Learning Through Policy Variation

Yale Law Journal

Volumn 118, Issue 3, 2008, Pages 480-553

  Listokin, Yair ^a  

^a YALE UNIVERSITY   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 62549085130     PISSN: 00440094     EISSN: None     Source Type: Journal    
DOI: 10.2307/20454719     Document Type: Review

References (231)

1. "Learning" about policies refers to the process through which the variability in outcomes associated with a policy is reduced. Before implementing a policy, policymakers may have only a dim idea about the effects of the policy. After implementing the policy and observing its effects, policymakers will often have a much better sense of the outcomes associated with the policy in current and future periods. Thus, policymakers will have "learned" about the policy and have a much greater ability to predict the policy's impacts. It may also be possible to "learn" about a policy without trying it. For example, one can learn about untried Policy A by extrapolating the effects of a similar but tested Policy B to Policy A. See infra Part III. This form of learning necessarily will be inferior to directly learning about Policy A by trying it.
2. See Martin L. Weitzman, Optimal Search for the Best Alternative, 47 ECONOMETRICA 641 (1979) (developing the mathematical basis of the optimal search approach) ;
3. see also Philippe Aghion et al., Optimal Learning by Experimentation, 58 REV. ECON. STUD. 621, 642-43 (1991) (same);
4. Volker Wieland, Learning by Doing and the Value of Optimal Experimentation, 24 J. ECON. DYNAMICS & CONTROL 501, 513-19 (2000) (characterizing the process of experimentation for economic agents, such as firms with unknown demand curves). This Article examines public policymaking as an optimal search process - a novel perspective in both the economics and legal literatures.
5. See, e.g., Ian Ayres & Eric Talley, Solomonic Bargaining: Dividing a Legal Entitlement To Facilitate Coasean Trade, 104 YALE L.J. 1027 (1995).
6. See, e.g., Louis Kaplow & Steven Shavell, Do Liability Rules Facilitate Bargaining? A Reply to Ayres and Talley, 105 YALE L.J. 221 (1995).
7. Default rules may fail because the assumptions underlying their supposed desirability prove faulty, or they may fail because of an unforeseen hindrance.
8. This is generally known as the "real options" approach to decisionmaking under uncertainty. For a comprehensive economic account, see AVINASH K. DIXIT & ROBERT S. PINDYCK, INVESTMENT UNDER UNCERTAINTY (1994).
9. For legal applications, see IAN AYRES, OPTIONAL LAW: THE STRUCTURE OF LEGAL ENTITLEMENTS (2005) ;
10. and Cass R. Sunstein, Irreversible and Catastrophic, 91 CORNELL L. REV. 841, 862-64 (2006). Irreversibility and the real options approach are discussed in detail in Part III.
11. This Part does not provide a full review of the political philosophies of classical liberals or Burkeans. Rather, it briefly sketches their attitudes toward policymaking to demonstrate how the optimal search approach relates to these two schools of thought. For an article applying Burkean and liberal political philosophies to methods of legal interpretation, see Ernest Young, Rediscovering Conservatism: Burkean Political Theory and Constitutional Interpretation, 72 N.C. L. REV. 619 (1994).
12. RICHARD HUDELSON, MODERN POLITICAL PHILOSOPHY 41 (1999).
13. If the impacts of a policy are uncertain, then reason should be able to predict the probabilities of certain outcomes.
14. See John C. Harsanyi, Cardinal Utility in Welfare Economics and in the Theory of Risk-Taking, 61 J. POL. ECON. 434, 434-35 (1953) (suggesting "a close affinity between the cardinal utility concept of welfare economics and the cardinal utility concept of the theory of choices involving risk [expected utility maximization]," thereby implying that policy decisions from a welfarist perspective should maximize expected value or utility).
15. See generally JEREMY BENTHAM, AN INTRODUCTION TO THE PRINCIPLES OF MORALS AND LEGISLATION 11-12 (J.H. Burns & H.L.A Hart eds., Oxford Univ. Press 1996) (1789) (describing the "principle of utility" as "that principle which approves or disapproves of every action whatsoever, according to the tendency which it appears to have to augment or diminish the happiness of the party whose interest is in question").
16. See, e.g., JOHN RAWLS, A THEORY OF JUSTICE (1971).
17. Some scholars argue that Burke was himself a classical liberal along many dimensions. Others claim that his reverence for tradition precluded his being a liberal. For a discussion of Burke's work, see CONOR CRUISE O'BRIEN, THE GREAT MELODY: A THEMATIC BIOGRAPHY AND COMMENTED ANTHOLOGY OF EDMUND BURKE 605-18 (1992). For an application of Burke's views to judicial decisionmaking, see Cass R. Sunstein, Burkean Minimalism, 105 MICH. L. REV. 353 (2006);
18. and Young, supra note 7. The discussion of Burke's thoughts in these articles influenced this Part.
19. EDMUND BURKE, REFLECTIONS ON THE REVOLUTION IN FRANCE 74 (Frank M. Turner ed., Yale Univ. Press 2003) (1790).
20. Edmund Burke, Speech on the Reform of the Representation of the Commons in Parliament (May 7, 1782), in 4 SELECT WORKS OF EDMUND BURKE: MISCELLANEOUS WRITINGS 15, 21 (E.J. Payne ed., 1990).
21. EDMUND BURKE, AN APPEAL FROM THE NEW TO THE OLD WHIGS 113 (London, J. Dodsley 1791). In this account, the value of past practice is very much empirical. Past practice has worked and is therefore owed deference. There are other values associated with Burke's fealty to tradition, of course. For a comprehensive discussion, see Anthony T. Kronman, Precedent and Tradition, 99 YALE L.J. 1029 (1990).
22. BURKE, supra note 14, at 143-44.
23. If some part of a policy reform fails, then policymakers should either return to the old status quo, or - if a partial reform yields better outcomes than the old status quo-maintain the partially reformed policy as a new status quo.
24. These benefits and costs can be quantified by any social welfare metric. For example, the benefits in dollars may measure the dollar benefits to the least well-off member of society (a social welfare function advocated by John Rawls) rather than measuring total wealth. More generally, the benefits and costs could account for risk aversion in a rationalist framework, as expected utility theory does within economic science. This would make the Burkean and rationalist frameworks more similar to each other, although Burke would presumably doubt the ability of the rationalist to appropriately measure risk aversion. Throughout this Article, I generally adopt a utilitarian framework.
25. Kaldor-Hicks efficiency proposes that a more socially efficient outcome can be reached by moving from the Pareto optimal level if those who would be made better off by the outcome compensate those who would be made worse off by the outcome. For a useful discussion of Kaldor-Hicks efficiency, see Guido Calabresi, The Pointlessness of Pareto: Carrying Coase Further, 100 YALE L.J. 1211 (1991).
26. An extremely risk-averse classical liberal might advocate cautious policymaking approaches that conform with the Burkean approach. Nevertheless, there is a difference between the sources of caution. For the liberal, caution is a product of a risk-averse response to a probability distribution of different policy outcomes. For the Burkean, caution is warranted by the tendency to underestimate the probability of bad outcomes.
27. See, e.g., Adrian Vermeule, Many-Minds Arguments in Legal Theory 11-12 (Harvard Law Sch. Pub. Law & Theory, Working Paper No. 08-02, 2008), available at http://ssrn.com/abstract=1087017.
28. See, e.g., Michael C. Dorf & Charles F. Sabel, A Constitution of Democratic Experimentalism, 98 COLUM. L. REV. 267 (1998);
29. Daniel A. Farber, Environmental Protection as a Learning Experience, 27 LOY. L.A. L. REV. 791 (1994) ;
30. Bradley C. Karkkainen, Environmental Lawyering in the Age of Collaboration, 2002 WIS. L. REV. 555.
31. For a review of this literature, see Orly Lobel, The Renew Deal: The Fall of Regulation and the Rise of Governance in Contemporary Legal Thought, 89 MINN. L. REV. 342 (2004).
32. Farber, supra note 23, at 791.
33. For a thorough examination of experimentalism in a federalist system, see infra Subsection II.C.4.
34. See, e.g., Farber, supra note 23, at 803 ("If a decision has irreparable consequences, then it may be worth delaying the decision in order to obtain new information. Taking an irreversible step forecloses the possibility of future learning, and therefore incurs an extra cost that does not show up in the usual cost-benefit analysis. In a formal sense waiting is equivalent to purchasing an option contract, and under many circumstances that option has positive value." (footnote omitted)). For a detailed discussion, see infra Section III.C.
35. For example, when new policies are irreversible and the outcome of B can be determined without actually implementing B, dynamic analysis suggests that policy A is preferred to policy B. See Sunstein, supra note 6; infra Part III.
36. Even if there is policy inertia, policies are not set in stone. For a detailed discussion, see infra Part III.
37. See, e.g., Dorf & Sabel, supra note 23, at 314 (advocating the "construct[ion of] the organizational rudiments of local, or, rather, subnational, pragmatist government, by transposing to the public sphere the institutions of benchmarking, simultaneous engineering, and error detection").
38. This example parallels the one found in Weitzman, supra note 2, at 641-43.
39. None of the results in this Article depends upon these assumptions, and they will be relaxed or altered in subsequent Sections.
40. This assumption is obviously too strong. For many reasons, including the political system and status quo bias, the existing policy is cheapest and easiest to implement in the next period. Previously implemented policies are the next cheapest, and brand new policies are presumably the most expensive. The consequences of the policy inertia that is imposed by this distribution of implementation costs are discussed in detail in Part III.
41. Alternatively, Policy A may be a very low-risk policy that has not been implemented before.
42. One might ask why Policy B has never been attempted before. There are several responses. One answer may be that Policy B had never been thought of before. A second may be that there has been some potentially permanent change in the environment that changes the payoff distribution of Policy B, making Policy B an essentially new policy.
43. Throughout the remainder of this Article, terms like "B being better than A" or "B beating A" will mean that B should be chosen before A. At no point do I mean to suggest that B is better than A in every period; I suggest only that B should be chosen first under many circumstances in spite of its inferiority in expected-value terms.
44. For a rigorous proof of this argument, see Weitzman, supra note 2.
45. This hypothetical is further examined in Section III.C.
46. Some of the most important complications are examined in Part III.
47. The problem of irreversible risks is discussed in detail in Part III.
48. See Douglas A Kysar, It Might Have Been: Risk, Precaution, and Opportunity Costs 3-4 (Cornell Law Sch. Legal Studies Research Paper Series, Research Paper No. 06-023, 2006), available at http://ssrn.com/abstract= 927995.
49. Cost-benefit analysis does not obviously distinguish between reversible and irreversible risks. In addition, "some formulations [of the precautionary principle] . . . apply to even theoretically reversible risks." David A. Dana, A Behavioral Economic Defense of the Precautionary Principle, 97 NW. U. L. REV. 1315, 1315-16 (2003).
50. See Amartya Sen, Discipline of Cost-Benefit Analysis, 29 J. LEGAL STUD. 931 (2000).
51. Kysar, supra note 40, at 3-4 (quoting Peter Montague, The Precautionary Principle, RACHEL'S ENV'T & HEALTH WKLY., Feb. 19, 1998, at 1, but mistranscribing "cause and effect" as "cause-and-effect"). Note that there are many different articulations of the precautionary principle.
52. See, e.g., Sunstein, supra note 6, at 848-50. The one stated here is merely one oftcited statement of the principle. A full characterization of the precautionary principle or of cost-benefit analysis is beyond the scope of this paper.
53. See C.A. BOWERS, MINDFUL CONSERVATISM: RETHINKING THE IDEOLOGICAL AND EDUCATIONAL BASIS OF AN ECOLOGICALLY SUSTAINABLE FUTURE 91-94 (2003) (drawing a link between environmental conservation and Burke's approach). Curiously, while legal scholars have spilled considerable ink on the precautionary principle (a search in Westlaw's Journals and Law Reviews database for ["precautionary principle"] on October 31, 2008 yielded over 2100 documents) and on Burke (a search in Westlaw's Journals and Law Reviews database for [Edmund /5 Burke!] on October 31, 2008 yielded over 1400 documents), they have not explored the seemingly obvious linkage between Burkean political philosophy and the precautionary principle. A search in Westlaw's Journals and Law Reviews database for ["precautionary principle" /50 Burke!] on October 31, 2008 yielded only six documents, five of which are related to a different Burke and the last of which makes only a casual allusion to the similarity between the precautionary principle and Burkean conservatism.
54. See Daniel A. Farber, From Here to Eternity: Environmental Law and Future Generations, 2003 U. ILL. L. REV. 289, 305-06 n.90. It is important to note that Westlaw's Journals and Law Reviews database neither catalogues all volumes of currently published law journals, nor includes scholarship in non-law disciplines.
55. Both issuing new regulations and allowing some polluters to change their technology for pollution reduction entail costs.
56. The benefits of either technology are realized in the form of reduced pollution.
57. See supra note 43 and accompanying text.
58. Exec. Order No. 12,866 § 6(a)(3)(c), 3 C.F.R. 638 (1993), reprinted in 5 U.S.C § 601 (2000).
59. OFFICE OF MGMT. & BUDGET, CIRCULAR A-4, REGULATORY ANALYSIS (2003), available at http://www.whitehouse.gov/omb/circulars/aoo4/a-4.pdf.
60. The OMB Circular makes no mention of measuring the information benefit estimates of a particular regulation under its guidelines. Id
61. Id. at 1.
62. The precautionary principle often assumes that probabilities cannot be estimated, a condition called uncertainty. See Kysar, supra note 40, at 10-12. So long as the risks involved are not irreversible, uncertainty does not require drastic changes in the lessons of the optimal search approach. With uncertain but reversible risks, one can assume the worstthat is, place heavy probability on bad outcomes - and still favor risky policies, rather than safe ones.
63. See Lisa Heinzerling, Regulatory Costs of Mythic Proportions, 107 YALE L.J. 1981, 2043-56 (1998).
64. See id.
65. See ROBERT E. SCOTT & JODY S. KRAUS, CONTRACT LAW AND THEORY 26-27 (4th ed. 2007).
66. Id. at 90.
67. SEE Ian Ayres & Robert Gertner, Filling Gaps in Incomplete Contracts: An Economic Theory of Default Rules, 99 YALE L.J. 87 (1989).
68. See Lucian Ayre Bebchuk & Steven Shavell, Information and the Scope of Liability for Breach of Contract: The Rule of Hadley v. Baxendale, 7 J.L. ECON. & ORG. 284 (1991).
69. Hadley v. Baxendale, (1854) 9 Ex. 341, 156 Eng. Rep. 145; see Barry E. Adler, The Questionable Ascent of Hadley v. Baxendale, 51 STAN. L. REV. 1547, 1547-48 (1999). Following the literature, this discussion will emphasize the Hadley rule. The same conclusions apply to other information-forcing or penalty default rules.
70. See Ayres & Gertner, supra note 57, at 112-15;
71. Bebchuk & Shavell, supra note 58, at 308. Even if the Hadley rule would be the choice of most parties considering the rule, the rule was not chosen for this reason.
72. See Ayres & Gertner, supra note 57, at 106-07.
73. See Eric A Posner, There Are No Penalty Default Rules in Contract Law, 33 FLA. ST. U. L. REV. 563 (2006) (arguing that penalty default rules do not exist in practice).
74. But see Ian Ayres, Ya-Huh: There Are and Should Be Penalty Defaults, 33 FLA. ST. U. L. REV. 589 (2006) (contradicting Posner's assertions and claiming that there are many examples of penalty default rules) .
75. See Bebchuk & Shavell, supra note 58, at 302 proposition 6(b) (demonstrating the inefficiency of the Hadley rule when communication between parties is undesirable).
76. See Adler, supra note 59, at 1547-48.
77. Id. at 1556.
78. See id. at 1559-70.
79. Id. at 1552.
80. Since most contract default rules have been chosen using the majoritarian paradigm, see SCOTT & KRAUS, supra note 55, at 89-90, the majoritarian rule is the existing default law in most cases.
81. Posner, supra note 62, at 565, 573-86.
82. Intriguingly, penalty default rules may entail less variance than majoritarian rules when the penalty rule is the existing default rule - as in the case of the Hadley damages rule. When the penalty rule is the status quo, majoritarian rules become more attractive, all else equal. When neither the majoritarian rule nor the penalty rule is the default rule, however, the majoritarian rule will generally entail less variance.
83. In Example 2, Policy B has a small probability of producing a favorable outcome. Penalty defaults probably have higher probabilities of success, but success is not quite as advantageous as in the example. If necessary, the example could be altered. The basic principle remains the same. So long as a penalty default rule provides some possibility of a better outcome while involving more risk than the majoritarian rule, then the optimal search approach dictates that the penalty default rule should be tried first in some cases.
84. Current law requires credit card issuers to prominently state their rates, but the issuers reserve the right to change the rates unilaterally. See Truth in Lending Act, 15 U.S.C §§ 1601-1693r (2000).
85. Critics of enhanced shareholder democracy may argue that policy variance is unnecessary in this context because federalism in corporate law already produces sufficient policy variance, see, e.g., Roberta Romano, Empowering Investors: A Market Approach to Securities Regulation, 107 YALE L.J. 2359, 2392 (1998), with the "race to the top" leading all states to adopt value-enhancing policy variants. If one accepts this proposition, then high-variance policies are unhelpful because the variance they provide is unlikely to find innovative policies missed by the states. Others, however, argue that state competition for corporate charters produces relatively little variance.
86. See, e.g., Lucian Arye Bebchuk & Assaf Hamdani, Vigorous Race or Leisurely Walk: Reconsidering the Competition over Corporate Charters, 112 YALE L.J. 553, 605-06 (2002). These arguments echo a broader debate about whether state competition for corporate charters produces a "race to the top" - with all states providing efficient and ever-improving laws - or a "race to the bottom" - with states undercutting each other by offering laws that favor management or other interested parties.
87. Compare Roberta Romano, Law As a Product: Some Pieces of the Incorporation Puzzle, 1 J.L. ECON. & ORG. 225, 280-81 (1985) (arguing that corporate law competition creates a race to the top),
88. and Ralph K. Winter, Jr., State Law, Shareholder Protection, and the Theory of the Corporation, 6 J. LEGAL STUD. 251 (1977) (characterizing corporate law competition as a race to the top), with Lucian Arye Bebchuk, Federalism and the Corporation: The Desirable Limits on State Competition in Corporate Law, 105 HARV. L. REV. 1435, 1441 (1992) (describing how state competition in corporate law does not invariably lead to a race to the top), and William L. Cary, Federalism and Corporate Law: Reflections upon Delaware, 83 YALE L.J. 663 (1974) (describing state-based corporate law as race to the bottom). The optimal search approach to corporate law described in this Article may be valuable to this debate for two reasons. First, if Bebchuk and Hamdani's lack-of-innovation argument is the better one, then high-variance optimal search policies at the federal level offer the potential for great benefits. Second, attempting new policies offers the possibility of resolving the seemingly never-ending debate about the value of state competition in corporate law. If innovative corporate law policies are attempted but consistently prove to be failures, then this provides strong evidence of a "race to the top." For a thorough discussion of the relation between federalism and the optimal search approach, see infra Subsection II.C.4 and Section IV.C.
89. See Lucian Arye Bebchuk, The Case for Increasing Shareholder Power, 118 HARV. L. REV. 833, 843-46 (2005).
90. See id.
91. See id. at 844-46.
92. See id.
93. See, e.g., Stephen M. Bainbridge, Director Primacy and Shareholder Disempowerment, 119 HARV. L. REV. 1735, 1758 (2006) [hereinafter Bainbridge, Shareholder Disempowerment];
94. Stephen M. Bainbridge, Director Primacy: The Means and Ends of Corporate Governance, 97 NW. U. L. REV. 547, 605-06 (2003);
95. Iman Anabtawi, Some Skepticism About Increasing Shareholder Power, 53 UCLA L. REV. 561, 598 (2006) ;
96. A. Gilchrist Sparks, III, Corporate Democracy: What It Is, What It Isn't, and What It Should Be, INSIGHTS, Mar. 2006, at 20, 23 (noting the "overwhelming historical success of the existing governance system") ;
97. see also Leo E. Strine, Jr., Toward a True Corporate Republic: A Traditionalist Response to Bebchuk's Solution for Improving Corporate America, 119 HARV. L. REV. 1759, 1759 (2006) (critiquing Bebchuk's proposal to increase shareholder power from the perspective of "traditionalist investors").
98. Bainbridge, Shareholder Disempowerment, supra note 78, at 1758. Note the Burkean allusions in Bainbridge's argument.
99. Id. at 1749 (quoting KENNETH J. ARROW, THE LIMITS OF ORGANIZATION 69 (1974)).
100. Blasius Indus., Inc. v. Atlas Corp., 564 A.2d 651,659 (Del. Ch. 1988).
101. Compare Lucian A. Bebchuk, The Myth of the Shareholder Franchise, 93 VA. L. REV. 675 (2007) (arguing that the shareholder franchise is illusory), with Martin Lipton & William Savitt, The Many Myths of Lucian Bebchuk, 93 VA. L. REV. 733, 734 (2007) (arguing "that the 'myth of the shareholder franchise' is no myth at all").
102. See Unocal Corp. v. Mesa Petroleum Co., 493 A.2d 946, 959 (Del. 1985) ("If the stockholders are displeased with the action of their elected representatives, the powers of corporate democracy are at their disposal to turn the board out.").
103. See Yair Listokin, Corporate Voting vs. Market Price Setting (July 3, 2008) (unpublished manuscript, on file with author), available at http://ssrn.com/abstract=1112671.
104. See id.
105. See Bebchuk, supra note 82, at 682-88 (arguing that the small number of contested elections indicates some problems with the shareholder franchise).
106. But see Jonathan R. Macey, Too Many Notes and Not Enough Votes: Lucian Bebchuk and Emperor Joseph II Kvetch About Contested Director Elections and Mozart's Seraglio, 93 VA. L. REV. 759 (2007) (arguing that Bebchuk has no baseline for determining the appropriate number of contested elections).
107. Letter from Wachtell, Lipton, Rosen & Katz to Jonathan G. Katz, Sec'y, SEC (June 11, 2003), available at http://www.sec.gov/rules/other/ s71003/wachtello61103.htm; see also Lipton & Savitt, supra note 82 (defending shareholder franchise).
108. See Bebchuk, supra note 82, at 688-94.
109. See generally Yair Listokin, Management Always Wins the Close Ones, 10 AM. L. & ECON. REV. 159 (2008) (discussing management advantages in corporate voting).
110. Bebchuk, supra note 82, at 688-94.
111. Id.
112. Id.
113. See,e.g., id.
114. Id. at 677.
115. Id.
116. See, e.g., Lucian Arye Bebchuk, The Case for Shareholder Access to the Ballot, 59 BUS. LAW. 43 (2003).
117. At present, dissident slates of directors must mail their own proxy forms to shareholders. See Bebchuk, supra note 82, at 696-98.
118. See, e.g., Lipton & Savitt, supra note 82;
119. see also Macey, supra note 86 (arguing that Bebchuk has no baseline for determining the appropriate number of contested elections).
120. Lipton & Savitt, supra note 82, at 733-34. Again, note the strong Burkean tone to the arguments in favor of the status quo.
121. See Annette L. Nazareth, Comm'r, SEC, Speech by SEC Commissioner: Remarks Before the International Corporate Governance Network (Oct. 29, 2007) (transcript available at http://www.sec.gov/news/speech/2007/spch102907aln.htm) [hereinafter Nazareth Speech].
122. See, e.g., Macey, supra note 86;
123. John F. Olson, Professor Bebchuk's Brave New World: A Reply to "The Myth of the Shareholder Franchise," 93 VA. L. REV. 773 (2007);
124. Lynn A Stout, The Mythical Benefits of Shareholder Control, 93 VA. L. REV. 789 (2007);
125. E. Norman Veasey, The Stockholder Franchise Is Not a Myth: A Response to Professor Bebchuk, 93 VA. L. REV. 811 (2007). It is extremely unlikely that the issue would have engendered so many comment letters if its proponents did not think it was likely to have significantly positive effects and its detractors did not think the reform would have negative implications.
126. WILLIAM H. RIKER, FEDERALISM: ORIGIN, OPERATION, SIGNIFICANCE 11 (1964).
127. Of course, there are many other arguments for federalism. For a comprehensive review of these arguments, see DAVID L. SHAPIRO, FEDERALISM: A DIALOGUE 58-106 (1995). This Article does not examine these powerful alternative justifications for federalism.
128. New State Ice Co. v. Liebmann, 285 U.S. 262, 311 (1932) (Brandeis, J., dissenting). Justice Brandeis's argument for federalism has been adopted by many scholars examining a wide variety of different contexts. A Westlaw search for law review articles quoting Justice Brandeis's federalism-as-laboratory quotation retrieved almost seven hundred articles.
129. Dorf & Sabel, supra note 23, at 287-88.
130. See generally id. at 314-24 (discussing how diversity can lead to local differences in policy and how policy should be "benchmarked" to produce new policies that are better than previous policies).
131. Id. at 287.
132. Susan Rose-Ackerman questions the argument that federalism promotes innovation. See Susan Rose-Ackerman, Risk Taking and Reelection: Does Federalism Promote Innovation?, 9 J. LEGAL STUD. 593 (1980). Public choice considerations, such as public policymakers' desire for reelection, may diminish incentives for risk taking by policymakers in a federal system. After a thorough analysis of these incentives, Rose-Ackerman concludes that "few useful experiments will be carried out in [state and local governments]."
133. Id. at 594.
134. See Dorf & Sabel, supra note 23.
135. For a thorough examination of preemption, see Caleb Nelson, Preemption, 86 VA. L. REV. 225 (2000).
136. Boggs v. Boggs, 520 U.S. 833, 844 (1997) (quoting Gade v. Nat'l Solid Wastes Mgmt. Ass'n, 505 U.S. 88, 98 (1921)).
137. See Nelson, supra note 110, at 229 n.16 (cataloguing and summarizing the primary critiques of preemption, including critiques found in KENNETH STARR ET AL., THE LAW OF PREEMPTION 47-56 (1991) ;
138. Betsy J. Grey, Make Congress Speak Clearly: Federal Preemption of State Tort Remedies, 77 B.U. L. REV. 559,561 (1997);
139. S. Candice Hoke, Preemption Pathologies and Civic Republican Values, 71 B.U. L. REV. 685, 687-88 (1991);
140. Donald P. Rothschild, A Proposed "Tonic" with Florida Lime To Celebrate Our New Federalism: How To Deal with the "Headache" of Preemption, 38 U. MIAMI L. REV. 829, 830 n.3 (1984);
141. Paul Wolfson, Preemption and Federalism: The Missing Link, 16 HASTINGS CONST. L.Q. 69, 114 (1988)).
142. See supra note 112.
143. Electric power production provides an excellent example. Coal-fired plants are considerably cheaper than gas-fired plants for producing electricity. Burning coal, however, produces more externalities than does gas. A national regulator might decide that these externalities outweigh the benefit of cheaper power production. If the acid rain harms other states but does not harm the state where the power is produced, however, then a state regulator may decide to approve coal power production.
144. Hines v. Davidowitz, 312 U.S. 52, 67 (1941).
145. This conclusion arises from the optimal search approach rather than the experimentalist approach to federalism. See infra Section IV.C.
146. This does not mean that the benefits of standardization should be ignored. Instead, they should be compared with the underappreciated informational benefits of variation.
147. Again, the argument will be presented in risk-neutral terms for the sake of clarity. Parallel arguments could be made in utility terms for a risk-averse utilitarian rationalist.
148. While Burke was opposed to "innovation" in the form of abrupt change in policy, he was more receptive to extremely gradual policy modifications. See Ian Harris, Edmund Burke, in STANFORD ENCYCLOPEDIA OF PHILOSOPHY §8 (Edward N. Zalta ed., 2004), http://plato.stanford.edu/entries/burke/ (last visited Oct. 27, 2008).
149. See supra Part I.
150. Young, supra note 7, at 646.
151. There is some rational judgment required for the Burkean approach as well. Unless Burkeans want to have rigid laws, which Burke (at least) did not, there must be some rational evaluation of new policies and debate about the scope by which they change existing policy.
152. That is, suppose that the range of possible outcomes is much wider than an expert thinks, but the increase in variability occurs on both the positive side and the negative side. If the negative consequences of a policy change increase but the positive consequences do not, then the expected value of a policy change goes down and the policy change becomes less attractive under the optimal search approach.
153. SEE Kronman, supra note 16, at 1066-68.
154. EDMUND BURKE, REFLECTIONS ON THE REVOLUTION IN FRANCE 81 (Frank M. Turner ed., 2003).
155. This does not mean that the present generation is behaving altruistically. The optimal search approach benefits all members of the present generation who will enjoy the potential benefits of policy innovation in future periods. Alternatively, the present generation could consume some of the surplus created by the optimal search approach for future generations by leaving less for future generations along other dimensions. Because the optimal search approach creates value, such gain sharing could leave all generations better off.
156. This will be discussed in detail in Section III.C.
157. For a detailed description of the "real options" approach, see infra Subsection III.C.1.
158. See Sunstein, supra note 6, at 860-64 (distinguishing between irreversibility as "seriousness" and irreversibility as "sunk costs"). The definition provided here resembles the sunk cost definition. A cost is sunk if it cannot be retrieved. A policy with long-lasting effects can therefore be represented as having a high initial cost-measured as the present discounted value of the future effects - or it can be represented as having costs distributed over a long time period. In the numerical examples, I choose to equate irreversibility with long-lasting effects, but the two methods are mathematically equivalent.
159. Almost all sources of irreversibility can be placed within these two broad categories. For example, inertia can be described as resulting from high formulation and implementation costs.
160. There are many policies that may have positive expected value, but cannot be implemented because of the scarcity of legislative time and effort. These time and effort expenditures are irretrievable opportunity costs of formulating a particular policy.
161. Daniel Kahneman, Jack L. Knetsch & Richard H. Thaler, Anomalies: The Endowment Effect, Loss Aversion, and Status Quo Bias, J. ECON. PERSP., Winter 1991, at 193, 197-99.
162. See DIXIT & PINDYCK, supra note 6, at 8-9. The effects of a policy in the current period are obviously irreversible. If a policy is a failure, these effects cannot be recouped by changing the policy in the future period. These costs are accounted for in both the optimal search approach and the real options approach, which have countervailing benefits that surpass these per-period costs in many cases.
163. See, e.g., NICHOLAS STERN, THE ECONOMICS OF CLIMATE CHANGE: THE STERN REVIEW 28 (2007) ("The impacts of climate change are persistent and develop over time. Once in the atmosphere, some [greenhouse gases] stay there for hundreds of years.").
164. See Sunstein, supra note 6, at 860-69.
165. Sunstein makes exactly this argument in his article, Irreversible and Catastrophic. See Sunstein, supra note 6, at 855-64. His article assumes that cost-benefit analysis is superior to the precautionary principle when risks are reversible and does not consider the possibility that reversibility may imply a deviation from expected-value maximization that is similar in logic but very different in outcome from the irreversible harm precautionary principle.
166. This discussion is not meant to minimize the harms caused by catastrophic occurrences, but to develop an analytical distinction between catastrophic and irreversible.
167. The choice of four periods is arbitrary, but this choice demonstrates how partial reversibility brings optimal policy choices closer to naïve expected-value maximization. Partial reversibility could also be denoted by imposing a significant switching cost on any change in policy. The two methods can be made mathematically equivalent, as the longevity of a policy that should be changed imposes a cost that is directly analogous to a high switching cost.
168. As in Example 4, the true impact of Policy B can be learned without actually implementing Policy B.
169. Policy A, like Policy B, can only be changed after four periods in this example. Policy A has no variance in outcomes, so the delay is not caused by the time needed to observe the effects of Policy A. The delay could be caused by a delay in the ability to evaluate Policy B when Policy B is not being implemented. (Recall that the only reason to change Policy A under the real options approach is after Policy B has been proven effective.)
170. This assumption is relaxed in later Sections.
171. U.S. CONST, art. V.
172. See Matthew C. Stephenson, The Price of Public Action: Constitutional Doctrine and the Judicial Manipulation of Legislative Enactment Costs, 118 YALE L.J. 2, 15-16 (2008).
173. It should be acknowledged, of course, that even constitutional policymaking is not impossible to reverse. The Eighteenth Amendment implementing Prohibition can be seen as an innovative constitutional policy - there are compelling theoretical arguments in favor of Prohibition - that was judged a failure and then repealed. See U.S. CONST, amends. XVIII, XXI. The long lag between Prohibition and repeal, however, demonstrates the inertia of constitutional policymaking.
174. U.S. CONST, art. I, § 7, cl. 2.
175. Judges are policymakers in the sense that they create rules affecting future behavior. See Republican Party of Minn. v. White, 536 U.S. 765, 784 (2002) (noting the power of state judges to "make" common law).
176. See Planned Parenthood of Se. Pa. v. Casey, 505 U.S. 833, 854-69 (1992).
177. This is not to say, however, that administrative agencies are inclined to change policies. Indeed, they may have incentives not to make frequent policy changes. These incentives will be examined in Part IV. See also JERRY L. MASHAW & DAVID L. HARFST, THE STRUGGLE FOR AUTO SAFETY (1990) (discussing inertia in administrative agencies).
178. Administrative Procedure Act of 1946, 5 U.S.C §§ 551-559 (2000).
179. Section 551 of the APA defines a rule as "the whole or a part of an agency statement of general or particular applicability and future effect designed to implement, interpret, or prescribe law or policy." Id. § 551(4).
180. See Anne Joseph O'Connell, Political Cycles of Rulemaking: An Empirical Portrait of the Modern Administrative State, 94 VA. L. REV. 889, 901-02 (2008) (providing a comprehensive description of notice-and-comment rulemaking).
181. See, e.g., Stephenson, supra note 143, at 15-16 (acknowledging that judicially imposed costs to the enactment of new legislation may lead to the prevention of "socially desirable, constitutionally justifiable legislation" if, for example, the judiciary has incomplete information).
182. See, e.g., Myers v. United States, 272 U.S. 52, 293 (1926) (Brandeis, J., dissenting) (recognizing that the separation of powers aims "not to promote efficiency but to preclude the exercise of arbitrary power");
183. Martin S. Flaherty, The Most Dangerous Branch, 105 YALE L.J. 1725, 1826-28 (1996) (praising separation of powers but recognizing that it hinders "action").
184. For an examination of legislative entrenchment doctrines, see Eric A Posner & Adrian Vermeule, Legislative Entrenchment: A Reappraisal, 111 YALE L.J. 1665, 1665-66 (2002).
185. See Newton v. Comm'rs, 100 U.S. 548, 559 (1879);
186. WILLIAM BLACKSTONE, COMMENTARIES *90.
187. Posner & Vermeule, supra note 154, at 1667.
188. Id. at 1676.
189. Id. at 1666.
190. See id. at 1670-73.
191. Id. at 1666.
192. See THE FEDERALIST NOS. 48, 51 (James Madison).
193. SEE Bruce Ackerman, The New Separation of Powers, 113 HARV. L. REV. 633, 650 (2000).
194. See THE FEDERALIST NO. 47 (James Madison).
195. See Douglas W. Kmiec, Of Balkanized Empires and Cooperative Allies: A Bicentennial Essay on the Separation of Powers, 37 CATH. U. L. REV. 73, 75-76 (1987).
196. See, e.g., supra note 153 (quoting Justice Brandeis).
197. Rachel E. Barkow, Separation of Powers and the Criminal Law, 58 STAN. L. REV. 989, 991 (2006).
198. For an illuminating discussion of stare decisis, see RICHARD A. POSNER, THE FEDERAL COURTS: CHALLENGE AND REFORM 371-82 (1996).
199. For one recent discussion, see Daniel A. Farber, The Rule of Law and the Law of Precedents, 90 MINN. L. REV. 1173 (2006).
200. Lewis F. Powell, Jr., Stare Decisis and Judicial Restraint, 47 WASH. & LEE L. REV. 281, 289-90 (1990).
201. South Carolina v. Gathers, 490 U.S. 805, 825 (1989) (Scalia, J., dissenting).
202. Judges can choose policies explicitly under the common law. See Republican Party of Minn. v. White, 536 U.S. 765, 784 (2002) (noting the power of state judges to "make" common law). When they are interpreting statutes, judges enjoy less discretion, as they are constrained by principles of statutory interpretation. Nevertheless, judges can choose interpretive methods to implement a variety of different outcomes when interpreting statutes.
203. See, e.g., Alexander Volokh, Choosing Interpretive Methods: A Positive Theory of Judges and Everyone Else, 83 N.Y.U. L. REV. 769 (2008). Thus, even when engaged in statutory interpretation, judges effectively "choose" policies to a certain degree, though they are constrained by principles of statutory interpretation.
204. The discussion in this Part is influenced by Rose-Ackerman, supra note 108.
205. See Susan Rose-Ackerman, Risktaking and Electoral Competition, 7 EUR. J. POL. ECON. 527, 527 (1991).
206. Id.
207. Cf Michael S. Kang, Race and Democratic Contestation, 117 YALE L.J. 734 (2008) (arguing that competitive electoral districts are less important than creating conditions of "democratic contestation").
208. But see Samuel Issacharoff & Richard H. Pildes, Politics as Markets: Partisan Lockups of the Democratic Process, 50 STAN. L. REV. 643, 646 (1998) ("Only through an appropriately competitive partisan environment can one of the central goals of democratic politics be realized: that the policy outcomes of the political process be responsive to the interests and views of citizens.").
209. For a thorough examination of the effects of federalism on innovation from a positive political theory perspective, see Rose-Ackerman, supra note 108.
210. To see this point, consider two jurisdictions that face the choice between Policy A and Policy B of Example 2. Section II.A demonstrated that in a uni-jurisdictional world, the optimal search approach dictates that Policy B is the efficient choice. When there are two jurisdictions, the choice of innovation (Policy B) versus no innovation (Policy A) can be modeled as the following game: The value of $628 in the game modeled above comes from retaining Policy A in the first period and observing the outcome of Policy B in the other jurisdiction. If Policy B fails in the other jurisdiction, then continue with Policy A in the second period. If Policy B succeeds in the other jurisdiction (and is observed), then switch to Policy B in the next period and enjoy the high upside of Policy B. In other words, one jurisdiction can free-ride off the other jurisdiction's experimentation with Policy B.
211. There is no pure strategy Nash Equilibrium to this game and therefore no unique solution. See, e.g., ANDREU MAS-COLELL, MICHAEL D. WHINSTON & JERRY R. GREEN, MICROECONOMIC THEORY 248-50 (1995). Choosing either policy is a rationalizable strategy for both jurisdictions, however. Therefore, we can make no definitive predictions about whether there will be innovation in this simple multi-jurisdictional world. The unijurisdictional world, by contrast, provides innovation using the optimal search approach. In this example, federalism can decrease innovation, but cannot increase it. When innovation is less attractive in the uni-jurisdictional world, however, there can be cases where federalism produces more innovation than does the optimal search approach.
212. See Michael Klausner, Corporations, Corporate Law, and Networks of Contracts, 81 VA. L. REV. 757, 764 (1995).
213. Corporate law has a longstanding debate about whether state competition for corporate charters entails a race to the bottom or a race to the top. For important contributions to this literature, see supra note 73.
214. See Kahneman et al., supra note 132.
215. Note, however, that there have been several examples of regulatory bodies running innovative policy experiments, such as the Moving to Opportunity experiment, see Jeffrey R. Kling, Jens Ludwig & Lawrence F. Katz, Neighborhood Effects on Crime for Female and Male Youth: Evidence from a Randomized Housing Voucher Experiment, 120 Q.J. ECON. 87 (2005), and the SECs pilot order restricting short sales of certain securities,
216. see Order Extending Term of Short Sale Pilot, Exchange Act Release No. 53,684, 71 Fed. Reg. 24,765 (Apr. 20, 2006).
217. For example, Judge Baron Alderson gained considerable notoriety for instituting an information-forcing default rule in Hadley v. Baxendale. An interesting avenue for future research would be to compare life-appointed judges with elected judges to determine whether electoral incentives promote innovation through variation in private law rulemaking. One might hypothesize that life-appointed judges, removed from electoral pressures to avoid negative-expected-value common law rules, are more willing to enact new legal doctrines with high variance. See David E. Pozen, The Irony of Judicial Elections, 108 COLUM. L. REV. 265, 324-28 (2008) (describing scholarly support for elected judges' more innovative or activist approaches and noting that state courts' "common-law lawmaking powers are broadly respected and [their] decisions are relatively easily reversed through constitutional amendment or legislative action"). I thank William J. Rinner for raising this point.
218. See Gratter v. Bollinger, 539 U.S. 306 (2003). In Grutter, the Court noted that it "expect[s] that 25 years from now, the use of racial preferences will no longer be necessary to further the interest approved today."
219. Id. at 343.
220. Note how international differences can also serve to produce useful variation in laws. One concern, of course, is that differences between nations are so great that one cannot infer the effects of a variation in Country X will be similar to the effects of the same variation in Country Y.
221. See Sourafel Girma, Steve Thompson & Peter W. Wright, Corporate Governance Reforms and Executive Compensation Determination: Evidence from the UK, 75 MANCHESTER SCH. 65, 66-67 n.5 (2007).
222. See Roei C. Campos, Comm'r, SEC, Speech by SEC Commissioner: The SECs Shareholder Access Proposal (Jan. 10, 2005) (transcript available at http://www.sec.gov/news/speech/spch011005rcc.htm) ("Based on these considerations, I believe that the [shareholder access] rule is fully within our authority. I am not alone in my view. Several academics have expressed support for the SECs authority to adopt the proxy access rule.").
223. But see Bus. Roundtable v. SEC, 905 F.2d 406 (D.C Cir. 1990) (suggesting limitations on the SECs power to regulate substantive aspects of corporate voting).
224. The SEC received over thirty-four thousand comments on the issue. See Nazareth Speech, supra note 100.
225. See, e.g., Broc Romanek, SEC Quickly Adopts Non-Shareholder Access Rule - Then Fireworks Ensue, TheCorporateCounsel.net Blog, Nov. 29, 2007, http://www.thecorporatecounsel.ner/blog/archive/oo1638.html ("The real fireworks began right after the meeting, when all sorts of investor groups, members of Congress, etc. issued statements disapproving the SECs rulemaking (eg. CalPERS; CII; RiskMetrics; AFL-CIO; Rep. Frank; Sen. Dodd) - and some approved (eg. Marty Lipton). It's notable that the opposition is fairly organized on this issue; I can't imagine something like this happening even five years ago.").
226. See Rose-Ackerman, supra note 108, at 594.
227. See id. at 615-16.
228. See id. at 616.
229. While random assignment may appear draconian, note that the SEC has randomly assigned corporations to different short-sale restriction regimes. See Order Extending Term of Short Sale Pilot, Exchange Act Release No. 53,684, 71 Fed. Reg. 24,765 (Apr. 20, 2006);
230. SEC OFFICE OF ECON. ANALYSIS, ECONOMIC ANALYSIS OF THE SHORT SALE PRICE RESTRICTIONS UNDER THE REGULATION SHO PILOT (2007), http://www.sec.gov/news/studies/2007/ regshopiloto20607.pdf.
231. See Ian Ayres & Yair Listokin, Randomizing Law (Sept. 2008) (unpublished manuscript, on file with author).