Making Federalism Great Again How the Trump Administration’s Attack on Sanctuary Cities Unintentionally Strengthened Judicial Protection for State Autonomy

Article - - Symposium Issue


Over the last few years, “sanctuary cities”—jurisdictions that refuse to assist federal government attempts to deport undocumented immigrants—have been at the center of the growing political conflict over immigration policy. Donald Trump targeted sanctuary cities for special opprobrium in his 2016 presidential campaign,[1] and his Administration has made a priority of forcing them to comply with federal dictates. The Trump Administration’s efforts to punish these jurisdictions have led to multiple legal battles over constitutional federalism.

The resulting court decisions have dealt a series of defeats for the Administration that have significant implications going beyond the field of immigration policy. This is the first academic article to attempt a comprehensive evaluation of the federalism issues at stake in the Trump-era litigation on sanctuary cities.[2] The Trump Administration’s crackdown on sanctuary jurisdictions has helped make federalism great again. It achieved this unintended outcome by generating a series of court decisions protecting state and local governments against federal coercion, and by leading many on the political left to take a more favorable view of judicial enforcement of constitutional limits on federal power.

Part I of the Article summarizes the three main sets of cases that have been the focus of sanctuary litigation under the Trump Administration: legal challenges to Trump’s January 2017 executive order on sanctuary cities;[3] challenges to then-Attorney General Jeff Sessions’s July 2017 policy of conditioning federal law enforcement grants on state and local government cooperation with federal efforts to deport undocumented immigrants;[4] and the Administration’s lawsuit against California’s “sanctuary state” law.[5]

All three sets of cases raise crucial issues for federalism. If the Administration were to prevail in the litigation over either Executive Order (EO) 13768 or the Sessions policy, the Executive would have broad power to impose new conditions on federal grants to state and local governments that have not been authorized by Congress. That would give the President immense leverage over states and localities, which are dependent on federal funds on a variety of issues. It would also undercut Congress’s control of the power of the purse.

Many of the decisions ruling against the Administration also address the constitutionality of 8 U.S.C. § 1373, a federal law barring state and local governments from instructing their employees to refuse to provide federal officials “information regarding the citizenship or immigration status, lawful or unlawful, of any individual.”[6] Entirely aside from its use as a grant condition in the Trump Executive Order and the Sessions policy, § 1373 may be unconstitutional under Supreme Court decisions barring federal “commandeering” of state and local governments under the Tenth Amendment. Should § 1373 be upheld by the courts, it would enable the federal government to use similar strategies to circumvent the
anti-commandeering rule in other contexts.

In Part II, I examine the results of the litigation to date. With rare exceptions, courts have ruled against the Administration on all the major federalism issues at stake in the sanctuary cases. I explain why these results are largely justified, and also why the Administration’s opponents deserve to prevail on the one key issue in the sanctuary state case that has so far gone in the Administration’s favor: the dispute over the constitutionality of Assembly Bill (AB) 450, a provision of California’s sanctuary state law restricting private-employer cooperation with federal Immigration and Customs Enforcement (ICE) raids.[7]

Part III considers the broader implications of the sanctuary litigation for federalism. Should the Administration prevail, it would have very dangerous implications for constitutional limits on federal power more generally. The consequences would extend far beyond the specific context of immigration policy. Even conservatives who sympathize with the Trump Administration’s campaign against sanctuary cities are likely to have reason to fear the ways in which future administrations—including liberal-Democratic ones—could use these same powers.

The sanctuary cities litigation is also a striking example of the potentially shifting ideological valence of judicial enforcement of federalism. For most of the period since the New Deal, such enforcement has often been viewed favorably by the political right and opposed by the left. One of the main reasons for left-of-center skepticism of federalism has been its apparent historic association with discrimination against unpopular racial and ethnic minorities. The sanctuary cities cases are among the recent developments indicating that we can no longer assume that federal power is necessarily the friend of minorities, and state and local authorities their enemies. In these cases, liberal “blue” jurisdictions have been relying on Supreme Court federalism precedents, authored by conservative Justices, to protect (primarily) Latino undocumented immigrants against a Republican administration asserting a broad view of federal power historically associated with the left.

The sanctuary cities cases suggest that the traditional ideological division over judicial review of federalism may be outdated. But the extent to which the left and right shift their positions on these issues remains to be seen.

In this Article, I do not consider issues raised in the sanctuary cities litigation that are not directly related to constitutional-federalism limits on federal power. Thus, I do not consider procedural questions such as standing and ripeness,[8] nor do I address the debate over whether nationwide injunctions are appropriate remedies in these cases—a hotly contested issue that goes beyond the federalism context.[9] I also do not consider issues of preemption raised by the California sanctuary state case, which do not directly pertain to the structural scope of federal power relative to the states, as opposed to interpretations of the scope of particular statutes.[10]

Finally, this Article does not attempt to ascertain the correct original meaning of the relevant constitutional provisions, though I do briefly explain how the sanctuary decisions are at least roughly consonant with it.[11]

I. Overview of the Sanctuary Jurisdiction Cases

Over the last two years, state and local governments have opposed the Trump Administration on three major sets of sanctuary cases: challenges to the President’s Executive Order on sanctuary cities; challenges to the Justice Department’s July 2017 policy imposing new immigration-enforcement conditions on Edward Byrne Memorial Justice Assistance Grants to states and localities; and the Administration’s lawsuit against California’s sanctuary state law. This Part gives a brief overview of the issues at stake in the three sets of cases.

A. Executive Order 13768

On January 25, 2017, President Trump issued EO 13768.[12] The order avows that sanctuary cities “that fail to comply with applicable Federal law do not receive Federal funds, except as mandated by law.”[13] More specifically, it mandates that:

[T]he Attorney General and the [Homeland Security] Secretary, in their discretion and to the extent consistent with law, shall ensure that jurisdictions that willfully refuse to comply with 8 U.S.C. § 1373 (sanctuary jurisdictions) are not eligible to receive Federal grants, except as deemed necessary for law enforcement purposes by the Attorney General or the Secretary.[14]

Section 1373 is a somewhat unusual federal law that mandates “a Federal, State, or local government entity or official may not prohibit, or in any way restrict, any government entity or official from sending to, or receiving from, the Immigration and Naturalization Service information regarding the citizenship or immigration status, lawful or unlawful, of any individual.”[15]

The text of EO 13768 is extraordinarily broad. It appears to cover all federal grants received by state and local governments that “willfully” violate § 1373, regardless of the size of the grant, the policy issue it relates to, or whether Congress has ever conditioned the grant on compliance with § 1373. This seems inconsistent with Supreme Court Spending Clause precedent requiring conditions attached to federal grants to be explicitly spelled out by Congress,[16] mandating that they must be “related” to the purposes of the grant,[17] and limiting their scope so as to avoid “coercion” of state and local governments.[18] The grants and conditions must also meet the Spending Clause’s requirement that the spending be “to pay the Debts and provide for the common Defense and general Welfare of the United States.”[19] But modern precedent defines “general welfare” so broadly that virtually any purpose passes muster.[20]

Pennhurst State School and Hospital v. Halderman,[21] the 1981 precedent mandating that conditions must be “unambiguously” stated by Congress, came in the context of a statute subjecting funding recipients to a private right of action.[22] But nothing in the reasoning of the Court suggests that the rule is limited to that context. To the contrary, the Court’s logic applies to all funding conditions. As the decision emphasizes, “legislation enacted pursuant to the spending power is much in the nature of a contract . . . . By insisting that Congress speak with a clear voice, we enable the States to exercise their choice knowingly, cognizant of the consequences of their participation.”[23] That reasoning applies to all conditional grants, whether they involve private rights of action or not.

In addition, § 1373 may itself be unconstitutional, because it runs afoul of Supreme Court precedents interpreting the Tenth Amendment as barring federal “commandeering” of state and local governments in order to force them to assist in enforcing federal law.[24]

As we shall see in Part II, both the Spending Clause issues and the commandeering question have been raised by local governments challenging the Executive Order. The resulting rulings have not gone well for the Trump Administration.

B. The Justice Department’s Imposition of New Conditions on Recipients of Edward Byrne Memorial Justice Assistance Grants

In July 2017, the Justice Department sought to impose three new conditions on state and local governments that receive Byrne Grants by mandating that they:

[1] certify compliance with section 1373, a federal statute applicable to state and local governments that generally bars restrictions on communications between state and local agencies and officials at the Department of Homeland Security;

[2] permit personnel of the U.S. Department of Homeland Security (“DHS”) to access any detention facility in order to meet with an alien and inquire as to his or her right to be or remain in the United States; and

[3] provide at least 48 hours advance notice to DHS regarding the scheduled release date and time of an alien in the jurisdiction’s custody when DHS requests such notice in order to take custody of the alien.[25]

The Justice Department recently added two new conditions to this list for potential grant recipients for the 2018 fiscal year. These conditions raise much the same sorts of legal issues as the original three mandated for 2017.[26]

The Byrne Grant program provides assistance to state and local law enforcement agencies, disbursing some $275 million in 2016.[27] The Justice Department’s attempts to attach new conditions raise many of the same issues as the § 1373 condition imposed by EO 13768.

Because the conditions are limited to one law enforcement grant program, the Byrne conditions do not raise the issue of “coercion” of states and localities that arises from the Executive Order’s sweeping targeting of virtually all federal grants to states and localities. But plaintiffs challenging it still argue that it violates the requirement of congressional authorization for grant conditions, that it fails the “relatedness” test, and that § 1373 is independently unconstitutional.[28]

C. The California Sanctuary State Laws

The litigation over the California sanctuary state legislation raises a wider range of issues than either of the other two sets of sanctuary cases. The lawsuit filed by the Justice Department challenges three state laws enacted to protect undocumented immigrants.

Senate Bill 54 restricts the range of information state and local governments are allowed to share with federal immigration enforcers[29] and bars state law enforcement agencies from “[t]ransfer[ring] an individual to immigration authorities unless authorized by a judicial warrant or judicial probable cause determination.”[30] Assembly Bill 103, enacted in June 2017, requires the state’s Attorney General to inspect and report on conditions in county, local, and private detention facilities in which noncitizens are detained for purposes of civil-immigration proceedings within the state.[31] The inspection requirements also extend to “the circumstances around [the] apprehension” of the detainees and their transfer to the facility in question.[32] The Attorney General was required to complete the review and file a report by March 1, 2019.[33]

Finally, Assembly Bill 450 bars employers from voluntarily allowing a federal immigration-enforcement agent to enter “nonpublic” areas of their workplaces or to access, review, or obtain the employees’ records.[34] It also requires employers to provide advance notice to their employees of “any inspections of I-9 Employment Eligibility Verification forms or other employment records conducted by an immigration agency within 72 hours of receiving notice of the inspection” and mandates that employers provide each affected, current employee the results of the inspection within seventy-two hours of receipt.[35]

The federal government argues that all three laws are preempted by federal legislation[36] and also that AB 450 and AB 103 violate the Supremacy Clause of the Constitution[37] because of the doctrine of “intergovernmental immunity,” which bars state laws that discriminate against the federal government and those who “deal” with it.[38]

II. A Preliminary Assessment of the Sanctuary Jurisdiction Decisions

Litigation over all three sets of sanctuary cases is ongoing. At this point, I can therefore offer only a preliminary assessment of the results. This is particularly true of the sanctuary state case, which has so far been the subject of only two preliminary injunction rulings.[39] By contrast, EO 13768 and the Justice Department’s Byrne Grant conditions have been the subject of multiple decisions by both trial and appellate courts.

Overall, state and local governments have so far prevailed on nearly all of the federalism issues at stake in these cases. These rulings are, in my view, correct. If they are not overruled by higher courts, the sanctuary litigation is likely to go down in history as a series of important victories for the enforcement of constitutional limitations on federal power.

In this Part, I assess the judicial decisions in each set of sanctuary cases in turn: starting with EO 13768 and continuing on to the Byrne Grant cases and the California sanctuary state cases. Because the issue of the constitutionality of 8 U.S.C. § 1373 occurs in all three sets of cases, I deal with it separately in the last section of this Part.

A. Litigation Over Executive Order 13768

Executive Order 13768 has been the subject of three federal district court rulings, all of which have concluded that it is unconstitutional.[40] The first and second are part of the same case; their reasoning on the federalism questions has been upheld on appeal by the U.S. Court of Appeals for the Ninth Circuit.[41]

In April 2017, Judge William Orrick of the U.S. District Court for the Northern District of California issued a preliminary injunction against EO 13768 in a lawsuit brought by the cities and counties of San Francisco and Santa Clara.[42] Judge Orrick concluded that:

Section 9 [of the Executive Order] purports to give the Attorney General and the Secretary the power to place a new condition on federal funds (compliance with Section 1373) not provided for by Congress. But the President does not have the power to place conditions on federal funds and so cannot delegate this power.[43]

This violates longstanding Supreme Court precedent mandating that conditions on federal grants must be “unambiguously” established by Congress.[44] In this case, Judge Orrick concluded there was no indication that Congress intended any of the federal grant programs covered by the Order to be conditioned on compliance with § 1373.[45]

A contrary decision would have undermined both federalism and the separation of powers, giving the president leverage to coerce state and local governments and usurping congressional control over the power of the purse.[46]

Perhaps because they recognized the weakness of their position, Justice Department lawyers tasked with defending the Order claimed it should not actually be interpreted to apply to all federal grants but merely to three federal law enforcement grants that, they contended, were already conditioned on compliance with § 1373.[47] This understanding of the Order has been endorsed by some legal commentators as well.[48]

Judge Orrick correctly responded that such a narrow interpretation of the EO makes little sense, and goes against its plain text:

[The government] explained for the first time at oral argument that the Order is merely an exercise of the President’s “bully pulpit” to highlight a changed approach to immigration enforcement. Under this interpretation, Section 9(a) applies only to three federal grants in the Departments of Justice and Homeland Security that already have conditions requiring compliance with 8 U.S.C. 1373. This interpretation renders the Order toothless; the Government can already enforce these three grants by the terms of those grants and can enforce 8 U.S.C. 1373 to the extent legally possible under the terms of existing law. Counsel disavowed any right through the Order for the Government to affect any other part of the billions of dollars in federal funds the Counties receive every year. . . . But Section 9(a), by its plain language, attempts to reach all federal grants, not merely the three mentioned at the hearing. The rest of the Order is broader still, addressing all federal funding.[49]

In interpreting statutes, the standard judicial practice is to avoid interpretations that would render them utterly ineffectual and superfluous,[50] or “toothless” in Judge Orrick’s words.[51] Admittedly, the Supreme Court has never clearly ruled on whether this traditional rule of interpretation applies to executive orders as well as statutes.[52] However, “[a]s is true of interpretation of statutes, the interpretation of an Executive Order begins with its text,” which “must be construed consistently with the Order’s ‘object and policy.’”[53]

In a forthcoming article discussing EO 13768, Professor Tara Leigh Grove suggests that interpreting an executive order in a way that renders it “toothless” might be permissible, because that toothlessness could be a result of interagency consultation that led the president to moderate his initial intentions.[54] But the content of an executive order more fully reflects the goals of a single decision-maker—the president—than does the text of a statute, which is usually the product of negotiation between multiple legislators, with divergent objectives. If it nonetheless fails to reflect the president’s priorities, it is easier for him to revise it. He need only issue a new order, or even just revise a section of the original one.[55] If there is nonetheless a presumption against interpreting statutes in a way that renders parts of them inoperative despite the possibility that such toothlessness may be the product of legislative negotiation, the same rule should apply to executive orders.

Moreover, the claim that the Trump Executive Order merely restates the requirements of existing law is undercut by the Administration’s own insistence that it is intended to deprive sanctuary jurisdictions that violate § 1373 of funds provided by three federal law enforcement grants.[56] The statutes authorizing the three grant programs identified by Administration lawyers also lack any requirement forcing recipients to adhere to § 1373.[57]

Defenders of the Administration’s position also point to the provision in Trump’s order that indicates funds can only be withheld “to the extent consistent with law.”[58] This could be interpreted as forswearing any denial of funds that violates constitutional constraints, including the requirement that grant conditions be authorized by Congress.

But, as Judge Orrick explained, this interpretation would make the Order utterly ineffectual and “toothless.”[59] It would not authorize the withholding of any funds that could not be withheld anyway. A more plausible interpretation of the Order is that it assumes, under existing law, there is some significant class of federal grants that could be withheld by executive order even if Congress did not authorize the conditions the Executive wants to impose.

The claim that EO 13768 is limited to authorizing withholding of grants that could already be denied under preexisting law was also undermined by the Administration’s decision to appeal Judge Orrick’s ruling, despite the fact that his opinion explicitly permitted withholding in such cases.[60] If the Administration only sought the right to withhold grants in accordance with preexisting law, there would have been no need to appeal the ruling, and Administration officials would not have vehemently denounced the decision, as they in fact did.[61] It is, to say the least, unusual for Administration officials to appeal rulings that give them the very thing they supposedly want.

Judge Orrick also ruled that EO 13768 violates the requirement that conditions on federal grants be “reasonab[ly] related” to the purposes of the grant program, a restriction that is clearly at odds with the Order’s requirement that nearly all federal grants be tied to compliance with § 1373.[62] In addition, he concluded that the Order’s grant conditions are “unconstitutionally coercive” because they threaten to deny sanctuary jurisdictions all federal grants, hundreds of millions of dollars on which the counties rely.[63]

The Court’s relatedness jurisprudence is often very lax.[64] But Judge Orrick was on strong ground in suggesting that EO 13768 violates even those weak constraints because “there is no nexus between Section 1373 and most categories of federal funding, including without limitation funding related to Medicare, Medicaid, transportation, child welfare services, immunization and vaccination programs, and emergency preparedness.”[65]

The Supreme Court’s jurisprudence on coercion in grant conditions is also far from a model of clarity. It is, to put it mildly, not easy to figure out the point at which “financial inducement” becomes “so coercive as to pass the point at which pressure turns into compulsion.”[66] But Judge Orrick was surely right to rely on Chief Justice John Roberts’s famous statement in NFIB v. Sebelius,[67] where the latter indicated that coercion is clearly present in a situation where the amount of federal funding in question is so large that the threat to remove it functions as a “gun to the head.”[68] If anything qualifies as a “gun to the head,” surely the threat to remove all federal grants—or even all not somehow specifically exempted by law—qualifies as such.

More dubiously, Judge Orrick concluded that the Executive Order violates the Tenth Amendment’s rules against commandeering, but without directly reaching the question of whether § 1373 is unconstitutional.[69] This issue is discussed in more detail below.[70]

Judge Orrick reaffirmed the reasoning of his initial ruling issuing a preliminary injunction against the Executive Order in a later opinion issuing a final judgment against it.[71] The latter ruling was later upheld by the Ninth Circuit.[72]

In between the two district court rulings, the Department of Justice (DOJ) issued a memorandum formally endorsing the limited reading of the Executive Order previously offered by DOJ lawyers in oral argument over the preliminary injunction.[73] Judge Orrick rightly rejected the memorandum for much the same reasons as he had previously rejected the less formal litigation position making the same claim:

The AG memorandum not only provides an implausible interpretation of Section 9(a) [of the order] but is functionally an “illusory promise” because it does not amend Section 9(a) and does not bind the Executive [B]ranch. It does not change the plain meaning of the Executive Order.

. . . .

The federal government attempts to read out all of Section 9(a)’s unconstitutional directives to render it an ominous, misleading, and ultimately toothless threat. . . . If Section 9(a) does not direct the Attorney General and [Homeland Security] Secretary to place new conditions on federal funds, then it only authorizes them to do something they already have the power to do, which is to enforce existing grant requirements; effectively, the federal government argues that Section 9(a) is “valid” and does not raise constitutional issues as long as it does nothing at all. But a construction so narrow that it renders a legal action legally meaningless cannot possibly be reasonable and is clearly inconsistent with the Executive Order’s broad intent.[74]

While the San Francisco and Santa Clara cases were being litigated, the cities of Seattle and Portland filed a similar challenge to EO 13768 in the U.S. District Court for the Western District of Washington.[75] In October 2017, Judge Richard Jones issued a ruling denying the federal government’s motion to dismiss.[76] His reasoning was very similar to Judge Orrick’s. He too ruled that the Order violates the Spending Clause and separation-of-powers principles because Congress had never authorized the imposition of the grant conditions required by the Order, because they are not “related” to the federal interest in the grants in question, and because they are unconstitutionally “coercive.”[77] Like Judge Orrick, Judge Jones also concluded that the DOJ memorandum is not binding and cannot override the plain text of the Order.[78]

The Ninth Circuit decision upholding Judge Orrick’s second ruling similarly relies on much the same reasoning as Judge Orrick and Judge Jones.[79] It too held that the Order violates the Spending Clause and separation of powers, and rejected the supposed authority of the DOJ memorandum.[80] The ruling did, however, narrow the scope of the nationwide injunction issued by Judge Orrick, rendering it applicable only to the parties in the case.[81]

A dissent by Judge Fernandez contended that the court should have granted deference to the DOJ memorandum[82] and, by implication, the Executive Order’s “savings clause” requiring withdrawal of federal funds only to the extent consistent with law. But Judge Fernandez’s extremely brief opinion does little to address the weighty objections to these conclusions. As Judge Thomas pointed out in his majority opinion:

Savings clauses are read in their context, and they cannot be given effect when the Court, by rescuing the constitutionality of a measure, would override clear and specific language.

. . . .

If “consistent with law” precludes a court from examining whether the Executive Order is consistent with law, judicial review is a meaningless exercise, precluding resolution of the critical legal issues.[83]

In this case, the savings clause cannot be read to render the Order meaningless or negate the clear language applying it to virtually all federal grants to state and local governments.[84]

If the President’s purpose really was limited to three law enforcement grants or to grants that Congress has previously conditioned on compliance with § 1373, he could easily have modified the Order at any time to clearly state as much.[85] The fact that he never did so, even as the litigation over the Order has gone on for almost two years, is a strong indication that the true purpose was broader.

Unlike in the case of a statute, which can only be modified by going through a complicated and difficult bicameral legislative process, an executive order can easily be modified simply by a stroke of one person’s pen. Thus, it makes more sense to draw inferences from a failure to modify in the latter case. Courts also have much less reason to adopt convoluted interpretations of executive orders to save them from unconstitutionality than might be true in the case of statutes, where there is a strong canon of interpreting statutes to avoid constitutional problems.[86] The difficulty of amending statutes is a possible justification for that canon which does not apply to executive orders.

And, as both Judge Orrick and the Ninth Circuit noted in their rulings, public statements by the President and then-Attorney General Jeff Sessions suggested that they sought to use the Executive Order to pull back a much wider range of grants.[87] Such statements are more probative in the case of an executive order than in the case of legislation because an executive order more clearly reflects a decision in which one person—the President—has the final say. By contrast, legislation is enacted by multimember bodies who often have divergent objectives.

In analyzing the rulings against EO 13768, it is important to recognize that the Order raises troubling issues even if its scope is limited to three federal law enforcement grants, as the DOJ memorandum contends. Even in that event, it still seeks to attach grant conditions that were never authorized by Congress.[88] And if it is possible to do so in this case, the Executive could readily do the same thing elsewhere.[89]

If, under the DOJ memorandum, the Executive Order can be interpreted to apply to these three federal grants, despite the lack of explicit congressional authorization, the Administration (or a successor) could easily broaden that theory to apply to other grants, especially ones that have some sort of connection to law enforcement. If, on the other hand, the Order really is limited to imposing conditions previously authorized by Congress, then it is indeed meaningless and “toothless,” as Judge Orrick put it.[90]

B. Litigation Over the DOJ Byrne Grant Conditions

The DOJ policy seeking to impose three new conditions on recipients of Edward Byrne law enforcement grants has been the subject of extensive litigation in multiple federal courts. All told, so far there have been nine district court rulings against the Byrne conditions.[91] Two of them have been affirmed by appellate courts—the Third Circuit and the Seventh Circuit.[92] The nine rulings address six separate cases: one brought by the City of Chicago; one by the City of Evanston, Illinois; one by the City of Philadelphia; one by the State of California and the County of San Francisco; one by the City of Los Angeles; and one brought by seven state governments and New York City.[93] Additionally, there have been two very recent rulings striking down the modified fiscal-year-2018 version of the Byrne Grant conditions, which added two new requirements to the three imposed for fiscal year 2017.[94]

1. Lack of Congressional Authorization.—With one exception, to be discussed shortly, all of these decisions struck down the three conditions because none of them were ever authorized by Congress. As the Seventh Circuit put it:

The Attorney General in this case used the sword of federal funding to conscript state and local authorities to aid in federal civil immigration enforcement. But the power of the purse rests with Congress, which authorized the federal funds at issue and did not impose any immigration enforcement conditions on the receipt of such funds.[95]

All but the most recent of these cases have already had two district court rulings each: one on a motion for a preliminary injunction against the three conditions and one issuing a final decision.[96]

For example, the most recent district court decision on the original 2017 conditions, addressing a lawsuit brought by seven state governments and New York City, concluded that the Justice Department was wrong to impose the three requirements based on vague statutes giving the Attorney General the authority to “exercise such other powers and functions as may be vested in the Assistant Attorney General pursuant to this chapter or by delegation of the Attorney General, including placing special conditions on all grants, and determining priority purposes for formula grants,” and requiring grant recipients to obey “applicable federal laws.”[97] Doing so violates the rule that grant conditions must be “unambiguously”[98] stated by Congress, so that “a state official would clearly understand that one of the obligations of the Act is the [purported] obligation.”[99] For that reason, the court concluded that “applicable federal laws” must be limited to those that specifically govern the duties of grant recipients, not all federal laws that constrain state and local governments in any way.[100] In addition, allowing the Attorney General to impose new conditions not authorized by Congress is clearly beyond the plain text of a law that merely authorizes him to impose conditions needed to exercise “such power[s]” as were vested in him elsewhere in the same law.[101] Other Byrne Grant decisions reach similar conclusions.[102]

This conclusion may be considered simply a minor example of statutory interpretation. But it has important broader implications. If vague formulations such as “applicable federal laws” can be used to impose new conditions on Byrne Grant recipients, the same is true of a wide variety of other federal grants. In this way, the Executive could apply a broad range of conditions to federal grants without any clear authorization by Congress.[103]

The one exception to the trend of federal courts holding that the Byrne Grant conditions were not authorized by Congress was a ruling by Judge Harry Leinenweber of the Northern District of Illinois, in City of Chicago v. Sessions I.[104] Judge Leinenweber recognized that nothing in the statute authorizing the Byrne Grants specifically mandates compliance with § 1373. But he concludes that § 1373 is nonetheless a condition of the grant because, as we have already seen, the authorizing statute states that recipients must “comply with all provisions of this part and all other applicable Federal laws.”[105]

As in other Byrne Grant cases, the Justice Department claimed that “all other applicable Federal laws” includes all laws that regulate the recipient jurisdictions in any way, while Chicago argued that “applicable” laws include only those that specifically regulate recipients of federal grants.[106] Judge Leinenweber correctly noted that both interpretations are “plausible.”[107] He could have saved himself considerable trouble if he had just stopped right there.

Longstanding Supreme Court precedent mandates that the federal government may not impose conditions on grants to states and localities unless the conditions are “unambiguously” stated in the text of the law, “so that the States can knowingly decide whether or not to accept those funds.”[108] If it is “plausible” to interpret the relevant law in a way that excludes the condition, then it is obvious that the condition is not unambiguously stated in the text of the law.

Instead, Judge Leinenweber concluded that a detailed analysis of the text leads to the conclusion that the federal government’s interpretation of the phrase “all other applicable Federal laws” is preferable to Chicago’s.[109] He then also had to address the issue of whether § 1373 is itself unconstitutional. While he initially answered that then-difficult question in the negative,[110] he ultimately reversed his position in Chicago v. Sessions II,[111] after the situation was altered by the Supreme Court’s intervening decision in Murphy v. NCAA.[112]

Professor Peter Margulies offers a different rationale for upholding the § 1373 condition. He argues that it fits the logic and purpose of the Byrne Grant program because, properly interpreted, it imposes only a fairly limited obligation on state and local governments, one that most already largely comply with.[113]

But even if the condition is indeed consistent with the logic and purpose of the program, such consistency is not an adequate substitute for actual, clear congressional authorization of the condition. If the Executive could impose any new condition on federal grants that is plausibly consistent with the logic and purpose of the program in question, that would open the door to enormous executive discretion to reshape federal grant programs.

Two very recent federal district court decisions have invalidated the new fiscal-year-2018 version of the Justice Department’s Byrne Grant policy.[114]

The new policy includes slightly revised versions of the original three conditions, plus two new ones. The “nondisclosure condition” bars “public disclosure . . . of any federal law enforcement information in a direct or indirect attempt to conceal, harbor, or shield from detection” any undocumented immigrants.[115]

The information condition . . . requires award recipients to collect certain information from sub-grant recipients. For example, California would not be able to authorize a sub-grant “unless it first obtains from the proposed subrecipient responses to the questions identified in the program solicitation as [‘]Information regarding Communication with the Department of Homeland Security (DHS) and/‌or Immigration and Customs Enforcement (ICE).[’] . . .”[116]

Two federal district courts have invalidated these two new conditions for much the same reasons as they previously invalidated the three original ones; they were never authorized by Congress, and only Congress has the power to authorize federal grants to state and local governments and impose conditions on recipients.[117]

2. Relatedness.—In addition to ruling against the federal government’s Byrne Grant conditions on grounds of lack of congressional authorization, two district court cases have also concluded that the conditions fail to meet the requirement of relatedness to the federal interest in the grant program at issue.[118] In Philadelphia I, Judge Baylson recognized that relatedness is far from a clear concept and that the Court has never struck down a grant condition for failing this requirement.[119] But he ultimately concluded that the Byrne Grant conditions fail because federal immigration enforcement is distinct from the conventional state law enforcement objectives the program is intended to promote:

[T]here is a seemingly endless list of areas of the law which can be said to be “related to” criminal justice and the local enforcement of criminal laws; it is not automatic, however, that these relationships operate in both directions. For example, while criminal law bears enormously on voting rights, voting laws don’t appear to have any impact on the criminal justice system.

Criminal law is integral to immigration law, specifying classes of noncitizens for high risk of removal, dictating procedures for detaining particular individuals pending removal proceedings, and defining who falls within the federal government’s priorities for immigration law enforcement. However, immigration law does not impact the criminal justice system.

. . . As the record has established, this is absolutely the case in Philadelphia. As Commissioner Ross testified, the criminal laws of Philadelphia are uniformly enforced across the city, without regard to the immigration status, whether lawful or unlawful, of individual residents, whether they come into contact with the criminal justice system as a witness, victim, or defendant. The City even has policies in place designed to remove immigration considerations entirely from the calculus of criminal law enforcement. . . . While federal immigration law officials care deeply about local criminal law outcomes, it simply is not the case that local criminal justice actors in Philadelphia care about federal immigration laws.

. . . [T]he fact that immigration enforcement depends on and is deeply impacted by criminal law enforcement does not mean that the pursuit of criminal justice in any way relies on the enforcement of immigration law. Realistically, it does not. Further, . . . the Byrne JAG statute is clearly designed for the purpose of enhancing local criminal justice. When considered at this level, the argument that enforcement of federal immigration laws is related to this objective is unsustainable . . . .[120]

Judge William Orrick reached a similar conclusion in a Byrne Grant case brought by the City and County of San Francisco.[121]

The obvious objection to this line of reasoning is that removal of undocumented immigrants might help reduce local crime by removing potential perpetrators of such crimes. As Judge Baylson notes, however, sanctuary jurisdictions argue that requiring local law enforcement to aid in deportation proceedings actually makes enforcement of laws against violent crime more difficult because it reduces the willingness of minority communities to cooperate with the police, and the available evidence seems to support that view.[122]

Because of the dubious and imprecise nature of the concept of relatedness, it is difficult to definitively determine whether Judge Baylson’s ruling on this point is correct. But he is surely right to point out that a relationship that runs in one direction (from criminal law enforcement to immigration law) does not necessarily go the other way, and that the latter is the connection relevant to the purposes of the Byrne Grant program.

There is little or no evidence backing the federal government’s claim that sanctuary city policies increase violent crime. To the contrary, the available social science studies all show that these policies either have no measurable impact on local crime or actually reduce it.[123]

And if an indirect relationship backed by little or no evidence is enough to meet the relatedness requirement, then almost any two policy areas can be shown to be related in much the same way. At the very least, anything that might reduce the number of people potentially likely to commit crimes would be related to the purposes of the Byrne Grant and other law enforcement programs.

For example, one can use similar reasoning to justify tying the Byrne Grants to establishing a federally mandated education curriculum in local public schools, funding federally approved health care programs, or even subsidizing the consumption of healthy food and discouraging the eating of “junk food.” It could be argued that people who get a better education have access to better health care, or those who consume healthier foods are less likely to turn to crime to sustain themselves, and therefore that these conditions are related to local law enforcement.

In a recent decision on the newly revised fiscal-year-2018 Byrne Grant conditions, Judge William Orrick concluded that the two new conditions added for that year also violate the relatedness standard.[124] He ruled that the “information” condition fails for much the same reasons as the original three conditions do: both condition an ordinary law enforcement grant on a requirement focused on immigration enforcement.[125] The “nondisclosure” requirement is potentially more defensible because it includes disclosures related to immigrants who have violated ordinary criminal law.[126] But it too ultimately fails the test because it “does not make a distinction between undocumented immigrants subject to potential criminal law enforcement and those who are not, even though ‘many immigration violations do not involve criminal law and are only violations of civil penalties.’”[127]

C. The California Sanctuary State Litigation

The litigation over California’s sanctuary state laws has not progressed nearly as far as that over EO 13768 and the DOJ Byrne Grant conditions. So far, we have one district court ruling on the subject, which is at the preliminary-injunction stage, and a court of appeals decision affirming the district court ruling on most points.[128] The questions involved are also closer calls than those at stake in the other two sets of sanctuary cases.

In a preliminary-injunction ruling issued in July 2018, Federal District Judge John Mendez ruled against the Trump Administration on most, but not all, of the issues at stake in the federal government’s high-profile lawsuit against California’s sanctuary state laws.[129] As Judge Mendez—a George W. Bush appointee—recognized, the case “presents unique and novel constitutional issues” involving federalism and immigration law.[130] The federal government has stronger claims here than in its efforts to cut federal grants to sanctuary cities by imposing conditions never authorized by Congress. Nonetheless, I believe the state ultimately deserves to prevail on all three issues involved.

Judge Mendez’s decision rejected the federal government’s request for an injunction blocking enforcement of SB 54 and AB 103.[131] But Judge Mendez did issue an injunction against the main provision of AB 450.[132]

The Trump Administration claims that SB 54 violates federal law because it conflicts with 8 U.S.C. § 1373,[133] the controversial federal law mandating that “a Federal, State, or local government entity or official may not prohibit, or in any way restrict, any government entity or official from sending to, or receiving from, the Immigration and Naturalization Service information regarding the citizenship or immigration status, lawful or unlawful, of any individual.”[134] Judge Mendez interpreted § 1373 narrowly, so that it does not conflict with SB 54.[135] He ruled that § 1373 does not require disclosure of information about immigrants’ addresses, release dates, and other matters, but focuses only on “immigration status,” narrowly construed.[136] Senate Bill 54, Judge Mendez concluded, does not cover the latter type of information.[137] But Judge Mendez additionally indicated that “the constitutionality of Section 1373 [is] highly suspect.”[138] I discuss this aspect of his ruling in more detail below.

Judge Mendez also rejected the federal government’s case against AB 103, the detention-facility inspection rule.[139] He concluded that AB 103 is not preempted by any provision of federal immigration law and noted that the inspections required under the bill are similar to those that apply to other law enforcement detention facilities in California, and that they impose little in the way of new burdens on the federal government.[140]

They therefore do not conflict with federal law and do not qualify as unconstitutional discrimination against federal facilities under the doctrine of “intergovernmental immunity,” which bars state discrimination against the federal government and its agents: “[T]he review appears no more burdensome than reviews required under California Penal Code §§ 6030, 6031.1. Thus, even if AB 103 treats federal contractors differently than the State treats other detention facilities, Plaintiff has not shown the State treats other facilities better than those contractors.”[141]

The federal government did, however, prevail on one important issue: Judge Mendez granted the request to issue an injunction against AB 450, the provision restricting employer cooperation with federal immigration-enforcement raids.[142] He did not issue any ruling on the issue of whether AB 450 is preempted by federal immigration law, though he strongly suggests it may not be because:

[I]n preemption analysis, the [Supreme] Court presumes “‘the historic police powers of the States’ are not superseded ‘unless that was the clear and manifest purpose of Congress.’” Laws governing labor relations and the workplace generally fall within the States’ police powers. Congress has not expressly authorized immigration officers to enter places of labor upon employer consent . . . .[143]

Nonetheless, Mendez ruled against California on AB 450 because he concluded it violates the doctrine of “intergovernmental immunity,” which bars state laws that “regulate the United States directly or discriminate against the Federal Government or those with whom it deals.”[144] Assembly Bill 450 runs afoul of this because it targets employers who assist the federal government in immigration enforcement.

In my view, Judge Mendez got this part of the decision wrong. As he noted, his ruling expands the definition of “dealing” from entities that have economic or contractual relationships with the federal government, to those that merely provide voluntary assistance.[145] Even under this broader definition, AB 450 does not truly “discriminate” against people who “deal” with the federal government because there is no meaningful private-sector analogue to federal immigration-enforcement raids.

The concept of discrimination implies treating similarly situated entities differently. The Supreme Court has held that a “State does not discriminate against the Federal Government and those with whom it deals unless it treats someone else better than it treats them.”[146] There is no way to prove the existence of such differential treatment unless there is a similarly situated “someone else” whose treatment by the state can be used as a baseline. The Supreme Court noted in Vacco v. Quill[147] that the antidiscrimination rule of the Equal Protection Clause of the Fourteenth Amendment “embodies a general rule that States must treat like cases alike but may treat unlike cases accordingly.”[148] Thus, for example, a state government discriminates against federal workers if it imposes taxes on their pensions that are not imposed on comparable pensions earned by state and local employees.[149] As Justice Neil Gorsuch pointed out in his opinion for the Court in a recent intergovernmental immunity case, discrimination is “something we’ve often described as treating similarly situated persons differently.”[150]

But in this case, there is no private or state activity that qualifies as a “like case” or is “similarly situated” to federal immigration-enforcement raids. There is no true counterpart to people who assist federal immigration raids because no private or state entity has the legal right to deport people with little due process, forcibly separate families, and confine people in cages, and none that can match ICE’s awful record of abusive treatment of detainees.[151]

Agency procedures are so defective that ICE has even mistakenly detained or deported thousands of American citizens.[152]

The impact of Mendez’s ruling against the main part of AB 450 was partly mitigated by his refusal to block implementation of a provision of the same law that requires employers to give employees warning of any planned federal inspection of their immigration records (which may include at least some on-site raids). Judge Mendez argues that the notice requirement is different from the anti-raid policy because:

Unlike the prohibitions on consent [to raids], violations of this provision do not turn on the employer’s choice to “deal with” (i.e., consent to) federal law enforcement. An employer is not punished for its choice to work with the Federal Government, but for its failure to communicate with its employees.[153]

Although this decision only addressed the federal government’s motion for a preliminary injunction temporarily blocking enforcement of the California laws, it nonetheless prefigures the court’s likely final decision on the merits. One of the criteria for securing a preliminary injunction is “likelihood of success on the merits,” and Judge Mendez explicitly ruled that the federal government was unlikely to succeed in its claims against SB 54 and AB 103 but was likely to prevail on AB 450.[154]

Most of Judge Mendez’s ruling was recently affirmed by the U.S. Court of Appeals for the Ninth Circuit.[155] Like Judge Mendez, the Ninth Circuit upheld AB 54 because it concluded that it did not conflict with federal law,[156] and was supported by the constitutional prohibition on federal commandeering of state governments.[157] The Ninth Circuit upheld AB 103 state inspections of federal immigration detention facilities on much the same basis as the district court: the inspections do not “discriminate” against the federal government because they are much the same as those that California requires for other prisons within the state.[158]

On SB 450, the Court of Appeals did not consider the one issue where the district court ruled in favor of the federal government: instituting an injunction against the part of the law that bars employers from voluntarily consenting to ICE raids. The case before the Ninth Circuit was an appeal by the United States against those parts of the district court ruling that went against it, so it does not raise the one issue on which the federal government won in the trial court.

The Ninth Circuit did reaffirm the district court’s ruling that SB 450’s worker notification requirement is constitutional. It emphasized that the requirement does not “discriminate” against the federal government because it does not treat its agents less favorably than similarly situated private parties:

The Supreme Court has clarified that a state “does not discriminate against the Federal Government and those with whom it deals unless it treats someone else better than it treats them.” AB 450 does not treat the federal government worse than anyone else . . . . Accordingly, the district court correctly concluded that AB 450’s employee-notice provisions do not violate the doctrine of intergovernmental immunity.[159]

I would add that the same reasoning should eventually lead the Ninth Circuit to overrule the District Court on the issue of SB 450’s bar on voluntary employer cooperation with ICE raids. There is no discrimination here because there is no private or state analogue to ICE that California treats better.

The Ninth Circuit did rule against California on one small issue where the trial court went the other way. It struck down a provision of AB 103 that required inspections of the circumstances of the detainees’ apprehension and transfer to the facility in question.[160] Unlike the rest of AB 103, “[t]his is a novel requirement, apparently distinct from any other inspection requirements imposed by California law,” and therefore qualifies as discrimination against the federal government, violating the doctrine of intergovernmental immunity.[161]

This part of the Ninth Circuit’s ruling seems flawed because ICE apprehension and detention of immigrants is not truly analogous to the detention of other kinds of prisoners, including those arrested by state law enforcement. The criticism I made against the district court’s ruling on AB 450 applies here too.[162] There is no meaningful state (or private) analogue to federal government detention of suspected illegal immigrants for deportation because no private or state agency has the power to deport people with only minimal due process, often so little that the government routinely detains and deports large numbers of people who are actually U.S. citizens.[163]

Like the district court decision this case largely upholds, the Ninth Circuit ruling only addresses the federal government’s motion for a preliminary injunction against the three state laws.[164] But, in both cases, the court’s ruling prefigures the likely outcome of a final judgment on the merits, since the federal government’s motion was largely rejected precisely because the court ruled that the United States had little chance of prevailing on the merits.[165]

D. The Legal Battle Over the Constitutionality of 8 U.S.C. § 1373

The constitutionality of 8 U.S.C. § 1373 is a question that cuts across all three types of Trump-era sanctuary-jurisdiction cases.[166] It is obviously at issue in the cases addressing EO 13768, which seeks to force recipients of federal grants to obey it. The same is true when it comes to the DOJ Byrne Grant policy, which outlines adherence to § 1373 as one of the three conditions it forces grant recipients to obey.[167] In the California sanctuary state case, the legality of SB 54, one of the three state laws challenged by the federal government, turned in large part on the constitutionality of § 1373, which the federal government claimed it violates.

As we shall see, the constitutionality of § 1373 was a difficult question when the sanctuary litigation began. In my view, the right conclusion was that it violated the anti-commandeering rules set out in New York v. United States[168] and Printz v. United States.[169] Still, the issue was a close one. It is understandable that lower courts were initially divided over it.

But it became much easier after the Supreme Court’s May 2018 decision in Murphy v. NCAA, which struck down a federal law banning state government “authorization” of sports gambling under their own state law.[170] Post-Murphy judicial rulings on the constitutionality of § 1373 have so far all come down against it—for good reason.

In this Article, I do not address the longstanding dispute over the correct interpretation of § 1373. The jurisdictions opposing the Trump Administration’s anti-sanctuary policies argue for a narrow interpretation, under which § 1373 is essentially limited to its text and only bars states and localities from instructing their employees to deny federal immigration-enforcement agencies information about the citizenship and residency status of individuals they have knowledge of.[171] By contrast, the Trump Administration argues for a much broader interpretation, under which § 1373 bars all restrictions on employee communications with federal immigration officials regarding immigration issues.[172] Federal courts have uniformly rejected the Department of Justice’s broad interpretation of § 1373 and adopted the narrow one.[173]

For purposes of this Article, I assume that the narrow interpretation is indeed correct but contend § 1373 is still unconstitutional. If so, then the same conclusion even more obviously holds true if the DOJ’s broad interpretation of the law is correct.

1. The Anti-Commandeering Doctrine and the Legal Status of § 1373 Before Murphy.—As expounded by the Supreme Court in Printz and New York, the anti-commandeering rule bars the “compelled enlistment of state executive officers for the administration of federal programs.”[174] Section 1373 attempts to circumvent this prohibition by forbidding higher-level state and local officials from mandating that lower-level ones refuse to help in enforcing federal policy. More specifically, it requires that “a Federal, State, or local government entity or official may not prohibit, or in any way restrict, any government entity or official from sending to, or receiving from, the Immigration and Naturalization Service information regarding the citizenship or immigration status, lawful or unlawful, of any individual.”[175]

This convoluted structure potentially avoids constitutional problems. The law does not directly require state and local governments to turn over any information to the federal government. It merely bars them from instructing their subordinates to refuse such assistance. The final decision on whether to help or not, however, is presumably left to lower-level state and local government employees. This consideration is what led the Second Circuit to uphold the constitutionality of § 1373 against a challenge brought by the city government of then-New York City Mayor Rudolph Giuliani.[176] Ironically, Giuliani is today serving as President Trump’s personal lawyer in his efforts to deal with legal issues arising from possible collusion with Russia during the 2016 election.[177] The Second Circuit concluded that § 1373 does not qualify as commandeering because its requirements “do not directly compel states or localities to require or prohibit anything. Rather, they prohibit state and local governmental entities or officials only from directly restricting the voluntary exchange of immigration information with the INS.”[178]

But the same principle that forbids direct commandeering also counts against the indirect approach adopted by § 1373. As Justice Antonin Scalia explained in his majority opinion in Printz, the purpose of the
anti-commandeering doctrine is the “[p]reservation of the States as independent and autonomous political entities.”[179] That independence and autonomy is massively undermined if the federal government can take away the states’ power to decide what state and local officials may do while on the job. As Scalia put it in the same opinion, federal law violates the Tenth Amendment if it “requires [state employees] to provide information that belongs to the State and is available to them only in their official capacity.”[180] The same is true if, as in the case of § 1373, the federal government tries to prevent states from controlling their employees’ use of information that “is available to them only in their official capacity.”

To be sure, Scalia also noted that the statute at issue in Printz imposed investigative and reporting requirements on states,[181] which are likely more onerous than the dictates of § 1373. But nothing in Scalia’s reasoning depends on the degree to which a statute undermines state autonomy or blocks states from controlling the flow of information acquired by state employees acting in “their official capacity.” The fact that a law imposes such requirements is itself sufficient to violate the anti-commandeering principle, even if some impositions of this type are more onerous than others.

Nonetheless, the issue was difficult enough that early Trump-era sanctuary cases split on the question. In City of Philadelphia v. Sessions I, the court ruled that § 1373 is likely unconstitutional, though its decision did not rely on that conclusion.[182] By contrast, in City of Chicago v. Sessions I, Judge Harry Leinenweber of the Northern District of Illinois upheld § 1373 based on reasoning similar to that of the Second Circuit.[183]

In Santa Clara I, the first ruling on the constitutionality of EO 13768, Judge William Orrick concluded that the order “attempts to use coercive methods to circumvent the Tenth Amendment’s direct prohibition against conscription,” and thereby violates the anti-commandeering rule.[184] This appears to suggest that § 1373 is unconstitutional; but Judge Orrick did not specifically address the issue. Instead, he seemed to suggest that “the threat” to pull federal funds from jurisdictions that do not obey § 1373 inherently violates the anti-commandeering rule.[185] This approach seems to conflate the anti-commandeering rule with the Spending Clause and separation-of-powers restrictions on the types of conditions the federal government can attach to federal grants. Judge Orrick might have done better to directly address the constitutionality of § 1373, thereby making the basis for this part of his ruling less confusing.

2. The Impact of Murphy v. NCAA.—The previously close question of the constitutionality of § 1373 was transformed by the Supreme Court’s May 2018 decision in Murphy v. NCAA.[186] Murphy invalidated a provision of the federal Professional and Amateur Sports Protection Act (PASPA), which mandates that states may not “sponsor, operate, advertise, promote, license, or authorize by law or compact” sports betting.[187] A coalition of sports leagues, including the NCAA, the NBA, the NFL, and Major League Baseball, filed a lawsuit arguing that New Jersey’s 2012 and 2014 laws partially legalizing sports gambling within the state qualifies as “authorization” and thus violates PASPA.[188] New Jersey, for its part, argued that PASPA violates the anti-commandeering rule.[189]

The leagues and the federal government argued that that there is a distinction between commandeering and blocking “affirmative authorization” of gambling under state law. The former “affirmatively commands” what states must do, while the latter merely prevents them from enacting laws of a particular type.[190] On this view, PASPA does not qualify as commandeering because it does not prevent complete legalization of sports gambling, but only state laws that affirmatively authorize gambling in some way, as New Jersey did by restricting it to some types of locations and limiting the range of teams that gamblers can bet on.

In his majority opinion for the Court, Justice Samuel Alito correctly concluded that this is a distinction without a difference:

The PASPA provision at issue here—prohibiting state authorization of sports gambling—violates the anti-commandeering rule. That provision unequivocally dictates what a state legislature may and may not do. . . .

Neither [the sports leagues] nor the United States contends that Congress can compel a State to enact legislation, but they say that prohibiting a State from enacting new laws is another matter. . . .

This distinction is empty. It was a matter of happenstance that the laws challenged in New York and Printz commanded “affirmative” action as opposed to imposing a prohibition. The basic principle—that Congress cannot issue direct orders to state legislatures—applies in either event.[191]

The parallels between PASPA and § 1373 are clear. Like PASPA, § 1373 is an attempt to circumvent the anti-commandeering rule’s strictures against federal mandates coercing states into helping to enforce federal law or enact a state law. Instead of directly ordering states to ban sports gambling, PASPA forbids them from repealing a prohibition of it in ways that “authorize” the activity under state law. But the Supreme Court saw through this subterfuge and struck down PASPA because it violated the anti-commandeering rule by putting state legislatures “under the direct control of Congress” and issuing “direct orders to state legislatures.”[192]

Section 1373 suffers from much the same flaw. Instead of directly ordering states and localities to divulge information to federal officials, it “merely” bars them from issuing orders to their subordinates forbidding such disclosure. But the practical effect is that states must comply with federal dictates.

Like PASPA, § 1373 is an “order” to state and local officials; it undermines states’ control over their governmental machinery and partially transfers it to the federal government. In this case, federal law prevents states and localities from directing their law enforcement officials to pursue state and local priorities rather than assist federal immigration enforcers. As legal scholar Garrett Epps puts it, “the federal government can’t order the states to dance to its tune; according to Murphy, it can’t tell the states they may not decide not to dance to the federal tune either. No double-negative tricks now!”[193] Murphy undercuts § 1373 in much the same way as it doomed PASPA.

Peter Margulies argues that § 1373 is distinguishable from PASPA because the latter:

[F]orced states to prohibit sports gambling . . . . In contrast, § 1373 deals only with the far more limited realm of contingent information-sharing by state and federal officials. Section 1373 does not require that a state or state official share information. It merely provides that if state officials possess information about immigration status, state or local law cannot restrict sharing of that information with federal authorities.[194]

But this seeming distinction collapses on inspection. PASPA does not in fact require states to ban sports gambling. It just requires them to legalize it in a way that avoids “authorization” of gambling under state law.[195] While the law does not directly require any action, it does restrict the state’s control over its own legislation.

In much the same way, § 1373 does not require state and local officials to provide information to the federal government. But it does constrain state and local governments’ ability to control their own employees by barring a specific type of assistance to federal law enforcement.

Margulies also tries to distinguish § 1373 from PASPA on the grounds that the former imposes less of a burden on state and local governments.[196] It does not require them to gather any information but merely forbids the barring of information sharing in cases where that information is already in the possession of state or local officials.[197] But given the vast amount of information in the hands of subnational governments, such an information-sharing exception to the anti-commandeering rule would in fact have great significance and potentially impose a serious burden on states and localities if similar tactics are used in other federal statutes.[198]

When Murphy was decided, both I and a number of other commentators predicted that it could be the death knell for § 1373.[199] And that is exactly what has happened in all of the post-Murphy cases addressing the issue so far.

In Philadelphia v. Sessions II,[200] a Byrne Grant-conditions case became the first ruling to address the constitutionality of § 1373 after the federal district court in Murphy invalidated § 1373. As Judge Michael Baylson explains, the same principle that invalidated PASPA also dooms § 1373:

8 U.S.C. §§ 1373(a) and 1373(b) by their plain terms prevent “Federal, State, or local government entit[ies] or official[s] from” engaging in certain activities. These provisions closely parallel the anti-authorization condition in PASPA which was at issue in Murphy. Specifically, the PASPA provision violated the Tenth Amendment because it “unequivocally dictates what a state legislature may and may not do.” . . . Sections 1373(a) and (b) do the same, by prohibiting certain conduct of government entities or officials.[201]

As Judge Baylson noted later in his opinion, the federal government can still restrict state and local officials when their conduct conflicts with federal laws regulating “private actors.”[202] It can also bar them from adopting policies that are unconstitutional, such as engaging in unconstitutional discrimination or violating the Bill of Rights. But it cannot dragoon state and local governments into using their resources to help enforce federal laws.[203] It cannot do it by simply ordering them to do so, and it cannot do it circuitously, as with § 1373.[204]

Other post-Murphy decisions reached similar conclusions. In its recent ruling on the Byrne Grant case brought by seven states and New York City, the district court for the Southern District of New York concluded that it was no longer bound by the 1999 Second Circuit opinion upholding § 1373 because the reasoning of that decision “cannot survive the Supreme Court’s decision in Murphy.[205] In July 2018, Judge Leinenweber of the Northern District of Illinois reversed his earlier preliminary ruling upholding the constitutionality of § 1373, on the grounds that Murphy has superseded it.[206]

In the California sanctuary state case, the court ultimately ruled that California Senate Bill 54 does not violate § 1373, thereby obviating the need to make a definitive ruling on the constitutionality of the latter.[207] But Judge Mendez also concluded that Murphy struck down PASPA because it “unequivocally dictates what a state legislature may and may not do,” thereby violating the anti-commandeering rule.[208] As Mendez explained, “Section 1373 does just what Murphy proscribes: it tells States they may not prohibit . . . the sharing of information regarding immigration status with the INS or other government entities.”[209]

Judge Mendez similarly rejected the federal government’s claims that SB 54 is preempted by federal laws facilitating the deportation of undocumented immigrants. As he explained, any such preemption would “likely” be unconstitutional because the federal government cannot force states to assist in federal law enforcement efforts: “a Congressional mandate prohibiting states from restricting their law enforcement agencies’ involvement in immigration-enforcement activities—apart from, perhaps, a narrowly drawn information-sharing provision—would likely violate the Tenth Amendment.”[210] By “a narrowly drawn information sharing provision,” he apparently meant one where state adherence to federal requests for information is not mandatory.[211] And, he further suggested that even § 1373’s seemingly non-mandatory scheme would likely be unconstitutional after Murphy.[212] I believe Judge Mendez could and should have gone further than this and ruled that there is no information-sharing exception to the anti-commandeering rule, not even a “narrowly drawn” one.[213]

Like Judge Mendez, the Ninth Circuit decision largely upholding his ruling concluded that SB 54 does not violate § 1373.[214] Thus, it did not reach the question of whether § 1373 is unconstitutional. But the Ninth Circuit ruling did emphasize that the anti-commandeering rule was an important factor in its decision to uphold SB 54:

The United States’ primary argument against SB 54 is that it forces federal authorities to expend greater resources to enforce immigration laws, but that would be the case regardless of SB 54, since California would still retain the ability to “decline to administer the federal program.” . . . As the Supreme Court recently rearticulated in Murphy, under the anti-commandeering rule, “Congress cannot issue direct orders to state legislatures[.]”

. . . .

SB 54 may well frustrate the federal government’s immigration enforcement efforts. However, whatever the wisdom of the underlying policy adopted by California, that frustration is permissible, because California has the right, pursuant to the anti-commandeering rule, to refrain from assisting with federal efforts.[215]

The Ninth Circuit upheld SB 54 for much the same reasons as other courts have struck down § 1373: the anti-commandeering rule gives states the power to instruct their employees not to cooperate with the federal government and bars the latter from overriding that authority.

3. Is There an Information-Sharing Exception to the Anti-Commandeering Rule?—Some defenders of the constitutionality of § 1373 argue that the anti-commandeering rule does not apply to information sharing. For example, prominent conservative lawyer David Rivkin and legal scholar Elizabeth Price Foley argue that Trump has broad power to force sanctuary cities to do his bidding because the Supreme Court’s precedents banning federal commandeering of state governments supposedly do not apply to cases “when Congress merely requests information.”[216] They claim that:

[I]n Reno v. Condon (2000), the Court unanimously rejected an anti-commandeering challenge to the Driver’s Privacy Protection Act, which required states under certain circumstances to disclose some personal details about license holders. The court concluded that, because the DPPA requested information and “did not require state officials to assist in the enforcement of federal statutes,” it was consistent with the New York and Printz cases.[217]

Rivkin and Foley conclude that Trump could force sanctuary cities to disclose the names and locations of undocumented immigrants, thereby facilitating deportation.[218]

Contrary to Rivkin and Foley’s assumptions, Reno v. Condon[219] does not give the federal government a blank check to compel state and local authorities to disclose information. Rather, the Court emphasized that the federal law in question was constitutional only because “[i]t does not require [states] to enact any laws or regulations, and it does not require state officials to assist in the enforcement of federal statutes regulating private individuals.”[220] The disclosures required in Reno were intended to help enforce a federal law that prevented states themselves from violating the privacy rights of citizen driver’s license holders.[221] By contrast, the whole point of forcing disclosure of information about undocumented migrants is precisely to “assist in the enforcement of federal statutes regulating private individuals.”[222]

It would be strange if there were a general information-gathering exception to New York and Printz, given that the latter case itself involved a federal law compelling state officials to gather and disclose to the federal government information about gun purchasers.[223] In striking down that law, the Court ruled that the federal government violates the Tenth Amendment if “[i]t requires [state employees] to provide information that belongs to the State and is available to them only in their official capacity.”[224] The Court did distinguish—and refused to specifically address—the status of federal statutes “which require only the provision of information to the Federal Government.”[225] But that presumably does not apply to situations where the information in question “belongs to the State and is available to [state employees] only in their official capacity.”[226] That will often be true of information about the citizenship status of persons detained by a state or local government or otherwise having contact with it.

The notion of a generalized information-disclosure exception to the anti-commandeering rule was recently rejected by the U.S. District Court for the Northern District of Illinois in one of the Byrne Grant decisions. As Judge Leinenweber put it, “[a] federal need for state information does not automatically free the federal government of the sometimes laborious requirement to acquire that information by constitutional means.”[227]

If there were indeed a general information-disclosure exception to the Tenth Amendment’s prohibition on commandeering, it would constitute a major expansion of federal power. State and local governments have extensive information about hundreds of millions of people that the federal government could abuse in many ways.

The extent of the potential danger was highlighted by the 2017 controversy surrounding President Trump’s abortive “Election Integrity Commission,” which demanded that state governments reveal a wide range of information from their voter registration databases, including the names, addresses, party registrations, and last four digits of Social Security numbers of all voters.[228] The Commission failed to achieve its goals and was eventually disbanded because numerous state governments refused to turn over the data it demanded, citing privacy and security concerns and fears that the Administration might misuse the information.[229] If there were a general information-gathering exception to the anti-commandeering rule, the federal government could force states and localities to hand over such sensitive data.

Such an exception would also potentially enable Congress to impose massive uncompensated burdens on state and local governments. Ultimately, mandatory information disclosure is a form of commandeering no less than any other. It makes little sense to forbid other types of commandeering of state and local resources but give the federal government a blank check to compel disclosure of information. Fortunately, there is no such exception in the Constitution or in any of the Supreme Court’s precedents.

III. Broader Implications of the Sanctuary Cases

The Trump-era sanctuary cases have broad implications for federalism that go well beyond the context of immigration policy. This is true of all three areas of recent sanctuary litigation: the EO 13768 cases, the Byrne Grant-conditions cases, and the California sanctuary state case. If state and local governments continue to prevail on all or most of the issues involved, it will reaffirm and strengthen important constitutional constraints on federal power.

The sanctuary cases also offer a striking illustration of the potentially shifting ideological valence of judicial enforcement of federalism. They feature liberal-Democratic “blue” jurisdictions appealing to federalism doctrines pioneered by conservative Supreme Court Justices and traditionally viewed with skepticism by most on the left. This shift may be an example of “fair-weather federalism,”[230] under which litigants and activists on both sides of the political spectrum opportunistically appeal to federalism whenever it might be politically advantageous but ignore its constraints whenever broad assertions of federal power might benefit their cause. But it could also be an indication of a more systematic change in attitudes.

A. Broader Implications for Federalism

All three sets of sanctuary cases have significant implications for federalism. If the Trump Administration prevails in some or all of them, it would greatly increase federal government leverage over state and local governments. In the case of the EO 13768 and the Byrne Grant cases, an Administration victory would also undermine separation of powers by enabling the Executive to attach conditions to federal grants that were not authorized by Congress. Upholding § 1373 against anti-commandeering
challenges would also have potentially dangerous effects.

1. Doctrinal Implications.—If the President can attach new conditions to federal grants, unauthorized by Congress, that would give the Executive enormous power to pressure states and localities. As of fiscal year 2016, federal grants accounted for 32.6% of all revenue in state budgets.[231] If the requirement of “unambiguous” congressional authorization for grant conditions is eliminated or seriously weakened, the President could potentially attach new conditions to a wide range of grants, making it very difficult for states and localities to resist the resulting pressure.

What the Trump Administration attempted to do in the area of immigration enforcement could just as easily be repeated elsewhere, in the case of environmental policy, education policy, health care, or other law enforcement issues. Conservatives who may cheer Trump’s efforts to pressure sanctuary cities would not be happy if a liberal-Democratic president adopts similar tactics to force states and localities to adopt progressive policies on gun control, transgender bathrooms, “Common Core” education,[232] or any number of other issues.

Partisan and ideological concerns about particular policies aside, giving the Executive broad power to impose new conditions on federal grants to states and localities would create a powerful weapon for imposing homogenization on state and local public policy. In a diverse and deeply polarized society, that would mean an increasing number of people forced to live under policies they oppose and a deepening of already severe ideological and partisan hostilities. We should instead maintain tighter constitutional constraints on federal power in order to protect political diversity.[233]

Protecting diversity by limiting federal power can also enhance our ability to choose the policies we prefer to live under by “voting with our feet.” Foot voting is often a more effective way of exercising political freedom than conventional ballot-box voting,[234] in part because it leads to better informed decision-making.[235] Because of the very low likelihood that any one vote will make a difference to electoral outcomes, most ballot-box voters are “rationally ignorant” and tend to know very little about the issues they vote on; by contrast, foot voters have much stronger incentives to become well-informed, because their individual decisions on where to live are highly likely to make a real difference to their lives.[236]

Even congressionally authorized grant conditions can undermine diversity in policy and reduce opportunities for foot voting. But giving the power to set grant conditions to the president greatly exacerbates the danger. It is much easier for the Executive to enact coercive grant conditions at odds with the preferences of numerous state and local governments than for Congress to do so. The latter generally has far greater partisan and ideological diversity than the former, and the prevalence of divided government also makes it harder for it to adopt dangerous new grant conditions.

In this way, the potential threat to federalism is heightened by the threat to separation of powers. By intruding on Congress’s power of the purse, the Administration also makes it easier for the federal government to coerce states and localities.

This danger is only slightly lessened if courts were to accept the Trump Administration’s attempts to use vaguely worded statutes as sources of authorization, as in the Byrne Grant cases, rather than give the Executive even more unconstrained discretion than that.[237] Such potential tools are common,[238] and future administrations could leverage them to impose new conditions on a variety of federal grant programs.

In this context, I should take the opportunity to criticize my own earlier, comparatively dismissive, attitude towards grant condition clear-statement rules in a 2006 article.[239] At that time, I did not sufficiently appreciate the potential dangers of opening the door to Executive Branch imposition of new conditions.

Broader risks to federalism also lurk should the Supreme Court ultimately uphold 8 U.S.C. § 1373, thereby reversing multiple lower-court decisions that struck it down, especially in the wake of Murphy v. NCAA.[240] The subterfuge by which the drafters of § 1373 sought to get around the anti-commandeering rule can also be used on many other issues. For example, instead of directly requiring state law enforcement agencies to assist in the enforcement of the federal law banning marijuana in states that have legalized it under state law, the federal government could adopt a law or regulation forbidding state governments from instructing their employees not to assist in federal enforcement efforts. It could then entice employees to go against state policy priorities in various ways, such as by promising them reciprocal assistance on other issues.

Such quid pro quos need not even necessarily require appropriations authorized by Congress. The Executive Branch could potentially offer its own enticements to state and local employees, utilizing the many programs under which state and local law enforcement officials cooperate with each other.[241]

The relatedness ruling in Philadelphia I also has potentially broader applications. As noted earlier, courts have been reluctant to strike down federal grant conditions on the basis of lack of relatedness.[242] But if Judge Baylson’s nuanced analysis of this issue is adopted by other courts, it could signal increased judicial enforcement of this restriction. Courts could potentially strike down other federal grant conditions where the issue covered by the condition is connected to the purpose of the grant, but not in a way that increases the likelihood of achieving that purpose. Further research is needed to determine how common such situations are. And, of course, it remains to be seen whether other judges follow Judge Baylson’s promising lead.

The potential implications of the California sanctuary state litigation are more difficult to foresee than those of EO 13768 and Byrne Grant cases because that legal battle is still at a relatively early stage.[243] But I would tentatively highlight at least two significant possible consequences.

First, should the courts ultimately uphold Assembly Bill 103, the provision requiring state inspection of federal immigration detention facilities, it could potentially pave the way for further state monitoring of potential abuses of federal power. That might provide a useful extra safeguard against such dangers.

Second, the litigation over Assembly Bill 450 might well set an important precedent on the question of whether states can violate the doctrine of intergovernmental immunity by “discriminating” against federal activities that have no true private-sector or state analogue.[244] If the answer is “no,” as I believe it should be,[245] states could potentially deny assistance of various kinds to unusual federal activities, even if similar assistance were offered to private parties or to other federal operations. This outcome might not be an unmitigated good. States could potentially use it to deny help to valuable federal programs. But it would enable them to deny cooperation to federal agencies—such as ICE—that engage in particularly heinous abuses.

2. A Broad Judicial Consensus.—Another striking aspect of the federalism rulings in the sanctuary cases so far is the broad agreement among judges across the political spectrum. Both Republican- and Democratic-appointed judges have almost uniformly ruled against the Trump Administration. The district judges in the Chicago and Philadelphia sanctuary cities cases—Harry Leinenweber and Michael Baylson—are both Republican appointees. Yet both ruled against the Administration on virtually all issues.[246] Judge Anthony Scirica, one of the three Third Circuit judges who upheld the district court ruling in the Philadelphia case,[247] is also a GOP appointee.[248] The same is true of all three Seventh Circuit judges who upheld the preliminary injunction against the Administration’s Byrne Grant policy[249]: William Joseph Bauer (appointed by Gerald Ford),[250] Daniel Manion (appointed by Ronald Reagan),[251] and Ilana Diamond Rovner (appointed by George H.W. Bush).[252] Judge John Mendez, the district court judge who ruled against the Administration on two of the three issues at stake in the California sanctuary state case,[253] is also a GOP appointee.[254] The same is true of Ninth Circuit Judge Milan Smith, Jr., author of that court’s opinion largely upholding Judge Mendez’s ruling.[255]

The other judges who ruled against the Administration in the sanctuary cases are all Democratic appointees.[256] Both the reasoning and conclusions they reached are remarkably similar to those of their Republican colleagues. However, the sole judge who voted to uphold EO 13768,[257] Judge Fernandez of the Ninth Circuit, is a Republican appointee.[258]

In November 2018, Supreme Court Chief Justice John Roberts issued an unusual public statement rebuking President Donald Trump for a crude attack against an “Obama judge” who had issued a decision against one of the President’s immigration policies: “We do not have Obama judges or Trump judges, Bush judges or Clinton judges,” Roberts emphasized, “What we have is an extraordinary group of dedicated judges doing their level best to do equal right to those appearing before them.”[259]

Critics pointed out that the Chief Justice was ignoring the reality that ideology and party often do correlate with judges’ rulings on controversial issues.[260] The criticism is not without some merit. But when it comes to the Trump-era sanctuary city cases, Chief Justice Roberts’s statement is largely accurate.[261] There is an impressively broad judicial consensus that the Administration’s policies are unconstitutional.

3. A Brief Note on Original Meaning.—In this Article, I cannot take up the longstanding debate over the text and original meaning of the Spending Clause, or over whether modern Supreme Court anti-commandeering decisions can be justified on originalist grounds. I do think, however, that the sanctuary decisions reviewed here are at least largely consonant with original meaning.

Elsewhere, I have outlined my views to the effect that the phrase “general Welfare” in the Spending Clause should be given a much narrower interpretation than under current Supreme Court precedent, one that imposes tight constraints on federal grants to state governments and reduces the federal government’s ability to use them as a tool to impose conformity.[262] From this perspective, making it harder for the Executive to attach new conditions to federal grants to state governments at least moves us closer to the original meaning, even though it certainly does not get us all the way there.

With regard to the relationship between original meaning and commandeering, I agree with critics of recent Supreme Court decisions who argue that the anti-commandeering rule cannot be derived from the original meaning of the Tenth Amendment alone.[263] But I am also persuaded by Professor Michael Rappaport’s argument that the rule can be derived from the original understanding of what it means to be a sovereign “state,” one whose administrative machinery cannot be appropriated by another level of government.[264] And if there is to be an anti-commandeering rule at all, there must be safeguards against circumventions of the sort attempted by the drafters of § 1373.[265]

B. Will the Left Rethink Its Approach to Judicial Enforcement of Federalism?

In addition to their doctrinal implications, the sanctuary cases also feature an unusual alignment of political forces: liberal–Democratic state and local governments are challenging a Republican Administration by relying on federalism doctrines traditionally associated with the political right.

At least until recently, judicial enforcement of federalism was a cause championed primarily by conservatives and libertarians, while progressives tended to be highly skeptical, if not downright hostile.[266] One of the principal reasons for that hostility was the long-standing conventional wisdom holding that federalism is a disaster for racial and ethnic minority groups, while the growth of federal power was a great benefit to them.[267] As the leading political scientist William Riker famously put it in 1964, “[t]he main beneficiary [of federalism] throughout American history has been the Southern whites, who have been given the freedom to oppress Negroes. . . . [I]f in the United States one approves of Southern white racists, then one should approve of American federalism.”[268]

This view of the history of American federalism contains an important measure of truth, though it also has notable shortcomings.[269] But, however accurate it may have been in earlier periods of American history, the sanctuary city litigation—and the Trump era more generally—highlights ways in which it no longer holds true. Sanctuary cities are a dramatic example of state and local governments protecting vulnerable racial and ethnic minorities against hostile federal policies.

It is, therefore, not surprising that the rise of Trump has led a number of prominent liberal legal commentators to take a more favorable view of judicial enforcement of federalism and embrace it as a strategy for resisting his policies, particularly those targeting sanctuary cities.[270] Harvard law professor Noah Feldman described the use of federalism for “[t]he protection of sanctuary cities [as] an example of how the Constitution protects minority rights—in this case the rights of cities that dissent on immigration policy.”[271]

Murphy v. NCAA, the May 2018 Supreme Court decision striking down a federal ban on state “authorization” of sports gambling,[272] got a more favorable reception on the left than any other Supreme Court decision limiting federal power in living memory, with commentators emphasizing the potential advantages for sanctuary cities.[273] Mark Joseph Stern, of the prominent liberal Slate website, praised conservative Justice Samuel Alito’s “fantastic opinion,” which shows that “federalism can be good for liberals” and “creates a precedent that could help liberals down the road by shielding state experimentation from federal intrusion.”[274] He pointed out that, in addition to protecting sanctuary cities, it could also benefit liberal states seeking to legalize marijuana and other drugs still banned by the federal government, and protect assisted suicide.[275]

Significantly, the five conservative Justices were joined in the majority by liberals Stephen Breyer and Elena Kagan.[276] This is an exception to the general trend under which federalism cases divide the Supreme Court along ideological lines, though there have been a few other recent exceptions.[277]

The work of Yale Law School Professor (now Dean) Heather Gerken, who has long urged her fellow progressives to take a more favorable view of federalism,[278] has enjoyed new prominence in the Trump era.[279] The rise of Trump and his Administration’s battles with sanctuary cities are dramatic illustrations of Gerken’s thesis that state and local governments are now often more supportive of minority groups than the federal government, in part because minorities now often have greater clout with the former than the latter.[280] Before she became Dean of Yale Law School in the summer of 2017, Professor Gerken also oversaw a Yale student clinic that, together with the City of San Francisco, helped litigate the case securing a nationwide preliminary injunction against EO 13768.[281]

Whether Trump-era legal battles over sanctuary cities herald a durable shift in liberal attitudes towards judicial enforcement of federalism remains to be seen. While some on the left now see greater virtue in limits on federal power, many others have rallied to the cause of “democratic socialism,” whose adherents seek vast expansions of federal control over economic and social policy.[282] If these ideas become the dominant view of the left, the sanctuary city cases could turn out to be another example of temporary “fair weather” federalism.[283]

On the other hand, recent political history suggests that it will not be easy for either major political party to maintain control over the federal government for long. Liberals may come to see judicial enforcement of federalism as valuable “insurance” against future GOP presidents and Congresses, who may well pursue policies similar to those of Trump.[284]

For now, all that we can safely say is that the sanctuary cases are a dramatic example of the use of judicial enforcement of federalism to protect an important liberal cause, one that has led some influential figures on the left to take a more favorable view of federalism. Whether that will help lead to a lasting change in the ideological valence of federalism remains to be seen.

IV. Conclusion

The Trump-era sanctuary city cases are notable both for their doctrinal importance to federalism and separation-of-powers doctrine, and for their potential political significance.

At this time, their ultimate long-term impact remains difficult to predict. The legal battles in question are not yet completely over, especially in the case of the litigation over California’s sanctuary state law. And it is also difficult to tell whether these cases are part of a watershed in liberal attitudes towards constitutional federalism.

But it is clear that the cases raise major issues about conditional grants to state governments, commandeering, and the scope of executive authority over federal spending. And, so far at least, the resulting litigation has given a valuable boost to the enforcement of constitutional limits on federal power. However unintentionally, Donald Trump’s assault on sanctuary cities has helped make federalism great again.

  1. .See, e.g., Tami Luhby, Trump Condemns Sanctuary Cities, But What Are They?, CNN (Sept. 1, 2016, 10:08 AM), [] (describing sanctuary cities and Donald Trump’s opposition to them). On the political movements for and against sanctuary jurisdictions, see Rose Cuison Villazor & Pratheepan Gulasekaram, The New Sanctuary and Anti-Sanctuary Movements, 52 U.C. Davis L. Rev. 549 (2018). For an overview of recent sanctuary policies adopted by state and local governments, see Christopher N. Lasch et al., Understanding “Sanctuary Cities, 59 B.C. L. Rev. 1703, 1736–52 (2018).
  2. .I offered a very brief, nonacademic overview of the Trump-era sanctuary cases in Ilya Somin, Fight Over Sanctuary Cities Is Also a Fight Over Federalism, Hill (Apr. 7, 2018, 11:30 AM), []. For a discussion that covers many of the relevant decisions through late 2017, see Annie Lai & Christopher N. Lasch, Crimmigration Resistance and the Case of Sanctuary City Defunding, 57 Santa Clara L. Rev. 539, 557–61, 572–84 (2017). However, this Article does not cover post-2017 rulings and therefore does not consider most of those that address the constitutionality of the Byrne Grant conditions imposed by the Department of Justice in July 2017. It also does not consider the debate over whether 8 U.S.C. § 1373 violates the anti-commandeering rule, and the litigation over California’s sanctuary state law. All of these are analyzed in this Article. A recent article by Peter Margulies analyzes several of the Byrne Grant decisions but does not consider the litigation over Executive Order 13768 and the sanctuary state cases. See Peter Margulies, Deconstructing “Sanctuary Cities”: The Legality of Federal Grant Conditions That Require State and Local Cooperation on Immigration Enforcement, 75 Wash. & Lee L. Rev. 1507, 1544–70 (2018). As discussed below, I also differ with Margulies on some key issues. Nelson Lund also briefly discusses some of the Byrne Grant cases but does not cover the other sanctuary cases. See Nelson Lund, The Constitutionality of Immigration Sanctuaries and Anti-Sanctuaries: Originalism, Current Doctrine, and a Second-Best Alternative, U. Penn. J. Const. L. (forthcoming) (manuscript at 24–27),

    3302818 []. For an overview of pre-Trump litigation and debates over constitutional federalism and state government efforts to refuse assistance to federal immigration-enforcement efforts, see Spencer E. Amdur, The Right of Refusal: Immigration Enforcement and the New Cooperative Federalism, 35 Yale L. & Pol’y Rev. 87 (2016).

  3. .See Exec. Order No. 13,768, 82 Fed. Reg. 8799 (Jan. 25, 2017) [hereinafter EO 13768].
  4. .See U.S. Dep’t of Justice, Backgrounder on Grant Requirements (2017), [] (outlining new conditions for recipients of Edward Byrne Memorial Justice Assistance Grants); City of Chicago v. Sessions (Chicago II), 321 F. Supp. 3d 855, 861–62 (N.D. Ill. 2018) (same).
  5. .See United States v. California, 314 F. Supp. 3d 1077 (E.D. Cal. 2018), aff’d in part, rev’d in part, 2019 WL 1717075 (9th Cir. Apr. 18, 2019) (granting the federal government’s motion for preliminary injunction against California in part and denying in part).
  6. .8 U.S.C. § 1373(a) (2012).
  7. .United States v. California, 314 F. Supp. 3d at 1094–97.
  8. .For a discussion on these issues, see City of San Francisco v. Trump, 897 F.3d 1225, 1231, 1235–38 (9th Cir. 2018) (ruling that the case against Trump’s Executive Order can proceed).
  9. .Two appellate courts have vacated nationwide injunctions in the sanctuary cities cases. See id. at 1231; City of Chicago v. Sessions, Nos. 17–2991 & 18–2649, 2018 WL 4268814, at *1–2 (7th Cir. Aug. 10, 2018). For a prominent critique of nationwide injunctions, see Samuel L. Bray, Multiple Chancellors: Reforming the National Injunction, 131 Harv. L. Rev. 417 (2017); for a leading defense, see Amanda Frost, In Defense of Nationwide Injunctions, 93 N.Y.U. L. Rev. 1065 (2018).
  10. .See, e.g., United States v. California, 314 F. Supp. 3d at 1105 (“SB 54 . . . directs the activities of state law enforcement, which Congress has not purported to regulate. Preemption is inappropriate here.”).
  11. .See discussion infra section III(A)(3).
  12. .EO 13768, supra note 3.
  13. .Id. § 2(c).
  14. .Id. § 9(a).
  15. .8 U.S.C. § 1373(a) (2012).
  16. .See, e.g., Pennhurst State Sch. & Hosp. v. Halderman, 451 U.S. 1, 17 (1981) (“[I]f Congress intends to impose a condition on the grant of federal moneys, it must do so unambiguously.”).
  17. .South Dakota v. Dole, 483 U.S. 203, 207–08 (1987).
  18. .Id. at 210–11; Nat’l Fed’n of Indep. Bus. v. Sebelius, 567 U.S. 519, 580–83 (2012) (plurality opinion).
  19. .U.S. Const. art. I, § 8, cl. 1.
  20. .See, e.g., Dole, 483 U.S. at 207 (holding that “[i]n considering whether a particular expenditure is intended to serve general public purposes, courts should defer substantially to the judgment of Congress”).
  21. .451 U.S. 1 (1981).
  22. .Id. at 5–8.
  23. .Id. at 17–18.
  24. .See Printz v. United States, 521 U.S. 898, 907 (1997) (barring commandeering of executive officials); New York v. United States, 505 U.S. 144, 179 (1992) (barring commandeering of state legislative authority).
  25. .U.S. Dep’t of Justice, supra note 4; cf. City of Chicago v. Sessions (Chicago II), 321 F. Supp. 3d 855, 861–62, 882 (N.D. Ill. 2018) (entering a nationwide injunction against DOJ grant conditions but staying the injunction for appellate review). On the origins and history of the Byrne Grant program, see Lai & Lasch, supra note 2, at 590–94.
  26. .See discussion infra section II(B)(1).
  27. .U.S. Dep’t of Justice, Justice Assistance Grant Program, 2016: Technical Report 1–3 (2016) (summarizing the program).
  28. .See infra Part II.
  29. .Cal. Gov’t Code Ann. §§ 7284.6(a)(1)(C)–(D) (West 2019).
  30. .Id. § 7284.6(a)(4).
  31. .Cal. Gov’t Code Ann. § 12532 (West 2018).
  32. .Id. § 12532(b).
  33. .Id. §§ 12532(b)(1)–(2). This report was in fact submitted in February 2019. See Cal. Dep’t of Justice, Immigration Detention in California (Feb. 9, 2019),

    files/agweb/pdfs/publications/immigration-detention-2019.pdf [].

  34. .Cal. Gov’t Code Ann. § 7285.1 (West 2019).
  35. .Cal. Lab. Code Ann. § 90.2(a)(1) (West 2019).
  36. .United States v. California, 314 F. Supp. 3d 1077, 1085 (E.D. Cal. 2018).
  37. .U.S. Const. art. VI, cl. 2.
  38. .United States v. California, 314 F. Supp. 3d at 1085, 1088.
  39. .Id. at 1085–86. The district court ruling was largely affirmed in United States v. California, No. 18-16496, 2019 WL 1717075 (9th Cir. Apr. 18, 2019).
  40. .County of Santa Clara v. Trump (Santa Clara I), 250 F. Supp. 3d 497 (N.D. Cal. 2017); County of Santa Clara v. Trump (Santa Clara II), 275 F. Supp. 3d 1196 (N.D. Cal. 2018), aff’d, City of San Francisco v. Trump, 897 F.3d 1225 (9th Cir.); City of Seattle v. Trump, No. 17-497-RAJ, 2017 WL 4700144 (W.D. Wash. Oct. 19, 2017).
  41. .San Francisco, 897 F.3d 1225.
  42. .Santa Clara I, 250 F. Supp. 3d at 539.
  43. .Id. at 531.
  44. .South Dakota v. Dole, 483 U.S. 203, 203 (1987); Pennhurst State Sch. & Hosp. v. Halderman, 451 U.S. 1, 17 (1981).
  45. .Santa Clara I, 250 F. Supp. 3d at 531.
  46. .Id. at 531–32.
  47. .Id. at 507–08.
  48. .See, e.g., David French, A Federal Judge Issues a Mostly Meaningless Ruling Against a Mostly Meaningless Executive Order, Nat’l Rev. (Apr. 26, 2017, 1:00 AM), [] (endorsing the narrow interpretation of the Order); Ilya Shapiro, The Sanctuary Cities Ruling Is Much Ado About Nothing, The Federalist (Apr. 26, 2017), [] (same).
  49. .Santa Clara I, 250 F. Supp. 3d at 507–08.
  50. .See, e.g., Colautti v. Franklin, 439 U.S. 379, 392 (1979) (noting “the elementary canon of construction that a statute should be interpreted so as not to render one part inoperative”).
  51. .Santa Clara I, 250 F. Supp. 3d at 508.
  52. .Cf. City of San Francisco v. Trump, 897 F.3d 1225, 1238 (9th Cir. 2018) (noting that “[i]n contrast to the many established principles for interpreting legislation, there appear to be few such principles to apply in interpreting executive orders”). For an important recent analysis of potential approaches to interpreting executive orders, see Tara Leigh Grove, Presidential Laws and the Missing Interpretive Theory, 168 U. Penn. L. Rev. (forthcoming 2020),

    sol3/papers.cfm?abstract_id=3338466 [].

  53. .Bassidji v. Goe, 413 F.3d 928, 934 (9th Cir. 2005) (quoting Nw. Forest Res. Council v. Glickman, 82 F.3d 825, 830 (9th Cir. 1996)).
  54. .Grove, supra note 52, at 36–38.
  55. . Grove in fact recognizes that executive orders are easier to revise than congressional legislation, noting that “the complexity of the process [or revising them] still pales in comparison to the veto gates of the bicameralism and presentment process of Article I.” Id. at 39–40.
  56. .See discussion supra p. 118.
  57. .For the relevant statutes, see 34 U.S.C. §§ 10151–10158 (2012) (Edward Byrne Memorial Justice Assistance Grant Program), 8 U.S.C. § 1231 (2012) (State Criminal Alien Assistance Grant Program), and 42 U.S.C. § 3796 (COPS Program). I discussed this issue more fully in Ilya Somin, Federal Court Rules Against Trump’s Executive Order Targeting Sanctuary Cities, Wash. Post: Volokh Conspiracy (Apr. 25, 2017), [].
  58. .See French, supra note 48 (citing EO 13768, § 9(a)); Shapiro, supra note 48 (explaining how the Executive Order directs compliance with funding requirements).
  59. .County of Santa Clara v. Trump (Santa Clara I), 250 F. Supp. 3d 497, 508 (N.D. Cal. 2017).
  60. .Id. at 540 (holding that “[t]his injunction does not impact the Government’s ability to use lawful means to enforce existing conditions of federal grants or 8 U.S.C. 1373”).
  61. .See Ilya Somin, The Court Decision Against Trump’s Sanctuary Cities Order Is Not “Much Ado About Nothing, Wash. Post: Volokh Conspiracy (Apr. 26, 2017), https://www. [https://perma.


  62. .Santa Clara I, 250 F. Supp. 3d at 532 (citing South Dakota v. Dole, 483 U.S. 203, 213 (1987) (O’Connor, J., dissenting)).
  63. .Id. at 533.
  64. .See, e.g., Sabri v. United States, 541 U.S. 600, 602 (2004) (upholding a grant condition requiring all governmental entities that receive at least $10,000 in federal funds to submit to a federal law permitting federal prosecution of bribery cases against their officials).
  65. .Santa Clara I, 250 F. Supp. 3d at 532–33.
  66. .Dole, 483 U.S. at 211 (quotations omitted).
  67. .Nat’l Fed’n of Indep. Bus. v. Sebelius, 567 U.S. 519 (2012).
  68. .Id. at 581 (plurality opinion).
  69. .Santa Clara I, 250 F. Supp. 3d at 533–34.
  70. .See infra section II(D)(1).
  71. .County of Santa Clara v. Trump (Santa Clara II), 275 F. Supp. 3d 1196, 1219 (N.D. Cal. 2018).
  72. .City of San Francisco v. Trump, 897 F.3d 1225 (9th Cir. 2018).
  73. .Santa Clara II, 275 F. Supp. 3d at 1201.
  74. .Id. at 1202, 1211–12.
  75. .City of Seattle v. Trump, No. 17-497-RAJ, 2017 WL 4700144 (W.D. Wash. Oct. 19, 2017).
  76. .Id. at *10.
  77. .Id. at *9.
  78. .Id. at *5.
  79. .City of San Francisco v. Trump, 897 F.3d 1225, 1234–35, 1242 (9th Cir. 2018).
  80. .Id.
  81. .Id. at 1245.
  82. .Id. at 1248–50 (Fernandez, J., dissenting).
  83. .Id. at 1239–40.
  84. .Id.
  85. .Cf. Grove, supra note 52, at 39–40 (explaining why executive orders are easier to modify than legislation).
  86. .See, e.g., Nat’l Fed’n of Indep. Bus. v. Sebelius, 567 U.S. 519, 562 (2012) (“[I]t is well established that if a statute has two possible meanings, one of which violates the Constitution, courts should adopt the meaning that does not do so.”).
  87. .San Francisco v. Trump, 897 F.3d at 1241; County of Santa Clara v. Trump (Santa Clara I), 250 F. Supp. 3d 497, 522–23, 537 (N.D. Cal. 2017).
  88. .See discussion supra subpart I(A).
  89. .See infra Part III.
  90. .Santa Clara I, 250 F. Supp. 3d at 507–08.
  91. .New York v. Dep’t of Justice, 343 F. Supp. 3d 213 (S.D.N.Y. 2018); City of San Francisco v. Sessions, 349 F. Supp. 3d 924 (N.D. Cal. 2018); City of Los Angeles v. Sessions, No. CV 17-7215-R, 2018 WL 6071072 (C.D. Cal. Sept. 13, 2018); City of Evanston v. Sessions, No. 18 C 4853, 2018 U.S. Dist. LEXIS 204500 (N.D. Ill. Aug. 9, 2018) (basing the decision largely on the same judge’s previous ruling in City of Chicago v. Sessions II); City of Chicago v. Sessions (Chicago II), 321 F. Supp. 3d 855 (N.D. Ill. 2018); City of Philadelphia v. Sessions (Philadelphia II), 309 F. Supp. 3d 289 (E.D. Pa. 2018), aff’d in part, vacated in part, City of Philadelphia v. Att’y Gen. of U.S., 916 F.3d 276 (3d Cir. 2019); State ex rel. Becerra v. Sessions, 284 F. Supp. 3d 1015 (N.D. Cal. 2018); City of Philadelphia v. Sessions (Philadelphia I), 280 F. Supp. 3d 579 (E.D. Pa. 2017); City of Chicago v. Sessions (Chicago I), 264 F. Supp. 3d 933 (N.D. Ill. 2017) aff’d, City of Chicago v. Sessions, 888 F.3d 272 (7th Cir. 2018).
  92. .City of Philadelphia v. Att’y Gen. of U.S., 916 F.3d 276 (3d Cir. 2019); City of Chicago v. Sessions, 888 F.3d 272 (7th Cir. 2018).
  93. .See supra note 91.
  94. .City of San Francisco v. Sessions, No. 18-cv-05146-WHO, 2019 WL 1024404, at *1–2 (N.D. Cal. Mar. 4, 2019); City of Los Angeles v. Sessions, No. CV 18-7347-R, slip op. at 10 (C.D. Cal. Feb. 15, 2019). See discussion infra section II(B)(1).
  95. .Chicago v. Sessions, 888 F.3d at 277.
  96. .See supra note 91.
  97. .State of New York v. Dep’t of Justice, 343 F. Supp. 3d 213, 227–31 (S.D.N.Y. 2018) (quoting 34 U.S.C. §§ 10102(a), 10153(a)(5)), appeal docketed sub nom. City of New York v. Whitaker, No. 19-275 (2d Cir. 2019).
  98. .Id. at 231 (quoting Pennhurst State Sch. & Hosp. v. Halderman, 451 U.S. 1, 17 (1981)).
  99. .Id. at 231 (quoting Arlington Cent. Sch. Dist. Bd. of Educ. v. Murphy, 548 U.S. 291, 296 (2006)).
  100. .Id. at 231.
  101. .Id. at 228.
  102. .See, e.g., City of Philadelphia v. Att’y Gen., 916 F.3d 276, 287–91 (3d Cir. 2019) (rejecting Attorney General’s reliance on §§ 10102(a), 10153(a)(5)(D) for imposition of Byrne Grant conditions); City of Chicago v. Sessions, 888 F.3d 272, 284–87 (7th Cir. 2018) (same as to § 10102(a)); City of San Francisco v. Sessions, 349 F. Supp. 3d 924, 947–48, 953–55 (N.D. Cal. 2018) (same), appeal docketed sub nom. City of San Francisco v. Whitaker, No. 18-17308 (9th Cir. 2018); City of Philadelphia v. Sessions, 280 F. Supp. 3d 579, 615–19 (E.D. Pa. 2017) (rejecting Attorney General’s reliance on § 10102(a) and declining to reach issue of authorization under § 10153(a)(5)(D)), appeal dismissed, No. 18-1103, 2018 WL 3475491 (3d Cir. 2018). For a more detailed analysis of the meaning of “applicable federal laws,” see Lai & Lasch, supra note 2, at 577–80.
  103. .For a more extensive explication of the potential risks of construing such vague formulations broadly in the context of the relationships between immigration enforcement and ordinary criminal law, see Lai & Lasch, supra note 2, at 581–84.
  104. .Chicago I, 264 F. Supp. 3d at 933.
  105. .Id. at 939–41, 944 (quoting 34 U.S.C. § 10153(a) (2012)) (emphasis omitted).
  106. .Id. at 943–44.
  107. .Id. at 944.
  108. .Pennhurst State Sch. & Hosp. v. Halderman, 451 U.S. 1, 24 (1981).
  109. .Chicago I, 264 F. Supp. 3d at 943–46.
  110. .Id. at 948–49.
  111. .321 F. Supp. 3d 855, 873 (N.D. Ill. 2018).
  112. .138 S. Ct. 1461 (2018). See infra sections II(D)(1)–(2).
  113. .Margulies, supra note 2, at 1566–69.
  114. .City of San Francisco v. Sessions, No. 18-cv-05146-WHO, 2019 WL 1024404, at *2 (N.D. Cal. Mar. 4, 2019); City of Los Angeles v. Sessions, No. 18-cv-7347-R, at *9 (C.D. Cal. Feb. 15, 2019). See discussion supra section II(B)(1).
  115. .San Francisco v. Sessions, 2019 WL 1024404, at *4 (quoting DOJ RJN Ex. D ¶ 44).
  116. .Id.
  117. .Id. at *2; Los Angeles v. Sessions, No. 18-cv-7347, at *6–8.
  118. .Philadelphia I, 280 F. Supp. 3d 579, 639–44 (E.D. Pa. 2017). Judge Baylson reaffirmed this and other Spending Clause rulings from Philadelphia I in Philadelphia II, 289 F. Supp. 3d 289, 344–45 (E.D. Pa. 2018).
  119. .Philadelphia I, 280 F. Supp. 3d at 639–40.
  120. .Id. at 641–42.
  121. .San Francisco v. Sessions, 349 F. Supp. 3d at 958–61.
  122. .Philadelphia I, 280 F. Supp. 3d at 590, 601–12.
  123. .See Daniel E. Martinez et al., Providing Sanctuary or Fostering Crime? A Review of the Research on “Sanctuary Cities” and Crime​, Soc. Compass, Jan. 2018, at 9, https://onlinelibrary [] (stating that the studies “have yielded an inverse or null relationship between limited cooperation policies and crime”). For an extensive recent discussion of the reasons why immigration enforcement undermines rather than enhances conventional criminal law enforcement, see Lai & Lasch, supra note 2, at 563–72.

  124. .City of San Francisco v. Sessions, No. 18-cv-05146-WHO, 2019 WL 1024404, at *11–12. See discussion supra section II(B)(1).
  125. .San Francisco v. Sessions, 2019 WL 1024404, at *12.
  126. .Id.
  127. .Id. at *13 (quoting City of San Francisco v. Sessions, 349 F. Supp. 3d 924, 960 (N.D. Cal. 2018)).
  128. .United States v. California, No. 18-16496, 2019 WL 1717075 (9th Cir. Apr. 18, 2019).
  129. .United States v. California, 314 F. Supp. 3d 1077 (E.D. Cal. 2018), aff’d in part, rev’d in part, 2019 WL 1717075 (9th Cir. Apr. 18, 2019).
  130. .Id.
  131. .Id. at 1093, 1111.
  132. .Id. at 1096, 1112.
  133. .Id. at 1099.
  134. .8 U.S.C. § 1373(a).
  135. .United States v. California, 314 F. Supp. 3d at 1102–04.
  136. .Id. at 1102.
  137. .Id. at 1104.
  138. .Id. at 1101; see also discussion infra subpart II(D).
  139. .United States v. California, 314 F. Supp. 3d at 1090–93.
  140. .Id.
  141. .Id. at 1093.
  142. .Id. at 1096, 1112.
  143. .Id. at 1095 (quoting Arizona v. United States, 567 U.S. 387, 400 (2012)).
  144. .Id. at 1088.
  145. .Id. at 1096.
  146. .Washington v. United States, 460 U.S. 536, 544–45 (1983).
  147. .521 U.S. 793 (1997).
  148. .Id. at 799.
  149. .Dawson v. Steager, 139 S. Ct. 698 (2019).
  150. .Id. at 705.
  151. .On ICE’s record of abuse, see, for example, Carlos Ballesteros, Trump’s Homeland Security Slammed for Ignoring Sexual Assaults in Immigration Detention Centers, Newsweek, (Dec. 18, 2017, 6:42 PM), [] (describing widespread incidence of alleged sexual assault in ICE facilities); see also Office of the Inspector Gen., Dep’t of Homeland Sec., Concerns About ICE Detainee Treatment and Care at Detention Facilities (2017) (documenting widespread “problems that undermine the protection of detainees’ rights, their humane treatment, and the provision of a safe and healthy environment”).
  152. .See Jacqueline Stevens, U.S. Government Unlawfully Detaining and Deporting U.S. Citizens as Aliens, 18 Va. J. Soc. Pol’y & L. 606, 608, 629–33 (2011) (estimating that there were 20,000 such cases between 2004 and 2010).
  153. .United States v. California, 314 F. Supp. 3d 1077, 1097 (E.D. Cal. 2018).
  154. .Id. at 1093, 1096, 1098, 1111.
  155. .United States v. California, No. 18-16496, 2019 WL 1717075, at *2 (9th Cir. Apr. 18, 2019).
  156. .Id. at *13–17.
  157. .On this latter point, see discussion infra subpart II(D).
  158. .United States v. California, 2019 WL 1717075, at *10–11.
  159. .Id. at *9 (citation omitted).
  160. .Cal. Gov’t Code § 12532(b)(1)(C).
  161. .United States v. California, 2019 WL 1717075, at *12.
  162. .See discussion supra pp. 147–48.
  163. .See Stevens, supra note 152, at 628–29.
  164. .United States v. California, 2019 WL 1717075, at *1–2.
  165. .Id. at *9–19.
  166. .On the history of § 1373 and its role in legal and political battles over sanctuary cities, see Lai & Lasch, supra note 2, at 550–63.
  167. .U.S. Dep’t of Justice, supra note 4 (outlining new conditions for recipients of Edward Byrne Memorial Justice Assistance Grants).
  168. .505 U.S. 144, 188 (1992) (barring commandeering of state legislative authority).
  169. .521 U.S. 898, 935 (1997) (barring commandeering of executive officials).
  170. .138 S. Ct. 1461, 1484–85 (2018).
  171. .See Margulies, supra note 2, at 1551–54.
  172. .See, e.g., United States v. California, 314 F. Supp. 3d 1077, 1099–1103 (E.D. Cal. 2018) (summarizing the federal government’s argument as to why the California law at issue conflicts with § 1373); Margulies, supra note 2, at 1551–54 (examining the implications under a broad reading of § 1373).
  173. .See, e.g., United States v. California, 314 F. Supp. 3d at 1101–04 (reaching this conclusion); City of Philadelphia v. Sessions (Philadelphia II), 289 F. Supp. 3d 289, 330–31 (E.D. Pa. 2018); Steinle v. City of San Francisco, 230 F. Supp. 3d 994, 1015 (N.D. Cal. 2017) (same). But see Bologna v. City of San Francisco, 121 Cal. Rptr. 3d 406, 413–15 (Cal. Ct. App. 2011) (endorsing the broad reading of § 1373).
  174. .Printz, 521 U.S. at 905.
  175. .8 U.S.C. § 1373(a).
  176. .See City of New York v. United States, 179 F.3d 29, 35 (2d Cir. 1999) (“We therefore hold that states do not retain under the Tenth Amendment an untrammeled right to forbid all voluntary cooperation by state or local officials with particular federal programs.”).
  177. .Maggie Haberman & Michael S. Schmidt, Giuliani to Join Trump’s Legal Team, N.Y. Times (Apr. 19, 2018), [].
  178. .City of New York v. United States, 179 F.3d at 35.
  179. .Printz, 521 U.S. at 928.
  180. .Id. at 932 n.17.
  181. .Id.
  182. .City of Philadelphia v. Sessions (Philadelphia I), 280 F. Supp. 3d 579, 651–52 (E.D. Pa. 2017). The Third Circuit did not review this part of the district court’s ruling because it considered the Spending Clause aspect of the decision sufficient to resolve the case. See City of Philadelphia v. Att’y Gen., 916 F.3d 276, 291 (3d Cir. 2019) (“Because the Attorney General exceeded his statutory authority in promulgating the Challenged Conditions, we needn’t reach Philadelphia’s other arguments.”).
  183. .City of Chicago v. Sessions (Chicago I), 264 F. Supp. 3d 933, 948–49 (N.D. Ill. 2017), aff’d, 888 F.3d 272 (7th Cir. 2018).
  184. .County of Santa Clara v. Trump (Santa Clara I), 250 F. Supp. 3d 497, 534 (N.D. Cal. 2017).
  185. .Id.
  186. .138 S. Ct. 1461 (2018).
  187. .Id. at 1466–67; 28 U.S.C. § 3702(1) (2012).
  188. .Murphy, 138 S. Ct. at 1465, 1473; Steve Ginsburg, NFL, Others File Lawsuit to Block New Jersey Sports Betting, Reuters, Oct. 20, 2014, [].
  189. .Murphy, 138 S. Ct. at 1472.
  190. .Id.
  191. .Id. at 1478.
  192. .Id.
  193. .Garrett Epps, The Supreme Court Says Congress Can’t Make States Dance to Its Tune, Atlantic (May 14, 2018), [].
  194. .Margulies, supra note 2, at 1559–60.
  195. .28 U.S.C. § 3702(1) (2012).
  196. .Margulies, supra note 2, at 1558–59.
  197. .Id. at 1558.
  198. .See discussion infra section II(D)(3) (noting the burden an information-sharing exception would impose on state and local governments).
  199. .See, e.g., Epps, supra note 193 (predicting that § 1373 would be doomed by Murphy); Ilya Somin, Federalism Comes Out as the Winner in Murphy v. NCAA, Reg. Rev. (July 10, 2018), [] (same); Ilya Somin, Broader Implications of the Supreme Court’s Sports Gambling Decision, Reason: Volokh Conspiracy (May 16, 2018, 3:30 PM), [] (same); Mark Joseph Stern, Three Cheers for Federalism, Slate (May 14, 2018, 5:50 PM), [] (same).
  200. .309 F. Supp. 3d 289 (E.D. Pa. 2018), vacated in part sub nom. City of Philadelphia v. Att’y Gen. of the U.S., 916 F.3d 276 (3d Cir. 2019).
  201. .City of Philadelphia v. Sessions (Philadelphia II), 309 F. Supp. 3d 289, 329 (E.D. Pa. 2018) (citation omitted).
  202. .Id. at 328.
  203. .Id.
  204. .Id. at 329.
  205. .New York v. U.S. Dep’t of Justice, 343 F. Supp. 3d 213, 234 (S.D.N.Y. 2018).
  206. .City of Chicago v. Sessions (Chicago II), 321 F. Supp. 3d 855, 873 (N.D. Ill. 2018).
  207. .United States v. California, 314 F. Supp. 3d 1077, 1101, 1104, 1110 (E.D. Cal. 2018).
  208. .Id. at 1089.
  209. .Id. at 1099.
  210. .Id. at 1109.
  211. .See id. at 1107 (concluding “it is highly unlikely that Congress could have made responses to requests seeking information and/or transfers of custody mandatory”).
  212. .Id. at 1107–09.
  213. .See generally infra section II(D)(3).
  214. .United States v. California, No. 18-16496, 2019 WL 1717075, at *17–19 (9th Cir. Apr. 18, 2019).
  215. .Id. at *16–17 (citation omitted).
  216. .David Rivkin & Elizabeth Price Foley, Can Trump Cut Off Funds for Sanctuary Cities? The Constitution Says Yes., L.A. Times (Dec. 7, 2016, 4:00 AM), /op-ed/la-oe-rivkin-foley-sanctuary-city-20161207-story.html [].
  217. .Id. (quoting Reno v. Condon, 528 U.S. 141, 151 (2000)).
  218. .Rivkin & Foley, supra note 216.
  219. .528 U.S. 141 (2000).
  220. .Id. at 141.
  221. .See id. at 143–45 (discussing Congress’s purposes in regulating state disclosure of drivers’ personal information).
  222. .Id. at 151.
  223. .Printz v. United States, 521 U.S. 898, 902–03 (1997).
  224. .Id. at 932 n.17.
  225. .Id. at 918.
  226. .Id. at 932 n.17.
  227. .City of Chicago v. Sessions (Chicago II), 321 F. Supp. 3d 855, 872 (N.D. Ill. 2018).
  228. .Liz Stark & Grace Hauck, Forty-four States and DC Have Refused to Give Certain Voter Information to Trump Commission, CNN (July 5, 2017, 5:49 AM), /2017/07/03/politics/kris-kobach-letter-voter-fraud-commission-information/index.html [].
  229. .See id. (describing the reasons for noncompliance by states); see also Ilya Somin, Demise of Trump Voter Fraud Commission Is a Victory for Federalism, Reason: Volokh Conspiracy (Jan. 4, 2018, 10:20 AM), [] (describing the reasons behind the disbandment of the Election Integrity Commission).
  230. .I recently decried this tendency in Ilya Somin, No More Fair-Weather Federalism, Nat’l Rev. (Aug. 18, 2017, 8:00 AM), [].
  231. .Anne Stauffer et al., Federal Share of State Revenue Rises for Third Year, Pew Charitable Tr. (July 24, 2018),

    2018/07/24/federal-share-of-state-revenue-rises-for-third-year [].

  232. .Cf. Scott Clement, Conservatives Hate Common Core. The Rest of America? Who Knows., Wash. Post (Jan. 28, 2015),

    conservatives-hate-common-core-the-rest-of-america-who-knows/?utm_term=.b7992bc76f57 [] (acknowledging conservative opposition to federal Common Core education standards).

  233. .I discuss this point in more detail in John O. McGinnis & Ilya Somin, Federalism vs. States’ Rights: A Defense of Judicial Review in a Federal System, 99 Nw. U. L. Rev. 89 (2004), and more recently in Ilya Somin, How Federalism Can Help Save the Failing “Marriage” Between Red and Blue States, Wash. Post: Volokh Conspiracy (Jan. 3, 2017), https://www. [].

  234. .E.g., Ilya Somin, Foot Voting, Federalism, and Political Freedom, in Federalism and Subsidiarity 83, 83, 90 (James E. Fleming & Jacob T. Levy eds., 2014).
  235. .For the information advantages of foot voting over ballot-box voting, see generally Ilya Somin, Democracy and Political Ignorance: Why Smaller Government Is Smarter ch. 5 (2d ed. 2016).
  236. .Id. at 138; Bryan Caplan, The Myth of the Rational Voter: Why Democracies Choose Bad Policies 94–95 (2007).
  237. .See supra section II(B)(1) (discussing the Byrne Grant cases).
  238. .Cf. City of Chicago v. Sessions (Chicago I), 264 F. Supp. 3d 933, 943–44 (listing examples).
  239. .See Ilya Somin, A False Dawn for Federalism: Clear Statement Rules after Gonzales v. Raich, in Cato Sup. Ct. Rev. 113 (2006) (arguing that clear-statement rules have little value). I do however still endorse one key point I made in that Article: that clear-statement rules are not a sufficient substitute for strong judicial enforcement of substantive constraints on the scope of federal power. Id. at 133.
  240. .See supra subpart II(D) (discussing lower-court decisions and impacts on federalism).
  241. .Cf. Peter J. Boettke et al., Federalism and the Police: An Applied Theory of “Fiscal Attention, 49 Ariz. St. L.J. 907, 920–25 (2017) (describing a pattern of increasing federal–state cooperation on law enforcement across a wide range of policies).
  242. .See supra section II(B)(2) (describing the relatedness requirement in the context of Byrne Grant conditions).
  243. .See infra subpart III(C).
  244. .See supra subpart II(C) (assessing the litigation surrounding Assembly Bill 450).
  245. .Id.
  246. .The exception is Judge Leinenweber’s initial decision on the § 1373 issue, which he later reversed after the Supreme Court decided Murphy v. NCAA. See discussion supra sections II(B)(1), II(D)(1) (discussing the sanctuary cities cases). Judge Leinenweber was also the judge in City of Evanston v. Sessions, No. 18 C 4853, 2018 U.S. Dist. LEXIS 204500, at *17 (N.D. Ill. Aug. 9, 2018), which he decided in accordance with his previous ruling in Chicago II, id. at *15.
  247. .City of Philadelphia v. Att’y Gen. of U.S., 916 F.3d 276, 278 (3d Cir. 2019). This decision did not, however, address the district court’s ruling that § 1373 violated the anti-commandeering rule.
  248. .See Scirica, Anthony Joseph, Fed. Jud. Ctr., [] (noting that Ronald Reagan nominated Judge Scirica to the Third Circuit in 1987).
  249. .City of Chicago v. Sessions (Chicago I), 888 F.3d 272, 276 (7th Cir. 2018). The ruling upheld the part of the district court decisions issuing a preliminary injunction against two of the three Byrne Grant conditions but did not address the § 1373 condition, which the initial lower-court ruling did not enjoin. See discussion supra section II(B)(1).
  250. .Bauer, William Joseph, Fed. Jud. Ctr., [].
  251. .Manion, Daniel Anthony, Fed. Jud. Ctr., [].
  252. .Rovner, Ilana Kara Diamond, Fed. Jud. Ctr., /rovner-ilana-kara-diamond [].
  253. .United States v. California, 314 F. Supp. 3d 1077, 1085, 1111–12 (E.D. Cal. 2018).
  254. .Mendez, John A., Fed. Jud. Ctr., [].
  255. .Smith, Milan Dale, Fed. Jud. Ctr., [].
  256. .The Democratic-appointed judges in question are Judge Orrick (Northern District of California); Judge Ramos (Southern District of New York); Judge Real (Central District of California); and Judges Thomas and Gould (9th Circuit). See their biographies in Biographical Directory of Article III Federal Judges, 1789 – Present, Fed. Jud. Ctr.,

    history/judges [].

  257. .See supra subpart II(A) (discussing Judge Fernandez’s problematic opinion).
  258. .Fernandez, Ferdinand Francis, Fed. Jud. Ctr.,

    fernandez-ferdinand-francis [].

  259. .Quoted in Robert Barnes, Rebuking Trump’s Criticism of ‘Obama Judge,’ Chief Justice Roberts Defends Judiciary as ‘Independent,Wash. Post (Nov. 21, 2018), https://www

    .html?utm_term=.791a3d3ba468 [].

  260. .See, e.g., S.A. Miller & Stephen Dinan, Trump More Right Than Roberts on ‘Obama Judges, Wash. Times (Nov. 25, 2018), /donald-trump-hit-judges-ideological-gap-correct/ [] (making this point).
  261. .Cf. id. (noting that “[c]ases dealing with sanctuary city laws are an exception. Both Democrat- and Republican-appointed judges alike have ruled illegal the Trump Administration’s attempts to crack down on jurisdictions that refuse cooperation with deportations”).
  262. .McGinnis & Somin, supra note 233, at 115–16; Ilya Somin, Closing the Pandora’s Box of Federalism: The Case for Judicial Restriction of Federal Subsidies to State Governments, 90 Geo. L.J. 461, 489–90 (2002); Ilya Somin, Putting the ‘General’ Back in the General Welfare Clause, Nat’l Const. Ctr., []; cf. John C. Eastman, Restoring the “General” to the General Welfare Clause, 4 Chap. L. Rev. 63, 65 (2001) (advancing a similar view).
  263. . See, e.g., H. Jefferson Powell, The Oldest Question of Constitutional Law, 79 Va. L. Rev. 633, 652–64 (1993) (considering historical bases for a principled law of federalism); Erik M. Jensen & Jonathan L. Entin, Commandeering, the Tenth Amendment, and the Federal Requisition Power: New York v. United States Revisited, 15 Const. Comment. 355 (1998) (arguing “the Court may well have gotten the original understanding wrong by reading too much into the historical evidence”).
  264. .See Michael B. Rappaport, Reconciling Textualism and Federalism: The Proper Textual Basis of the Supreme Court’s Tenth and Eleventh Amendment Decisions, 93 Nw. U. L. Rev. 819, 821 (1999) (arguing for an original understanding of “state” sovereignty).
  265. .See supra subpart II(A) and section II(D)(1).
  266. .For a discussion of the relevant history, see generally Ilya Somin, The Supreme Court of the United States: Promoting Centralization More than State Autonomy, in Courts in Federal Countries: Federalists or Unitarists? 440 (Nicholas Aroney & John Kincaid eds., 2017).
  267. .For a good example of this conventional wisdom, see Douglas Laycock, Protecting Liberty in a Federal System: The US Experience, in Patterns of Regionalism and Federalism: Lessons for the UK 137 (Jörg Fedtke & Basil S. Markesisinis eds., 2006).
  268. .William H. Riker, Federalism: Origin, Operation, Significance 152–55 (1964).
  269. .For a discussion of the latter, see Somin, supra note 234, at 102 (recognizing that foot voting “may have less to offer minority groups in the many federal systems where they are actually the majority in a few regions, but widely despised”). Cf. Desmond King, Separate and Unequal: African Americans and the US Federal Government 205–06 (rev. ed. 2007) (describing history of oppression of African-Americans by the federal government).
  270. .E.g., Noah Feldman, Sanctuary Cities Are Safe, Thanks to Conservatives, Bloomberg (Nov. 29, 2016, 1:52 PM), []; Heather Gerken, We’re About to See States’ Rights Used Defensively Against Trump, Vox (Jan. 20, 2017, 2:14 PM), []; Jeffrey Rosen, Federalism for the Left and the Right, Wall St. J. (May 19, 2017), [].
  271. .Feldman, supra note 270.
  272. .Murphy v. Nat’l Collegiate Athletic Ass’n, 138 S. Ct. 1461, 1468, 1485 (2018). For a discussion of Murphy, see infra subpart II(D).
  273. .E.g., Epps, supra note 193; Stern, supra note 199.
  274. .Stern, supra note 199.
  275. .Id.
  276. .See Murphy, 138 S. Ct. at 1468 (listing the Justices joining the opinion).
  277. .The best-known example is Breyer and Kagan’s decision to join the conservatives in partially invalidating the Affordable Care Act’s expansion of Medicaid grant conditions in Nat’l Fed’n of Indep. Bus. v. Sebelius, 567 U.S. 519 (2012). For a discussion of gradually increasing support for federalism by liberal Supreme Court Justices in recent years, see Ilya Somin, Federalism and the Roberts Court, 46 Publius: J. of Federalism 441, 456–57 (2016).
  278. .E.g., Heather K. Gerken, Foreword: Federalism All the Way Down, 124 Harv. L. Rev. 4, 74 (2010); Heather K. Gerken, A New Progressive Federalism, Democracy J. (2012), [].
  279. .See Michael Jonas, Progressive Politics from the Ground Up: With the Election of Donald Trump, Progressive Federalism’s Moment May Have Arrived, CommonWealth (July 11, 2017), [] (discussing the growing appeal of Gerken’s theories).
  280. .E.g., Gerken, New Progressive Federalism, supra note 278.
  281. .See Jonas, supra note 279 (describing Gerken’s involvement in the litigation).
  282. .See Rise of the Democratic Socialists, The Week (July 30, 2018),

    articles/786937/rise-democratic-socialists [] (“[A] growing number of progressives argue that the party can win back the working and middle classes only by moving to the left.”).

  283. .Cf. Somin, supra note 230 (arguing Trump is not respecting limits on federal power because it is not politically convenient).
  284. .The idea of federalism as “insurance” is adapted from NYU law professor and prominent federalism scholar, Rick Hills. Roderick Hills, Message to Trump-Anxious Decentralizers: Is Your Federalism Insurance Premium Paid Up?, PrawfsBlawg (Dec. 18, 2016, 9:42 AM), []; cf. Ilya Somin, Federalism as Insurance, Wash. Post: Volokh Conspiracy (Dec. 20, 2016), [] (agreeing with Professor Hills but concluding that federalism arguments against Trump’s policies are valid despite “fair weather” federalism).