Administrative Discretion in Criminal and Immigration Enforcement

Essay - by Online Edition - Volume 103

Separation-of-powers scholars and the Supreme Court are obsessed with administrative discretion these days. Essentially, discretion is “the power to choose between two or more courses of action, each of which is thought of as permissible.”[1] Agencies exercise discretion when they engage in behavior that is not specified by or falls outside the requirements of the legislation that empowers them to act.[2] Academics and lawyers concerned about the scope of administrative discretion from a constitutional perspective, including much of today’s Court, seek to limit administrative capacity to implement statutes flexibly and responsively.[3] Others are convinced of bureaucrats’ essential regulatory competence,[4] and of the impossibility of defining legislative standards “to the extent that the exercise of judgment and discretion by administrative agencies would no longer be desirable or necessary,”[5] which renders these scholars deeply committed to expansive administrative power to imp/lement the law. For some in this community, the internal (or administrative) separation of powers—whereby the public, civil servants, and political officials temper one another’s potential for the excessive exercise of power—sufficiently balances the executive branch.[6]

And yet, this debate glosses over the anxieties of those who observe the exercise of discretion outside the confines of administrative statutory interpretation that occurs while regulating. Law enforcement officials hold an immense amount of discretionary authority. Furthermore, bureaucrats’ individualized exercise of enforcement power can have a systemic impact[7]—and arguably, perhaps even more so than the use of discretion that undergirds administrative statutory interpretation, which is by nature less frequent and requires consensus among the administrators involved in rulemaking.[8]

Just about all outcomes of the enforcement of law, whether beneficial or concerning, are to some extent the result of administrative discretion. In her insightful new article, Misdemeanor Declination: A Theory of Internal Separation of Powers, Professor Alexandra Natapoff expertly illustrates that this is as true in criminal administration as anywhere else.[9] In particular, Natapoff identifies and evaluates an important moment of discretion in the administration of criminal law: the declination decision—that is, a prosecutor’s “all-important decision whether to decline or file formal criminal charges after police have made an arrest.”[10] Notably, her concerns lie not with agencies’ potential aggrandizement of authority as a result of engaging in statutory interpretation, but with the exercise of enforcement discretion and its potential impact on the “excesses of the domestic criminal Leviathan.”[11]

Natapoff describes the act of discretion inherent in the declination decision as of “institutional and constitutional” significance.[12] “[F]ar from a mere formalism,” the decision to prosecute “is the starting point of our whole system of adversary criminal justice.”[13] Per Natapoff’s thoughtful conceptualization, weak rates of declination amount to an abdication of the prosecutorial role and even a failure of due process.[14] At the very least, a decision to pursue a misdemeanor case trigger costs for “defendants, public defenders, prosecutors, jails, courts, and taxpayers” that might have otherwise been avoided.[15] Moreover, low declination may indicate that prosecutors are failing to make choices that could mitigate the extent to which criminalization is motivated by bias, flows from shoddy policing, or results in disparate punishment for minorities.[16]

Natapoff does not clarify when low declination rates are necessarily the result of a failure to screen, or whether they may, in fact, indicate a proactive decision to prosecute. But to the extent low declination is indeed an abdication, it might be understood as “an absence of prosecutorial gatekeeping at the outset of the adversarial process,” the result of which is that “police in effect get to decide who will become a defendant simply by arresting them.”[17] Given the importance of the declination decision to limiting the excesses of policing, Natapoff characterizes it as an intrabranch checking measure that “offers decisional friction, oversight, and accountability within the executive at precisely the moment when good law enforcement decision-making makes a big difference for millions of people.”[18] In this way, she successfully applies the internal separation of powers framework, which originated in the administrative law and structural constitutionalism scholarship,[19] to the criminal administration and policing context.

This invited Response both appreciates Natapoff’s enthusiasm for enhancing prosecutorial discretion in criminal administration and maintains a bit of skepticism regarding its efficacy. More specifically, it draws on the immigration context in order to explore the possibilities and hazards of internal administrative checks for constraining excessive policing in both the criminal and immigration environments. First, this Response suggests that directives ensuring uniformity are important to ensuring high-quality prosecutorial discretion. Second, it argues that agency culture is an important player in the internal separation of powers that reduces prosecutors’ overarching potential to constrain law enforcement. Third, it observes that the declination decision can serve a gatekeeping function that limits desirable access to adjudication, particularly for communities with fewer resources and reduced participation in democratic process. Finally, this Response notes the possibilities of institutional design and public oversight for improving law enforcement accountability.

Criminal and immigration enforcement evince key similarities and meaningful differences. As in the criminal legal system, an excess of “policing” in immigration can lead to significant and dire results.[20] Accordingly, prosecutorial discretion may reduce the consequences of brutal law enforcement, including against noncitizens.[21] Nonetheless, publicized examples of immigration prosecutorial discretion differ from declination in criminal administration, in that the former is instigated by top-down orders directing how administrators are supposed to enforce legal punishment, while the latter continues to remain primarily at the sole discretion of each prosecutor.

Indeed, objectives for the enforcement of immigration consequences have shifted at the ground level, ostensibly with some consistency, in response to presidential orders. For instance, the Obama Administration directed the Department of Homeland Security (DHS) to implement a temporary policy known as the Deferred Action for Childhood Arrivals program, identifying people who were “low priority” for Immigration and Customs Enforcement (ICE) in order both to focus limited enforcement resources[22] and to allow noncitizens who know only the United States as their home to stay in the country.[23] In contrast, Natapoff observes, “most declination policies are not blanket refusals to enforce legislation,”[24] but rather case-by-case decisions that are not channeled by enforcement priorities.

But perhaps misdemeanor declination policies should be more uniform, as in the immigration setting. After all, there are “numerous jurisdictions where police have been shown to systemically engage in baseless misdemeanor arrests that lack probable cause.”[25] Using precise guidelines to constrain the scope of prosecution[26] could be a meaningful way to systematize declination’s amelioration of the consequences of poor policing. This could happen through internal guidances issued by prosecutors’ offices and spearheaded by elected district attorneys,[27] which are not dissimilar to policies regarding immigration enforcement priorities issued via political directives.[28] Furthermore, Natapoff notes that state legislatures sometimes limit misdemeanor arrests to certain circumstances, including in one state[29] notoriously plagued by police brutality.[30] One note of caution is that in criminal policing,[31] as in immigration,[32] systematic enforcement priorities may instigate backlash from arresting officers themselves.

If detailed criteria are difficult to enact, the simple fact that the arrests at issue concern misdemeanors should be enough to put prosecutors on alert that there may be good reason to decline to prosecute these charges. This could offset the practice of those who reflexively approve of what Natapoff describes as “obviously invalid arrest decisions” based in police malice.[33] To this end, Natapoff suggests a heightened charging standard to “compensate for the psychological tendency to water down the probable cause standard in misdemeanor cases,”[34] and points to a sample proposal from the American Bar Association “to affirmatively require dismissal [of misdemeanor arrests] if prosecutors ‘reasonably believe[] that proof of guilt beyond reasonable doubt is lacking.’”[35] While a compelling idea, it seems like raising the standard for prosecution in abstraction may not reduce the real-world tendency to apply the standard in a manner that favors prosecution.

Specificity in directives that tangibly reduces the scope of discretion in the declination decision may be more promising.[36] The best approach for reaching the core goals of policing reform may be to put a fine point on limits to prosecution. For instance, if “police are known to arrest people of color in numbers disproportionate to underlying offense rates,” perhaps “those arrests should be presumptively suspect.”[37] The immigration context, too, might benefit from contending directly with “the racism, xenophobia, and Islamophobia that drive . . . enforcement.”[38] In any case, the optimal level of tension between actors in the internal separation of powers requires precisely what Natapoff explores in Misdemeanor Declination: how to both empower and constrain administrative actors in order to achieve a sustainably progressive criminal legal system.

Overall, Natapoff is optimistic about the capacity of individualized acts of prosecutorial discretion, be they guided or not, to restrain the consequences of poor policing, but there is reason to temper this faith. As an initial matter, the idiosyncrasies of declination decisions, including that they require no paperwork;[39] may be made “on the fly”;[40] are admittedly lax in their analysis[41] unless, perhaps, prosecutors are intentional; tend to be sloppy;[42] and can revert to rubber stamping[43] suggest that it would be difficult to induce uniformity or systemic accountability across these decisions in fact, even in the event that they are expected to follow a specific set of requirements. This weakens the potential for these decisions to constitute a dependable intrabranch check.

Furthermore, sometimes prosecutors “affirmatively and intentionally decide to screen very little.”[44] In immigration, decisions regarding the validity of deportation arrests are left to ICE officers themselves[45]—an example of the “egregious” practice of “letting police act as prosecutors on their own arrests.”[46] As Natapoff herself observes, even where DOJ has discretion not to bring criminal charges against noncitizens, “almost all cases brought by ICE are prosecuted by DOJ; the fact that almost none are dismissed later on suggests that the initial charging decisions are intentional and that full enforcement is a matter of substantive immigration policy.”[47]

Natapoff asserts that “[s]uch heavy handed enforcement policies pose substantive and normative problems but not necessarily intrabranch checking ones.”[48] But to the contrary, the lack of functional separation between law enforcement and prosecutors suggests that what appears to be an intrabranch check is merely a “parchment barrier”[49] obscuring the influence of political or other aims in administration that merge the incentives of different administrative actors. One such source of influence could be presidential authority.[50] Another, far less studied, is the agency’s own political and institutional aims.

“Agencies’ mandates and structures are often established by the priorities and political leanings of their enacting legislature, and may be influenced by the President. Nonetheless, these mandates become institutional, baked into the agency after a certain point, and therefore distinct from the agendas established by temporary political leadership.”[51] Furthermore, agency mission is arguably another “power” in the internal separation of powers. On the one hand, it may play a unifying role: an agency “can develop a strong sense of internal culture and character” that “provides it with a rudder that steers its decisions about available policy options.”[52] On the other hand, the weight of an agency’s overarching mission can obscure important ways in which various bureaucrats’ goals and mandates diverge,[53] which then reduces the fruitful friction between intra-agency actors that sustains consistent and reliable intrabranch checks.

Natapoff does not grapple with the fact that the exercise of individualized discretion in criminal law enforcement is impacted by prosecutors’ and law enforcement’s joint emphasis on meeting the government’s aims. As Natapoff notes, there is a heavy emphasis on “getting the bad guys” in criminal administration as whole.[54] Therefore, despite the apparent differences in responsibilities and interests between prosecutors and the police,[55] their kinship, based in the transcending mission of enforcing criminal consequences against the public, seems likely to result in the conflation of their interests.

Incidentally, this dynamic resembles the overlapping priorities of immigration prosecution and adjudication,[56] which may lead to poor outcomes for noncitizens, or at least a regrettable reduction in administrative due process.[57] Likewise, prosecutors and police are “professional allies” that are governed by, are loyal to, and cooperatively support the criminal apparatus,[58] whose well-worn pathways advance the interests of the police. This is illustrated by situations in which “inexperienced prosecutors tend to take everything the police say ‘as holy writ’.”[59] The resulting thumb on the scale in favor of prosecution may be difficult to release through intraagency checking mechanisms, because all of the administrative actors involved in the criminal legal system are rooted in and shaped by related norms and expectations. Complementarily, shared loyalties may also help to explain prosecutors’ unforgivable failure to prosecute the crimes of police themselves, including those that involve the killing of Black people.[60]

In addition, it is worth noting that prosecutorial discretion, while holding promise for constraining law enforcement, may serve a concerning gatekeeping function as well. While Natapoff’s advocacy for nonprosecution is well-taken,[61] one wonders whether criminal administration would benefit from less initial screening and more adjudication in some cases. As Natapoff observes, “prosecutors who uncritically charge all police arrests are postponing the substantive merits decision of whether charges should be filed at all.”[62] In doing so, she seems to critique the possible default of the charging decision to judges, which could lead to decisions additionally biased in favor of the criminal apparatus by these “prosecutors in robes.”[63] However, on the flip side,

[h]igh declination rates might also reflect inappropriate prosecutorial bias or underenforcement. For example, DOJ has been criticized for its high declination rates in criminal civil rights cases on the theory that this particular exercise of prosecutorial discretion impedes the enforcement of civil rights laws against police. Similar arguments have been made regarding excessive corporate crime declinations. Rape cases have infamously high declination rates. In other words, prosecutors can screen too much.[64]

Even under far more justifiable circumstances for high declination rates, prosecutors who decline to charge arrests are using discretion to engage in a substantive decision of sorts regarding the merits of the case, rather than providing access to more fulsome adjudication and its attendant proceduralism and accessibility. More specifically, the discretionary gatekeeping of access to administrative procedure often focuses on preserving institutional resources and improving efficiency.[65] In these contexts, a “declination” decision, such as it is, may bring about subpar process and outcomes, particularly for those without the monetary resources or political power to demand better. For instance, prosecutorial discretion can restrict access to administrative hearings, resulting in decisions not to adjudicate claims in matters of public land use, civil rights, and labor that may hinder the claims of under-resourced people.[66] In both criminal law and immigration enforcement, the gatekeeping of access to procedure can also be shaped by political interests that taint the quality of administrative decisionmaking, especially for marginalized communities.[67] Like any mechanism of intrabranch constraint, the benefits and drawbacks of prosecutorial discretion reveal themselves in different proportion depending on which stakeholders and priorities the agency tends to emphasize.

Finally, any suggestions for criminal reform must contend with the limitations to prosecutorial discretion discussed in this Response. Doing so requires acknowledging a fact generally overlooked by both those who believe the administrative state operates in clear service of good government and those who insist that pathways of democratic accountability ensure better administration: that administrative actors sometimes harm people in the course of meeting institutional priorities.[68] Arguably, fiddling with prosecutors’ individualized exercises of discretion in isolation will not improve their capacity to check the police, particularly when the police wield power over those with the fewest opportunities for participation in democratic process (for instance, racial minorities and noncitizens). The solution requires, if not substantial shifts to the “tough on crime” legislation animating both criminal administration and immigration, then at least intentional shifts to administrative institutional design.[69] Another possible corrective is emphasizing oversight by the public, which is another “branch” of the internal separation of powers,[70] and one that could feasibly hold criminal and immigration enforcement more accountable if offered accessible ways to draw the legislature’s attention to harm caused by the bureaucracy.[71]

Ultimately, Natapoff’s article is of a piece with the literature advocating for prosecutorial discretion to check aggression in immigration enforcement.[72] And even more, her work takes advantage of the fact that progressive objectives are perhaps more acceptable in discussions of criminal procedure in the wake of the racial reckoning of the past five years, while immigration scholars are still expected to emphasize other justifications for their proposed reforms.[73] Nonetheless, like in immigration, the promise of administrative self-constraint in the criminal legal space may be more difficult to realize than Natapoff suggests, particularly given both the institutional values that shape the criminal and immigration bureaucracies and the vulnerability of the populations subject to prosecution and punishment.

 

  1. Henry M. Hart, Jr. & Albert M. Sacks, The Legal Process: Basic Problems in the Making and Application of Law 144 (William N. Eskridge, Jr. & Philip P. Frickey eds., 1994). ↑
  2. See Robert M. Cooper, Administrative Justice and the Role of Discretion, 47 Yale L.J. 577, 582–83 (1938). ↑
  3. See Bijal Shah, A Critical Analysis of Separation-of-Powers Functionalism, 84 Ohio St. L. J. 1007, 1023–29 (2024) (discussing how the major questions doctrine allows the judiciary to usurp administrative policymaking power); Bijal Shah, Statute-Focused Presidential Administration, 90 Geo. Wash. L. Rev. 1165, 1212–14 (2022) (illustrating how the major questions doctrine constrains environmental policy); Bijal Shah, Judicial Administration, 11 U.C. Irvine L. Rev. 1119, 1176–79 (2021) (arguing that both the major questions doctrine and eschewing Chevron are judicial tools limiting administrative policymaking). ↑
  4. See generally, e.g., Anya Bernstein & Cristina Rodríguez, Working with Statutes, 103 Texas L. Rev. (forthcoming 2025). ↑
  5. Cooper, supra note 2, at 582–83. ↑
  6. See generally Bijal Shah, Toward an Intra-Agency Separation of Powers, 92 N.Y.U. L. Rev. 101 (2017) (discussing the concept of the administrative separation of powers): see also id. at 102–4 (mentioning scholars who have advanced this idea, a number of whom advocate for a strong administrative state). ↑
  7. See, e.g., Bijal Shah, Procedural Administrative Discretion (work in progress) [hereinafter Shah, Procedural Discretion]; Bijal Shah, Administrative Subordination, 91 U. Chi. L. Rev. 1603, 1619 (2024) [hereinafter Shah, Administrative Subordination]. ↑
  8. For an explanation of notice-and-comment rulemaking in practice, see Informal Rulemaking, Legal Info. Inst., https://www.law.cornell.edu/wex/informal_rulemaking (last visited Oct. 10, 2024). ↑
  9. Alexandra Natapoff, Misdemeanor Declination: A Theory of Internal Separation of Powers, 102 Texas L. Rev. 937, 939 (2024) (“[T]he executive Leviathan exercises highly discretionary, largely unreviewed power over nearly all aspects of the criminal system.”). ↑
  10. Id. at 937. ↑
  11. Id. at 962. ↑
  12. Id. at 937. ↑
  13. Id. at 1007 (quoting Kirby v. Illinois, 406 U.S. 682, 689 (1972)). ↑
  14. Id. at 989 (“And the abdication of the prosecutorial quasi-judicial screening obligation is a watering-down of due process itself, the formal guarantees built into the adjudication process meant to offer protection to the defendant and legitimacy to the system.”). ↑
  15. Id. at 988. ↑
  16. Id. at 945–46. ↑
  17. Id. at 975 (“Prosecutorial abdication at this stage thus undermines the integrity and logic of the entire executive penal process.”). ↑
  18. Id. at 937. ↑
  19. See Shah, supra note 7, at 102–4 (discussing the literature on this concept). ↑
  20. See Lindsay Nash, The Immigration Subpoena Power, 125 Colum. L. Rev. (forthcoming 2025); Emily R. Chertoff, Violence in the Administrative State, 112 Calif. L. Rev. (forthcoming 2024). ↑
  21. See generally Shoba Sivaprasad Wadhia, Beyond Deportation: The Role of Prosecutorial Discretion in Immigration Cases (2015). ↑
  22. Memorandum from John Morton, Dir., U.S. Immigr. & Customs Enf‘t, to Agency Pers. (June 17, 2011), https://www.ice.gov/doclib/secure-communities/pdf/prosecutorial-discretion-memo.pdf. ↑
  23. Memorandum from Janet Napolitano, Sec’y of Homeland Sec. (June 15, 2012), https://www.dhs.gov/xlibrary/assets/s1-exercising-prosecutorial-discretion-individuals-who-came-to-us-as-children.pdf. ↑
  24. Natapoff, supra note 9, at 963. ↑
  25. Id. at 1003 (“In Baltimore, police routinely arrest young Black men for loitering under circumstances where the men are clearly innocent. In New York, public housing police engaged for years in stop-and-frisk and arrest policies that swept up thousands of people of color who clearly were not trespassing.”) (internal citations omitted). ↑
  26. See, e.g., id. and accompanying text; see also Natapoff, supra note 9, at 1000 (“Kentucky law prohibits police from making a physical arrest in most misdemeanor cases unless there are reasonable grounds to believe that the defendant will not appear in response to a citation. Virginia law requires police to issue a summons and release most of the people they initially detain in connection with low level misdemeanors.”) (internal citations omitted). ↑
  27. Natapoff, supra note 9, at 993. ↑
  28. See supra notes 22–23 and accompanying text. ↑
  29. Id. at 1000. For example, Minnesota Rule of Criminal Procedure 6.01 requires police to issue a citation in all misdemeanor cases “unless it reasonably appears: (1) the person must be detained to prevent bodily injury to that person or another; (2) further criminal conduct will occur; or (3) a substantial likelihood exists that the person will not respond to a citation.” Minn. R. Crim. P. 6.01. ↑
  30. See, e.g., Madison Park, The 62-Second Encounter Between Philando Castile and the Officer Who Killed Him, CNN (May 30, 2017, 12:10 PM), https://www.cnn.com/2017/05/30/us/philando-castile-shooting-officer-trial-timeline/index.html; Ray Sanchez, George Floyd Killing Latest in String of Police Actions to Stoke Public Anger in Minnesota, CNN (May 30, 2020; 3:56 PM), https://www.cnn.com/2020/05/30/us/minnesota-police-actions-public-anger-trnd/index.html. ↑
  31. Natapoff, supra note 9, at 993. ↑
  32. Bijal Shah, Civil Servant Alarm, 94 Chi.-Kent L. Rev. 627, 638–39 (2019). ↑
  33. Natapoff, supra note 9, at 1003. ↑
  34. Id. at 998. ↑
  35. Id. at 1000 (quoting Sanford H. Kadish, Stephen J. Schulhofer & Rachel E. Barkow, Criminal Law and Its Processes 84 & n.28 (11th ed. 2022) (alteration in original)). ↑
  36. See, e.g., id. at 995 (noting a directive from a prosecutor’s office “instruct[ing] prosecutors that a specified list of thirteen misdemeanors ‘shall be declined or dismissed before arraignment and without conditions unless “exceptions” or “factors for consideration” exist’”). ↑
  37. Id. at 998. ↑
  38. Shah, Administrative Subordination, supra note 7, at 1607. ↑
  39. Natapoff, supra note 9, at 942. ↑
  40. Id. at 944. ↑
  41. Id. ↑
  42. Id. at 945. ↑
  43. Id. ↑
  44. Id. at 974. ↑
  45. Lindsay Nash, Inventing Deportation Arrests, 121 Mich. L. Rev. 1301, 1310–12 (2023). ↑
  46. Cf. Natapoff, supra note 9, at 1004. ↑
  47. Id. at 974. ↑
  48. Id. ↑
  49. Gillian E. Metzger, The Interdependent Relationship Between Internal and External Separation of Powers, 59 Emory L.J. 423, 437 (2009) (quoting The Federalist No. 48, at 308 (Madison) (Clinton Rossiter ed., 1961)). ↑
  50. Id. at 437, 441–42; see also Bijal Shah, Deploying the Internal Separation of Powers Against Racial Tyranny, 116 Nw. U. L. Rev. Online 244, 259 (2021) (arguing that the internal separation of powers can check the president most effectively only when the judiciary and legislature empower administrative actors). ↑
  51. Shah, Administrative Subordination, supra note 7, at 1648–49. ↑
  52. Bernstein & Rodríguez, supra note 4, at 51, 53. ↑
  53. Shah, Administrative Subordination, supra note 7, at 1657–58. ↑
  54. Natapoff, supra note 9, at 957 (quotations omitted). ↑
  55. Id. at 974. ↑
  56. Shah, Administrative Subordination, supra note 7, at 1649; see also Rachel E. Barkow, Institutional Design and the Policing of Prosecutors: Lessons from Administrative Law, 61 Stan. L. Rev. 869, 871 (2009). ↑
  57. See Bijal Shah, Uncovering Coordinated Interagency Adjudication, 128 Harv. L. Rev. 805, 838–40 (2015). ↑
  58. Natapoff, supra note 9, at 955 (quoting Kate Levine, Who Shouldn’t Prosecute the Police, 101 Iowa L. Rev. 1447, 1450–51 (2016)). ↑
  59. Id. at 985 (quoting Kay L. Levine & Ronald F. Wright, Prosecutor Risk, Maturation, and Wrongful Conviction Practice, 42 Law. & Soc. Inquiry 648, 658 (2017)). ↑
  60. See id. at 991–92 (“Racial disparities in misdemeanor policing and prosecution deform the entire criminal system.”). ↑
  61. “A first-of-its kind study on the long-term effects of declination in Boston found that nonprosecution of nonviolent misdemeanor offenses led to large reductions in recidivism with no negative effect on public safety.” Natapoff, supra note 9, at 977. ↑
  62. Id. at 967. ↑
  63. Cf. Jacob Schuman, Prosecutors in Robes, 77 Stan. L. Rev. (forthcoming 2025). ↑
  64. Natapoff, supra note 9, at 974 (internal citations omitted). ↑
  65. Id. at 974, 983. ↑
  66. Shah, Procedural Discretion, supra note 7, at 22–25. ↑
  67. Id.; see also Bijal Shah, The President’s Fourth Branch?, 92 Fordham L. Rev. 499, 524–32 (2023). ↑
  68. See generally Shah, Administrative Subordination, supra note 7. ↑
  69. See, e.g., id. at 1686–92 (offering prescriptions for institutional reform to improve bureaucrats’ ability to meet their goals despite the impact of agency culture). ↑
  70. See supra note 7 and accompanying text. ↑
  71. Cf. Shah, supra note 42, at 655–57. ↑
  72. See generally, e.g., Wadhia, supra note 29. ↑
  73. See, e.g., Nash, supra note 28 (manuscript at 46-56) (emphasizing federalism, privacy, and free speech as reasons to limit the power of immigration officials to issue intrusive subpoenas).↑

Download PDF