Auditing executive discretion

Notre Dame Law Review

Volumn 82, Issue 1, 2006, Pages 227-311

  Cuéllar, Mariano Florentino ^a,b  

^a STANFORD LAW SCHOOL   (United States)

^b STANFORD UNIVERSITY   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 33846645025     PISSN: 07453515     EISSN: None     Source Type: Journal    
DOI: None     Document Type: Review

References (364)

1. For insightful discussions of how concerns about discretion led to the enactment of much of modern administrative law, see McNollgast, The Political Origins of the Administrative Procedure Act, 15 J.L. ECON. & ORG. 180, 195-201 (1999);
2. Peter L. Strauss, Changing Times: The APA at Fifty, 63 U. CHI. L. REV. 1389, 1399-1400 (1996).
3. See, e.g., Citizens to Pres. Overton Park, Inc. v. Volpe, 401 U.S. 402, 411 (1971) (subjecting an informal, discretionary decision of the Secretary of Transportation to judicial review on the basis of statutory language prohibiting federal aid for highways through public parks unless "no feasible and prudent alternative" existed). Overton Park set the stage for a substantial expansion in the availability (and stringency) of judicial review governing informal, discretionary decisions.
4. Id. at 410. But review remains either unavailable or fairly cursory for a massive range of discretionary decisions involving national security, foreign policy, immigration, domestic regulatory enforcement, public benefits, and investigation or prosecution.
5. See KENNETH CULP DAVIS, DISCRETIONARY JUSTICE 151-55 (1969).
6. Regarding the trope that judicial review should have an exalted role in constraining arbitrary bureaucratic action, see Rachel E. Barkow, More Supreme Than Court? The Fall of the Political Question Doctrine and the Rise of Judicial Supremacy, 102 COLUM. L. REV. 237, 317-19 (2002);
7. David Cole, Judging the Next Emergency: Judicial Review and Individual Rights in Times of Crisis, 101 MICH. L. REV. 2565, 2576-77 (2003);
8. Lisa Schultz Bressman, Beyond Accountability: Arbitrariness and Legitimacy in the Administrative State, 78 N.Y.U. L. REV. 461, 473-74 (2003);
9. James E. Pfander, Article I Tribunals, Article III Courts, and the Judicial Power of the United States, 118 HARV. L. REV. 643, 653 (2004).
10. See infra Part I.B.
11. See, e.g., City of Chi. v. Int'l Coll. of Surgeons, 522 U.S. 156, 184 (1997) (summarizing previous interpretation of federal jurisdictional statute by emphasizing the Court's conclusion that "Congress meant to hold federal agencies accountable by making their actions subject to judicial review");
12. Rebecca L. Brown, Accountability, Liberty, and the Constitution, 98 COLUM. L. REV. 531, 535 (1998) (emphasizing the alleged role of judicial review in promoting accountability);
13. Jerry L. Mashaw, Reinventing Government and Regulatory Reform: Studies in the Neglect and Abuse of Administrative Law, 57 U. PITT. L. REV. 405, 414 (1996) ("The long-term commitment of American administrative law has been to assure that administrative discretion is structured, checked, and balanced. Administrative efficacy must be weighed against demands for liberty and legality, as well as political accountability."). Understandably, judicial review is considered important both because of its role in generating information about, and promoting the accountability of, executive authorities, and because of its direct remedial role.
14. Regarding "accountability" see, for example, Michael Asimow. The Scope of Judicial Review of Decisions of California Administrative Agencies, 42 UCLA L. REV. 1157, 1239 (1995);
15. Jeffrey J. Rachlinski & Cynthia R. Farina, Cognitive Psychology and Optimal Government Design, 87 CORNELL L. REV. 549, 587-89 (2002);
16. Mark Seidenfeld, Cognitive Loafing, Social Conformity, and Judicial Review of Agency Rulemaking, 87 CORNELL L. REV, 486, 509-12 (2002);
17. William Christian, Note, Normalization as a Goal: The Americans With Disabilities Act and Individuals With Mental Retardation, 73 TEX. L. REV. 409, 417 (1994).
18. The importance of providing a hearing and a remedy for government actions affecting a protected interest, meanwhile, is firmly lodged at the core of modern procedural due process doctrine. See Mathews v. Eldridge, 424 U.S. 319, 333 (1976).
19. Voices on one side of the debate emphatically insist on greater opportunities for highly stringent judicial review of executive branch actions. See, e.g., DAVIS, supra note 2, at 216 ("The vast quantities of unnecessary discretionary power that have grown up in our system should be cut back, and the discretionary power that is found to be necessary should be properly confined, structured, and checked.");
20. Cole, supra note 2, at 2567;
21. Nicole Nice-Petersen, Note, Justice for the "Designated": The Process That is Due to Alleged U.S. Financiers of Terrorism, 93 GEO. L.J. 1387, 1420 (2005) ("The courts should not hesitate to act solely because those stripped of their rights are accused terrorists."). Similarly emphatic voices take the position in equipoise.
22. See, e.g., Eric A. Posner & Adrian Vermeule, Accommodating Emergencies, 56 STAN. L. REV. 605, 644 (2003) ("Judicial scrutiny can only interfere with forceful executive action.");
23. Ruth Wedgwood, Al Qaeda, Terrorism, and Military Commissions, 96 AM. J. INT'L L. 328, 332-35 (2002).
24. Similar debates play out in the context of constitutional torts. See, e.g., James J. Park, The Constitutional Tort Action as Individual Remedy, 38 HARV. C.R.-C.L. L. REV. 393, 395 (2003). At least some of the debate turns on differing views about the extent to which a larger "political process" promotes "accountability." I discuss this infra Part III.
25. Although this Article does not directly address judicial review's role as provider of individual remedies, the argument developed here is nonetheless relevant to the provision of remedies by either courts or political actors. See infra Part II.C (discussing the relevance of the argument to court review of individual cases); Part II.D (discussing the implications of the argument for how political actors deliver remedies to aggrieved individuals or groups).
26. The focus here is primarily on the type of routine executive discretion, such as that vested in a prosecutorial authority, to impose costs on discrete individuals or groups with minimal judicial intervention. Cf. Heckler v. Chaney, 470 U.S. 821, 837-38 (1985) (holding that in the absence of a specific statutory requirement to the contrary, regulatory agency's decision not to exercise authority in a particular context where such authority could be exercised is committed to agency discretion). For a discussion of routine executive discretion, see infra notes 29-34 and accompanying text.
27. Cf. Cass R. Sunstein, Administrative Law Goes to War, 118 HARV. L. REV. 2663, 2672 (2005) ("In war no less than in peace, the inquiry into presidential authority can be organized and disciplined if it is undertaken with close reference to standard principles of administrative law.").
28. See infra text accompanying notes 105-07.
29. For a discussion of the proverbial "guarding the guardians" problem, see infra notes 211-13 and accompanying text.
30. Compare Webster v. Doe, 486 U.S. 592, 602-04 (1988) (finding CIA director's power to fire employees on national security grounds committed by law to agency discretion), with Klamath-Siskiyou Wildlands Ctr. v. Bureau of Land Mgmt., 387 F.3d 989, 997 (9th Cir. 2004) (finding, under an arbitrary and capricious standard that the court understood to require "hard look" review, that the Bureau of Land Management's environmental assessments of two timber sales, conducted pursuant to the National Environmental Policy Act, were inadequate because they failed to consider the cumulative impact of the sales). I do not mean to minimize the subtleties of the variegated constitutional, statutory, and prudential doctrines on which courts (and even legislatures) draw when they decide on how much discretion to grant. Separation of powers, deference to national security and foreign policy decisions, judicial deference to expert determinations of government agencies, and statutory interpretation techniques all figure in this process. Even the two cases cited represent extraordinarily different contexts, and the kinds of discretion involved in the decisions are also different. The point is that nearly any plausible applications of such doctrines require (or, at the very least, allow) some consequentialist balancing of the costs and benefits associated with discretion, and different ways of striking that balance are associated with distinct degrees of stringency in the courts' review of some executive decision. As these two cases show, courts indeed strike different balances when applying these doctrines, and in the process, they set different degrees of stringency for the review of discretionary executive decisions.
31. See, e.g., Brief for the Respondents at 11-12, Hamdi v. Rumsfeld, 542 U.S. 507 (2004) (No. 03-6696), 2004 WL 724020 (arguing that further factual development of the circumstances surrounding an alleged enemy combatant's designation as such "would divert the military's attention from the ongoing conflict in Afghanistan");
32. Reply Brief for the Petitioner at 4, Heckler v. Chaney, 470 U.S. 821 (1985) (No. 83-1878), 1984 WL 566059 ("[R]espondents' submission, if accepted, would allow anyone to seek judicial review of the agency's decision not to bring enforcement proceedings under any portion of the Act." (emphasis added)).
33. Moralistic intuitions about the importance of horizontal equity combine with the content of legal doctrines such as stare decisis to complicate the possibility of using differing degrees of stringency to review cases in the same class. See infra Part I.B.
34. See infra text accompanying notes 109-10.
35. See infra Part I.D.1.
36. See infra Part I.A.
37. Of course, the mere creation of some auditing system does not automatically solve organizational learning and accountability problems. As Part II explains, a great deal depends on details of institutional design. The impact of an audit system also depends on the public's response, and the institutional dynamics affecting that response. Audit systems can be counterproductive if they merely provide a false sense of security-which is in some sense precisely my criticism of judicial review in many of the contexts I discuss in this Article. Nonetheless, the status quo seems even more likely to provide precisely that false sense of security because it lacks many of the potential advantages that a carefully-structured audit system could generate. For a thoughtful discussion of the role of audits and the pitfalls in designing them, see generally MICHAEL POWER, THE AUDIT SOCIETY: RITUALS OF VERIFICATION 123 (1999) ("The audit society is a society that endangers itself because it invests too heavily in shallow rituals of verification at the expense of other forms of organizational intelligence.").
38. See infra Part I.D.1 (discussing the potential for dynamic interaction between audits and judicial review).
39. As Part I explains, this means that the audit bureaucracies generally appear not to: (1) take random samples of decisions in a legal domain involving decisions directly affecting individuals or groups, (2) assess decisions in accordance with a defensible pre-defined standard, and (3) release the results publicly.
40. Among the nuances distinguishing different types of discretion, one might recognize a distinction between the actual, de jure amount of flexibility the law permits (as in the examples of routine executive discretion), and the de facto amount of discretion that results from the fact that review of a given decision is limited or almost entirely absent (as may occur in the context of police searches involving situations that do not give rise to criminal proceedings and therefore are not policed through the exclusionary rule). Although the primary focus of the argument is on the former type of discretion, the present analysis also has implications for the latter-because expanding the range of available review mechanisms may change policymakers' and legal interpreters' willingness to impose review that could change the extent of de facto discretion.
41. See, e.g., Abbott Labs. v. Gardner, 387 U.S. 136, 141 (1967) (relying on presumption of reviewability in federal administrative law).
42. For examples from the states, see State Dep't of Fish & Game v. Meyer, 906 P.2d 1365, 1370 (Alaska 1995) (acknowledging that final agency action is presumed reviewable);
43. Minn. Pub. Interest Research Group v. Minn. Envtl. Quality Council, 237 N.W.2d 375, 379 (Minn. 1975) (recognizing presumption of reviewability for state administrative decisions);
44. Pisano v. Shillinger, 835 P.2d 1136, 1138-39 (Wyo. 1992) (presuming reviewability of state administrative action).
45. For a discussion of how the institutional design of judicial review in the European Union provides for the review of administrative action, see generally Jurgen Schwarze, Judicial Review in EC Law-Some Reflections on the Origins and the Actual Legal Situation, 51 INT'L & COMP. L.Q. 17, 22-27 (2002).
46. 401 U.S. 402 (1971).
47. Id. at 410.
48. See, e.g., Marbury v. Madison, 5 U.S. (1 Cranch) 137, 170 (1803). As I acknowledge below, there may indeed be some contexts where discretion does (and ought to) mean no review-from courts, auditors, or anyone else. My suggestion is simply that we ought to proceed with great caution in considering the rationales for that kind of discretion. In particular, we ought to recognize that the rationales for limiting or even barring judicial review in some contexts do not necessarily speak to the question of whether we should also bar an alternative form of review. Put differently, perhaps one might envision an array of review mechanisms to address the fact that the two kinds of discretion are on a continuum, and at intermediate places on the continuum we might do well to use audits (or some other system of review) instead of leaving the executive unrestrained.
49. 470 U.S. 821 (1985).
50. Id. at 831.
51. Id. at 834-35. The analogy bolstered the Court's case for limiting judicial interference in the Chaney context (involving the Food and Drug Administration) because the perception was already so deeply rooted among courts that judicial regulation of prosecutorial discretion would unduly burden the administration of justice.
52. Cf. Bordenkircher v. Hayes, 434 U.S. 357, 364 (1978) ("In our system, so long as the prosecutor has probable cause to believe that the accused committed an offense defined by statute, the decision whether or not to prosecute, and what charge to file or bring before a grand jury, generally rests entirely in his discretion.").
53. See, e.g., Bressman, supra note 2, at 473-74 (noting that judicial review is effective as the main protection against abuse of executive discretion).
54. See infra Part I.D.2 for a discussion of those risks. Although cases like Chaney may seem less troubling to some because they involve a government decision to restrain from acting, such a decision still counts as action under the Administrative Procedure Act,
55. see 5 U.S.C. § 551 (13) (2000), and is potentially no less coercive to individuals whose desired outcomes depend on government action. Part II, infra, discusses how the practical problems of reviewing executive decisions not to act may be solved through alternatives to traditional judicial review.
56. 356 U.S. 309 (1958).
57. Id. at 310.
58. Id. at 317 (citing New York v. United States, 331 U.S. 284, 335 (1947)).
59. See, e.g., Brief for the Respondents, supra note 12, at 25. The government's language in the brief is typical of the positions that lawyers for the executive branch have taken in this Administration-and not dramatically different (on the core issue of deference)-from that taken by lawyers for other presidential administrations. It states: As this Court has observed, "courts traditionally have been reluctant to intrude upon the authority of the Executive in military and national security affairs." The customary deference that courts afford the Executive in matters of military affairs is especially warranted in this context. A commander's wartime determination that an individual is an enemy combatant is a quintessentially military judgment, representing a core exercise of the Commander-in-Chief authority.... Especially in the course of hostilities, the military through its operations and intelligence-gathering has an unmatched vantage point from which to learn about the enemy and make judgments as to whether those seized during a conflict are friend or foe.
60. Id. (citations omitted) (quoting Dep't of the Navy v. Egan, 484 U.S. 518, 530 (1988)).
61. This statement should not obscure the massive extent of variation among legal systems, many of which assign quite different roles to judicial institutions. The point is that it is quite common for those different systems to assign considerable importance to the goal of reviewing executive action through courts. See generally Anne-Marie Slaughter, A Typology of Transjudicial Communication, 29 U. RICH. L. REV. 99, 101 (1994) (recognizing differences in the "form, function, and degree of reciprocal engagement" among legal systems).
62. Actually measuring the precise impact of review with some analytical clarity is enormously complex, but a number of scholars have made convincing arguments to this effect using qualitative or quantitative methodologies in different contexts. For some cogent examples, see JERRY L. MASHAW & DAVID HARFST, THE STRUGGLE FOR AUTO SAFETY 147-71 (1990) (suggesting that the National Highway Traffic Safety Administration's reliance on costly recalls of questionable safety effects rather than prospective rulemaking has in part been driven by the impact of intrusive judicial review in rulemaking);
63. Brandice Canes-Wrone, Bureaucratic Decisions and the Composition of the Lower Courts, 47 AM. J. POL. SCI. 205, 212-13 (2003) (analyzing whether changes in the ideological composition of lower courts affected decisions of the U.S. Army Corps of Engineers to grant permits for development of wetlands, and finding that a standard deviation increase in estimated pro-environmental ideology of the lower courts decreased the probability that the Corps would grant a permit by fourteen percent);
64. Thomas O. McGarity, The Courts and the Ossification of Rulemaking: A Response to Professor Seidenfeld, 75 TEX. L. REV. 525, 528 (1997) (discussing the impact of the "ossification" of rulemaking, where judicial review, among other factors, shapes agencies' willingness to use regulatory authority).
65. Moreover, there is no good reason to expect that courts or legislators routinely strike the proper balance (under almost any defensible definition of "proper") when they police discretion outside the context of separation of powers and the political question doctrine. My claim is this: at a minimum, Chaney-type discretion directly affecting individuals and justified on consequentialist grounds concentrates great power in executive authorities and raises questions about the appropriate extent of bureaucratic flexibility that should be permitted.
66. The invocation of such doctrines could also be scrutinized and subjected to criticism, but such an inquiry is beyond the scope of the present Article.
67. Heckler v. Chaney, 470 U.S. 821, 823-28 (1985).
68. See, e.g., Ronald M. Levin, Understanding Unreviewability in Administrative Law, 74 MINN. L. REV. 689, 692 (1990);
69. Cass R. Sunstein, Reviewing Agency Inaction After Heckler v. Chaney, 52 U. CHI. L. REV. 653, 654 (1985).
70. For an insightful pre-Chaney perspective on reviewability, see generally Harvey Saferstein, Nonreviewability: A Functional Analysis of "Committed to Agency Discretion," 82 HARV. L. REV. 367 (1968).
71. See Chaney, 470 U.S. at 831-32 ("[A]n agency decision not to enforce often involves a complicated balancing of a number of factors which are peculiarly within its expertise.... An agency generally cannot act against each technical violation of the statute it is charged with enforcing. The agency is far better equipped than the courts to deal with the many variables involved in the proper ordering of its priorities."). From the Court's perspective, matters would have been different if legislators explicitly crafted statutes to require such review.
72. See id. at 832-33 ("[T] he decision is only presumptively unreviewable; the presumption may be rebutted where the substantive statute has provided guidelines for the agency to follow in exercising its enforcement powers."). Absent some exceptional circumstances, they would be loath to do so.
73. Cf. Bowen v. Mich. Acad. of Family Physicians, 476 U.S. 667, 675 (1986) (bolstering a conclusion that Congress had deliberately precluded judicial review of the amount of Medicare Part B claims in order to "'avoid overloading the courts' with 'trivial matters,' a consequence which would 'unduly ta[x]' the federal court system with 'little real value' to be received by participants in the program" (quoting United States v. Erika, Inc., 456 U.S. 201, 210 n.13 (1982))).
74. See Chaney, 470 U.S. at 831 ("This Court has recognized on several occasions over many years that an agency's decision not to prosecute or enforce...is a decision generally committed to an agency's absolute discretion.");
75. id. at 834 (concluding that the court of appeals erred in finding inapplicable the "principle of absolute prosecutorial discretion").
76. See, e.g., United States v. Armstrong, 517 U.S. 456, 464 (1996);
77. United States v. Nixon, 418 U.S. 683, 693 (1974);
78. Vaca v. Sipes, 386 U.S. 171, 182 (1967).
79. For an insightful doctrinal review, see Lara Beth Sheer, Prosecutorial Discretion, 86 GEO. L.J. 1353 (1998).
80. Armstrong, 517 U.S. at 464. As Davis cogently observed, "[a] judicial trial is an acceptance of a prosecutor's decision to prosecute, not a review of it. Even a quick finding of not guilty may leave untouched the harms that flow from the prosecution."
81. DAVIS, supra note 2, at 209 n.21.
82. 50 U.S.C.A. §§ 403-404a(e) (West Supp. 2006).
83. Webster v. Doe, 486 U.S. 592, 602-04 (1988).
84. A variety of practical problems for the claimant, beginning with discovery and exacerbated by courts' tendency to defer to executive branch national security decisions, make it doubtful this avenue for review would exert much of an effect on the behavior of the agency's leadership.
85. See, e.g., Peter M. Shane, Interbranch Accountability in State Government and the Constitutional Requirement of Judicial Independence, LAW & CONTEMP. PROBS., Summer 1998, at 21, 25;
86. sources cited supra note 2.
87. See Steven G. Calabresi & Saikrishna B. Prakash, The President's Power to Execute the Laws, 104 YALE L.J. 541, 570-99 (1994);
88. Saikrishna B. Prakash & Michael D. Ramsey, The Executive Power Over Foreign Affairs, 111 YALE L.J. 231, 295-355 (2001);
89. John Yoo, War, Responsibility, and the Age of Terrorism, 57 STAN. L. REV. 793, 810 (2004).
90. 5 U.S.C. §§ 551-559, 701-706 (2000 & Supp. IV 2004).
91. See Hamdi v. Rumsfeld, 542 U.S. 507, 532 (2004) (plurality opinion).
92. Why this state of affairs may be problematic for dominant legislative coalitions without prompting them to fix it is discussed in Part IV.
93. 50 U.S.C. §§ 1701-1706 (2000).
94. See generally James J. Savage, Executive Use of the International Emergency Economic Powers Act-Evolution Through the Terrorist and Taliban Sanctions, CURRENTS: INT'L TRADE L.J., Winter 2001, at 28, 28-31 (describing the history and effect of IEEPA).
95. 50 U.S.C. § 1701 (a).
96. See Global Relief Found., Inc. v. O'Neill, 315 F.3d 748, 752-53 (7th Cir. 2002).
97. See 50 U.S.C. § 1701 (a).
98. Id.
99. Id.
100. See generally Jason Luong, Note, Forcing Constraint: The Case for Amending the International Emergency Economic Powers Act, 78 TEX. L. REV. 1181, 1197 (2000) ("Congress has consistently acquiesced to the economic regulations enacted by the president under the IEEPA because of inadequate legislative drafting, lack of political will, and popular support for the most common of the IEEPA regulations-economic sanctions and embargoes.").
101. Holy Land Found, for Relief & Dev. v. Ashcroft, 333 F.3d 156, 160 (D.C. Cir. 2003).
102. 50 U.S.C.A. § 1705(a) (West Supp. July 2006).
103. Id. § 1705.
104. 18 U.S.C. § 2339 (2000 & Supp. III 2003).
105. 333 F.3d 156.
106. Id. at 159-60.
107. Id. at 162.
108. Id.
109. Id.
110. Id. at 161.
111. Id. (quoting Holy Land Found, for Relief & Dev. v. Ashcroft, 219 F. Supp. 2d 57, 69 (D.D.C. 2002)).
112. Id. at 161-63.
113. Id. at 163 (applying these standards in reviewing the actions and record at hand).
114. Id. at 162.
115. Ironically enough, the Holy Land appeals panel suggested that the government's position was strengthened by the fact that "[t]here was no plausible evidence presented which showed that [ties to Hamas] had been severed." Id.
116. Manipulation of the record, moreover, need not be conscious or explicit. A number of pressures could affect the considerable number of investigators, analysts, spies, lawyers, and higher level officials whose work influences the record that the court reviews. As long as they feel at least some subtle pressure to support the conclusion that a designation should be made, they may fail to consider countervailing arguments, or the potential consequences of an "erroneous" designation (i.e., erroneous in the sense of not complying with the statute, the President's executive order, the "arbitrary and capricious" standard, or the executive branch's stated goals for using the IEEPA emergency powers).
117. See 29 U.S.C. § 666(a)-(c) (2000).
118. 42 U.S.C. §§ 7401-7671q (2000 & Supp. III 2003).
119. Id. § 7413(a)(5).
120. See id. § 7413(b)-(c).
121. For a detailed discussion, see generally Jason D. Nichols, Towards Reviving the Efficacy of Administrative Compliance Orders: Balancing Due Process Concerns and the Need for Enforcement Flexibility in Environmental Law, 57 ADMIN. L. REV. 193, 197-99 (2005) (discussing the EPA's use of ACOs to enforce the Clean Air Act).
122. See 42 U.S.C. § 7413(b)(2).
123. See Tenn. Valley Auth. v. Whitman, 336 F.3d 1236, 1258 (11th Cir. 2003).
124. See Alaska Dep't of Envtl. Conservation v. EPA. 540 U.S. 461, 496 (2004);
125. Alaska Dep't of Envtl. Conservation v. EPA, 298 F.3d 814, 822 (9th Cir. 2002), aff'd 540 U.S. 461.
126. 42 U.S.C. § 7413(a)(5).
127. See Justin Florence, Note, Making the No Fly List Fly: A Due Process Model for Terrorist Watchlists, 115 YALE L.J. 2148, 2157-59 (2006).
128. See generally Gillian E. Metzger, Privatization As Delegation, 103 COLUM. L. REV. 1367, 1400-10 (2003) (arguing that privatization would lessen judicial review);
129. Steven L. Schooner, Contractor Atrocities at Abu Ghraib: Compromised Accountability in a Streamlined, Outsourced Government, 16 STAN. L. & POL'Y REV. 549, 562-64 (2005) (discussing the dangers of contractors who receive no guidance or management).
130. See Hamdi v. Rumsfeld, 542 U.S. 507, 518-19 (2004) (plurality opinion).
131. See Jody Freeman, The Contracting State, 28 FLA. ST. U. L. REV. 155, 165 (2000) (noting that, despite the "highly technocratic approach to contract design" prevalent in federal procurement law, the existing framework is "too limited to address the much more substantial issues that arise" in some contracts).
132. See cases cited supra note 11.
133. Indeed, casual observers may be forgiven for assuming (heroically) that courts (or legislatures, when they directly impose limits on review) are balancing the marginal costs and benefits of greater stringency of review. See infra Part I.D.2 for a discussion.
134. See, e.g., SECTION OF ADMIN. LAW AND REGULATORY PRACTICE, AM. BAR ASS'N, A BLACKLETTER STATEMENT OF FEDERAL ADMINISTRATIVE LAW 34 (2004). The Blacketter Statement provides a cogent and revealing synthesis: The court may set aside an agency action as an abuse of discretion... on any of several grounds. In practice, application of these grounds varies according to the nature and magnitude of the agency action. Thus, a court will typically apply the criteria rigorously during judicial review of high-stakes rulemaking proceedings (a practice commonly termed "hard look" review), but much mort leniently when reviewing a routine, uncomplicated action.
135. Id. (emphasis added).
136. See Hamdi v. Rumsfeld, 316 F.3d 450, 473 (4th Cir. 2003), vacated, 542 U.S. 507 (2004).
137. See Hamdi, 542 U.S. at 509 (plurality opinion).
138. For some examples from different regulatory contexts, see Bowen v. Mich. Acad. of Family Physicians, 476 U.S. 667, 675 (1986) (discussing legislative reluctance to tax judicial resources by allowing routine review of Medicare Part B claims);
139. Bush v. Lucas, 462 U.S. 367, 388 (1983) (noting, in the context of reviewing an alleged retaliatory demotion and defamation claim against a federal official, that "[t]he costs associated with the review of disciplinary decisions are already significant-not only in monetary terms, but also in the time and energy of managerial personnel who must defend their decisions");
140. Langevin v. Chenango Court, Inc., 447 F.2d 296, 303 (2d Cir. 1971) (finding that judicial review of the Federal Housing Agency's discretionary actions resulting in rent increase approvals in part on the basis that an unacceptably high number of rent increases would be subject to review). Here and in similar cases, the court's discussion of costs assumes (either implicitly or explicitly) that whatever costs are generated by the stringency of review the court adopts in the present case will be applied to future cases with similar characteristics.
141. See supra note 12 for examples of briefs making this argument.
142. See Cole, supra note 2, at 2567-77.
143. The reference to "traditional judicial review" implies that executive decisions can be reviewed through judicial fora, where judges tend to believe that court decisions governing how stringent the review of discretionary executive decisions is will affect every future case (or nearly so) in a particular class, and where litigants seek, and the court can deliver, some kind of relief, such as vacating a particular government action or providing an injunctive remedy.
144. Cf. CASS R. SUNSTEIN, AFTER THE RIGHTS REVOLUTION 219 (1990) ("[A]n unduly aggressive judicial posture may increase delays and paperwork in a way that threatens implementation.").
145. Robust ombuds systems, such as those used in Scandinavian countries, may represent another alternative. Nonetheless, if an ombuds system functioned as it is traditionally understood in being driven by public complaints, it would have some of the same strengths and weaknesses that the litigation process would, and would thereby provide a somewhat distorted picture of bureaucratic activity.
146. Each of these features interacts to give the proposal developed here its unique characteristics. (1) Random selection assures a representative picture of decisions and provides a mixed strategy approach to deterrence that is difficult to evade. See infra Part II.A. (2) Fixing some standard in advance reduces risk-normalization dynamics described by scholars of bureaucracy, making it more likely that problems will be detected by audit bureaucracies that are not immune from the organizational and political problems affecting the agencies they are auditing. See infra note 136 and accompanying text (discussing Vaughan's work on the "normalization of risk"). (3) Public announcement (including provisions for protecting classified information) allows audits to play a role in a larger system of accountability defined by representative politics and public reactions. See infra Part II.D.
147. Hence, the audits of executive discretion discussed here differ from the more informal program evaluation audits often undertaken by the GAO and Inspectors General. The incentives of federal audit bureaucracies to perform different types of audits are discussed infra Part IV.
148. James G. March et al., Learning From Samples of One or Fewer, 2 ORG. SCI. 1, 2 (1991).
149. See generally JOEL SLEMROD & JON BAKIJA, TAXING OURSELVES 180-82 (3d ed. 2004) (discussing how the IRS performs audits).
150. This last point is at the core of the explanation for why a "mixed strategy" is so valuable in the framework of game theory. See
151. SLEMROD & BAKIJIA, supra note 100, at 174.
152. Id. at 175.
153. Id. ("[T]he estimates are based on data that is now over fifteen years old. But these are the best numbers around.").
154. See Chaves County Home Health Serv., Inc. v. Sullivan, 931 F.2d 914. 915 (D.C. Cir. 1991).
155. Id. at 916.
156. See, e.g., McComber v. Travelers Prop. & Cas., 37 Conn. L. Rptr. 349, 350-51 (Conn. Super. Ct. 2004).
157. Cf. ALBERT O. HIRSCHMAN, EXIT, VOICE, AND LOYALTY 24-25 (1970) (discussing the risks to organizations of obtaining performance signals by simply waiting for market-driven "exit" responses among customers, employees, or other constituencies).
158. See J. David Cummins & Sharon Tennyson, Controlling Automobile Insurance Costs, J. ECON. PERSP., Spring 1992, at 95, 110-11.
159. For a more general discussion of the value of, and incentives for use of, audits in the private sector, see Ross L. Watts & Jerold L. Zimmerman, Agency Problems, Auditing, and the Theory of the Firm: Some Evidence, 26 J.L. & ECON. 613, 626-33 (1983).
160. See KREPS, supra note 101, at 763-64.
161. 406 U.S. 311 (1972).
162. Id. at 312.
163. Id. at 316.
164. See Exec. Order No. 12,564, 3 C.F.R. 224 (1987), reprinted in 5 U.S.C. § 7301 note (2000) (subjecting numerous categories of federal employees to random drug tests): Editorial, Guilty Until Proven Innocent, N.Y. TIMES, Jan. 20, 2006, at A16 (describing how the IRS Office of the Taxpayer Advocate audited a random sample of 500 tax returns where refunds had been frozen and taxpayers complained, and found that sixty-six percent of taxpayers deserved a full refund and another fourteen percent deserved a partial refund).
165. See John T. Rourke, The GAO: An Evolving Role, 38 PUB. ADMIN. REV. 453, 453-54 (1978).
166. Id. at 455.
167. See PAUL LIGHT, MONITORING: GOVERNMENT 42 (1993).
168. See infra Part III.
169. The proposal is, therefore, somewhat reminiscent of one that Mashaw offered for social security benefit determinations in the concluding pages of his exhaustive study of the Social Security Administration. See JERRY L. MASHAW, BUREAUCRATIC JUSTICE 226 (1983).
170. For instance, in the criminal context, the auditor's determination of a standard would be shaped by statements of prosecutors regarding the purposes of the prosecutions. See infra notes 204-05 and accompanying text.
171. Cf. United States v. Isa, 923 F.2d 1300, 1306 (8th Cir. 1991) (using in camera review of information obtained by federal agents through wiretaps authorized under the Foreign Intelligence Surveillance Act and denying aggrieved party's motion for suppression);
172. JOHN HART ELY, WAR AND RESPONSIBILITY 105-14 (1993) (discussing how fear of leaking is overblown).
173. See infra Part III.A.
174. Depending on the assumptions made about the political system, legislators and the public might respond to the audits in ways that would provide relief to all or some of the people aggrieved by problematic applications of targeted executive discretion.
175. See Water Keeper Alliance v. U.S. Dep't of Def., 271 F.3d 21, 30-35 (1st Cir. 2001).
176. See CORNELIUS KERWIN, RULEMAKING 174 (3d ed. 2003);
177. see also McGarity, supra note 35, at 21-23 (discussing the effect of judicial review on agencies' willingness to use regulatory authority).
178. See supra Part I.C.
179. See Perez v. FBI, 71 F.3d 513, 514 n.6 (5th Cir. 1995).
180. But see N. Jersey Media Group, Inc. v. Ashcroft, 308 F.3d 198, 219-20 (3d Cir. 2002).
181. See INS v. Lopez-Mendoza, 468 U.S. 1032, 1048 (1984);
182. Vt. Yankee Nuclear Power Corp. v. Natural Res. Def. Council, Inc., 435 U.S. 519, 545-48 (1978).
183. Cf. KEITH KREHBIEL, INFORMATION AND LEGISLATIVE ORGANIZATION 265 (1991) (arguing that "informational concerns-in the sense of how politicians can be provided with incentives to study public problems and formulate public policy-are at the heart of legislative organization").
184. See Greenbaum v. EPA, 370 F.3d 527, 533-34 (6th Cir. 2004).
185. See, e.g., Chevron v. Natural Res. Def. Council, Inc., 467 U.S. 837, 865 (1984) ("While agencies are not directly accountable to the people, the Chief Executive is ....");
186. Motor Vehicle Mfrs. Ass'n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 59 (1983) (Rehnquist, J., concurring in part and dissenting in part).
187. Newman v. United States, 382 F.2d 479, 482 (D.C. Cir. 1967).
188. See, e.g., MASHAW & HARFST, supra note 35, at 111-23 (using the example of NHTSA to show that agencies react to judicial review).
189. See ARTHUR LUPIA & MATHEW MCCUBBINS, THE DEMOCRATIC DILEMMA 101-48 (1998).
190. This is certainly true in the case of people who work in offices whose broad performance is reviewed by the Inspectors General or the GAO. See infra Part III.A.
191. DIANE VAUGHAN, THE CHALLENGER LAUNCH DECISION 409-10 (1996) (emphasis added).
192. See March et al., supra note 99, at 2-3;
193. Michal Tamuz, Learning Disabilities for Regulators: The Perils of Organizational Learning in the Air Transportation Industry, 33 ADMIN. & SOC'Y 276, 295-99 (2001).
194. See MASHAW & HARFST, supra note 35, at 111-23, for a discussion of how a regulatory agency (in that case, NHTSA) learned to use alternative policymaking strategies to avoid the costs associated with judicial review. The example serves to emphasize two points, both of which are relevant to the present discussion: (1) that agencies appear to react to judicial review, and (2) that organizational learning is not necessarily associated with learning to achieve the most valuable organizational goals better. Learning can be a bad thing; no doubt that organizations making large numbers of discretionary decisions that rarely if ever get reviewed (or, if reviewed, rarely get scrutinized carefully) probably learn that they can shift resources, time, attention, quality control, and strict adherence to legal or aspirational goals away from those decisions and towards other pursuits. The question is how to encourage the most desirable kinds of learning.
195. See Mariano-Florentino Cuéllar, The Mismatch Between State Power and State Capacity in Transnational Law Enforcement, 22 BERKELEY J. INT'L L. 15, 44-48 (2004).
196. See id. at 49-53.
197. See THE ABU GHRAIB INVESTIGATIONS 1-20 (Steven Strasser ed., 2004).
198. See Hart Seely, The Poetry of D.H. Rumsfeld, SLATE, Apr. 2, 2003, http://www.slate.com/id/2081042.
199. See, e.g., LUPIA & MCCUBBINS, supra note 134, at 17-38 (1998) (emphasizing citizens' limited knowledge about politics);
200. Richard R. Lau & David P. Redlawsk, Voling Correctly, 91 AM. POL. SCI. REV. 585, 587-93 (1997) (using survey data to construct a statistical model estimating the proportion of voters that vote "correctly" given their attitudes and impressions, and finding that one-in-four voters voted incorrectly in the five American presidential elections between 1972 and 1988).
201. Cf. Mariano-Florentino Cuéllar, Rethinking Regulatory Democracy, 57 ADMIN. L. REV. 411, 412-16 (2005) (finding that comments made by citizens during the notice and comment phase of a new administrative regulation were less likely to be incorporated into the final regulation than comments made by administrative attorneys because the citizens' comments lacked technical sophistication even though they offered constructive insights relevant to the agencies' legal mandates).
202. See Kenneth E. Scott & Barry R. Weingast, Banking Reform: Economic Propellants, Political Impediments, in REFORMING FINANCIAL INSTITUTIONS AND MARKETS IN THE UNITED STATES 19, 27-36 (George G. Kaufman ed., 1994).
203. See Barry R. Weingast, Caught in the Middle: The President, Congress, and the. Political-Bureaucratic System, in THE EXECUTED BRANCH 312, 338 (Joel D. Aberbach & Mark A. Peterson eds., 2005) ("Bureaucracy is a pejorative term in modern America, and rightly so.... The political compromises necessitated by the American constitutional system mean that legislation rarely attacks problems in a straightforward manner, but typically through political compromise that combines multiple and conflicting goals ....").
204. This assumes, plausibly, that for most domains the gains from greater stringency in review of the limited sample of cases outweighs cursory review of all cases in a particular class (which is already available, in most cases, with existing deferential forms of judicial review). Even if judicial review is entirely precluded, stringent review of a small sample may prove vastly more informative than dividing a scarce review budget among an entire pool of potential cases. For instance, a non-linear function mapping stringency to a given probability of discovering problems may doom deferential review to uncover nothing; meanwhile, a small percentage sample of a larger population may reveal almost as much as review of the entire population. See infra Part II.B for an elaboration of the last point.
205. In any event, some observers suggest the political question doctrine itself is becoming increasingly irrelevant. For an insightful discussion of the doctrine's evolution, its relative decline, and the extent to which its theoretical bases are compelling, see Barkow, supra note 2. As Part I.B notes, no similar growth in judicial assertiveness is discernible in judicial supervision of more routine instances of executive discretion.
206. See infra notes 190-91 and accompanying text (emphasizing the potential value, in social research design, of using scarce resources to generate more information about individual cases rather than increasing sample size).
207. See supra text accompanying notes 105-07 (discussing instances of sample adjudication).
208. See supra note 92 (referencing cases discussing the costs of review).
209. See supra note 12 (referencing federal government arguments discussing the dangers associated with consuming executive branch resources during the course of the review process).
210. Cf. PETER H. SCHUCK, SUING GOVERNMENT 59-81 (1983) (analyzing the over-deterrence problem in the context of damage actions against the government).
211. Court opinions often include references to the efficiency of the particular degree of stringency they have chosen (or the legislature has) for a given class of cases. See supra note 92. But courts may fail at conducting the analysis. They may be concerned about workload pressures to a degree that overwhelms more nuanced evaluation of marginal costs and benefits. Cf. Theodore Eisenberg & Stewart Schwab, The Reality of Constitutional Tort Litigation, 72 CORNELL L. REV. 641, 693-95 (1987) (discussing the disjuncture between judicial workload concerns and the actual extent of litigation involving constitutional torts). They may fail to analyze the full range of benefits and costs associated with more (or less) stringent review.
212. Cf. Cass R. Sunstein & Adrian Vermeule, Interpretation and Institutions, 101 MICH. L. REV. 885, 890 (2003) (arguing that theories of interpretation that ignore institutional capacities and the dynamic effects of any particular approach to interpretation are inadequate). In any case, the argument for audits developed above is strengthened when it turns out that courts and legislatures fail at balancing marginal costs and benefits when fixing the level of stringency for a particular class of cases.
213. For a discussion of sample size, see infra Part II.B.
214. For a discussion of what standard a reviewer would apply in an environment where discretion is often defined precisely by the absence of standards, see infra Part II.C.
215. This property also generates its own costs, and litigants may face a variety of standing, redressability, and remedy-related hurdles if they choose to litigate their cases. But for simplicity, these details need not be addressed here.
216. See infra Part II.B for details. Sampling theory provides one important rationale for randomization-as opposed to an intuitive but potentially misleading focus on "problem" cases. Although audits can be adapted to focus on reviewing a subset of decisions that seem from surface indicators to be especially likely to be problematic, such an approach dilutes what can be learned about the entire pool of cases. To the extent the relevant bureaucracies and associated organizational leaders know the basis for selection, it also makes it easier to evade review. See supra note 101 and accompanying text (discussing mixed strategies).
217. And suppose the right stringency-problem detection curve were present-making the new resources critical for each case reviewed. Such a scenario would increase the desirability of litigation, further increasing the costs associated with traditional review.
218. There may be both overbreadth and under-inclusiveness problems. For an early discussion of selection effects, see George L. Priest & Benjamin Klein, The Selection of Disputes for Litigation, 13 J. LEGAL STUD. 1, 1-6 (1984).
219. For an illuminating discussion of how different institutional rules can shape selection effects, see Peter H. Huang, Lawsuit Abandonment Options in Possibly Frivolous Litigation Games, 23 REV. LITIG. 47, 78-85 (2004).
220. See William H. Simon, Rights and Redistribution in the Welfare System, 38 STAN. L. REV. 1431, 1467-68 (1986).
221. If determined advocates of executive branch discretion choose to treat the potential loss of public trust arising from audits as the most important cost to be borne, then the very fact that audits might be effective in changing public perceptions about executive discretion might allegedly still make them too costly. This argument may not always be entirely disingenuous, but it is a harder one to defend than one rooted in a host of costs involving bureaucratic attention, resources, frivolous claims, and similarly tangible costs. In any case, the specificity gained in discussing costs may be at least a partial reward from contemplating audits.
222. See, e.g., VAUGHAN, supra note 136, at 63 ("[T]He consequence of professional training and experience is itself a particularistic world-view comprising certain assumptions, expectations, and experiences that become integrated with the person's sense of the world.");
223. see also CHARLES PERROW, NORMAL ACCIDENTS (1984) ("A warning...is only effective if it fits into our mental model of what is going on. As with the 'warnings' of Pearl Harbor, it can get swamped by the multitude of signals that fit our expectations, and thus be discounted as 'noise' in the system.");
224. SCOTT D. SAGAN, THE LIMITS OF SAFETY (1995) (noting that adding redundant safety devices does not always increase safety because human operators can often work around these redundancies if the devices do not serve immediate interests).
225. And obviously, whatever the benefits of specialization, they need to be balanced against the risks.
226. See generally Terry M. Moe, The New Economics of Organization, 28 AM. J. POL. SCI. 739 (1984) (arguing that theories of hierarchy developed in the field of organizational economics can contribute much to the understanding of political control over bureaucracies).
227. See Weingast, supra note 146, at 318-31.
228. For examples of work analyzing the implications of divergent preferences on specific issues among members of broad governing coalitions, see David P. Baron, A Noncooperative Theory of Legislative Coalitions, 33 AM. J. POL. SCI. 1048 (1989);
229. Michael Laver & Kenneth A. Shepsle, Government Coalitions and Intraparty Politics, 20 BRIT. J. POL. SCI. 489 (1990).
230. For an analysis of when coalition governments have especially pronounced incentives to monitor the bureaucratic implementation of policies at the ministerial level, see Micahel F. Thies, Keeping Tabs on Partners: The Logic of Delegation in Coalition Governments, 45 AM. J. POL. SCI. 580 (2001).
231. See Bruce Bimber, information as a Factor in Congressional Politics, 16 LEGIS. STUD. Q. 585, 596 (1991).
232. Indeed, although the incidence of legislative proposals drawing unanimous or near-unanimous support varies, such proposals are not rare. For a discussion of empirical evidence demonstrating the recurring importance of universal or near-universal legislative coalitions, see Melissa P. Collie, The Legislature and Distributive Policy Making in Formal Perspective, 13 LEGIS. STUD. Q. 427, 445-49 (1988).
233. For an illuminating account of how common such lapses are, see SAGAN, supra note 165, at 251.
234. See supra Part I.D.2 for a discussion of the risks of bureaucratic malfeasance and related shortcomings.
235. See Jonathan Bendor & Terry M. Moe, An Adaptive Model of Bureaucratic Politics, 79 AM. POL. SCI. REV. 755, 761 (1985) (discussing variations in the extent of politicians' knowledge of bureaucratic performance).
236. Audits' chilling effect on agency activities found desirable by a subset of the political coalition would arise as a result of the information they might reveal to pivotal legislators or organized interests, who could in turn shape the perceptions of mass political audiences.
237. Cf. William Bernhard, A Political Explanation of Variations in Central Bank Independence, 92 AM. POL. SCI. REV. 311, 312-15 (1998) (discussing how legislators may create an independent central bank as a commitment device not only in spite of, but because of, politicians' differing incentives on monetary policy);
238. Matthew Potoski, Managing Uncertainty Through Bureaucratic Design: Administrative Procedures and State Air Pollution Control Agencies, 9 J. PUB. ADMIN. RES. & THEORY 623, 626 (1999) (discussing the information-generating benefits of specific procedural mechanisms and institutional designs for state legislators with differing political goals);
239. Douglas H. Shumavon, Policy Impact of the 1974 Congressional Budget Act, 41 PUB. ADMIN. REV. 339, 346-47 (1981) (discussing legislators' incentive for the creation of the Congressional Budget Office on the basis of the value of unbiased information to politicians with differing political goals).
240. See Rui J.P. de Figueiredo, Jr., Electoral Competition, Political Uncertainty, and Policy Insulation, 96 AM. POL. SCI. REV. 321, 362-28 (2002) (demonstrating the extent to which, from legislators' perspective, the desirability of mechanisms to constrain the executive branch depends on beliefs about future control of the executive branch).
241. The existence of an audit program-run by an external private-sector company-associated with the Treasury Department's anti-terrorism wire-transfer screening program involving the SWIFT system may serve as an example. See Greg Miller & Josh Meyer, Officials Defend Bank Data Tracking, L.A. TIMES, June 24, 2006, at A1.
242. Some politicians opposed to the program would still prefer that, if the program is to operate, audits be included in its institutional design. See, e.g., id. But see infra Part IV.A for a discussion of why some opponents of the wire-transfer screening program might have strategic reasons to oppose audits.
243. See infra Part IV.C.
244. See KREHBIEL, supra note 129, at 264-65;
245. see also Thomas W. Gilligan & Keith Krehbiel, Organization of Informative Committees by a Rational Legislature, 34 AM. J. POL. SCI. 531, 546 (1990) (noting that committees are superfluous if they possess no expertise with which to reduce uncertainty regarding the effects of legislation that they sponsor).
246. DAVIS, supra note 2, at 216.
247. Cf. Ronald J. Krotoszynski, Jr., Taming the Tail That Wags the Dog: Ex Post and Ex Ante Constraints on Informal Adjudication, 56 ADMIN. L. REV. 1057, 1069 (2004) ("Taken to its logical extreme, a decision to turn on the lights or lower an agency's office building temperature two degrees constitutes 'informal adjudication.'").
248. Routine executive discretion involving the targeted application of standards to individuals or groups bears some resemblance to the concept of "informal adjudication" long discussed by administrative law scholars. For a brief discussion of the definitions associated with "informal adjudication" and some of the doctrinal problems implicit in this category of administrative action, see id. at 1069-75. Obtaining review of such decisions under the arbitrary and capricious standard, or on the basis of some constitutional theory, is unlikely to provide adequate oversight for these decisions because of the specter of the review costs problem. But one should nonetheless beware of the term "informal adjudication" as a description of the full scope of discretionary decisions that merit new review mechanisms. Many discretionary decisions are neither informal in the sense that they are bereft of some alleged procedural safeguards (for example, review of administrative compliance orders), and in the case of many others, the term "adjudication" is but a euphemistic concession to an almost entirely unrealized aspiration.
249. See Brief for the Respondents, supra note 12, at 11-12.
250. Moreover, such judgments are more likely to garner media and legislative attention even in the absence of audits. And existing investigations performed by audit bureaucracies and congressional committees are more likely to analyze broad policy judgments rather than routine instances of executive discretion applying standards to individuals or groups. See infra Part III. Hence, when resource constraints force a choice between using audits to screen instances of routine discretion involving applications of law to the facts and broader policy judgments, the former probably merits greater attention. More generally, the arguments for deferring to the executive branch when an agency writes a rule or a president signs an executive order seem weaker when the executive branch claims to be applying a given standard to the facts. In the latter case the implicit claim is: "we may have to apply some judgment, but when we detain someone as an enemy combatant, freeze assets, or inspect an industrial plant, there's no question about the purpose we are serving. We're enforcing the law."
251. 50 U.S.C. §§ 1701-1706 (2000).
252. See Doherty v. U.S. Dep't of Justice, 775 F.2d 49, 52 (2d Cir. 1985).
253. For a skeptical view of such contentions, see VAUGHAN, supra note 136, at 389-90.
254. Cf. Marbury v. Madison, 5 U.S. (1 Cranch) 137, 170 (1803) (explaining that although the Court could not review decisions of the executive branch that involved political discretion, it could enforce the rights of individuals who had been harmed by executive actions that did not involve political discretion).
255. The contention may be that, in addition to being free of any sort of formal review, certain national security decisions-like a presidential choice to bomb a potentially threatening target in the Sudan-should not conform to any standard at all. This position seems to confound the question of whether we should avoid setting a standard because it is too difficult to monitor, or whether in an ideal world the president should never rely on a standard at all. The former is easier to justify than the latter.
256. Small samples are not a fatal problem when Bayesian techniques are applied and the analysis is accompanied by appropriate assumptions. See, e.g., M. Elisabeth Paté-Cornell, Organizational Aspects of Engineering System Safety: The Case of Offshore Platforms, 250 SCIENCE 1210, 1212 (1990). Regarding policy discretion: although policy discretion arguably raises different problems, it may also be worth scrutinizing, case-by-case, the basis for application of an overtly stated policy; for example, how specific permissible exposure limits get set when courts allow rules that fix multiple-which they now discourage, though audits might be a promising alternative to simply banning such rules-instances of information dissemination and use of regulatory agencies (as an alternative to the potentially cumbersome data quality act rules).
257. See SAMPLE SIZE CALCULATION IN CLINICAL RESEARCH 28-29 (Shein-Chung Chow et al. eds., 2003).
258. For example, if a sample of 1000 cases is drawn randomly from an infinitely large population where one-half of the population hold a characteristic, the researcher has a 95% chance (or better) of obtaining a result that is plus or minus 3.1 percentage points of the actual distribution in a population. If a sample of 5000 is drawn under the same conditions, the 95% confidence interval would be plus or minus 1.4 percentage points. And if a sample of 10,000 is drawn, the 95% confidence interval would be plus or minus 1 percentage point. See ALAN D. MONROE, ESSENTIALS OF POLITICAL RESEARCH 69 (2000). While increasing the sample size has diminishing marginal returns in large populations, things are more complicated with smaller sample sizes.
259. See CAROLE SUTTON & MATTHEW DAVID, SOCIAL RESEARCH 154 (2004).
260. An exhaustive analysis of twenty-five cases of Guantanamo detainees, for example, constitutes about 5% of the total population. But in orthodox statistical terms it yields far less reliable information than a sample of 1% of the total population of 100,000. Cf. MONROE, supra note 191, at 69.
261. Even in such situations, the concept of statistical confidence ought to be treated as a species of metaphor, and applied with some nuance, since the auditor's role here is not to estimate a precise quantitative parameter but to conduct a reasoned analysis of the decision in accordance with a defensible standard. For an interesting effort to address the problem of tailoring qualitative concepts of reliability in the analogous context of property value appraisals, see Nathan Berg, A Simple Bayesian Procedure for Sample Size Determination in an Audit of Property Value Appraisals, REAL. EST. ECON., Spring 2006, at 133.
262. Cf. SUTTON & DAVID, supra note 191, at 154 (noting that even a sample as small as thirty can yield useful statistical information). Moreover, it may be possible to deal with domains involving extremely low numerosity by creating a reliable system to randomize the probability that a decision will be reviewed at all (or, to put it differently, by lumping together several classes of related, low-numerosity decisions and then auditing some proportion of them).
263. Cf. March et al., supra note 99, at 2 (noting that sometimes "[o]rganizations attempt such pooling" of cases to create a larger sample from which to assess performance).
264. For a somewhat stylized argument that the Supreme Court (or, more generally, appellate courts) already engage in something akin to audits, see Matt Spitzer & Eric Talley, Judicial Auditing, 29 J. LEGAL STUD. 649 (2000). It should be noted, however, that Spitzer and Talley are using the concept of auditing differently in that context, to describe a court's decision to review a particular case at a high level of stringency rather than a decision to select a subset of executive branch discretionary decisions at random for more stringent review.
265. See id. at 652.
266. They would have to do this in a way that avoids running afoul of the Supreme Court's decision in Vermont Yankee Nuclear Power Corp. v. Natural Resources Defense Council, Inc., 435 U.S. 519 (1978).
267. Id. at 557-58.
268. Audits would not deprive courts of alternative means of adjusting stringency of review even within a particular class of cases. Spitzer and Talley, for example, discuss how reviewing courts could combine their beliefs about a lower court and observable characteristics of a lower court's decision to determine whether to review that decision (or to review it more stringently). Spitzer & Talley, supra note 195, at 667 (indicating how appellate courts, functioning as "auditors" of lower courts, might adopt "an auditing strategy that can be either even-handed or asymmetric, depending on whether the dominant reason for reviewing is (respectively) imprecision or ideology"). The authors recognize, however, that such an approach represents a departure from the principle that like cases should be treated alike by courts, and acknowledge that external observers expect courts to adhere to that principle.
269. See id. at 675 ("One of the basic principles of Anglo-American jurisprudence is that like cases should be treated alike.").
270. See John F. Duffy, The Death of the Independent Regulatory Commission (and the Birth of a New Independence?), 54-57 (June 9, 2006) (on file with author).
271. Cf. SUNSTEIN, supra note 95, at 45 ("In some circumstances, [government] interference may produce significant benefits for trivial costs.... [R]egulatory programs are necessary to solve collective action and coordination problems.").
272. Under existing law, judges sometimes handle sensitive information and use it to make decisions in cases. See United States v. Nixon, 418 U.S. 683, 715-16 (1974). It is hard to see why audits of targeted discretion should be avoided on the grounds that the underlying discretionary decision depends on sensitive information. If the argument is that it is dangerous or problematic to share the results of audits of targeted discretion because the policy domain requires complete secrecy even of the quality of decisions being made, then that argument should be advanced on its own merits and it should have to overcome a high barrier.
273. See, e.g., Ex parte Quirin, 317 U.S. 1, 24-25 (1942).
274. Even in the prosecution of alleged Al Qaeda terrorist Zacharias Moussaoui, the federal government provided Moussaoui's defense team with the opportunity to review classified information in the context of the criminal discovery process. See A. John Radsan, The Moussaoui Case: The Mess From Minnesota, 31 WM. MITCHELL L. REV. 1417, 1433 (2005).
275. See, e.g., Jifry v. FAA, 370 F.3d 1174 (D.C. Cir. 2004).
276. See NAT'L COMM'N ON TERRORIST ATTACKS UPON THE U.S., THE 9/11 COMMISSION REPORT (2004).
277. Deriving such implicit standards is exactly what the Supreme Court has discouraged in cases such as Herkler v. Chaney, 470 U.S. 821 (1985).
278. There the Court held that agency decisions not to use their enforcement powers are almost always "committed to agency discretion" under the terms of the Administrative Procedure Act, 5 U.S.C. § 701 (a) (2) (2000).
279. Chaney, 470 U.S. at 837-38. The exception is where a statute provides "clear guidelines" that a court could use as a standard against which to judge agency decisions. Indeed, when no such standard is apparent on the face of the statute, then courts tend to find that the absence of such a standard overcomes what is otherwise a presumption of reviewability.
280. See Lincoln v. Vigil, 508 U.S. 182, 192-94 (1993). What these cases sometimes gloss over (but seem to recognize far more explicitly in cases involving the non-delegation doctrine) is the extent to which the existence of standards is on a continuum, where virtually any government action or inaction (including decisions not to prosecute) could be evaluated in accordance with some defensible criterion.
281. For an exception, see Adams v. Richardson, 480 F.2d 1159, 1162 (D.C. Cir. 1973) (finding an exception where an agency has "consciously and expressly adopted a general policy" that is so extreme as to amount to an abdication of its statutory responsibility). This is not to suggest that courts should be the primary actors in conducting targeted discretion audits (perhaps some independent auditing authority including a mixture of judges and non-judges would have more flexibility to articulate standards). It is, rather, to point out that courts' reluctance to articulate standards when they find them missing on the face of a statute should not be taken as an indication that such an enterprise is fruitless.
282. For example, if statutes say that the CIA Director can fire someone for being a national security risk and let him define what that means, then the audit could review the definition with particular care to see if it is plausible and announce the results, or it could rely on agency definitions of "national security" in related contexts. Cf. Chaney, 470 U.S. at 836 (suggesting that an agency decision not to enforce could be challenged if the agency itself has committed to act in specified circumstances).
283. See Asher Arian & Sigalit Olzaeker, Political and Economic Interactions with National Security Opinion: The Gulf War Period in Israel, 43 J. CONFLICT RESOL. 58, 65-71 (1999);
284. David O. Sears et al., Political System Support and Public Response to the Energy Crisis, 22 AM. J. POL. SCI. 56, 66-75 (1978).
285. See infra notes 225-28 and accompanying text.
286. Although I do not develop the argument here, one might imagine a situation where the designers of the audit system would tradeoff the ability to grant relief in exchange for the political and economic resources to audit more cases or to do so more intensely. Truth commissions, for example, reflect an equilibrium where supporters have likely traded off explicit punitive power in exchange for political and economic resources.
287. See generally LAWRENCE FRIEDMAN, AMERICAN LAW IN THE 20TH CENTURY (2002) (noting that important legal decisions in various contexts, but particularly those implicating race relations, did not produce corresponding social conformity therewith);
288. Barry Weingast, The Political Foundations of Democracy and the Rule of Law, 91 AM. POL. SCI. REV. 245 (1997) (employing game theory to find that democratic stability depends on self-enforcing equilibrium).
289. See JAMES HAMILTON, ALL THE NEWS THAT'S FIT TO SELL 15-17 (2004) (discussing the economics of the media and their incentives against resource-intensive forms of news generation).
290. See FRIEDMAN, supra note 209, at 280-90 (discussing the role of civil rights court cases in generating information that can galvanize political and journalistic responses to public officials' decisions).
291. 5 U.S.C. §§ 551-557 (2000).
292. See, e.g., Christina E. Wells, "National Security" Information and the Freedom of Information Act, 56 ADMIN. L. REV. 1195 (2004);
293. Amy E. Rees, Note, Recent Developments Regarding the Freedom of Information Act: A "Prologue to a Farce or a Tragedy; or, Perhaps Both," 44 DUKE L.J. 1183 (1995).
294. 31 U.S.C. § 712(1) (2000).
295. See Inspector General Act of 1978, Pub. L. No. 95-452, 92 Stat. 1101 (codified as amended in scattered sections of 5 U.S.C.).
296. See, e.g., LIGHT, supra note 117;
297. ROGER R. TRASK, DEFENDER OF THE PUBLIC INTEREST: THE GENERAL ACCOUNTING OFFICE, 1921-1966 (1996).
298. See Mariano-Florentino Cuéllar, GAO and IG Report Pilot Study: A Preliminary Analysis of the Methodologies Used by the GAO and Inspectors General to "Audit" Federal Agencies (Apr. 1, 2006) (on file with author).
299. The analysis employs a stratified sampling design focusing on these agencies because some of the goals of the larger empirical project depend on closely studying the interactions between the audit bureaucracies and a small number of specific agencies with differing functions. Id. At 1-2.
300. Id. at 1.
301. For a rare example of an Inspector General report that comes close to achieving the goal of systematically analyzing executive discretion, see OFFICE OF THE INSPECTOR GEN., U.S. DEP'T OF JUSTICE, SUPPLEMENTAL REPORT ON SEPTEMBER 11 DETAINEES' ALLEGATIONS OF ABUSE AT THE METROPOLITAN DETENTION CENTER IN BROOKLYN, NEW YORK (2003).
302. Although this report focuses primarily on describing the treatment of immigration detainees at two government facilities, it did undertake a thorough review of the files describing the status and treatment of the 762 aliens held in connection with the September 11 investigation. Id. at 1.
303. While the report does not leverage the insight that fewer cases reviewed more thoroughly can sometimes yield greater information, its searing analysis proves so informative in part (in all likelihood) because the investigators conceived of their task in terms of generating information, rather than in terms of resolving particular cases. Id. at 5-7. The vast majority of other reports surveyed involve either the use of interviews and non-random analysis of documents to make broad policy recommendations, or traditional financial audits. Of course, in some sense the GAO and Inspector General reports that monitor any aspect of government performance are examining the application of executive discretion because they are examining how agencies perform in some (ordinarily quite general) aspect of a mission they are lawfully empowered to undertake. While all of these studies expand the scope of information (and in some cases generate considerable media attention), there is a striking gap in the methodology and subject matter coverage of these reports substantially overlapping with the audits of executive discretion recommended here. Specifically, what the reports almost never encompass are systematic audits of repetitive, discrete applications of executive power affecting individuals or groups using sampling.
304. Cuéllar, supra note 217, at 21.
305. Four trained research assistants coded the GAO and IG reports. The stratified random sample was drawn from among reports completed between March 1999 and March 2005, and focused on the following five departments (or their components): the Department of Homeland Security, the Department of Justice, the Environmental Protection Agency, the Labor Department, and the Department of Defense. Id. at 2.
306. Of the sample of 400 reports analyzed, about sixty-eight (or seventeen percent) used some form of random sampling. Id. at 21.
307. The bulk of these did so in the context of a traditional financial audit. Only seven reports in the total sample undertook audits of executive discretion. Id.
308. These covered matters such as the proportion of background checks conducted by the Transportation Security Agency on prospective airport screeners that met statutory standards and the extent of FEMA payouts for certain disaster relief programs. A separate comparison sample of fifty reports randomly selected from among those dealing with departments not in the stratified sample (covering similar dates) yielded similar results. Id.
309. U.S. GEN. ACCOUNTING OFFICE, GAO-04-385, REPORT TO CONGRESSIONAL COMMITTEES, AVIATION SECURITY 14-17 (2004).
310. Id.
311. For a study of Inspectors General revealing that only about ten percent of their audits sought to measure program results as opposed to financial, procurement, or management performance, see Kathryn E. Newcomer, Opportunities and Incentives for Improving Program Quality: Auditing and Evaluation, 54 PUB. ADMIN. REV. 147, 152 (1994).
312. Moreover, the audit bureaucracies have incentives to distort their portfolio of activities towards a focus on procurement, contracting, grants, and financial audits and away from auditing executive discretion (particularly in the absence of congressional pressure to do the latter), relative to what is likely to be most socially desirable. The former activities are more likely to produce unambiguous financial results that can more easily support arguments for preserving and expanding agency resources. The latter may exacerbate the political animosity between the audit bureaucracies and ordinary executive agencies. See infra Part IV. Accordingly, changing the existing distribution of activity among the audit bureaucracies to favor audits of executive discretion is likely to yield a more defensible distribution of activity from a social welfare perspective.
313. A search ran on October 26, 2006 on Westlaw's New York Times database for the time period of January 2002 to January 2005 using terms "'Inspector General' or 'GAO' or 'General Accounting Office' or 'Government Accountability Office'" resulted in 1354 documents. The same search ran in Westlaw's United States Broadcast Transcripts database resulted in 528 documents.
314. See supra note 225.
315. Id.
316. For example, the Department of Energy is mentioned only half as often (230 times) during the same period in broadcast transcripts; the Department of Transportation is mentioned only in 181 transcripts. Many other cabinet agencies are mentioned even less often.
317. The agency (quite plausibly) assumes that if a recommendation is not implemented within four years of being made, it probably will not be. Telephone Interview with U.S. Gen. Accounting Office official, in Wash., D.C. (Jan. 6, 2005).
318. See JAMES Q. WILSON, BUREAUCRACY 235-44 (1989).
319. See Mathew McCubbins & Thomas Schwartz, Congressional Oversight Overlooked: Police Patrols Versus Fire Alarms, 28 AM. J. POL. SCI. 165, 166-68 (1984).
320. Id. at 166.
321. Id.
322. Id.
323. 5 U.S.C. §§ 551-559, 701-706 (2000 & Supp. IV 2004).
324. Id. § 551-557.
325. See, e.g., LIGHT, supra note 117, at 42.
326. Obviously there are still other players who matter. There may be moderate legislators, for example, who have an interest in restraining abuses but fundamentally accept the argument that intrusive judicial review may be too costly. I suspect both that these actors could play a crucial and constructive role in promoting bureaucratic changes that could contribute to improvements in how we oversee targeted discretion. I also imagine that their numbers have been thinned by the increasing polarization of politics described by some recent work among political scientists. See Nolan McCarty et al., Political Polarization and Income Inequality 5-11 (Jan. 27, 2003) (on file with author).
327. See supra Part I.D.1.
328. Sometimes things are not equal, as when leaders in the executive branch seem to avoid responsibility for politically controversial functions holding little prospect of generating a public reward, such as choosing a site for a nuclear waste dump.
329. Unless, of course, their support of executive branch power is overwhelmed by incentives to support (for example) legislative power.
330. For a discussion of the possibility that executive authorities could limit their discretion as a means of sending a costly signal of their competence to the public, see Cuéllar, supra note 139, at 57-58.
331. The legislators favoring the investigative provisions limit the risk that the authority will be used in an embarrassing manner. The legislators opposed obtain a partial victory valued by supporting constituencies and a means of generating information that could change the course of future political debates about the relevant legal provisions. See generally Bendor & Moe, supra note 173.
332. See Michael Sandler, Proposed Changes to the Anti-Terrorism Law, CONG., Q. WKLY. REP., Dec. 26, 2005, at 3394.
333. Of course, audits may in fact embarrass the executive branch. That depends on what the opponents think the audits will reveal. But whatever political benefits audits can provide must be weighed against the opportunity costs that I discuss above.
334. Hamdi v Rumsfeld, 542 U.S. 507 (2004).
335. Because these observations raise valid concerns, the approach described here is designed to work in tandem with, rather than to entirely supplement, judicial review. See supra note 164.
336. See Daryl J. Levinson, Rights Essentialism and Remedial Equilibration, 99 COLUM. L. REV. 857, 926 (1999).
337. See supra Part I.C.
338. 435 U.S. 519 (1978).
339. Cf. Bressman, supra note 2, at 555 (laying out a new theory for "examining more directly the concern for arbitrariness").
340. Vermont Yankee, 435 U.S. at 524 (holding that courts are generally not free to impose additional procedural requirements beyond those established by the APA).
341. 424 U.S. 319, 323-28 (1976).
342. See id. at 334-35 (establishing a three-part test focusing on increasing accuracy in administrative decisionmaking). Indeed, the frequent relative absence of an empirical basis informing courts' determinations of what process is due in Eldridge-like cases raises the question of how courts are currently making these determinations, and how these might be best evaluated.
343. LIGHT, supra note 117, at 42 (quoting GEN. ACCOUNTING OFFICE, FINANCIAL AUDITS IN FEDERAL EXECUTIVE BRANCH AGENCIES (1978)).
344. Id. at 43-45.
345. See, e.g., Seven Years of GPRA: Has the Results Act Provided Results? 106th Cong. (2000) (testimony of the Hon. Pete Sessions), available at 2000 WL 1211285 ("GAO has identified waste, fraud and abuse in government programs in amount of over $800 Billion in the last 5 years.").
346. See, e.g., Anne Margaret Joseph, Political Appointees and Auditors of Politics 8 (May 2002) (unpublished Ph.D. thesis, Harvard University) (on file with author).
347. See LIGHT, supra note 117, at 44-45 (noting how some "member[s] of Congress... use... IG input to frame issues for resolution. However, some members of Congress have been less concerned with policy reform, using the IGs instead to score short-term political successes.").
348. Id. at 84-85 (providing a table summarizing the backgrounds of IGs during the Carter and Reagan adminisrations).
349. See, e.g., Joseph, supra note 258, at 146 ("Auditors-whether in economic or political institutions-choose what to observe and report in order to advance their objectives.").
350. Id. at 179 (statutory duties);
351. id. 181-82 (describing the areas of work of the GAO);
352. id. 209 (describing the high portion of reports requested by Congress in recent decades).
353. Cf. DANIEL CARPENTER, THE FORGING OF BUREAUCRATIC AUTONOMY 353 (2001) (noting that "[b]ureacratic autonomy is politically forged").
354. The auditor's capacity to undertake audits in a manner that embodies some principled standard therefore depends on two critical factors. The first is whether the auditor manages to create a sufficient degree of autonomy from the ordinary political pressures that might keep the entity from being purely a creature of the legislature (or the executive). As Carpenter notes, this tends to turn on the capacity of the agency to forge ties to policymaking elites that support its mission and favorable impressions among the mass public (which suggests that, in a sense, it is these segments of the public guarding the proverbial guardians). Id. at 353-54.
355. The second is whether, given such autonomy, the agency's leaders and staff sufficiently value their information-generating mission to expend the necessary effort to generate material information. Cf. Mariano-Florentino Cuéllar, Refugee Security and the Organizational Logic of Legal Mandates, 38 GEO. J. INT'L. L. (forthcoming Nov. 2006).
356. The example of independent electoral and redistricting commissions in other countries is instructive. As Pildes notes, a number of democracies, including the United Kingdom, Australia, Canada, and New Zealand, use nonpartisan commissions to draw political boundaries, where "politicians typically have no role." See Richard Pildes, The Constitutionalization of Democratic Politics, 118 HARV. L. REV. 28, 78 n.211 (2004).
357. Given that these commissions exist in the context of a democratic political system, political actors may face electoral sanctions offsetting the advantages of interfering with these commissions. See also Christopher Elmendorf, Representation Reinforcement Through Advisory Commissions: The Case of Election Law, 80 N.Y.U. L. REV. 1366, 1432 (2005) ("The AC [Electoral Advisory Commission] that wants to enact a reform that most legislators disfavor must establish a base of public support for its plan."). In a similar sense, the formal legal autonomy of an agency such as the Federal Reserve Board is best understood (in the absence of compelling theory and data to the contrary) as a reflection rather than a cause of political circumstances (sometimes critically shaped by the agency itself) promoting an agency's autonomy.
358. See CARPENTER, supra note 263 (providing a series of specific examples of agency involvement in policymaking).
359. Id. at 123 (discussing the implementation of rural free delivery).
360. Id. at 282 (discussing vast changes within the Forest Service policy without congressional approval).
361. Id. at 366 (discussing the Food and Drug Administration's "tremendous power over the pharmaceutical industry").
362. See also Daniel Carpenter & Colin D. Moore, Robust Action and the Strategic Ambiguity in a Bureaucratic Cohort: FDA Officers and the Evolution of New Drug Regulations, 1950-1970, at 8 (Harvard Univ. Dep't of Gov't, Working Paper, 2006) ("If rulemaking officials (or cohorts) are able to secure public legitimacy, are able to assemble "strange bedfellows" coalitions behind their activities, and are able to capitalize upon political opportunities, they can achieve the translation of divergent bureaucratic goals into policy while bypassing entirely the legislative process.").
363. See James L. Sundquist, Reflections on Watergate: Lessons for Public Administration, 34 PUB. ADMIN. REV. 453, 456 (1974).
364. See REACHING FOR GLORY 245 (Michael R. Beschloss ed. 2001).