Harvard Law Review Harvard Law Review Harvard Law Review

Immigration Law

States’ Commandeered Convictions: Why States Should Get a Veto Over Crime-Based Deportation

Federal immigration enforcement turns increasingly on criminal law outcomes, with criminal convictions triggering deportability,1×1. 8 U.S.C. § 1227(a)(2)(A)–(B) (2012) (making noncitizens deportable for “[c]rimes of moral turpitude,” id. § 1227(a)(2)(A)(i), aggravated felonies, id. § 1227(a)(2)(A)(iii), controlled substances offenses, id. § 1227(a)(2)(B)(i), firearms offenses, id. § 1227(a)(2)(C), other miscellaneous crimes, id. § 1227(a)(2)(D), and crimes related to domestic violence, stalking, violations of protection orders, or child abuse, id. § 1227(a)(2)(E)); id. § 1182(a)(2) (establishing similar grounds of inadmissibility). detention,2×2. Id. § 1226(c). and ineligibility for discretionary relief.3×3. Id. § 1229b(a)–(b)(1). In fiscal year 2018, the United States deported 145,000 noncitizens who were convicted of some crime, comprising 57% of deportations that year.4×4. U.S. Immigration & Customs Enf’t, Fiscal Year 2018 ICE Enforcement and Removal Operations Report 11 fig.11 (2018), https://www.ice.gov/doclib/about/offices/ero/pdf/eroFY2018Report.pdf [https://perma.cc/2GSC-99ZH] [hereinafter ICE 2018 Report]. This Note uses the term “deportation” to refer to instances in which noncitizens currently present in the United States are made to leave — what current law labels “removal.” 8 U.S.C. § 1229. This includes both instances in which individuals who lawfully entered the United States become “deportable,” id. § 1227(a), and instances when individuals present in but not formally admitted to the United States are deemed “inadmissible,” id. § 1182(a)(2). See T. Alexander Aleinikoff et al., Immigration and Citizenship 895 (8th ed. 2016). Because the vast majority of U.S. criminal convictions are for state offenses,5×5. In 2016, only 5% of the United States’s “total correctional population,” which includes individuals on parole or probation as well as the currently incarcerated, belonged to the federal system (320,000 out of 6.6 million). The federal share of incarcerated people was slightly higher, at 9% (188,000 out of 2.13 million). Danielle Kaeble & Mary Cowhig, Bureau of Justice Statistics, U.S. Dep’t of Justice, Correctional Populations in the United States, 2016, at 11 app. tbl.1 (2018), https://www.bjs.gov/content/pub/pdf/cpus16.pdf [https://perma.cc/S994-L5JK]. conviction-based immigration enforcement depends crucially on state and local law enforcement to investigate and prosecute noncitizens.

This creates tensions when federal, state, and local actors disagree on how immigration laws should be enforced. Such tensions come in various permutations: when states seek to enforce immigration laws more harshly than the federal government,6×6. See, e.g., Arizona v. United States, 567 U.S. 387, 393–95, 416 (2012) (determining such efforts preempted by federal law). when states force cities to cooperate with federal immigration authorities,7×7. See, e.g., City of El Cenizo v. Texas, 890 F.3d 164, 173–75, 182 (5th Cir. 2018). or when local actors, cities, or states curtail such cooperation,8×8. Several jurisdictions avoid collecting the immigration status of those they arrest or prosecute, so as to avoid having to share it with federal authorities. See, e.g., S.F., Cal., Administrative Code §§ 12H.2, 12I.1 (2018); Chi., Ill., Munipality Code § 2-173-020 (2012). either because they think it will impair local law enforcement9×9. See City of New York v. United States, 179 F.3d 29, 36 (2d Cir. 1999) (explaining the basis for New York’s anticooperation policy as promoting communication with law enforcement). or because they disagree with it.10×10. See Press Release, Kevin de León, President Pro Tempore, Cal. State Senate, SB 54 (de León) The California Values Act (Sept. 11, 2017), https://www.documentcloud.org/documents/3990887-SB-54-De-Leon-CA-Values-Act-FACT-SHEET.html [https://perma.cc/9D4P-Q8GK].

Though such tensions turn increasingly into lawsuits,11×11. See infra section I.B, pp. 2327–28. no state has yet challenged the central premise of this system: that the federal government may condition deportation on the bare fact of a state conviction — even when the state that obtained it disagrees. This Note argues that such involuntary “triggering” effects an unconstitutional commandeering of the state’s criminal law enforcement.12×12. The anticommandeering principle forbids the federal government from conscripting unwilling states into enforcing federal regulatory programs. See infra section I.A, pp. 2324–27. To be sure, the federal government itself detains and deports noncitizens, demanding no state assistance at the final step. But the process begins earlier — sometimes irrevocably — with state police, prosecutors, judges, and juries. By fulfilling their responsibilities under state law, these actors trigger significant federal sanctions on the individuals the state seeks to regulate — whether or not the state believes those sanctions to be warranted.

This amounts to commandeering. Formally, crime-based deportations don’t normally result from a federal determination that a noncitizen committed crimes; they result rather from the federal determination that a state adjudicated the noncitizen guilty of (state) crimes. Under the Court’s commandeering cases, that should make all the difference: the federal government has elected to base deportation not on direct regulation of individuals, but rather by operating through states. Practically, the case is stronger still: by triggering deportation based exclusively on state convictions, the federal government outsources the tasks of investigating criminal conduct of noncitizens, apprehending them, and adjudicating their guilt. The federal government thereby gets the benefits of the state’s criminal justice system without sharing in its costs. And because the federal government’s (in)action flows inexorably from the state’s, no one can be quite sure who to blame when the federal government detains and deports someone (or fails to do so). Political accountability — the commandeering cases’ lodestar — is eviscerated.

Of course, the federal government maintains broad immigration powers, including the power to detain and deport noncitizens guilty of reprehensible conduct. It may, if it chooses, ascertain that conduct on its own and act accordingly. If it hopes to outsource such determinations, it may ask for and receive help from consenting states’ law enforcement. And it may coax or coerce states disinclined to help in myriad ways — including by conditioning federal funds on their cooperation. It simply cannot deny states all opportunity to (withhold) consent to their criminal justice systems’ serving federal immigration enforcement.

Given the pervasiveness with which federal action turns on state law classifications, this Note’s argument may appear disruptive, perhaps radically so. But even assuming that its central premise (that triggering unwelcome federal action based on a state’s regulation presents a commandeering problem) extends to contexts less core to state sovereignty than criminal law, its effect will likely prove practically limited for the same reasons that the original commandeering cases proved practically limited. Because commandeering issues disappear when the state consents to its participation in federal regulation and the federal government retains powerful political tools to secure that consent, accepting this Note’s analysis will only bring disruption in areas like immigration enforcement: where the federal government seeks to condition especially controversial federal action on especially important forms of state regulation. These are, of course, the cases where the federalism concerns which animate commandeering doctrine count the most.

This Note proceeds in four parts. Part I canvasses the anticommandeering cases and their application so far to immigration enforcement. Part II explains how that enforcement works, including the extent to which it operates through state law enforcement. It argues that until recently states could control the immigration outcomes of state convictions, making the current regime’s inattention to state preferences relatively novel. Part III advances the primary argument: triggering deportation exclusively based on state convictions commandeers states’ criminal justice enforcement. Part IV considers and finds wanting two objections: first, that applying commandeering principles to triggering underweights federal interests or threatens the immigration system’s workability, and second, that states retain a choice whether to prosecute.

I. Commandeering Doctrine

A. Development by the Supreme Court

This section charts the commandeering doctrine’s development, which began in 1992 in New York v. United States.13×13. 505 U.S. 144 (1992). There, the Court struck down a federal requirement that states “take title” to nuclear waste produced in their boundaries or enact a congressionally dictated regulatory program.14×14. Id. at 175–76. Unlike other carrot-and-stick provisions, the take-title provision “directly compel[led states] to enact and enforce a federal regulatory program,” thereby “commandeer[ing their] legislative processes.”15×15. Id. at 176 (quoting Hodel v. Va. Surface Mining & Reclamation Ass’n, 452 U.S. 264, 288 (1981)). This Congress could not do: “[w]hether one views the take title provision as lying outside Congress’ enumerated powers, or as infringing upon the core of state sovereignty reserved by the Tenth Amendment, the provision [was] inconsistent with the federal structure of our Government established by the Constitution.”16×16. Id. at 177.

Acknowledging that this principle had little basis in constitutional text, the Court turned instead to history and structure, reasoning that the whole point of replacing the Articles of Confederation with the Constitution was to allow the federal government to operate directly on citizens, rather than through sometimes recalcitrant states.17×17. Id. at 163–65. This observation led to the debatable inference that by empowering the federal government to regulate individuals, the Constitution implicitly rescinded its prior powers to regulate states as states.18×18. Id. at 163–66. Why the Constitution might not have given Congress new powers without implicitly rescinding old ones is unclear, especially since any commandeering would still have needed to be a necessary and proper exercise of an enumerated power. Id. at 210 (Stevens, J., concurring in part and dissenting in part). The Court further elucidated two substantive goals served by anticommandeering: First, by enforcing a strict separation between federal and state power, it served as a structural guarantor of individual liberty.19×19. Id. at 181 (majority opinion). Second, it promoted political accountability, as the federal government could not take credit for a program’s benefits while shifting unpopular aspects to states.20×20. Id. at 182–83.

In Printz v. United States,21×21. 521 U.S. 898 (1997). its second anticommandeering case, the Court held that the federal government not only can’t “compel the States to enact or enforce a federal regulatory program . . . [but also] cannot circumvent that prohibition by conscripting the States’ officers directly.”22×22. Id. at 935 (rejecting provisions of the Brady Handgun Violence Protection Act giving state police officers certain verification responsibilities). Printz also expanded upon the anticommandeering rule’s goals, glossing both the protection of liberty and political accountability rationales with a particular emphasis on the problem of regulatory cost-shifting.23×23. Id. at 930. On individual liberty, the Court maintained that “[t]he power of the Federal Government would be augmented immeasurably if it were able to impress into its service — and at no cost to itself — the police officers of the 50 states.” Id. at 922. With respect to political accountability, it warned against a regime in which “Members of Congress can take credit for ‘solving’ problems without having to ask their constituents to pay for the solutions with higher federal taxes,” or where states are “put in the position of taking the blame for [a federal program’s] burdensomeness and for its defects.” Id. at 930 (citing Deborah Jones Merritt, Three Faces of Federalism: Finding a Formula for the Future, 47 Vand. L. Rev. 1563, 1580 & n.65 (1994)). In response to evidence that the Framers intended state officers to administer federal programs, Printz noted the “critical point . . . [that such intent failed to demonstrate] that Congress could impose these responsibilities without the consent of the States.”24×24. Id. at 910–11.

Most recently, the Court in Murphy v. National Collegiate Athletic Ass’n25×25. 138 S. Ct. 1461 (2018). struck down a provision of the Professional and Amateur Sports Protection Act26×26. Pub. L. No. 102-559, 106 Stat. 4227 (1992) (codified at 28 U.S.C. §§ 3701–3704 (2012)). (PASPA) forbidding states that did not already permit sports gambling from “authoriz[ing]” it.27×27. Murphy, 138 S. Ct. at 1468. Emphasizing the three substantive values protected by anticommandeering as laid out in New York and Printz — protecting individual liberty by diffusing power, promoting political accountability, and preventing regulatory cost-shifting28×28. Id. at 1477. In New York and Printz, regulatory cost-shifting was probably best understood as one element of the accountability prong, rather than a wholly separate justification. — the Court held that Congress could no more “forbid” states from enacting laws than it could command them to enact laws.29×29. Id. at 1478. The United States defended PASPA as a preemption provision, but the Court disagreed, reasoning that preemption merely displaces contrary state law by its direct conferral of rights or prohibitions on individuals, without demanding any particular form of (or forbearance from) state regulation. Id. at 1479–81. PASPA, in seeking to prevent sports betting not through criminalization or direct prohibition but through ordering states to prohibit or refrain from authorizing it, impermissibly outsourced federal regulation to states.30×30. Id. at 1478.

The anticommandeering rule has attracted its share of critics: it relies on general invocations of “federalism” unmoored from specific constitutional provisions,31×31. See John F. Manning, Federalism and the Generality Problem in Constitutional Interpretation, 122 Harv. L. Rev. 2003, 2029–32 (2009). its historical analysis of the Constitutional Convention and early practice is contested,32×32. See Vicki C. Jackson, Federalism and the Uses and Limits of Law: Printz and Principle?, 111 Harv. L. Rev. 2180, 2186–91 (1998); Saikrishna Bangalore Prakash, Field Office Federalism, 79 Va. L. Rev. 1957, 1995–2004 (1993). and its accountability arguments are riddled with speculation on how voters assign credit or blame33×33. Matthew D. Adler & Seth F. Kreimer, The New Etiquette of Federalism: New York, Printz, and Yeskey, 1998 Sup. Ct. Rev. 71, 98–101; Roderick M. Hills, Jr., The Political Economy of Cooperative Federalism: Why State Autonomy Makes Sense and “Dual Sovereignty” Doesn’t, 96 Mich. L. Rev. 813, 824–30 (1998); Jackson, supra note 32, at 2200–05; Mark Tushnet, Globalization and Federalism in a Post-Printz World, 36 Tulsa L.J. 11, 28–29 (2000). and fail to distinguish garden-variety preemption.34×34. See Adler & Kreimer, supra note 33, at 83–88; Evan H. Caminker, State Sovereignty and Subordinacy: May Congress Commandeer State Officers to Implement Federal Law?, 95 Colum. L. Rev. 1001, 1070–71 (1995); Neil S. Siegel, Commandeering and Its Alternatives: A Federalism Perspective, 59 Vand. L. Rev. 1629, 1673 (2006). But see Hills, supra note 33, at 899–900 (distinguishing commandeering from preemption as unneeded and ignoring states’ continued role). But the rule has achieved qualified endorsement by “process federalists,” who, while less concerned than the Court with sovereignty, endorse the rule as a modest but effective safeguard of pluralistic state policymaking.35×35. Jessica Bulman-Pozen & Heather K. Gerken, Uncooperative Federalism, 118 Yale L.J. 1256, 1297–98 (2009); Adam B. Cox, Expressivism in Federalism: A New Defense of the Anti-Commandeering Rule?, 33 Loy. L.A. L. Rev. 1309, 1312 (2000); Hills, supra note 33, at 855–91; Ernest A. Young, The Rehnquist Court’s Two Federalisms, 83 Tex. L. Rev. 1, 34–35, 127–28 (2004).

B. Anticommandeering in Immigration

The Court’s decision in Murphy triggered immediate speculation over how anticommandeering principles might apply to the standoff between the Trump Administration and “sanctuary” jurisdictions opposed to its immigration policies.36×36. See, e.g, Mark Sherman, Sanctuary Cities Could Get Boost from Sports Betting Ruling, AP News (May 15, 2018), https://www.apnews.com/f47fba5c501a459787a3f9210adaf70b [https://perma.cc/54D2-DRW2]; Richard Wolf, Sports Betting Decision May Empower States in Other Areas, from Marijuana Laws to Sanctuary Cities, USA Today (May 16, 2018, 7:29 AM), https://www.usatoday.com/story/news/politics/2018/05/16/sports-betting-ruling-impact-marijuana-immigration/612122002/ [https://perma.cc/7JGF-GLRA]. Commandeering concerns had already led courts to construe federal immigration “detainers” — directives that state law enforcement hold noncitizens forty-eight hours beyond the time they would otherwise be released — as requests, rather than mandates.37×37. See Galarza v. Szalczyk, 745 F.3d 634, 640–41, 645 (3d Cir. 2014) (“All Courts of Appeals to have commented on the character of ICE detainers refer to them as ‘requests’ or as part of an ‘informal procedure.’” Id. at 640.) (citations omitted). Since these decisions, the Department of Homeland Security itself restyled “detainers” as requests. See Recent Case, Lunn v. Commonwealth, 78 N.E.3d 1143 (Mass. 2017), 131 Harv. L. Rev. 666, 666 n.4 (2017). Since Murphy, courts have grown more willing to question federal statutes prohibiting states from restricting the flow of immigration-related information between state officers and the federal government; while the Second Circuit upheld such a provision in the 1990s,38×38. City of New York v. United States, 179 F.3d 29, 31 (2d Cir. 1999). the distinction it drew between “affirmative[] conscript[ion]” and “prohibit[ions]”39×39. Id. at 35. is roughly the one Murphy rejected.40×40. Murphy v. Nat’l Collegiate Athletic Ass’n, 138 S. Ct. 1461, 1478 (2018). As a result, at least three district courts since Murphy, including one in the Second Circuit, have concluded that federal efforts to stop states from regulating their employees’ use of immigration information constitute commandeering.41×41. See New York v. Dep’t of Justice, 343 F. Supp. 3d 213, 233–38 (S.D.N.Y. 2018); City of Chicago v. Sessions, 321 F. Supp. 3d 855, 866–73 (N.D. Ill. 2018); City of Philadelphia v. Sessions, 309 F. Supp. 3d 289, 329–31 (E.D. Pa. 2018), aff’d in part and vacated in part on other grounds sub nom. City of Philadelphia v. Attorney Gen., 916 F.3d 276 (3d Cir. 2019); see also United States v. California, 314 F. Supp. 3d 1077, 1100–01 (E.D. Cal. 2018) (suggesting though not deciding that such a law would be unconstitutional).

The frequent fights over how federalism in general and anticommandeering in particular translate to the immigration context seem likely to grow in number and intensity,42×42. For scholarly treatment of “immigration federalism,” see generally Spencer E. Amdur, The Right of Refusal: Immigration Enforcement and the New Cooperative Federalism, 35 Yale L. & Pol’y Rev. 87, 88–94 (2016); Barbara E. Armacost, “Sanctuary” Laws: The New Immigration Federalism, 2016 Mich. St. L. Rev. 1197, 1198–1205; Ming H. Chen, Essay, Immigration and Cooperative Federalism: Toward a Doctrinal Framework, 85 U. Colo. L. Rev. 1087, 1087–88 (2014); Pratheepan Gulasekaram & S. Karthick Ramakrishnan, Immigration Federalism: A Reappraisal, 88 N.Y.U. L. Rev. 2074, 2075–83 (2013); Clare Huntington, The Constitutional Dimension of Immigration Federalism, 61 Vand. L. Rev. 787, 788–95 (2008); Philip L. Torrey, Principles of Federalism and Convictions for Immigration Purposes, 36 Immigr. & Nat’lity L. Rev. 3, 5–8 (2015). given immigration’s increasing political valence,43×43. Even now, as so-called “blue states” fight to extricate themselves from the Trump Administration’s immigration policies, see cases cited supra note 41, “red states” are still moving to challenge more lenient federal immigration policies left over from the Obama administration, see Texas v. United States, 328 F. Supp. 3d 662, 734, 743 (S.D. Tex. 2018) (noting that the states’ challenge to the Obama-era Deferred Action for Childhood Arrivals program was likely to succeed on the merits). the dependence of federal immigration enforcement on state criminal law enforcement,44×44. See infra section II.B, pp. 2331–33. and the extent to which state and local interests may conflict with federal immigration priorities.45×45. See, e.g., Jessica Bulman-Pozen, Essay, Preemption and Commandeering Without Congress, 70 Stan. L. Rev. 2029, 2038–41 (2018); Annie Lai & Christopher N. Lasch, Crimmigration Resistance and the Case of Sanctuary City Defunding, 57 Santa Clara L. Rev. 539, 540–62 (2017). Already, some states seek to insulate state criminal law enforcement from federal immigration enforcement in whatever ways they can — by selectively choosing which charges to bring,46×46. See Richard Gonzales, Prosecutors Protect Immigrants from Deportation for Minor Crimes, NPR (May 31, 2017, 4:31 PM), https://www.npr.org/2017/05/31/530929936/prosecutors-protect-immigrants-from-deportation-for-minor-crimes [https://perma.cc/EB6L-YHQ5]; Corinne Ramey, Some Prosecutors Offer Plea Deals to Avoid Deportation of Noncitizens, Wall St. J. (July 7, 2017, 9:53 AM), https://www.wsj.com/articles/some-prosecutors-offer-plea-deals-to-avoid-deportation-of-noncitizens-1499419802 [https://perma.cc/6DCF-PUTP]; Vivian Yee, Prosecutor’s Dilemma: Will Conviction Lead to ‘Life Sentence of Deportation’?, N.Y. Times (July 31, 2017), https://nyti.ms/2udQXih [https://perma.cc/RX3H-543U]. refusing to collect and share data on detainees’ immigration status,47×47. See cases cited supra note 41. declining to inform federal immigration authorities when jailed noncitizens are due for release,48×48. Christopher N. Lasch et al., Understanding “Sanctuary Cities,” 59 B.C. L. Rev. 1703, 1745–46 (2018). and denying federal immigration authorities access to state correctional facilities.49×49. Id. at 1743–45.

II. Immigration Enforcement and State Criminal Law

While commandeering doctrine plays an increasing role in state challenges to federal immigration policies, no state has yet used the doctrine to challenge the federal government’s conviction-based detention and deportation regime. This Part explains that regime. It demonstrates that the federal government has only recently sought to condition so much immigration enforcement on state criminal law enforcement, even while it dispensed with mechanisms that allowed states to control the immigration consequences of state criminal law. It then describes how federal immigration enforcement turns exclusively on state criminal convictions — rather than noncitizens’ actual criminal conduct. Though primarily descriptive, this Part helps establish that conviction-based deportation triggering cannot be shielded from commandeering scrutiny because of longstanding practice; that it really does operate through states, rather than on individuals; and that a system that protects states’ sovereignty is not only possible, but in fact existed until 1990.

A. History

The federal government asserted little control over immigration until the late nineteenth century, when it sought to restrict Chinese immigration in particular.50×50. See Gerald L. Neuman, Strangers to the Constitution 19–55 (1996); Sarah H. Cleveland, Powers Inherent in Sovereignty: Indians, Aliens, Territories, and the Nineteenth Century Origins of Plenary Power over Foreign Affairs, 81 Tex. L. Rev. 1, 112–23 (2002). Before then, it was unclear whether the federal government was even thought to possess immigration powers,51×51. See Neuman, supra note 50, at 19–51; Cleveland, supra note 50, at 81–98. let alone the near-plenary power it exercises today.52×52. See, e.g., Arizona v. United States, 567 U.S. 387, 394 (2012). In its first forays into immigration enforcement, it denied entry to noncitizens with criminal convictions (as states and colonies had done before).53×53. Neuman, supra note 50, at 21–22; Juliet Stumpf, The Crimmigration Crisis: Immigrants, Crime, and Sovereign Power, 56 Am. U. L. Rev. 367, 380–81 (2006). But only in 1917 did Congress make noncitizens deportable for postentry criminal conduct: namely, for committing “crimes involving moral turpitude” (CIMTs).54×54. See Immigration Act of 1917, ch. 29, § 19, 39 Stat. 874, 889 (repealed 1952); see also Neuman, supra note 50, at 22–23. Courts have proved relatively unwilling to fill in this open-ended and values-dependent deportability ground. See Julia Ann Simon-Kerr, Moral Turpitude, 2012 Utah L. Rev. 1001, 1047–55.

At the time — and for the next seventy years — state or federal sentencing judges could issue binding Judicial Recommendations Against Deportation (JRADs) for any CIMT conviction.55×55. Immigration Act § 19. This despite increasing perceptions that control over immigration was a federal prerogative;56×56. See, e.g., Chae Chan Ping v. United States (The Chinese Exclusion Case), 130 U.S. 581, 604–09 (1889) (“The power of exclusion of foreigners being an incident of sovereignty belonging to the government of the United States, as a part of those sovereign powers delegated by the Constitution, the right to its exercise at any time when, in the judgment of the government, the interests of the country require it, cannot be granted away or restrained on behalf of any one.” Id. at 609.); Cleveland, supra note 50, at 10–11, 158–63. Congress nonetheless believed that the sentencing judge, federal or state, should be able to prevent convictions from triggering deportation.57×57. Margaret H. Taylor & Ronald F. Wright, The Sentencing Judge as Immigration Judge, 51 Emory L.J. 1131, 1144–45 (2002). While courts disagreed as to whether JRADs precluded any consideration of CIMT convictions or merely prevented them from conclusively establishing deportability,58×58. Id. at 1148–49, 1149 n.65. Compare Giambanco v. INS, 531 F.2d 141, 149 (3d Cir. 1976) (disallowing any consideration of a CIMT conviction subject to JRAD), with Hassan v. INS, 66 F.3d 266, 269 (10th Cir. 1995) (permitting consideration of a CIMT conviction, notwithstanding a JRAD, as one of several considerations), and Delgado-Chavez v. INS, 765 F.2d 868, 869–70 (9th Cir. 1985) (per curiam) (same). the question remained relatively inconsequential given how rarely conviction-based deportability grounds were invoked: in fiscal year 1980, for example, only 206 noncitizens were deported for criminal violations (about one percent of deportations that year).59×59. See Immigration & Naturalization Serv., U.S. Dep’t of Justice, 1999 Statistical Yearbook of the Immigration and Naturalization Service 227 tbl.65 (2002), https://www.dhs.gov/sites/default/files/publications/Yearbook_Immigration_Statistics_1999.pdf [https://perma.cc/VDM3-88BP] [hereinafter 1999 Statistical Yearbook]. The table treats “[n]arcotics violations” as a separate category; in 1980, 188 noncitizens were deported for narcotics violations. Id.

Though Congress at various points augmented deportability grounds with a limited set of automatic-weapon and controlled-substance offenses, CIMTs remained the primary (if rare) conviction-based deportability grounds through the early 1980s.60×60. Teresa A. Miller, Citizenship & Severity: Recent Immigration Reforms and the New Penology, 17 Geo. Immigr. L.J. 611, 622 (2003). Though courts initially assumed JRADs were available for any conviction-based deportation, Congress in 1956 removed their availability in narcotics cases. Narcotic Control Act of 1956, Pub. L. No. 84-728, § 301(b)–(c), 70 Stat. 567, 575. See generally Philip L. Torrey, The Erosion of Judicial Discretion in Crime-Based Removal Proceedings, Immigr. Briefings, Feb. 2014, at 1, 4–5. That changed in the decade’s second half: first in 1986, when Congress vastly expanded drug-related deportation grounds to include all convictions related to a federally controlled substance,61×61. The Anti-Drug Abuse Act of 1986, Pub. L. No. 99-570, § 1751, 100 Stat. 3207, 3247 (codified as amended in scattered sections of 21 U.S.C.); see César Cuauhtémoc García Hernández, Creating Crimmigration, 2013 BYU L. Rev. 1457, 1468 & n.58 (detailing the scope of this change). and second in 1988, when Congress made noncitizens deportable for “aggravated felonies.”62×62. The Anti-Drug Abuse Act of 1988, Pub. L. No. 100-690, § 7342, 102 Stat. 4181, 4469 (codified as amended in scattered sections of the U.S. Code). At the time, aggravated felonies involved convictions for mostly violent offenses with sentences greater than five years, but due to Congress’s repeated expansion of the category throughout the 1990s and early 2000s, they now reach many violent or nonviolent crimes for which the sentence exceeds one year (suspended or not).63×63. Miller, supra note 60, at 633–35. What’s more, controlled-substance convictions, aggravated felonies, or two or more CIMTs now trigger mandatory federal detention throughout the course of deportation proceedings.64×64. 8 U.S.C. § 1226(c) (2012); see Jennings v. Rodriguez, 138 S. Ct. 830, 848–51 (2018). As a result of these changes, conviction-based deportations skyrocketed from fewer than 1000 through 198465×65. 1999 Statistical Yearbook, supra note 59, at 227 tbls.65 & 66. to almost 70,000 by 199966×66. Id. at 222 tbl.63. and more than twice that today.67×67. See ICE 2018 Report, supra note 4, at 11 fig.11.

Equally important for this Note’s purposes, Congress also severely curtailed discretionary pathways to relief. In particular, just two years after making JRADs available for the new aggravated-felony category in 1988, it eliminated them altogether.68×68. Torrey, supra note 60, at 5 (first citing Anti-Drug Abuse Act of 1988, Pub. L. No. 100-690, § 7344, 102 Stat. 4181, 4470–71 (codified as amended in scattered sections of the U.S. Code); and then citing Immigration Act of 1990, Pub. L. No. 101-649, § 505(a), 104 Stat. 4978, 5050 (codified as amended in scattered sections of 8 U.S.C.)). Congress also severely restricted the Attorney General’s authority to grant waivers for conviction-based deportation, making deportation a more or less mandatory consequence for, among other things, aggravated-felony convictions.69×69. Id. at 6–7.

B. The Categorical Approach

Throughout, Congress’s focus remained on the fact of a criminal conviction, rather than the associated misconduct. Federal immigration authorities do not ask what a noncitizen actually did; they ask instead whether the state grounds of conviction “match” the federal deportation grounds.70×70. See, e.g., Moncrieffe v. Holder, 569 U.S. 184, 190 (2013) (describing the categorical approach); Alina Das, The Immigration Penalties of Criminal Convictions: Resurrecting Categorical Analysis in Immigration Law, 86 N.Y.U. L. Rev. 1669, 1688–1702 (2011) (explaining its history). This “categorical approach” is anchored in the Immigration and Nationality Act’s71×71. Pub. L. No. 89-236, 79 Stat. 911 (1965) (codified as amended in scattered sections of 8 U.S.C.). (INA) specification that deportability and inadmissibility are triggered when a noncitizen is “convicted” of certain crimes,72×72. See statutes cited supra note 1 (detailing conviction-based deportation and inadmissibility grounds); Moncrieffe, 569 U.S. at 191. The categorical approach is also employed in deciding whether defendants should receive sentencing enhancements under the Armed Criminal Career Act (ACCA), 18 U.S.C. § 924 (2012), based on their past convictions; indeed, the ACCA is responsible for the birth of the modern categorical approach in Taylor v. United States, 495 U.S. 575, 588–89 (1990). See also Mathis v. United States, 136 S. Ct. 2243, 2248–49 (2016). and promotes (in theory) efficient, predictable enforcement.73×73. Moncrieffe, 569 U.S. at 200–01; see Das, supra note 70, at 1733–42.

To illustrate the categorical approach in action, consider what happens when Congress ties deportation to “burglary” convictions.74×74. 8 U.S.C. § 1101(a)(43)(G) (2012). The ACCA targets burglary as well. 18 U.S.C. § 924(e). The Supreme Court has interpreted Congress’s references to “burglary” to mean “an unlawful or unprivileged entry into, or remaining in, a building or other structure, with intent to commit a crime.”75×75. Taylor, 495 U.S. at 598 (citations omitted) (describing burglary in the ACCA); see also Moncrieffe, 569 U.S. at 190 (stating that the same approach used to ascertain the “generic” version of offenses under the ACCA generally applies to the INA). If a state’s burglary statute criminalizes only conduct incorporated by generic burglary, then a conviction under the state statute qualifies as grounds for deportation. If, however, the state adopts a broader burglary definition — say it criminalizes breaking or entering a car, as well as a building or structure — then a conviction under its burglary statute may not establish generic burglary, and therefore deportability, under the INA.76×76. The inquiry does not end there. If the state treats the type of thing entered as a necessary “element” of its burglary offense — if, in other words, the prosecutor must prove or the defendant plead that she entered either a house, or a car, or some other structure, then, under the “modified categorical approach,” a court or immigration official may “peek” at the defendant’s record of conviction solely to establish which prong of the state’s burglary statute the defendant was convicted under. Mathis, 136 S. Ct. at 2256–57 (quoting Rendon v. Holder, 782 F.3d 466, 473 (9th Cir. 2015) (Kozinski, J., dissenting from the denial of rehearing en banc)). If, on the other hand, the state treats the fact of entry into a house, car, or other structure as mere “means” of committing burglary — if, in other words, it is enough that the jury agrees that the defendant unlawfully entered one of those things, without necessarily agreeing as to which — then a conviction under the state’s burglary statute fails to establish whether the defendant was “convicted” of burglary for INA purposes. Id. at 2257.

The noncitizen’s actual conduct is irrelevant; what matters is what the state thought about her conduct, as expressed through its conviction.77×77. Specific (rare) provisions in the INA require a more fact-based inquiry. In Nijhawan v. Holder, 557 U.S. 29 (2009), the Court held that a fraud deportability ground requiring losses greater than $10,000 necessitated a generic inquiry into whether the crime was fraud and a fact-based inquiry into the loss amount (as state fraud convictions rarely establish by themselves a particular dollar loss). Id. at 33–38. To take the most obvious example, if one state (State A) criminalizes conduct corresponding to a federal deportability ground, while another state (State B) does not, then a noncitizen in State A faces deportation where a noncitizen in State B who engaged in the same conduct does not. But differences can be more subtle, or even arbitrary; perhaps States A and B each criminalize distribution of any of 101 substances, 100 of which are federally controlled. If State A requires proof as to which of the 100 drugs a defendant distributed, then convictions under State A’s statute may trigger deportation under the INA (unless the defendant is convicted of distributing the single substance not on the federal schedule). But imagine State B does not require proof of which substance was distributed, allowing convictions so long as the defendant is proved to have possessed any controlled substance — then, assuming that there is a “realistic probability” that someone in State B might be prosecuted for distributing the single drug not on the federal schedules,78×78. See Gonzales v. Duenas-Alvarez, 549 U.S. 183, 193 (2007) (reasoning that a mismatch under the categorical approach requires “a realistic probability, not a theoretical possibility, that the State would apply its statute to conduct that falls outside the [federal] definition of a crime”). any distribution conviction in State B will fail to establish deportability, even though the vast majority of its distribution cases likely involve a federally controlled substance.79×79. See Mellouli v. Lynch, 135 S. Ct. 1980, 1990 (2015). The point here is that the federal immigration consequence depends crucially on the state classification, rather than the noncitizen’s conduct.

This increased entwinement of federal immigration law with state criminal law blurs boundaries between the two, leading to increased scholarly treatment of both systems as parts of a unified whole.80×80. See, e.g., García Hernández, supra note 61, at 1467–85; Stephen H. Legomsky, The New Path of Immigration Law: Asymmetric Incorporation of Criminal Justice Norms, 64 Wash. & Lee L. Rev. 469, 475–500 (2007); Miller, supra note 60, 611–14; Stumpf, supra note 53, at 376 (coining the term “crimmigration” to refer to the intersection of criminal and immigration law). The Supreme Court’s landmark 2010 decision in Padilla v. Kentucky81×81. 559 U.S. 356 (2010). recognized this reality: in holding that effective assistance of criminal defense counsel under the Sixth Amendment requires at least some warning of a conviction’s likely immigration consequences, the Court acknowledged both that deportation may constitute a harsher sanction than criminal punishment and that its likelihood turns fundamentally on the precise resolution of state criminal charges.82×82. See id. at 365, 368–69.

The result is that states that wish to avoid triggering federal immigration consequences with which they disagree are forced to distort or refrain from exercising their criminal enforcement powers. As detailed above, several jurisdictions prohibit gathering information relating to immigration status so as to avoid any legal obligation to share that status with federal authorities.83×83. See Lasch et al., supra note 48, at 1745–48; sources cited supra note 8. Others adjust the criminal charges they bring or even refrain from bringing them altogether in order to avoid the most damaging immigration consequences.84×84. See sources cited supra note 46. Such tactics have generated enormous controversy, with the current administration criticizing them as fundamentally unfair to citizens who receive harsher criminal treatment for the same conduct.85×85. See id. And yet these distortions seem the inevitable consequence of tethering federal immigration decisions to those of states that may substantively disagree with those decisions.

III. State Crime–Based Deportation as Commandeering of State Criminal Enforcement

This Part makes the case that deporting a noncitizen on the sole basis of a state criminal conviction constitutes an impermissible commandeering of that state’s criminal enforcement process where the state does not want its conviction to be so used. It first argues that doing so involves acting on states, rather than individuals: both formally, in the sense that federal action turns exclusively on decisions by state government, rather than individuals’ actions, and practically, in the sense that criminal justice bureaucracies of all fifty states are being transformed, whether they like it or not, into tools of immigration enforcement. It then considers how the functional considerations invoked by the Court in its commandeering cases — protecting individual liberty, promoting political accountability, and preventing regulatory cost-shifting — are implicated in the context of immigration enforcement.

Before laying out this argument, a fundamental point bears mentioning: the unconstitutional commandeering occurs only when the state objects to its criminal justice system being used to effect deportation. In many, perhaps most, cases, the state may have no such objection.86×86. Indeed, when state sentencing judges were empowered to veto deportation with a JRAD, the procedure was used relatively rarely. Taylor & Wright, supra note 57, at 1143, 1148–51. As Printz made clear, such cases simply do not bear on what “Congress could impose . . . without the consent of the States.”87×87. Printz v. United States, 521 U.S. 898, 910–11 (1997).

A. Conviction-Based Immigration Enforcement Operates Through States, Not Individuals

1. The INA Targets “Convictions,” Not Conduct. — As outlined in section I.A, the basic structural and historical argument underlying the prohibition on commandeering is the idea that in trading the Articles of Confederation for the Constitution, we traded a federal government empowered to operate exclusively on states for one that could operate exclusively on their citizens.88×88. See supra section I.A, pp. 2324–27; see also New York v. United States, 505 U.S. 144, 164–66 (1992). Congress may regulate states incidentally, the same way it regulates similarly situated entities. See Garcia v. San Antonio Metro. Transit Auth., 469 U.S. 528, 531 (1985) (determining that states as employers enjoyed no special carveout from generally applicable regulations such as the Fair Labor Standards Act). Under that framing, federal deportation decisions — even those based on state criminal convictions — might seem to pose no commandeering problem, because the federal government clearly acts upon an individual by deporting him. But, as explained above, this action does not result from federal determinations about a noncitizen’s conduct; it rather is triggered when the federal government determines that the state acted a certain way by “convict[ing]” the noncitizen.89×89. See supra section II.B, pp. 2331–33.

The result is that a harsh federal penalty90×90. See Sessions v. Dimaya, 138 S. Ct. 1204, 1213 (2018) (plurality opinion). turns exclusively on the actions of state officers who investigate the crime, determine responsibility, and prove guilt to state judges or juries. Such “triggering” is not quite the direct impressment of state officers that the Court forbade in Printz, but it involves the same central feature of commandeering cases: the federal government’s operating on or through states, rather than directly on individuals.

To see why, consider the case where the federal government chooses to base federal action entirely on some quintessentially state or local legal classification that a state government can’t very well avoid making.91×91. Putting aside, until infra section IV.B, pp. 2342–43, whether a state’s ability to avoid the action changes the analysis. Imagine too that the triggered federal action is something which the federal government indisputably has power to do independently but which is only loosely related to the state’s triggering action, and which the state might not like. For example, the federal government might choose to ban from driving on any interstate highway all drivers who receive any driving citation under state law, no matter the reason for the citation, the severity of any associated penalty, or the procedures the state might have in place for restoring the driver’s privileges (if they were ever even suspended).92×92. Though such a law might seem punitive or arbitrary, the absence of invidious discrimination and the presence of a rational (if poorly tailored) connection to interstate safety would likely doom an individual rights challenge. Cf. Armour v. City of Indianapolis, 566 U.S. 673, 682 (2012) (holding that administrative costs can justify arbitrary and discriminatory treatment of a nonsuspect class). Such a law would necessarily and substantially affect how states go about issuing driving citations, because the state knows any effort it makes to regulate driving will result in drastic federal consequences to the regulated individuals. And while the federal government may be interested in regulating driving by individuals, it has chosen to do so through the state’s regulatory process — whether the state likes it or not.

One might respond that such “triggering,” to the extent it involves acting through the state, still falls short of outright impressment as contemplated in the Court’s commandeering cases. The federal government regularly conditions federal action on state action: for instance, when it offers contingent federal funds — an offer it can make subject to only weak constitutional constraints, far from the commandeering cases’ flat prohibition.93×93. Compare South Dakota v. Dole, 483 U.S. 203, 207–08 (1987) (laying out relevant considerations when determining the appropriateness of conditions on federal funds), and Nat’l Fed’n of Indep. Bus. v. Sebelius (NFIB), 567 U.S. 519, 576–78 (2012) (same), with New York, 505 U.S. at 177–78 (rejecting any sort of balancing of the federal interest in commandeering against the state’s interest in sovereignty). But as the Court itself noted in New York, the primary distinction between the spending and commandeering cases is consent: in the former but not the latter cases, the “residents of the State retain the ultimate decision as to whether or not the State will comply” with the federal regulatory objective.94×94. New York, 505 U.S. at 168. This is why conditions on federal funds must be expressed unambiguously and in advance,95×95. See Dole, 483 U.S. at 207 (citing Pennhurst State Sch. & Hosp. v. Halderman, 451 U.S. 1, 17 (1981)). and the state must be actually, not just theoretically, equipped to say no.96×96. NFIB, 567 U.S. at 581 (determining that Congress cannot use its spending power as a “gun to the head”). What the federal government does in the driving hypothetical or in immigration enforcement is different; it is not offering an incentive or threatening a penalty, but rather tying its own regulation to the state’s in such a way as to deny the state any opportunity to opt out of the federal regime to the extent it hopes to regulate at all.

Of course, in tying federal immigration enforcement to state convictions, the federal government doubtless is trying to get at individuals’ underlying criminal conduct. Given its power to condition immigration consequences on such conduct, it may seem a little silly to ask it to close its eyes to the reality of a state conviction just because the state asks it to.97×97. Precisely this issue came up when sentencing judges were empowered to issue JRADs. Some courts reasoned that federal immigration authorities could legitimately consider state convictions, notwithstanding JRADs, as relevant information, while others reasoned that the whole point of the JRAD was to negate the conviction’s relevance. See cases cited supra note 58. But that is essentially a criticism of commandeering doctrine itself, which does not allow the federal government to dictate that federal problems be solved through state governments even where it determines that doing so would be most sensible. And, to the extent the federal government really does find it most convenient to operate through state police, prosecutors, judges, and juries, rather than bear the cost of conducting independent investigations and adjudications, it retains powerful tools to entice states to go along.98×98. See infra section IV.A, pp. 2340–41. What it can’t do is hijack the state’s criminal law enforcement process, upsetting the state’s attempts to calibrate punishment to conduct, or perhaps even preventing it from exercising its criminal enforcement powers altogether.99×99. See sources cited supra note 46.

2. State Law Enforcements Are Transformed — Perhaps Unwillingly — into Agents of Federal Immigration Law. If the INA’s convictions provisions operate formally through states, the argument that the INA practically transforms state law enforcement into federal immigration agents seems even stronger. Federal law enforcement rises and falls with the decisions of state law enforcement; that is why recent fights over sanctuary policies are so contested.100×100. See, e.g., cases cited supra note 41; Lasch et al., supra note 48, at 1713–19. As explained in section II.A, this is not how immigration enforcement used to work: though federal immigration law paid some attention to state criminal outcomes for the last century, only recently did so many detention and deportation determinations come to turn exclusively on the fact of a state conviction. Simultaneously, Congress removed most pathways for canceling deportation based on the facts of a particular case.101×101. See Padilla v. Kentucky, 559 U.S. 356, 363–64 (2010); Legomsky, supra note 80, at 524. Most important here, state judges formerly could but no longer can control the immigration consequences of state convictions with JRADs.102×102. See supra section II.A, pp. 2329–31.

What’s more, Congress now sweeps up a much broader range of state crimes as potential predicates for deportation,103×103. Legomsky, supra note 80, at 482–86. from the single “crime of moral turpitude,” codified in 1917, to crimes of violence, crimes involving fraud or deceit, crimes involving firearms, and even essentially all state drug crimes involving a federally controlled substance.104×104. 8 U.S.C. § 1227(a)(2)–(3) (2012). The INA’s threshold for aggravated felonies — which link mandatory detention, deportation, and bars to future entry — has been defined downward to include broad swaths of state convictions.105×105. See id. § 1227(a)(2)(A)(iii); id. § 1101(a)(43). And be-cause this harshening has occurred alongside the harshening of state criminal laws,106×106. See generally Michelle Alexander, The New Jim Crow 40–58 (rev. ed. 2012). more and more state crimes trigger immigration consequences even holding federal law constant. This is reflected in the immigration arrest and deportation statistics, with about 57% of total deportations driven by a criminal offense,107×107. ICE 2018 Report, supra note 4, at 11 fig.11. This figure includes only those deportations where the noncitizen has actually been convicted of a crime; a further 25% of deportations involve individuals who are suspected of or have been charged with one. Id. up from about 1% in fiscal year 1980.108×108. See 1999 Statistical Yearbook, supra note 59, at 227 tbl.65.

State law enforcement therefore plays a crucial role in federal immigration outcomes, whether or not the state consents. As a result, the current system does not differ so sharply from the “any driving citation” hypothetical posited in section A.1.109×109. See supra p. 2335. To be sure, unlike that hypothetical, deportations are triggered only by state convictions corresponding to particular federal deportation grounds, not by all criminal convictions under state law. But those grounds are pervasive, and fundamental to state criminal law enforcement — analogous to a driving-citation trigger that targeted all citations for “speeding.”110×110. Congress could specify a speeding threshold, or leave it to the courts to discern the term’s “generic” meaning. Cf. Esquivel-Quintana v. Sessions, 137 S. Ct. 1562, 1569–70 (2017) (discerning likewise for sexual abuse of a minor). The problem is still the same: the state cannot regulate “speeding” without the federal government seizing (and free riding) on the state’s regulatory action.

In fact, the Supreme Court already recognizes the extent to which state criminal law enforcement operates as an agent of federal immigration enforcement, determining that all state criminal defense counsels are constitutionally obligated to warn their clients of the adverse immigration consequences likely to be triggered by a state conviction.111×111. Padilla v. Kentucky, 559 U.S. 356, 374 (2010). But because those consequences may be complex or arbitrary — turning on the nuances of how a particular state statute, as written, interpreted, and applied, intersects with federal removal grounds112×112. See supra section II.B, pp. 2331–33. — the Court acknowledges that state defense counsel (not to mention state prosecutor or sentencing judge) cannot always predict those consequences,113×113. Padilla, 559 U.S. at 369. let alone control them. The result is an immigration enforcement system that employs state law enforcement throughout its administration, while denying the state any role in its outcomes.

B. Substantive Goals Behind Anticommandeering

As elaborated by the Supreme Court, anticommandeering serves three substantive values: the preservation of liberty through the diffusion of state power, political accountability, and preventing the federal government from regulatory cost-shifting.114×114. Murphy v. Nat’l Collegiate Athletic Ass’n, 138 S. Ct. 1461, 1477 (2018). The third of these points seems in many ways to be a subset of the second. Unsurprisingly, given the formal and practical senses in which federal immigration enforcement operates through state criminal law enforcement, crime-based deportations clearly implicate these values. To take the first point, commandeering doctrine purports to preserve individual liberty by preventing the federal government (whose powers are enumerated) from assuming and centralizing powers not delegated to it.115×115. See id.; Printz v. United States, 521 U.S. 898, 921 (1997). Why commandeering doctrine is necessary to achieve this goal is unclear, as enumeration would presumably limit the purposes and scope of commandeering just as it limits direct regulation. Such a story applies clearly to the INA’s conviction provisions, which allow the federal government to regulate huge swaths of potential and actual conduct by millions of people in areas core to state police powers, where Congress’s own regulatory powers are bounded.116×116. Cf. Bond v. United States, 134 S. Ct. 2077, 2088–92 (2014) (construing a statutory ban on “chemical weapons” implausibly narrowly, so as to avoid constitutional questions regarding Congress’s Treaty Clause power to regulate in areas of core state police powers); United States v. Morrison, 529 U.S. 598, 617–19 (2000) (holding that Congress lacks powers under the Fourteenth Amendment or the Commerce Clause to authorize a civil cause of action for all instances of gender-based violence); United States v. Lopez, 514 U.S. 549, 564–67 (1995) (holding a ban on guns near schools beyond Congress’s Commerce Clause power).

The Court’s concern about accountability is less focused on how much power gets exercised than with ensuring that citizens know who exercises it.117×117. New York v. United States, 505 U.S. 144, 168–69 (1992). But see sources cited supra notes 33–34. Under this framework, the problem with triggering deportation based on state convictions is obvious, as it is difficult to imagine a system where responsibility for a noncitizen’s ultimate deportation is more diffuse or obfuscated. State police or prosecutors can legitimately claim that their decisions were based on local enforcement priorities and that to act differently would favor noncitizens over citizens.118×118. See sources cited supra note 46. State criminal defense counsel, though constitutionally obliged to warn about a conviction’s likely immigration consequences, may be unable to foresee those consequences — as the Court acknowledged in the same opinion recognizing the obligation.119×119. See Padilla v. Kentucky, 559 U.S. 356, 369, 374 (2010). Federal immigration agents may see their hands tied by state convictions, which may mandate a noncitizen’s detention and ultimate deportation.120×120. 8 U.S.C. §§ 1226(c), 1227 (2012). Perhaps the President might impose some order on a statutory hodgepodge of duties to detain or deport, but the extent of any discretion to tailor enforcement efforts is unclear,121×121. Cf., e.g., Texas v. United States, 809 F.3d 134, 146 (5th Cir. 2015), aff’d by an equally divided court, 136 S. Ct. 2271 (2016) (upholding by a 4–4 vote a preliminary injunction against the Obama Administration’s efforts to categorically defer deportation for broad classes of noncitizens). But see Bulman-Pozen, supra note 45, at 2047 & n.75; Gillian E. Metzger, The Constitutional Duty to Supervise, 124 Yale L. J. 1836, 1929 (2015) (“Precluding prospective and categorical articulation of immigration enforcement policy and priorities is tantamount to insisting that nonenforcement decisions be made by lower-level officials, a requirement as much at odds with constitutional structure as a presidential dispensation power.”). and her priorities may, at any rate, conflict with those of the states.122×122. See, e.g., Texas, 809 F.3d at 149; Bulman-Pozen, supra note 45, at 2039–41. All these actors might wave their hands and point to Congress, but Congress can fairly kick the ball back to states or the President, claiming that it is perfectly reasonable to base immigration status on “crimes of violence,” “firearms offenses,” or “drug offenses,” and that any error in application is the fault of over- or under-aggressive state or federal enforcement. This blame merry-go-round pretty well describes the sanctuary city debate, with all levels of government criticizing each other for unfairly putting them into a bind.123×123. See, e.g., Tanvi Misra, To Combat “Sanctuary Cities” the Department of Justice Has Opted for Shame and Blame, Pac. Standard (Aug. 16, 2018), https://psmag.com/social-justice/the-doj-plays-the-shame-and-blame-game [https://perma.cc/9U2P-X7RS] (recounting a spat beginning with a tweet by the Philadelphia Mayor’s Deputy Chief of Staff, which the Department of Justice countered with a press release blaming the city for a heinous crime).

As for the final rationale behind the anticommandeering rule — the prevention of regulatory cost-shifting — it is hard to imagine a regime that shifts costs more than the present one. The federal government’s system of immigration regulation targets an expansive set of potentially criminal conduct by noncitizens, but state police, prosecutors, judges, and juries are the ones who must actually investigate, prosecute, and adjudicate such conduct. The federal government therefore reaps the benefits of state criminal regulation without sharing its costs.

IV. Objections and Responses

Applying a commandeering analysis to conviction-based deportation triggering raises two serious concerns. First, applying commandeering principles to federal action triggered by state action — as opposed to direct federal compulsion — undervalues federal interests by granting states a trump over the sorts of action necessary to any system of dual government. Second, states retain the choice to refrain from prosecuting, and in that sense are not compelled to participate in any federal program with which they disagree. This Part explains why neither concern is convincing.

A. Federal Interests and Workability

The nature of our system of dual sovereignty is that federal law pervasively presumes and relies on the backdrop of state regulation.124×124. To take some obvious examples: property rights protected by the Takings Clause are defined by state law, see Webb’s Fabulous Pharmacies, Inc. v. Beckwith, 449 U.S. 155, 161 (1980); the federal bankruptcy code is riddled with the incorporation of “applicable law,” often the states’, in defining contractual and familial obligations, see, e.g., 11 U.S.C. § 101(11)(C), (14A) (2012), as are the Treasury regulations interpreting the Internal Revenue Code, see, e.g., 26 C.F.R. 301.7701-18(b)(1) (2018) (defining marriages for federal tax purposes with reference to state law). Given that backdrop, this Note’s central argument — that federal reliance on state convictions without state consent constitutes commandeering — should engender a certain amount of skepticism. At least, it may systematically overweight state sovereignty while undervaluing any interest in effective federal action; at most, it seems to strike at the workability of dual governance.

To see why these concerns are unfounded, it is helpful to return to Printz’s explanation for why the Founding-era expectation that state officers would enforce federal law did not doom its anticommandeering rule.125×125. See 521 U.S. 898, 910–11 (1997). Here, as there, the determinative factor remains state consent.126×126. See id. States presumably have little objection to ordinary federal bankruptcy, tax, or other provisions that depend incidentally on state classifications. And any concern that this Note’s analysis would give states a tool for blackmail — that would enable them to threaten to gum up routine federal governance as a way to extract concessions — neglects the same powerful tools the federal government maintains for incentivizing state cooperation that prevented the original anticommandeering rule from proving particularly disruptive.127×127. See Larry D. Kramer, Putting the Politics Back into the Political Safeguards of Federalism, 100 Colum. L. Rev. 215, 278–87 (2000) (noting the myriad political mechanisms available to states and the federal government to safeguard their interests and encourage the other’s cooperation); Hills, supra note 33, at 855–91; Jackson, supra note 32, at 2211–12; Young, supra note 35, at 141–47.

In particular, federal power to condition federal funds on state compliance should take care of bald recalcitrance.128×128. See Hills, supra note 33, at 858–65; Young, supra note 35, at 141–47. What this Note’s framework instead leaves open is the possibility that states might, at some cost to themselves, decline to allow their governmental actions to serve as predicates for federal action with which they strongly disagree — opting out, in other words, at precisely those times when their interest in not advancing a federal objective is the strongest.129×129. An analogous case may be that of the Affordable Care Act’s Medicaid expansion: since NFIB gave states the choice whether to go along, 567 U.S. 519, 587–88 (2012), states mostly but by no means uniformly responded to powerful incentives embodied in a controversial law by participating in the federal program. See Abbe R. Gluck & Nicole Huberfeld, What Is Federalism in Healthcare For?, 70 Stan. L. Rev. 1689, 1733–57 (2018). In such cases, federal regulation hardly need grind to a halt, as Congress could simply regulate individual conduct directly. In the immigration context, for instance, Congress has already authorized detention or deportation based on a noncitizen’s acts of sabotage or espionage, whether or not a state has obtained a related conviction.130×130. 8 U.S.C. § 1227(a)(4)(A)(i) (2012). To the extent it hopes to reach a broader variety of criminal conduct and proves unable to entice states to allow their law enforcement to be used in federal enforcement, Congress could authorize immigration penalties based on direct federal investigation and adjudication of noncitizens’ criminal conduct.

Even if this Note fails to persuade that commandeering theory generally requires states to retain at least some control over federal use of state law, there remains a powerful case for curtailing federal authority to rely on nonconsenting states’ convictions to effect deportation. State criminal law enforcement, effectively impressed into federal service in the ways described above, is particularly essential to state sovereignty; what’s more, the federal government’s broad powers to regulate immigration directly make such impressment particularly unnecessary. That state sentencing judges were empowered between 1917 and 1990 to veto the immigration consequences of state convictions should alleviate concerns about such a system’s workability. To the extent courts were more inclined to balance state interests against federal needs — as opposed to adopting the commandeering cases’ categorical rule131×131. See Jackson, supra note 32, at 2257–58 (advocating a more flexible approach to addressing commandeering concerns).  — the current immigration system seems like a clear case of state interests being underweighted.132×132. This analysis likewise rebuts another potential argument in favor of the current system of triggering based on state convictions: that structural constitutional rules like anticommandeering apply with less force in the immigration context. Cf. United States v. Curtiss-Wright Export Corp., 299 U.S. 304, 319–22 (1936) (considering nondelegation strictures less binding with respect to foreign relations); Ann Althouse, The Vigor of Anti-Commandeering Doctrine in Times of Terror, 69 Brook. L. Rev. 1231, 1261–73 (2004) (considering but rejecting an “emergency” exception to commandeering doctrine in light of the War on Terror). Whether or not some areas involving immigration are so core to federal sovereignty and peripheral to states’ as to allow exceptions to the commandeering rule, the INA’s reliance on states to investigate and adjudicate quotidian criminal conduct seems like a poor fit.

B. Do States Have a Choice?

One last, powerful objection to applying a commandeering analysis to conviction-based deportation triggering is the basic intuition that the state may choose not to convict — in contrast to the commandeering cases, where the state’s action was compulsory. Under the current regime, this may in fact be states’ only way out of enforcing federal immigration law.133×133. See sources cited supra note 46. But this falls short of the type of choice the Court’s federalism cases require: a meaningful opportunity not to participate in a federal regulatory program with which states disagree.

For one thing, conviction-based deportation triggering seems to violate the combination of two explicit constraints articulated in the commandeering cases: first, New York’s explanation that Congress does not make coercion constitutionally palatable by giving states a choice between two unacceptable options,134×134. 505 U.S. 144, 175–76 (1992). and second, Murphy’s rule against federal commands that states not regulate.135×135. Murphy v. Nat’l Collegiate Athletic Ass’n, 138 S. Ct. 1461, 1478–79 (2018). Of course, given Congress’s broad latitude to preempt state laws that conflict with federal ones, it is unclear whether this constraint bars anything more than the formal ban on the use of particular words that “State A may not pass X law.” Compare id., with sources cited supra note 34. Under the INA, the state’s “choice” is between triggering unwelcome federal immigration enforcement, on the one hand, or being forced to not enforce its own criminal laws, on the other. That choice is especially problematic because of state criminal regulation’s unique role in state sovereignty,136×136. As part of the same federalism revival to which New York and Printz belonged, the Rehnquist Court also held that Congress’s powers under the Commerce Clause do not extend to quintessentially noneconomic areas of traditional state regulation, including, especially, ordinary criminal law enforcement. See Young, supra note 35, at 2 & n.1, 23–39. which makes it altogether unclear whether the federal government could displace state criminal law enforcement even if it wished.137×137. Professor Neil Siegel has argued that commandeering and preemption doctrines should turn at least in part on the feasibility of a federal substitute for the prohibited state action. Siegel, supra note 34, at 1635. And whether or not it constitutionally could fill the gap, that’s of course not the deal it currently offers the states, which must either go along with the federal immigration consequences of their convictions or refrain from criminally regulating at all.

More fundamentally, the “choice” posed by the INA fails to solve the commandeering problem because it eliminates any way for state governments to remain accountable for their participation in a federal regulatory program. This requirement is precisely how New York distinguished between (acceptable) spending incentives or “cooperative federalism” models,138×138. These give states a choice between administering a program in service of a federal objective or allowing the federal government to do so. See New York, 505 U.S. at 167–68 (describing such programs in the Clean Water Act, the Resource Conservation and Recovery Act, the Occupational Safety and Health Act, and the Alaska National Interest Lands Conservation Act); Gluck & Huberfeld, supra note 129, at 1769–72 (describing features of the Affordable Care Act fitting this model). and (unacceptable) commandeering.139×139. 505 U.S. at 168. And it is entirely absent in the case of conviction-based deportation triggering; because states can only avoid enforcing federal immigration law by not enforcing their own criminal laws, accountability for one choice is inextricably bound up with accountability for the other. States thereby lack an ability to opt into or out of the federal immigration scheme.140×140. See South Dakota v. Dole, 483 U.S. 203, 207–08 (1987) (emphasizing the importance of clearly stated federal conditions and state consent to such conditions); see also Nat’l Fed’n of Indep. Bus. v. Sebelius, 567 U.S. 519, 581 (2012) (clarifying that consent requires an actual ability to say no). Even leaving aside fairness concerns raised by a state’s choice to not prosecute noncitizens where it would prosecute similarly situated citizens,141×141. See sources cited supra note 46. the state’s option to not prosecute is therefore not a sufficient constitutional alternative to triggering unwanted immigration consequences.

V. Conclusion

In denying states the opportunity to opt out of the severe immigration consequences of their convictions, the federal government impresses the police, prosecutors, judges, and juries of nonconsenting states into its service. That impressment — historically recent but massive in scale — goes against the commandeering cases’ doctrinal analysis and flouts their spirit. The resulting system obscures accountability while leading federal, state, and local government actors to work around, not with, each other. Resolving these concerns necessitates neither radical innovation nor disruption; it simply requires restoring a mechanism for states to keep their convictions from triggering deportation.