 2256 (“So applied, the provision ‘impose[s] special disabilities on the basis of religious status.’”) (internal citation omitted).

[28] Espinoza, 140 S. Ct. at 2257 (“None of this is meant to suggest that we agree with the Department . . . that some lesser degree of scrutiny applies to discrimination against religious uses of government aid.”).

[29] Carson, 142 S. Ct. at 2001.

[30] Id. (alteration in original).

[31] Id. (alteration in original).

[32] Watson v. Jones, 80 U.S. 679, 727 (1871). See also Kedroff v. St. Nicholas Cathedral of Russian Orthodox Church in N. Am., 344 U.S. 94, 115 (1952); Kreshik v. St. Nicholas Cathedral, 363 U.S. 190, 191 (1960).

[33] Larson v. Valente, 456 U.S. 228, 244 (1982) (alteration in original).

[34] Hosanna-Tabor Evangelical Lutheran Church & Sch. v. EEOC, 565 U.S. 171, 190 (2012) (alteration in original). See also Our Lady of Guadalupe Sch. v. Morrissey-Berru, 140 S. Ct. 2049, 2062 (2020).

[35] Town of Greece v. Galloway, 572 U.S. 565, 566 (2014).

[36] Am. Legion v. Am. Humanist Ass’n, 139 S. Ct. 2067, 2096–97 (2019) (Thomas, J., concurring) (cleaned up). See also Walz v. Tax Comm’n of City of N.Y., 397 U.S. 664, 675 (1970) (explaining that government programs requiring “official and continuing surveillance” to administer can lead “to an impermissible degree of entanglement”).

[37] Our Lady, 140 S. Ct. at 2061.

[38] Id. at 2060. See also Christian Legal Soc’y Chapter of the Univ. of Cal., Hastings Coll. of L. v. Martinez, 561 U.S. 661, 680 (“[W]ho speaks . . . colors what concept is conveyed.”).

[39] Our Lady, 140 S. Ct. at 2063.

[40] Brief of Respondent at 19, Carson v. Makin, 142 S. Ct. 1987 (2022) (No. 20-1088).

[41] Transcript of Oral Argument at 87, Carson v. Makin, 142 S. Ct. 1987 (2022) (No. 20-1088).

[42] Id at 87–88 (alteration in original).

[43] Carson v. Makin, 142 S. Ct. 1987, 2008–09 (2022) (Breyer, J., dissenting) (emphasis added) (citing Plyler v. Doe, 457 U.S. 202, 221 (1982)).

[44] Statement of Maine Attorney General Aaron Frey on Supreme Court Decision in Carson v. Makin, Office of the Maine Attorney General (June 21, 2022) (alteration in original) (emphasis added),

[45] Eric Rassbach, The End of the Anti-Religion Blaine Amendments is a Victory for Religious Freedom, Real Clear Religion (July 9, 2020),

[46] 540 U.S. 712 (2004).

[47] Id. at 725.

[48] Carson, 142 S. Ct. at 2002.

[49] Id.

[50] Statement of Maine Attorney General Aaron Frey on Supreme Court Decision in Carson v. Makin, Office of the Maine Attorney General (June 21, 2022),

[51] Id.

[52] Id.

[53] See, e.g., Bus. Leaders in Christ v. Univ. of Iowa, 991 F.3d 969 (8th Cir. 2021); Intervarsity Christian Fellowship/USA v. Univ. of Iowa, 5 F.4th 855 (8th Cir. 2021); InterVarsity Christian Fellowship/USA v. Bd. of Governors of Wayne State Univ., 534 F. Supp. 3d 785 (E.D. Mich. 2021).

[54] See, e.g., Fulton v. City of Philadelphia, 141 S. Ct. 1868, 1878 (rejecting argument that government contracting programs are subject to lesser scrutiny under the Free Exercise Clause).

[55] Id.

[56] See Masterpiece Cakeshop, Ltd. v. Colo. C.R. Comm’n, 138 S. Ct. 1719, 1739 (2018). See also Agency for Int’l Dev. v. All. for Open Soc’y Int’l, Inc., 570 U.S. 205, 215 (2013); Walz v. Tax Comm’n of City of N.Y., 397 U.S. 664, 696 (1970) (Harlan, J., concurring) (“The Court must survey meticulously the circumstances of governmental categories to eliminate, as it were, religious gerrymanders.”). Maine even conceded in its briefing that it defined its funding program specifically to exclude “sectarian” schools. Brief of Respondent at 2, Carson v. Makin, 142 S. Ct. 1987 (2022) (No. 20-1088) (“As long as the school provides a nonsectarian (i.e., public) education, it may receive public funds.”).

[57] Tandon v. Newsom, 141 S. Ct. 1294, 1296 (2021); Me. Revised Statute Annotated, Title 5, §4553(2-A) (exempting single-sex pri­vate schools).

[58] Aaron Tang, There’s a Way to Outmaneuver the Supreme Court, and Maine Has Found It, NY Times (June 23, 2022) (alteration in original),

[59] New survey finds widespread support for letting Church, not State, control internal religious direction, Becket (June 17, 2020),

Read More »