Controlling corporate misconduct: An analysis of corporate liability regimes

New York University Law Review

Volumn 72, Issue 4, 1997, Pages 687-779

  Arlen, Jennifer ^a,b,c   Kraakman, Reinier ^b,d,e  

^a UNIVERSITY OF SOUTHERN CALIFORNIA   (United States)

^b HARVARD UNIVERSITY   (United States)

^c NEW YORK UNIVERSITY   (United States)

^d HARVARD LAW SCHOOL   (United States)

^e YALE UNIVERSITY   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 0031512311     PISSN: 00287881     EISSN: None     Source Type: Journal    
DOI: None     Document Type: Article

References (265)

1. See generally Pamela H. Bucy, White Collar Crime: Cases and Materials 192-93 (1992); Kathleen F. Brickey, Corporate Criminal Liability: A Primer for Corporate Counsel, 40 Bus. Law. 129, 131 (1984) (stating that corporation may be held liable for criminal acts of even menial workers).
2. See generally Pamela H. Bucy, White Collar Crime: Cases and Materials 192-93 (1992); Kathleen F. Brickey, Corporate Criminal Liability: A Primer for Corporate Counsel, 40 Bus. Law. 129, 131 (1984) (stating that corporation may be held liable for criminal acts of even menial workers).
3. See, e.g., United States v. Basic Constr. Co., 711 F.2d 570, 573 (4th Cir. 1983) (holding that "corporation may be held criminally responsible for antitrust violations committed by its employees if they were acting within the scope of their authority, or apparent authority, and for the benefit of the corporation, even if . . . such acts were against corporate policy or express instructions"); United States v. Hilton Hotels Corp., 467 F.2d 1000, 1004 (9th Cir. 1972) (holding that criminal liability of corporation may arise from acts of agents without proof that conduct was within agent's actual authority); see also United States v. Twentieth Century Fox Film Corp., 882 F.2d 656, 660 (2d. Cir. 1989) (holding that corporate compliance program - "however extensive" - will not shield the company from criminal liability for its employees' actions); cf. Yates v. Avco Corp., 819 F.2d 630, 636 (6th Cir. 1987) (observing that supervisor's sexual harassment was foreseeable because company had adopted policy to address problem).
4. Generally, the benefit requirement is imposed only when the crime requires a specific mental state. See 10 William Meade Fletcher et al., Fletcher Cyclopedia of the Law of Private Corporations § 4944 (1986). Moreover, the benefit requirement does not require proof that the corporation actually received any benefit; all that is necessary is that the agent intended to further a corporate interest. See United States v. Carter, 311 F.2d 934, 943 (6th Cir. 1963) (holding that corporation may be found criminally liable for act of its president when he acted in course of employment and in furtherance of business interests of corporation); Bucy, supra note 1, at 201.
5. For legal scholarship applying agency cost analysis to the issue of corporate versus individual liability, see generally Christopher Stone, Where the Law Ends: The Social Control of Corporate Behavior (1975); Bruce Chapman, Corporate Tort Liability and the Problem of Overcompliance, 69 S. Cal. L. Rev. 1679 (1996); Stephen P. Croley, Vicarious Liability in Tort: On the Sources and Limits of Employee Reasonableness, 69 S. Cal. L. Rev. 1705 (1996); Lewis Kornhauser, An Economic Analysis of the Choice Between Enterprise and Personal Liability for Accidents, 70 Cal. L. Rev. 1345 (1982); Reinier H. Kraakman, Corporate Liability Strategies and the Costs of Legal Controls, 93 Yale L.J. 857 (1984); Jonathan R. Macey, Agency Theory and the Criminal Liability of Organizations, 71 B.U. L. Rev. 315 (1991); A. Mitchell Polinsky & Steven Shavell, Should Employees Be Subject to Fines and Imprisonment Given the Existence of Corporate Liability?, 13 Int'l Rev. L. & Econ. 239 (1993); Gary T. Schwartz, The Hidden and Fundamental Issue of Employer Vicarious Liability, 69 S. Cal. L. Rev. 1739 (1996); Kathleen Segerson & Tom Tietenberg, The Structure of Penalties in Environmental Enforcement: An Economic Analysis, 23 J. Envtl. Econ. & Mgmt. 179 (1992); Alan O. Sykes, The Economics of Vicarious Liability, 93 Yale L.J. 1231 (1984). Note that neither these articles nor our own discussion assumes that the firm is always a literal corporation. Rather, our conclusions extend to any principal-agent relationship. We focus our inquiry - and our terminology - on corporate misconduct, because the corporation is the most common form of the firm as principal.
6. For legal scholarship applying agency cost analysis to the issue of corporate versus individual liability, see generally Christopher Stone, Where the Law Ends: The Social Control of Corporate Behavior (1975); Bruce Chapman, Corporate Tort Liability and the Problem of Overcompliance, 69 S. Cal. L. Rev. 1679 (1996); Stephen P. Croley, Vicarious Liability in Tort: On the Sources and Limits of Employee Reasonableness, 69 S. Cal. L. Rev. 1705 (1996); Lewis Kornhauser, An Economic Analysis of the Choice Between Enterprise and Personal Liability for Accidents, 70 Cal. L. Rev. 1345 (1982); Reinier H. Kraakman, Corporate Liability Strategies and the Costs of Legal Controls, 93 Yale L.J. 857 (1984); Jonathan R. Macey, Agency Theory and the Criminal Liability of Organizations, 71 B.U. L. Rev. 315 (1991); A. Mitchell Polinsky & Steven Shavell, Should Employees Be Subject to Fines and Imprisonment Given the Existence of Corporate Liability?, 13 Int'l Rev. L. & Econ. 239 (1993); Gary T. Schwartz, The Hidden and Fundamental Issue of Employer Vicarious Liability, 69 S. Cal. L. Rev. 1739 (1996); Kathleen Segerson & Tom Tietenberg, The Structure of Penalties in Environmental Enforcement: An Economic Analysis, 23 J. Envtl. Econ. & Mgmt. 179 (1992); Alan O. Sykes, The Economics of Vicarious Liability, 93 Yale L.J. 1231 (1984). Note that neither these articles nor our own discussion assumes that the firm is always a literal corporation. Rather, our conclusions extend to any principal-agent relationship. We focus our inquiry - and our terminology - on corporate misconduct, because the corporation is the most common form of the firm as principal.
7. For legal scholarship applying agency cost analysis to the issue of corporate versus individual liability, see generally Christopher Stone, Where the Law Ends: The Social Control of Corporate Behavior (1975); Bruce Chapman, Corporate Tort Liability and the Problem of Overcompliance, 69 S. Cal. L. Rev. 1679 (1996); Stephen P. Croley, Vicarious Liability in Tort: On the Sources and Limits of Employee Reasonableness, 69 S. Cal. L. Rev. 1705 (1996); Lewis Kornhauser, An Economic Analysis of the Choice Between Enterprise and Personal Liability for Accidents, 70 Cal. L. Rev. 1345 (1982); Reinier H. Kraakman, Corporate Liability Strategies and the Costs of Legal Controls, 93 Yale L.J. 857 (1984); Jonathan R. Macey, Agency Theory and the Criminal Liability of Organizations, 71 B.U. L. Rev. 315 (1991); A. Mitchell Polinsky & Steven Shavell, Should Employees Be Subject to Fines and Imprisonment Given the Existence of Corporate Liability?, 13 Int'l Rev. L. & Econ. 239 (1993); Gary T. Schwartz, The Hidden and Fundamental Issue of Employer Vicarious Liability, 69 S. Cal. L. Rev. 1739 (1996); Kathleen Segerson & Tom Tietenberg, The Structure of Penalties in Environmental Enforcement: An Economic Analysis, 23 J. Envtl. Econ. & Mgmt. 179 (1992); Alan O. Sykes, The Economics of Vicarious Liability, 93 Yale L.J. 1231 (1984). Note that neither these articles nor our own discussion assumes that the firm is always a literal corporation. Rather, our conclusions extend to any principal-agent relationship. We focus our inquiry - and our terminology - on corporate misconduct, because the corporation is the most common form of the firm as principal.
8. For legal scholarship applying agency cost analysis to the issue of corporate versus individual liability, see generally Christopher Stone, Where the Law Ends: The Social Control of Corporate Behavior (1975); Bruce Chapman, Corporate Tort Liability and the Problem of Overcompliance, 69 S. Cal. L. Rev. 1679 (1996); Stephen P. Croley, Vicarious Liability in Tort: On the Sources and Limits of Employee Reasonableness, 69 S. Cal. L. Rev. 1705 (1996); Lewis Kornhauser, An Economic Analysis of the Choice Between Enterprise and Personal Liability for Accidents, 70 Cal. L. Rev. 1345 (1982); Reinier H. Kraakman, Corporate Liability Strategies and the Costs of Legal Controls, 93 Yale L.J. 857 (1984); Jonathan R. Macey, Agency Theory and the Criminal Liability of Organizations, 71 B.U. L. Rev. 315 (1991); A. Mitchell Polinsky & Steven Shavell, Should Employees Be Subject to Fines and Imprisonment Given the Existence of Corporate Liability?, 13 Int'l Rev. L. & Econ. 239 (1993); Gary T. Schwartz, The Hidden and Fundamental Issue of Employer Vicarious Liability, 69 S. Cal. L. Rev. 1739 (1996); Kathleen Segerson & Tom Tietenberg, The Structure of Penalties in Environmental Enforcement: An Economic Analysis, 23 J. Envtl. Econ. & Mgmt. 179 (1992); Alan O. Sykes, The Economics of Vicarious Liability, 93 Yale L.J. 1231 (1984). Note that neither these articles nor our own discussion assumes that the firm is always a literal corporation. Rather, our conclusions extend to any principal-agent relationship. We focus our inquiry - and our terminology - on corporate misconduct, because the corporation is the most common form of the firm as principal.
9. For legal scholarship applying agency cost analysis to the issue of corporate versus individual liability, see generally Christopher Stone, Where the Law Ends: The Social Control of Corporate Behavior (1975); Bruce Chapman, Corporate Tort Liability and the Problem of Overcompliance, 69 S. Cal. L. Rev. 1679 (1996); Stephen P. Croley, Vicarious Liability in Tort: On the Sources and Limits of Employee Reasonableness, 69 S. Cal. L. Rev. 1705 (1996); Lewis Kornhauser, An Economic Analysis of the Choice Between Enterprise and Personal Liability for Accidents, 70 Cal. L. Rev. 1345 (1982); Reinier H. Kraakman, Corporate Liability Strategies and the Costs of Legal Controls, 93 Yale L.J. 857 (1984); Jonathan R. Macey, Agency Theory and the Criminal Liability of Organizations, 71 B.U. L. Rev. 315 (1991); A. Mitchell Polinsky & Steven Shavell, Should Employees Be Subject to Fines and Imprisonment Given the Existence of Corporate Liability?, 13 Int'l Rev. L. & Econ. 239 (1993); Gary T. Schwartz, The Hidden and Fundamental Issue of Employer Vicarious Liability, 69 S. Cal. L. Rev. 1739 (1996); Kathleen Segerson & Tom Tietenberg, The Structure of Penalties in Environmental Enforcement: An Economic Analysis, 23 J. Envtl. Econ. & Mgmt. 179 (1992); Alan O. Sykes, The Economics of Vicarious Liability, 93 Yale L.J. 1231 (1984). Note that neither these articles nor our own discussion assumes that the firm is always a literal corporation. Rather, our conclusions extend to any principal-agent relationship. We focus our inquiry - and our terminology - on corporate misconduct, because the corporation is the most common form of the firm as principal.
10. For legal scholarship applying agency cost analysis to the issue of corporate versus individual liability, see generally Christopher Stone, Where the Law Ends: The Social Control of Corporate Behavior (1975); Bruce Chapman, Corporate Tort Liability and the Problem of Overcompliance, 69 S. Cal. L. Rev. 1679 (1996); Stephen P. Croley, Vicarious Liability in Tort: On the Sources and Limits of Employee Reasonableness, 69 S. Cal. L. Rev. 1705 (1996); Lewis Kornhauser, An Economic Analysis of the Choice Between Enterprise and Personal Liability for Accidents, 70 Cal. L. Rev. 1345 (1982); Reinier H. Kraakman, Corporate Liability Strategies and the Costs of Legal Controls, 93 Yale L.J. 857 (1984); Jonathan R. Macey, Agency Theory and the Criminal Liability of Organizations, 71 B.U. L. Rev. 315 (1991); A. Mitchell Polinsky & Steven Shavell, Should Employees Be Subject to Fines and Imprisonment Given the Existence of Corporate Liability?, 13 Int'l Rev. L. & Econ. 239 (1993); Gary T. Schwartz, The Hidden and Fundamental Issue of Employer Vicarious Liability, 69 S. Cal. L. Rev. 1739 (1996); Kathleen Segerson & Tom Tietenberg, The Structure of Penalties in Environmental Enforcement: An Economic Analysis, 23 J. Envtl. Econ. & Mgmt. 179 (1992); Alan O. Sykes, The Economics of Vicarious Liability, 93 Yale L.J. 1231 (1984). Note that neither these articles nor our own discussion assumes that the firm is always a literal corporation. Rather, our conclusions extend to any principal-agent relationship. We focus our inquiry - and our terminology - on corporate misconduct, because the corporation is the most common form of the firm as principal.
11. For legal scholarship applying agency cost analysis to the issue of corporate versus individual liability, see generally Christopher Stone, Where the Law Ends: The Social Control of Corporate Behavior (1975); Bruce Chapman, Corporate Tort Liability and the Problem of Overcompliance, 69 S. Cal. L. Rev. 1679 (1996); Stephen P. Croley, Vicarious Liability in Tort: On the Sources and Limits of Employee Reasonableness, 69 S. Cal. L. Rev. 1705 (1996); Lewis Kornhauser, An Economic Analysis of the Choice Between Enterprise and Personal Liability for Accidents, 70 Cal. L. Rev. 1345 (1982); Reinier H. Kraakman, Corporate Liability Strategies and the Costs of Legal Controls, 93 Yale L.J. 857 (1984); Jonathan R. Macey, Agency Theory and the Criminal Liability of Organizations, 71 B.U. L. Rev. 315 (1991); A. Mitchell Polinsky & Steven Shavell, Should Employees Be Subject to Fines and Imprisonment Given the Existence of Corporate Liability?, 13 Int'l Rev. L. & Econ. 239 (1993); Gary T. Schwartz, The Hidden and Fundamental Issue of Employer Vicarious Liability, 69 S. Cal. L. Rev. 1739 (1996); Kathleen Segerson & Tom Tietenberg, The Structure of Penalties in Environmental Enforcement: An Economic Analysis, 23 J. Envtl. Econ. & Mgmt. 179 (1992); Alan O. Sykes, The Economics of Vicarious Liability, 93 Yale L.J. 1231 (1984). Note that neither these articles nor our own discussion assumes that the firm is always a literal corporation. Rather, our conclusions extend to any principal-agent relationship. We focus our inquiry - and our terminology - on corporate misconduct, because the corporation is the most common form of the firm as principal.
12. For legal scholarship applying agency cost analysis to the issue of corporate versus individual liability, see generally Christopher Stone, Where the Law Ends: The Social Control of Corporate Behavior (1975); Bruce Chapman, Corporate Tort Liability and the Problem of Overcompliance, 69 S. Cal. L. Rev. 1679 (1996); Stephen P. Croley, Vicarious Liability in Tort: On the Sources and Limits of Employee Reasonableness, 69 S. Cal. L. Rev. 1705 (1996); Lewis Kornhauser, An Economic Analysis of the Choice Between Enterprise and Personal Liability for Accidents, 70 Cal. L. Rev. 1345 (1982); Reinier H. Kraakman, Corporate Liability Strategies and the Costs of Legal Controls, 93 Yale L.J. 857 (1984); Jonathan R. Macey, Agency Theory and the Criminal Liability of Organizations, 71 B.U. L. Rev. 315 (1991); A. Mitchell Polinsky & Steven Shavell, Should Employees Be Subject to Fines and Imprisonment Given the Existence of Corporate Liability?, 13 Int'l Rev. L. & Econ. 239 (1993); Gary T. Schwartz, The Hidden and Fundamental Issue of Employer Vicarious Liability, 69 S. Cal. L. Rev. 1739 (1996); Kathleen Segerson & Tom Tietenberg, The Structure of Penalties in Environmental Enforcement: An Economic Analysis, 23 J. Envtl. Econ. & Mgmt. 179 (1992); Alan O. Sykes, The Economics of Vicarious Liability, 93 Yale L.J. 1231 (1984). Note that neither these articles nor our own discussion assumes that the firm is always a literal corporation. Rather, our conclusions extend to any principal-agent relationship. We focus our inquiry - and our terminology - on corporate misconduct, because the corporation is the most common form of the firm as principal.
13. For legal scholarship applying agency cost analysis to the issue of corporate versus individual liability, see generally Christopher Stone, Where the Law Ends: The Social Control of Corporate Behavior (1975); Bruce Chapman, Corporate Tort Liability and the Problem of Overcompliance, 69 S. Cal. L. Rev. 1679 (1996); Stephen P. Croley, Vicarious Liability in Tort: On the Sources and Limits of Employee Reasonableness, 69 S. Cal. L. Rev. 1705 (1996); Lewis Kornhauser, An Economic Analysis of the Choice Between Enterprise and Personal Liability for Accidents, 70 Cal. L. Rev. 1345 (1982); Reinier H. Kraakman, Corporate Liability Strategies and the Costs of Legal Controls, 93 Yale L.J. 857 (1984); Jonathan R. Macey, Agency Theory and the Criminal Liability of Organizations, 71 B.U. L. Rev. 315 (1991); A. Mitchell Polinsky & Steven Shavell, Should Employees Be Subject to Fines and Imprisonment Given the Existence of Corporate Liability?, 13 Int'l Rev. L. & Econ. 239 (1993); Gary T. Schwartz, The Hidden and Fundamental Issue of Employer Vicarious Liability, 69 S. Cal. L. Rev. 1739 (1996); Kathleen Segerson & Tom Tietenberg, The Structure of Penalties in Environmental Enforcement: An Economic Analysis, 23 J. Envtl. Econ. & Mgmt. 179 (1992); Alan O. Sykes, The Economics of Vicarious Liability, 93 Yale L.J. 1231 (1984). Note that neither these articles nor our own discussion assumes that the firm is always a literal corporation. Rather, our conclusions extend to any principal-agent relationship. We focus our inquiry - and our terminology - on corporate misconduct, because the corporation is the most common form of the firm as principal.
14. For legal scholarship applying agency cost analysis to the issue of corporate versus individual liability, see generally Christopher Stone, Where the Law Ends: The Social Control of Corporate Behavior (1975); Bruce Chapman, Corporate Tort Liability and the Problem of Overcompliance, 69 S. Cal. L. Rev. 1679 (1996); Stephen P. Croley, Vicarious Liability in Tort: On the Sources and Limits of Employee Reasonableness, 69 S. Cal. L. Rev. 1705 (1996); Lewis Kornhauser, An Economic Analysis of the Choice Between Enterprise and Personal Liability for Accidents, 70 Cal. L. Rev. 1345 (1982); Reinier H. Kraakman, Corporate Liability Strategies and the Costs of Legal Controls, 93 Yale L.J. 857 (1984); Jonathan R. Macey, Agency Theory and the Criminal Liability of Organizations, 71 B.U. L. Rev. 315 (1991); A. Mitchell Polinsky & Steven Shavell, Should Employees Be Subject to Fines and Imprisonment Given the Existence of Corporate Liability?, 13 Int'l Rev. L. & Econ. 239 (1993); Gary T. Schwartz, The Hidden and Fundamental Issue of Employer Vicarious Liability, 69 S. Cal. L. Rev. 1739 (1996); Kathleen Segerson & Tom Tietenberg, The Structure of Penalties in Environmental Enforcement: An Economic Analysis, 23 J. Envtl. Econ. & Mgmt. 179 (1992); Alan O. Sykes, The Economics of Vicarious Liability, 93 Yale L.J. 1231 (1984). Note that neither these articles nor our own discussion assumes that the firm is always a literal corporation. Rather, our conclusions extend to any principal-agent relationship. We focus our inquiry - and our terminology - on corporate misconduct, because the corporation is the most common form of the firm as principal.
15. See, e.g., Jennifer Arlen, The Potentially Perverse Effects of Corporate Criminal Liability, 23 J. Legal Stud. 833 (1994); John C. Coffee, Jr., Does "Unlawful" Mean "Criminal"?: The Disappearing Tort/Crime Distinction in American Law, 71 B.U. L. Rev. 193 (1991) (discussing blurring of line between civil and criminal law); David A. Dana, The Perverse Incentives of Environmental Audit Immunity, 81 Iowa L. Rev. 969 (1996) (arguing against environmental audit liability); Daniel R. Fischel & Alan O. Sykes, Corporate Crime, 25 J. Legal Stud. 319 (1996) (arguing for strict corporate liability); Eric W. Orts & Paula C. Murray, Environmental Disclosures and Evidentiary Privilege, 1997 III. L. Rev. 1.
16. See, e.g., Jennifer Arlen, The Potentially Perverse Effects of Corporate Criminal Liability, 23 J. Legal Stud. 833 (1994); John C. Coffee, Jr., Does "Unlawful" Mean "Criminal"?: The Disappearing Tort/Crime Distinction in American Law, 71 B.U. L. Rev. 193 (1991) (discussing blurring of line between civil and criminal law); David A. Dana, The Perverse Incentives of Environmental Audit Immunity, 81 Iowa L. Rev. 969 (1996) (arguing against environmental audit liability); Daniel R. Fischel & Alan O. Sykes, Corporate Crime, 25 J. Legal Stud. 319 (1996) (arguing for strict corporate liability); Eric W. Orts & Paula C. Murray, Environmental Disclosures and Evidentiary Privilege, 1997 III. L. Rev. 1.
17. See, e.g., Jennifer Arlen, The Potentially Perverse Effects of Corporate Criminal Liability, 23 J. Legal Stud. 833 (1994); John C. Coffee, Jr., Does "Unlawful" Mean "Criminal"?: The Disappearing Tort/Crime Distinction in American Law, 71 B.U. L. Rev. 193 (1991) (discussing blurring of line between civil and criminal law); David A. Dana, The Perverse Incentives of Environmental Audit Immunity, 81 Iowa L. Rev. 969 (1996) (arguing against environmental audit liability); Daniel R. Fischel & Alan O. Sykes, Corporate Crime, 25 J. Legal Stud. 319 (1996) (arguing for strict corporate liability); Eric W. Orts & Paula C. Murray, Environmental Disclosures and Evidentiary Privilege, 1997 III. L. Rev. 1.
18. See, e.g., Jennifer Arlen, The Potentially Perverse Effects of Corporate Criminal Liability, 23 J. Legal Stud. 833 (1994); John C. Coffee, Jr., Does "Unlawful" Mean "Criminal"?: The Disappearing Tort/Crime Distinction in American Law, 71 B.U. L. Rev. 193 (1991) (discussing blurring of line between civil and criminal law); David A. Dana, The Perverse Incentives of Environmental Audit Immunity, 81 Iowa L. Rev. 969 (1996) (arguing against environmental audit liability); Daniel R. Fischel & Alan O. Sykes, Corporate Crime, 25 J. Legal Stud. 319 (1996) (arguing for strict corporate liability); Eric W. Orts & Paula C. Murray, Environmental Disclosures and Evidentiary Privilege, 1997 III. L. Rev. 1.
19. See, e.g., Jennifer Arlen, The Potentially Perverse Effects of Corporate Criminal Liability, 23 J. Legal Stud. 833 (1994); John C. Coffee, Jr., Does "Unlawful" Mean "Criminal"?: The Disappearing Tort/Crime Distinction in American Law, 71 B.U. L. Rev. 193 (1991) (discussing blurring of line between civil and criminal law); David A. Dana, The Perverse Incentives of Environmental Audit Immunity, 81 Iowa L. Rev. 969 (1996) (arguing against environmental audit liability); Daniel R. Fischel & Alan O. Sykes, Corporate Crime, 25 J. Legal Stud. 319 (1996) (arguing for strict corporate liability); Eric W. Orts & Paula C. Murray, Environmental Disclosures and Evidentiary Privilege, 1997 III. L. Rev. 1.
20. U.S. Sentencing Guidelines Manual, ch. 8, 393-433 (1993) [hereinafter Sentencing Guidelines].
21. As of the end of 1996, at least 18 states had some form of environmental audit privilege. See Louis M. Brown et al., The Legal Audit: Corporate Internal Investigation, § 9.10[5] (1997) (detailing various state audit privileges); David R. Erickson & Sarah D. Matthews, Environmental Compliance Audits: Analysis of Current Law, Policy and Practical Considerations to Best Protect Their Confidentiality, 63 Mo. L. Rev. 491, 517 (1995); infra Part IV (discussing environmental audit privileges).
22. As of the end of 1996, at least 18 states had some form of environmental audit privilege. See Louis M. Brown et al., The Legal Audit: Corporate Internal Investigation, § 9.10[5] (1997) (detailing various state audit privileges); David R. Erickson & Sarah D. Matthews, Environmental Compliance Audits: Analysis of Current Law, Policy and Practical Considerations to Best Protect Their Confidentiality, 63 Mo. L. Rev. 491, 517 (1995); infra Part IV (discussing environmental audit privileges).
23. U.S. Environmental Protection Agency, Final Policy Statement, Incentives for Self-Policing: Discovery, Disclosure, Correction and Prevention of Violations, 60 Fed. Reg. 66,706 (1995) [hereinafter EPA Guidelines]; Antitrust Division, U.S. Dep't of Justice, Corporate Leniency Policy (Aug. 10, 1993), 328 Trade Reg. Rep. 20,649-21, at ¶ 13,113 (1994) [hereinafter Antitrust Division Guidelines].
24. See Richard Gruner, Corporate Crime and Sentencing § 8.5.2 (Supp. 1995).
25. See, e.g., Fischel & Sykes, supra note 5, at 327-30; Polinsky & Shavell, supra note 4, at 250-53.
26. See, e.g., Fischel & Sykes, supra note 5, at 327-30; Polinsky & Shavell, supra note 4, at 250-53.
27. In a later paper we will consider a second class of incentive regimes that reach inside the firm to structure the incentives of corporate managers or employees directly. We term these incentives "targeted incentives." See Jennifer Arlen & Reinier Kraakman, Controlling Corporate Misconduct: The Role of Supervisory Incentive Regimes (draft in progress). Corporate liability and targeted incentive regimes are partial substitutes because the enforcement rationale for both regimes is to mobilize the firm's resources to prevent misconduct by subordinate employees.
28. We do not consider the issue of whether this liability should be criminal or civil. For a discussion of this question, see, e.g., Fischel & Sykes, supra note 5 (arguing there is no need for corporate criminal liability in legal system with appropriate civil remedies); V.S. Khanna, Corporate Criminal Liability: What Purpose Does it Serve?, 109 Harv. L. Rev. 1477 (1996) (favoring corporate civil liability over corporate criminal liability); Jeffrey S. Parker, Doctrine for Destruction: The Case of Corporate Criminal Liability, 17 Managerial and Decision Econ. 381 (1996) (same).
29. We do not consider the issue of whether this liability should be criminal or civil. For a discussion of this question, see, e.g., Fischel & Sykes, supra note 5 (arguing there is no need for corporate criminal liability in legal system with appropriate civil remedies); V.S. Khanna, Corporate Criminal Liability: What Purpose Does it Serve?, 109 Harv. L. Rev. 1477 (1996) (favoring corporate civil liability over corporate criminal liability); Jeffrey S. Parker, Doctrine for Destruction: The Case of Corporate Criminal Liability, 17 Managerial and Decision Econ. 381 (1996) (same).
30. Although we focus on intentional wrongdoing, our conclusions about the basic structure of an optimal corporate liability regime apply as well to unintentional wrongs when liability for the underlying activity is governed by a strict liability rule. Optimal sanctions may differ from liability targeting intentional misconduct, however, because firms are likely to bear their agents' expected individual liability for unintentional wrongdoing in the form of higher wages. See infra note 27. Also, application of our analysis to private actions would raise the further issue, not addressed here, of ensuring that damages do not induce insufficient caretaking by victims or frivolous or inefficient lawsuits.
31. Corporate liability may serve other aims. Deterrence, however, is generally recognized to be the central goal of corporate liability. Moreover, to the extent policymakers wish to pursue other aims, our analysis reveals when pursuit of these aims comes at the expense of increased corporate wrongdoing.
32. See infra text accompanying notes 20-23.
33. See, e.g., Arlen, supra note 5; Polinsky & Shavell, supra note 4; Segerson & Tietenberg, supra note 4.
34. See, e.g., Arlen, supra note 5; Polinsky & Shavell, supra note 4; Segerson & Tietenberg, supra note 4.
35. See, e.g., Arlen, supra note 5; Polinsky & Shavell, supra note 4; Segerson & Tietenberg, supra note 4.
36. See Steven Shavell, Strict Liability Versus Negligence, 9 J. Legal Stud. 1, 2 (1980) (noting negligence rule is inefficient because actors will choose too high an activity level); see also Polinsky & Shavell, supra note 4, at 241, 252 (stressing that duty-based corporate liability will induce excessive activity levels); Alan O. Sykes, The Boundaries of Vicarious Liability: An Economic Analysis of the Scope of Employment Rule and Related Legal Doctrines, 101 Harv. L. Rev. 563, 579-81 (1988) (same).
37. See Steven Shavell, Strict Liability Versus Negligence, 9 J. Legal Stud. 1, 2 (1980) (noting negligence rule is inefficient because actors will choose too high an activity level); see also Polinsky & Shavell, supra note 4, at 241, 252 (stressing that duty-based corporate liability will induce excessive activity levels); Alan O. Sykes, The Boundaries of Vicarious Liability: An Economic Analysis of the Scope of Employment Rule and Related Legal Doctrines, 101 Harv. L. Rev. 563, 579-81 (1988) (same).
38. See Steven Shavell, Strict Liability Versus Negligence, 9 J. Legal Stud. 1, 2 (1980) (noting negligence rule is inefficient because actors will choose too high an activity level); see also Polinsky & Shavell, supra note 4, at 241, 252 (stressing that duty-based corporate liability will induce excessive activity levels); Alan O. Sykes, The Boundaries of Vicarious Liability: An Economic Analysis of the Scope of Employment Rule and Related Legal Doctrines, 101 Harv. L. Rev. 563, 579-81 (1988) (same).
39. Private sanctions imposed by the firm may be superior to state imposed sanctions when the firm can determine guilt more accurately, or has lower administrative and sanctioning costs, and is not more restricted than the state in the sanction it can impose. Cf. Polinsky & Shavell, supra note 4, at 240 (emphasizing that state-imposed sanctioning of employees improves social welfare if employers have limited sanctioning power).
40. See infra Part I.D. and text accompanying note 74. These market forces include the reputational penalties that firms bear when their agents commit certain types of wrongs. See Jonathan M. Karpoff & John R. Lott, Jr., The Reputational Penalty Firms Bear from Committing Criminal Fraud, 36 J.L. & Econ. 757, 758-59 (1993) (analyzing the reputational impact of firm wrongdoing). Indeed, for this reason, entity liability of the right sort actually benefits firms that the market fully punishes for agents' misbehavior by providing the private benefit of a low-cost commitment device.
41. See Gary S. Becker, Crime and Punishment: An Economic Approach, 76 J. Pol. Econ. 169, 190-93 (1968). This Article focuses on intentional wrongdoing. In the case of unintentional wrongs, corporate liability might be justifiable, even in a world of solvent and savvy individuals, for collective torts for which no individual agent could be made personally responsible - although even there, many wrongs probably could be ultimately attributable to company managers, who would be indemnified by their firms.
42. Absent insolvency concerns, the state could deter socially undesirable misconduct by relying solely on individual liability, with the expected sanction set equal to the social cost of the harm. See id. To minimize enforcement costs, the state could impose large sanctions on a relatively small number of wrongdoers who would face a low probability of detection. Since enforcement costs would be de minimis, there would then be no need to consider whether an entity other than the state - for example, the firm - might be able to deter wrongdoing at lower cost. This low-probability - high-sanction strategy will not work, however, if agents lack the wealth to pay large fines. In this case, agents' expected sanctions will be less than the social cost of wrongdoing and too many wrongs will result. As we will discuss, corporate liability can increase agents' expected liability by increasing the probability of detection. Of course, the legal system can also supplement monetary fines with nonmonetary sanctions, such as prison sentences, when agents lack the wealth to pay large monetary fines. Yet these nonmonetary penalties are, at best, only a partial solution to the problem of agent insolvency: they generally do not eliminate the need to make substantial enforcement expenditures to deter wrongdoing. First, imprisonment is very expensive. Many crimes can be deterred more efficiently by increased expenditures on "enforcement." See infra. Second, the use of nonmonetary sanctions is limited by "marginal deterrence" concerns - the state is limited in the sanction it can impose for relatively minor crimes by the need to impose greater sanctions on more serious crimes. Finally, normative considerations other than efficiency may limit the use of nonmonetary sanctions. For example, the state may be unwilling to impose a long jail term on someone who committed a relatively minor crime with a very low probability of detection because it seems unjust. See John Coffee, Jr., "No Soul to Damn; No Body to Kick": An Unscandalized Inquiry Into the Problem of Corporate Punishment, 79 Mich. L. Rev. 386, 401 (1981) (stating that imprisonment imposes externalities); Reinier Kraakman, The Economic Functions of Corporate Liability, in Corporate Governance and Directors' Liabilities 178, 194-97 (Klaus J. Hopt & Gunther Teubner eds., 1985) (discussing constraints on the magnitude of enterprise sanctions). Thus, even when the state can and does use imprisonment as a sanction, it is unlikely to deter all wrongdoing, particularly if enforcement expenditures (and thus the probability of detection) are very small. In addition, the state may need to increase enforcement expenditures (and thus turn to the firm for help) for other reasons. If individuals are risk averse, a low-probability - high-penalty strategy increases the likelihood that an individual accused of wrongdoing will plead guilty to that wrong even though he is in fact innocent. The state can reduce this risk of false convictions by increasing enforcement expenditures, thereby reducing the sanction. See Bruce H. Kobayashi & John R. Lott, Jr., Low-Probability - High-Penalty Enforcement Strategies and the Efficient Operation of the Plea-Bargaining System, 12 Int'l Rev. L. & Econ. 69, 70 (1992) (discussing how large fines may cause innocent people to plead guilty). Once significant enforcement expenditures are required, however, it will often be optimal for the state to induce firms to incur some of these enforcement expenditures by employing corporate liability in addition to individual liability. See infra note 22 and accompanying text.
43. Absent insolvency concerns, the state could deter socially undesirable misconduct by relying solely on individual liability, with the expected sanction set equal to the social cost of the harm. See id. To minimize enforcement costs, the state could impose large sanctions on a relatively small number of wrongdoers who would face a low probability of detection. Since enforcement costs would be de minimis, there would then be no need to consider whether an entity other than the state - for example, the firm - might be able to deter wrongdoing at lower cost. This low-probability - high-sanction strategy will not work, however, if agents lack the wealth to pay large fines. In this case, agents' expected sanctions will be less than the social cost of wrongdoing and too many wrongs will result. As we will discuss, corporate liability can increase agents' expected liability by increasing the probability of detection. Of course, the legal system can also supplement monetary fines with nonmonetary sanctions, such as prison sentences, when agents lack the wealth to pay large monetary fines. Yet these nonmonetary penalties are, at best, only a partial solution to the problem of agent insolvency: they generally do not eliminate the need to make substantial enforcement expenditures to deter wrongdoing. First, imprisonment is very expensive. Many crimes can be deterred more efficiently by increased expenditures on "enforcement." See infra. Second, the use of nonmonetary sanctions is limited by "marginal deterrence" concerns - the state is limited in the sanction it can impose for relatively minor crimes by the need to impose greater sanctions on more serious crimes. Finally, normative considerations other than efficiency may limit the use of nonmonetary sanctions. For example, the state may be unwilling to impose a long jail term on someone who committed a relatively minor crime with a very low probability of detection because it seems unjust. See John Coffee, Jr., "No Soul to Damn; No Body to Kick": An Unscandalized Inquiry Into the Problem of Corporate Punishment, 79 Mich. L. Rev. 386, 401 (1981) (stating that imprisonment imposes externalities); Reinier Kraakman, The Economic Functions of Corporate Liability, in Corporate Governance and Directors' Liabilities 178, 194-97 (Klaus J. Hopt & Gunther Teubner eds., 1985) (discussing constraints on the magnitude of enterprise sanctions). Thus, even when the state can and does use imprisonment as a sanction, it is unlikely to deter all wrongdoing, particularly if enforcement expenditures (and thus the probability of detection) are very small. In addition, the state may need to increase enforcement expenditures (and thus turn to the firm for help) for other reasons. If individuals are risk averse, a low-probability -high-penalty strategy increases the likelihood that an individual accused of wrongdoing will plead guilty to that wrong even though he is in fact innocent. The state can reduce this risk of false convictions by increasing enforcement expenditures, thereby reducing the sanction. See Bruce H. Kobayashi & John R. Lott, Jr., Low-Probability - High-Penalty Enforcement Strategies and the Efficient Operation of the Plea-Bargaining System, 12 Int'l Rev. L. & Econ. 69, 70 (1992) (discussing how large fines may cause innocent people to plead guilty). Once significant enforcement expenditures are required, however, it will often be optimal for the state to induce firms to incur some of these enforcement expenditures by employing corporate liability in addition to individual liability. See infra note 22 and accompanying text.
44. Absent insolvency concerns, the state could deter socially undesirable misconduct by relying solely on individual liability, with the expected sanction set equal to the social cost of the harm. See id. To minimize enforcement costs, the state could impose large sanctions on a relatively small number of wrongdoers who would face a low probability of detection. Since enforcement costs would be de minimis, there would then be no need to consider whether an entity other than the state - for example, the firm - might be able to deter wrongdoing at lower cost. This low-probability - high-sanction strategy will not work, however, if agents lack the wealth to pay large fines. In this case, agents' expected sanctions will be less than the social cost of wrongdoing and too many wrongs will result. As we will discuss, corporate liability can increase agents' expected liability by increasing the probability of detection. Of course, the legal system can also supplement monetary fines with nonmonetary sanctions, such as prison sentences, when agents lack the wealth to pay large monetary fines. Yet these nonmonetary penalties are, at best, only a partial solution to the problem of agent insolvency: they generally do not eliminate the need to make substantial enforcement expenditures to deter wrongdoing. First, imprisonment is very expensive. Many crimes can be deterred more efficiently by increased expenditures on "enforcement." See infra. Second, the use of nonmonetary sanctions is limited by "marginal deterrence" concerns - the state is limited in the sanction it can impose for relatively minor crimes by the need to impose greater sanctions on more serious crimes. Finally, normative considerations other than efficiency may limit the use of nonmonetary sanctions. For example, the state may be unwilling to impose a long jail term on someone who committed a relatively minor crime with a very low probability of detection because it seems unjust. See John Coffee, Jr., "No Soul to Damn; No Body to Kick": An Unscandalized Inquiry Into the Problem of Corporate Punishment, 79 Mich. L. Rev. 386, 401 (1981) (stating that imprisonment imposes externalities); Reinier Kraakman, The Economic Functions of Corporate Liability, in Corporate Governance and Directors' Liabilities 178, 194-97 (Klaus J. Hopt & Gunther Teubner eds., 1985) (discussing constraints on the magnitude of enterprise sanctions). Thus, even when the state can and does use imprisonment as a sanction, it is unlikely to deter all wrongdoing, particularly if enforcement expenditures (and thus the probability of detection) are very small. In addition, the state may need to increase enforcement expenditures (and thus turn to the firm for help) for other reasons. If individuals are risk averse, a low-probability - high-penalty strategy increases the likelihood that an individual accused of wrongdoing will plead guilty to that wrong even though he is in fact innocent. The state can reduce this risk of false convictions by increasing enforcement expenditures, thereby reducing the sanction. See Bruce H. Kobayashi & John R. Lott, Jr., Low-Probability - High-Penalty Enforcement Strategies and the Efficient Operation of the Plea-Bargaining System, 12 Int'l Rev. L. & Econ. 69, 70 (1992) (discussing how large fines may cause innocent people to plead guilty). Once significant enforcement expenditures are required, however, it will often be optimal for the state to induce firms to incur some of these enforcement expenditures by employing corporate liability in addition to individual liability. See infra note 22 and accompanying text.
45. See Steven Shavell, Economic Analysis of Accident Law 172-74 (1987) (discussing benefit of imposing vicarious liability on firm for its employees' actions when firms can better identify wrongdoers and evaluate their actions); Stephen Salzburg, The Control of Criminal Conduct in Organizations, 71 B.U. L. Rev. 421, 428 (1991) (same). Sanctioning costs justify corporate liability only if the sanction the state would need to impose if it spent virtually nothing on enforcement exceeds agents' wealth. Absent wealth constraints, sanctioning costs would not justify corporate liability because the government could optimally deter wrongdoing by spending almost nothing on enforcement and imposing large sanctions on those few individual wrongdoers that it managed to catch. This low-probability - high-penalty strategy will not necessarily work if agents are insolvent, however. In this case, private sanctioning reduces enforcement costs if it is less expensive and as effective as public sanctioning. The benefit of lower cost and more frequent sanctions would be even greater if, as evidence suggests, individuals are not rational utility maximizers, but rather are more deterred by a high probability of a relatively low sanction than a low probability of a very high sanction. See James Q. Wilson & Richard J. Herrnstein, Crime and Human Nature 397-401 (1985).
46. See Steven Shavell, Economic Analysis of Accident Law 172-74 (1987) (discussing benefit of imposing vicarious liability on firm for its employees' actions when firms can better identify wrongdoers and evaluate their actions); Stephen Salzburg, The Control of Criminal Conduct in Organizations, 71 B.U. L. Rev. 421, 428 (1991) (same). Sanctioning costs justify corporate liability only if the sanction the state would need to impose if it spent virtually nothing on enforcement exceeds agents' wealth. Absent wealth constraints, sanctioning costs would not justify corporate liability because the government could optimally deter wrongdoing by spending almost nothing on enforcement and imposing large sanctions on those few individual wrongdoers that it managed to catch. This low-probability - high-penalty strategy will not necessarily work if agents are insolvent, however. In this case, private sanctioning reduces enforcement costs if it is less expensive and as effective as public sanctioning. The benefit of lower cost and more frequent sanctions would be even greater if, as evidence suggests, individuals are not rational utility maximizers, but rather are more deterred by a high probability of a relatively low sanction than a low probability of a very high sanction. See James Q. Wilson & Richard J. Herrnstein, Crime and Human Nature 397-401 (1985).
47. See Steven Shavell, Economic Analysis of Accident Law 172-74 (1987) (discussing benefit of imposing vicarious liability on firm for its employees' actions when firms can better identify wrongdoers and evaluate their actions); Stephen Salzburg, The Control of Criminal Conduct in Organizations, 71 B.U. L. Rev. 421, 428 (1991) (same). Sanctioning costs justify corporate liability only if the sanction the state would need to impose if it spent virtually nothing on enforcement exceeds agents' wealth. Absent wealth constraints, sanctioning costs would not justify corporate liability because the government could optimally deter wrongdoing by spending almost nothing on enforcement and imposing large sanctions on those few individual wrongdoers that it managed to catch. This low-probability - high-penalty strategy will not necessarily work if agents are insolvent, however. In this case, private sanctioning reduces enforcement costs if it is less expensive and as effective as public sanctioning. The benefit of lower cost and more frequent sanctions would be even greater if, as evidence suggests, individuals are not rational utility maximizers, but rather are more deterred by a high probability of a relatively low sanction than a low probability of a very high sanction. See James Q. Wilson & Richard J. Herrnstein, Crime and Human Nature 397-401 (1985).
48. For example, it appears that, holding the expected sanction constant, individuals are deterred more by a high probability of paying a relatively low fine than the relatively low probability of paying a high fine. See, e.g., Wilson & Herrnstein, supra note 22, at 397-401. This might justify imposing corporate liability to induce firms to raise the probability of detection, even if it would not be justifiable were individuals risk neutral and utility maximizers.
49. Corporate liability is not the only third-party liability regime capable of achieving these goals. Other possibilities include regimes designed to induce third parties within the firm to monitor firm agents and report agents, such as supervisory liability and bounty regimes, see Arlen & Kraakman, supra note 11; Jennifer Arlen, Commentary on Rewarding Whistleblowers: The Costs and Benefits of an Incentive-Based Compliance Strategy, in Corporate Decisionmaking in Canada 635 (Ronald J. Daniels & Randall Morck eds., 1995); Ronald J. Daniels & Robert Howse, Rewarding Whistleblowers: The Costs and Benefits of an Incentive-Based Compliance Strategy, in Daniels & Morck, supra, at 525-49; and those designed to induce outsiders to monitor and report, such as accountant liability, see Reinier H. Kraakman, Gatekeepers: The Anatomy of a Third-Party Enforcement Strategy, 2 J.L. Econ. & Org. 53, 53-54 (1986) (discussing "gatekeeper liability" -"liability imposed on private parties who are able to disrupt misconduct by withholding their cooperation from wrongdoers").
50. Corporate liability is not the only third-party liability regime capable of achieving these goals. Other possibilities include regimes designed to induce third parties within the firm to monitor firm agents and report agents, such as supervisory liability and bounty regimes, see Arlen & Kraakman, supra note 11; Jennifer Arlen, Commentary on Rewarding Whistleblowers: The Costs and Benefits of an Incentive-Based Compliance Strategy, in Corporate Decisionmaking in Canada 635 (Ronald J. Daniels & Randall Morck eds., 1995); Ronald J. Daniels & Robert Howse, Rewarding Whistleblowers: The Costs and Benefits of an Incentive-Based Compliance Strategy, in Daniels & Morck, supra, at 525-49; and those designed to induce outsiders to monitor and report, such as accountant liability, see Reinier H. Kraakman, Gatekeepers: The Anatomy of a Third-Party Enforcement Strategy, 2 J.L. Econ. & Org. 53, 53-54 (1986) (discussing "gatekeeper liability" -"liability imposed on private parties who are able to disrupt misconduct by withholding their cooperation from wrongdoers").
51. Corporate liability is not the only third-party liability regime capable of achieving these goals. Other possibilities include regimes designed to induce third parties within the firm to monitor firm agents and report agents, such as supervisory liability and bounty regimes, see Arlen & Kraakman, supra note 11; Jennifer Arlen, Commentary on Rewarding Whistleblowers: The Costs and Benefits of an Incentive-Based Compliance Strategy, in Corporate Decisionmaking in Canada 635 (Ronald J. Daniels & Randall Morck eds., 1995); Ronald J. Daniels & Robert Howse, Rewarding Whistleblowers: The Costs and Benefits of an Incentive-Based Compliance Strategy, in Daniels & Morck, supra, at 525-49; and those designed to induce outsiders to monitor and report, such as accountant liability, see Reinier H. Kraakman, Gatekeepers: The Anatomy of a Third-Party Enforcement Strategy, 2 J.L. Econ. & Org. 53, 53-54 (1986) (discussing "gatekeeper liability" -"liability imposed on private parties who are able to disrupt misconduct by withholding their cooperation from wrongdoers").
52. Corporate liability is not the only third-party liability regime capable of achieving these goals. Other possibilities include regimes designed to induce third parties within the firm to monitor firm agents and report agents, such as supervisory liability and bounty regimes, see Arlen & Kraakman, supra note 11; Jennifer Arlen, Commentary on Rewarding Whistleblowers: The Costs and Benefits of an Incentive-Based Compliance Strategy, in Corporate Decisionmaking in Canada 635 (Ronald J. Daniels & Randall Morck eds., 1995); Ronald J. Daniels & Robert Howse, Rewarding Whistleblowers: The Costs and Benefits of an Incentive-Based Compliance Strategy, in Daniels & Morck, supra, at 525-49; and those designed to induce outsiders to monitor and report, such as accountant liability, see Reinier H. Kraakman, Gatekeepers: The Anatomy of a Third-Party Enforcement Strategy, 2 J.L. Econ. & Org. 53, 53-54 (1986) (discussing "gatekeeper liability" -"liability imposed on private parties who are able to disrupt misconduct by withholding their cooperation from wrongdoers").
53. Alternatively, in theory the state could employ payment regimes, which grant rewards to firms to induce the desired behavior. Like liability regimes, payment regimes may be strict (outcome-based) or duty-based. In theory, payment regimes are functional substi-tutes for liability regimes in many respects. Moreover, payment regimes - such as bounty regimes - are currently employed. See, e.g., Securities Exchange Act of 1934 § 21A(e), 15 U.S.C. § 78u-1(e) (1994) (bounty provision for information on insider trading); False Claims Act, 31 U.S.C. § 3730(d) (1994) (qui tam provisions). Nevertheless, for several reasons payment regimes are not plausible entity-level incentive regimes. Cf. infra note 30 (discussing targeted incentive regimes). To begin, payment regimes are very expensive to administer because every firm would have to receive a properly determined payment regardless of whether a wrong occurred. Moreover, this regime would impose additional social costs if the government collected the revenues required for the payment through a suboptimal tax system. In addition, rewarding firms for thwarting misconduct or discharging enforcement duties would distort activity levels, at least in circumstances where intentional misconduct is appropriately treated as a production cost, because such a regime would not ensure the demise of firms that create excessive risks of wrongdoing; indeed, it might even keep alive firms that are inefficient for other reasons. Finally, payment incentives offered to firms are open to a peculiar kind of moral hazard, especially when agent misconduct benefits firms as well. Firms that could earn rewards by providing enforcement services, such as reporting the wrongdoing of their own agents, might induce misconduct in the hope of benefitting once or even twice - first from the misconduct itself, and subsequently from reporting it.
54. For analysis of which harms can be said to be "caused" by a firm for purposes of cost internalization, see Sykes, supra note 17, at 571-81 (discussing "enterprise causation," which relates existence of business to wrongs by employees). Professor Sykes observes that a harm can be said to be "fully caused" by an enterprise where dissolution of the enterprise would reduce its probability of occurrence to zero. "Partial causation" is defined similarly as a partial reduction in the probability of a harm's occurrence following dissolution of the enterprise. See id. at 572.
55. See Polinsky & Shavell, supra note 4, at 240; Shavell, supra note 17, at 3-4. Strict vicarious liability can induce optimal activity levels if agents are strictly liable for the underlying wrong, as they are for intentional misconduct. By contrast, this rule does not induce efficient activity levels when the underlying activity is governed by a negligence standard. In this case, firms escape liability if agents take due care and activity levels are too high. Precisely for this reason, Polinsky and Shavell propose expanding vicarious liability to make the firm strictly liable for harms resulting from activities that are governed by a negligence standard for purposes of determining individual liability, even if the agent was not negligent. See Polinsky & Shavell, supra note 4, at 252-53 (arguing that "it is socially desirable to make firms liable according to a strict liability rule but to impose employee sanctions according to a negligence rule"). Expected corporate liability for intentional misconduct must equal the full social cost of wrongdoing to others even if employees also are held liable because, absent a risk of court error, a firm only bears its own expected liability for these wrongs. A firm will not compensate employees for their expected liability for intentional wrongs if the firm does not benefit from the wrong (because it is liable for the full social cost) and the employee can prevent the wrong. See Arlen, supra note 5, at 852 n.59 (noting that "corporations will not compensate agents ex ante for their expected criminal liability from" intentional wrongful acts that firms do not want committed). By contrast, in the case of accidental harms, a firm must reimburse employees for their expected liability, either ex ante through higher wages or ex post by indemnifying them. Thus, the firm will undertake optimal activity levels if its expected sanction equals the social cost of the harm minus any expected employee liability. See, e.g., Polinsky & Shavell, supra note 4, at 241 (noting that firms effectively pay their employees' liability for unintentional wrongdoing through higher wages).
56. See Polinsky & Shavell, supra note 4, at 240; Shavell, supra note 17, at 3-4. Strict vicarious liability can induce optimal activity levels if agents are strictly liable for the underlying wrong, as they are for intentional misconduct. By contrast, this rule does not induce efficient activity levels when the underlying activity is governed by a negligence standard. In this case, firms escape liability if agents take due care and activity levels are too high. Precisely for this reason, Polinsky and Shavell propose expanding vicarious liability to make the firm strictly liable for harms resulting from activities that are governed by a negligence standard for purposes of determining individual liability, even if the agent was not negligent. See Polinsky & Shavell, supra note 4, at 252-53 (arguing that "it is socially desirable to make firms liable according to a strict liability rule but to impose employee sanctions according to a negligence rule"). Expected corporate liability for intentional misconduct must equal the full social cost of wrongdoing to others even if employees also are held liable because, absent a risk of court error, a firm only bears its own expected liability for these wrongs. A firm will not compensate employees for their expected liability for intentional wrongs if the firm does not benefit from the wrong (because it is liable for the full social cost) and the employee can prevent the wrong. See Arlen, supra note 5, at 852 n.59 (noting that "corporations will not compensate agents ex ante for their expected criminal liability from" intentional wrongful acts that firms do not want committed). By contrast, in the case of accidental harms, a firm must reimburse employees for their expected liability, either ex ante through higher wages or ex post by indemnifying them. Thus, the firm will undertake optimal activity levels if its expected sanction equals the social cost of the harm minus any expected employee liability. See, e.g., Polinsky & Shavell, supra note 4, at 241 (noting that firms effectively pay their employees' liability for unintentional wrongdoing through higher wages).
57. See Polinsky & Shavell, supra note 4, at 240; Shavell, supra note 17, at 3-4. Strict vicarious liability can induce optimal activity levels if agents are strictly liable for the underlying wrong, as they are for intentional misconduct. By contrast, this rule does not induce efficient activity levels when the underlying activity is governed by a negligence standard. In this case, firms escape liability if agents take due care and activity levels are too high. Precisely for this reason, Polinsky and Shavell propose expanding vicarious liability to make the firm strictly liable for harms resulting from activities that are governed by a negligence standard for purposes of determining individual liability, even if the agent was not negligent. See Polinsky & Shavell, supra note 4, at 252-53 (arguing that "it is socially desirable to make firms liable according to a strict liability rule but to impose employee sanctions according to a negligence rule"). Expected corporate liability for intentional misconduct must equal the full social cost of wrongdoing to others even if employees also are held liable because, absent a risk of court error, a firm only bears its own expected liability for these wrongs. A firm will not compensate employees for their expected liability for intentional wrongs if the firm does not benefit from the wrong (because it is liable for the full social cost) and the employee can prevent the wrong. See Arlen, supra note 5, at 852 n.59 (noting that "corporations will not compensate agents ex ante for their expected criminal liability from" intentional wrongful acts that firms do not want committed). By contrast, in the case of accidental harms, a firm must reimburse employees for their expected liability, either ex ante through higher wages or ex post by indemnifying them. Thus, the firm will undertake optimal activity levels if its expected sanction equals the social cost of the harm minus any expected employee liability. See, e.g., Polinsky & Shavell, supra note 4, at 241 (noting that firms effectively pay their employees' liability for unintentional wrongdoing through higher wages).
58. See Polinsky & Shavell, supra note 4, at 240; Shavell, supra note 17, at 3-4. Strict vicarious liability can induce optimal activity levels if agents are strictly liable for the underlying wrong, as they are for intentional misconduct. By contrast, this rule does not induce efficient activity levels when the underlying activity is governed by a negligence standard. In this case, firms escape liability if agents take due care and activity levels are too high. Precisely for this reason, Polinsky and Shavell propose expanding vicarious liability to make the firm strictly liable for harms resulting from activities that are governed by a negligence standard for purposes of determining individual liability, even if the agent was not negligent. See Polinsky & Shavell, supra note 4, at 252-53 (arguing that "it is socially desirable to make firms liable according to a strict liability rule but to impose employee sanctions according to a negligence rule"). Expected corporate liability for intentional misconduct must equal the full social cost of wrongdoing to others even if employees also are held liable because, absent a risk of court error, a firm only bears its own expected liability for these wrongs. A firm will not compensate employees for their expected liability for intentional wrongs if the firm does not benefit from the wrong (because it is liable for the full social cost) and the employee can prevent the wrong. See Arlen, supra note 5, at 852 n.59 (noting that "corporations will not compensate agents ex ante for their expected criminal liability from" intentional wrongful acts that firms do not want committed). By contrast, in the case of accidental harms, a firm must reimburse employees for their expected liability, either ex ante through higher wages or ex post by indemnifying them. Thus, the firm will undertake optimal activity levels if its expected sanction equals the social cost of the harm minus any expected employee liability. See, e.g., Polinsky & Shavell, supra note 4, at 241 (noting that firms effectively pay their employees' liability for unintentional wrongdoing through higher wages).
59. Cost internalization also promotes optimal activity levels by ensuring the demise of those firms whose activities produce excessive costs, once the cost of expected wrongdoing is taken into account. See Shavell, supra note 17, at 3-4.
60. See supra note 21.
61. The assumption that shareholders control the firm's enforcement policy (directly or indirectly) is reasonable when intentional wrongdoing is committed by agents of closely held firms, since the shareholders of these firms exercise managerial power directly. Cf. Mark Cohen, Corporate Crime and Punishment: An Update on Sentencing Practices in the Federal Courts, 71 B.U. L. Rev. 247, 251-52 (1991) (more than 95% of firms convicted between 1984 and 1988 were closely held). Shareholders of publicly held firms may be less able to rely on managers to implement optimal preventive and policing measures because managers bear much of the cost of prevention and policing but do not directly bear the firm's expected liability for any wrongdoing that occurs. Thus, in these cases all corporate liability regimes become less attractive on the margin. Notwithstanding these agency costs, however, our analysis should apply to the proper design of corporate liability regimes for publicly held firms. Managers of publicly held firms do place considerable weight on firm profits, in which case firms generally will behave as we describe. In addition, to the extent that management discretion exists, it should not affect the optimal design of corporate liability regimes - only their effectiveness. Solving the agency cost problem will not require a better corporate liability regime, although it may require other enforcement regimes, including targeted incentives aimed at individuals in the firm, such as supervisory liability or bounty provisions for employees who report wrongdoing ("targeted incentives"). See Arlen & Kraakman, supra note 11.
62. See supra note 22. Inducing firms to sanction privately is entirely consistent with sound enforcement policy, since an effective corporate liability regime should also induce firms to report wrongdoing. Thus, the government also can impose a public sanction on wrongful agents if this is appropriate. Firms should be able to sanction their agents because they bear the costs of misconduct. Moreover, firms cannot impose private sanctions that exceed those permitted by law because agents have no incentive to agree to excessive sanctions. Finally, in the case of unintentional wrongs, market forces will ensure that firms do not impose excessive sanctions since wages will reflect workers' expected liability. See supra note 27.
63. Of course, strict liability could generate perverse incentives if corporate sanctioning increases the probability that firms themselves will be held liable. See infra Part I.C. Private sanctioning will not affect the firm's probability of detection in many important situations - for example, if the firm's responsibility for the wrong is obvious (and liability is certain), even if the identities of its culpable agents are not. As long as the firm's efforts to sanction wrongful employees do not affect its probability of facing liability, there is no danger of perverse incentives. If sanctioning does affect the firm's probability of being found liable, then the issues we raise concerning inducing optimal policing measures will apply. See infra Parts I.C. & D.
64. Although we distinguish between prevention measures that do not affect the probability of detection and policing measures that do, we recognize that many measures are both preventive and policing measures. To the extent a prevention measure also affects the probability of detection, it is, for our purposes, partially a policing measure and our discussion of the problems of inducing policing measures will apply. See infra Parts I.C. & D. Because the distinguishing feature of policing measures is whether they affect the probability the firm is detected, when the firm's own liability for a harm is clear, the firm's efforts to determine which agent committed the wrong can properly be treated as a prevention measure, because these efforts will not affect the firm's expected liability.
65. For development of the gatekeeper metaphor in the context of official, as distinct from private, enforcement measures, see Kraakman, supra note 4, at 888-96 (discussing gatekeeper liability for outsiders who can discover and prevent wrongdoing). A gate-keeper interdicts misconduct by withholding critical approval or support ex ante. While gatekeepers who undertake extensive monitoring might also increase the probability that wrongdoing will be detected ex post, many internal gatekeeper strategies - including those listed in the text - are unlikely to increase the probability of detecting misconduct ex post. To the extent they do, however, our discussion of policing measures applies.
66. Empirical evidence suggests that the incidence of certain corporate crimes is higher when agents' compensation or performance evaluations are based largely on their employers' rate of return or short-run profits, as opposed to long-run profits. See Mark A. Cohen & Sally S. Simpson, The Origins of Corporate Criminality: Rational Individual and Organizational Actors, in Debating Corporate Crime: An Interdisciplinary Examination of the Causes and Control of Corporate Misconduct 33 (William S. Lofquist et al. eds., 1997); Charles W. L. Hill et al., An Empirical Examination of the Causes of Corporate Wrongdoing in the United States, 45 Hum. Rel. 1055, 1069-70 (1992); John R. Lott, Jr. & Tim C. Opler, Testing Whether Predatory Commitments Are Credible, 69 J. Bus. 339, 367 (1996) (concluding that firms focusing on short-run profits are more likely to be accused of predatory pricing). Even shareholders of publicly held firms (particularly institutional investors) probably can often obtain sufficient information about a firm's compensation policies to determine whether these policies encourage or deter misconduct.
67. Empirical evidence suggests that the incidence of certain corporate crimes is higher when agents' compensation or performance evaluations are based largely on their employers' rate of return or short-run profits, as opposed to long-run profits. See Mark A. Cohen & Sally S. Simpson, The Origins of Corporate Criminality: Rational Individual and Organizational Actors, in Debating Corporate Crime: An Interdisciplinary Examination of the Causes and Control of Corporate Misconduct 33 (William S. Lofquist et al. eds., 1997); Charles W. L. Hill et al., An Empirical Examination of the Causes of Corporate Wrongdoing in the United States, 45 Hum. Rel. 1055, 1069-70 (1992); John R. Lott, Jr. & Tim C. Opler, Testing Whether Predatory Commitments Are Credible, 69 J. Bus. 339, 367 (1996) (concluding that firms focusing on short-run profits are more likely to be accused of predatory pricing). Even shareholders of publicly held firms (particularly institutional investors) probably can often obtain sufficient information about a firm's compensation policies to determine whether these policies encourage or deter misconduct.
68. Empirical evidence suggests that the incidence of certain corporate crimes is higher when agents' compensation or performance evaluations are based largely on their employers' rate of return or short-run profits, as opposed to long-run profits. See Mark A. Cohen & Sally S. Simpson, The Origins of Corporate Criminality: Rational Individual and Organizational Actors, in Debating Corporate Crime: An Interdisciplinary Examination of the Causes and Control of Corporate Misconduct 33 (William S. Lofquist et al. eds., 1997); Charles W. L. Hill et al., An Empirical Examination of the Causes of Corporate Wrongdoing in the United States, 45 Hum. Rel. 1055, 1069-70 (1992); John R. Lott, Jr. & Tim C. Opler, Testing Whether Predatory Commitments Are Credible, 69 J. Bus. 339, 367 (1996) (concluding that firms focusing on short-run profits are more likely to be accused of predatory pricing). Even shareholders of publicly held firms (particularly institutional investors) probably can often obtain sufficient information about a firm's compensation policies to determine whether these policies encourage or deter misconduct.
69. See Gary S. Becker & George J. Stigler, Law Enforcement, Malfeasance, and Compensation of Enforcers, 3 J. Legal Stud. 1, 6-13 (1974);
70. Steven Shavell, The Optimal Level of Corporate Liability Given the Limited Ability of Corporations to Penalize Their Employees, 17 Int'l Rev. L. & Econ. 203, 204 (1997). Indeed, it might appear that firms could prevent wrongdoing entirely - thereby eliminating the need for enforcement measures - by paying supercompensatory wages, since in theory these wages can increase an agent's wealth enough to eliminate the insolvency problem. Despite their initial theoretical appeal, however, supercompensatory wages cannot be relied upon exclusively to solve the problem of corporate wrongdoing. First, supercompensatory wages are expensive because they must be paid to all agents engaged in particular activities who do not commit a wrong; other prevention and enforcement mechanisms may prove to be more effective.
71. See Becker & Stigler, supra, at 13-16;
72. William T. Dickens et al., Employee Crime and the Monitoring Puzzle, 7 J. Lab. Econ. 331, 343-44 (1989);
73. B. Curtis Eaton & William D. White, Agent Compensation and the Limits of Bonding, 20 Econ. Inquiry 330, 342 (1982). Second, they will not deter wrongdoing motivated by an agent's fear of impending job loss - for example, fraud concerning the stock price of publicly held firms - because agents who will lose their jobs if they do not engage in misconduct are not going to be deterred by the risk of losing the supercompensatory wage should they commit the crime and get caught.
74. See Jennifer Arlen & William Carney, Vicarious Liability for Fraud on Securities Markets: Theory and Evidence, 1992 Ill. L. Rev. 691, 708-09. Nevertheless, the possibility of super-compensatory wages may affect the optimal residual sanction. See Shavell, supra, at 203-04.
75. See Robert Cooter, Prices and Sanctions, 84 Colum. L. Rev. 1523, 1539 (1984); Shavell, supra note 17. The standard result is that strict liability, with an expected sanction set equal to the social cost of the harm, can induce an actor to take due care, where "due care" here is defined as prevention measures designed to deter wrongdoing. Cf. supra note 30 and accompanying text (discussing agency costs). See generally Shavell, supra note 22, at 5-46.
76. See Robert Cooter, Prices and Sanctions, 84 Colum. L. Rev. 1523, 1539 (1984); Shavell, supra note 17. The standard result is that strict liability, with an expected sanction set equal to the social cost of the harm, can induce an actor to take due care, where "due care" here is defined as prevention measures designed to deter wrongdoing. Cf. supra note 30 and accompanying text (discussing agency costs). See generally Shavell, supra note 22, at 5-46.
77. See Robert Cooter, Prices and Sanctions, 84 Colum. L. Rev. 1523, 1539 (1984); Shavell, supra note 17. The standard result is that strict liability, with an expected sanction set equal to the social cost of the harm, can induce an actor to take due care, where "due care" here is defined as prevention measures designed to deter wrongdoing. Cf. supra note 30 and accompanying text (discussing agency costs). See generally Shavell, supra note 22, at 5-46.
78. As can be the case under a composite regime.
79. Here, as elsewhere, wrongdoing is defined as conduct for which the marginal social benefit is less than the marginal social cost.
80. Even when the agent benefits from a wrong primarily as a result of the effect of the wrong on the firm's profits, he will not necessarily be deterred from misconduct by a corporate liability regime that ensures that the firm bears the full social cost of wrongdoing. A firm's compensation and promotion policies may reward employees when the firm's short-run profits increase as a result of the wrong, without necessarily ensuring that all employees bear their proportionate share of any corporate liability should wrongdoing be detected. Thus, if the firm cannot necessarily determine who committed a wrong, a wrongdoer may expect to get a raise or promotion if the wrong increases profits, without expecting to be demoted or to have his salary fall if the wrong is detected and the firm is sanctioned. Firms also may be unable to link agents' compensation to long-run profits if there is a substantial likelihood of employee turnover or if other concerns - such as excessive managerial risk aversion - militate against such policies. Similarly, firm liability will not necessarily eliminate agents' incentives to commit unintentional wrongs if firms cannot monitor agents' caretaking perfectly and agents cannot pay the optimal sanction. This is because "caretaking" often imposes a private cost on agents. Thus, wrongdoing which reduces care costs may benefit the agent even if the firm does not benefit. Finally, despite corporate liability, managers of publicly held firms will have an incentive to commit wrongs intended to secure their positions if the misconduct helps the manager secure his job but its detection does not significantly increase his risk of being fired (either because his position is insecure if he does not commit the wrong, or because, by then, the manager is likely to have retired or moved to another firm). Thus, managers may benefit from wrongdoing even if the firm does not. Indeed, existing empirical evidence suggests that agency costs may explain most wrongdoing by publicly held firms. See Cindy Alexander & Mark Cohen, Why Do Corporations Become Criminals? Ownership, Hidden Action, and Crime as an Agency Cost (Mar. 1997) (Working Paper, Owen Graduate School of Management, Vanderbilt University) [hereinafter Alexander & Cohen, Working Paper] (publicly held firms are more likely to engage in crime the smaller management's ownership stake); see also Cindy R. Alexander & Mark A. Cohen, New Evidence on the Origins of Corporate Crime, 17 Managerial & Decision Econ. 421 (1996) [hereinafter Alexander & Cohen, Corporate Crime] (criminal behavior is more likely the larger the firm and the lower shareholders' ability to monitor managers).
81. In addition, this regime would be less effective than strict liability at reducing agents' benefits from wrongdoing because firms that did not violate a duty would not be liable. In these cases, the firm would still get the full benefit of the crime, and thus agents who benefit when the firm benefits still would have an incentive to commit the wrong.
82. See Cooter, supra note 37, at 1539-40 (noting that strict liability is superior where it is very costly for firms to determine due care); cf. Richard Craswell & John E. Calfee, Deterrence and Uncertain Legal Standards, 2 J.L. Econ. & Org. 279 (1986) (noting that where court error renders the legal standard uncertain, duty-based liability will not necessarily cause firms to take optimal care, even if on average courts are correct).
83. See Cooter, supra note 37, at 1539-40 (noting that strict liability is superior where it is very costly for firms to determine due care); cf. Richard Craswell & John E. Calfee, Deterrence and Uncertain Legal Standards, 2 J.L. Econ. & Org. 279 (1986) (noting that where court error renders the legal standard uncertain, duty-based liability will not necessarily cause firms to take optimal care, even if on average courts are correct).
84. Indeed, duty-based regimes may be superior if either (i) there is a risk of firm insolvency, see Arlen, supra note 5, at 886, or (ii) if the precaution is unobservable and thus is plagued with possible "credibility problems," see infra Part I.D. (discussing duty-based regimes as a means of reducing the credibility problem). A duty-based regime also may have lower administrative costs because there will be fewer cases than under a strict liability regime.
85. See Cooter, supra note 37, at 1539-40 (arguing strict liability is superior where it is very costly for firms to determine due care). Duty-based regimes are particularly susceptible to error where prevention involves "nondurable" activities (such as those involving human action), as opposed to installing "durable" technologies (such as locking certain cabinets). For cases of "nondurable" controls, the risk arises not only that the court will set the standard incorrectly, but that it may be unable to determine whether the firm has adhered to that standard. Cf. Mark F. Grady, Why Are People Negligent? Technology, Nondurable Precautions, and the Medical Malpractice Explosion, 82 Nw. U. L. Rev. 293, 317 (1988) (making this distinction in torts context); see also infra Part I.C.
86. See Cooter, supra note 37, at 1539-40 (arguing strict liability is superior where it is very costly for firms to determine due care). Duty-based regimes are particularly susceptible to error where prevention involves "nondurable" activities (such as those involving human action), as opposed to installing "durable" technologies (such as locking certain cabinets). For cases of "nondurable" controls, the risk arises not only that the court will set the standard incorrectly, but that it may be unable to determine whether the firm has adhered to that standard. Cf. Mark F. Grady, Why Are People Negligent? Technology, Nondurable Precautions, and the Medical Malpractice Explosion, 82 Nw. U. L. Rev. 293, 317 (1988) (making this distinction in torts context); see also infra Part I.C.
87. There are important distinctions among monitoring, auditing, and investigating that are not fully addressed in the present analysis. For example, monitoring must be done ex ante, before the wrong has occurred, and thus before the firm knows the seriousness of the wrong, whereas investigating occurs when more information is available as to the seriousness of the wrong. Thus, all else equal, investigating often may be superior to monitoring because the firm (and society) can concentrate enforcement expenditures on the most serious wrongs. See Dilip Mookherjee & I.P.L. Png, Monitoring vis-à-vis Investigation in Enforcement of Law, 82 Am. Econ. Rev. 556, 556-57 (1992) (discussing optimal choice for government officials between monitoring and investigating). Yet, in some circumstances, monitoring nevertheless may be superior if it is observable because it is undertaken ex ante and thus is less likely to be subject to a "credibility problem." See infra Part I.D.
88. Corporate liability should ensure that firms invariably report detected wrongdoing. See Louis Kaplow & Steven Shavell, Optimal Enforcement with Self-Reporting of Behavior, 102 J. Pol. Econ. 583 (1994) (arguing that liability should induce individual wrongdoers to report their own wrongdoing). Firms should report even when the firm is the best party to sanction the wrongdoer because reporting is the lowest cost method for informing the government about wrongdoing. The government thus can ensure that firms have adequately sanctioned wrongdoers (increasing the credibility of firms' threats to do so). Also, even if the wrongdoer is sanctioned, the government should hold the firm liable in order to induce optimal activity levels, prevention measures, and policing. In addition, if principals are risk averse, self-reporting reduces risk-bearing costs because those who report wrongdoing pay a lower amount with certainty, which is less costly to them than an equivalent expected sanction based on a lower risk of detection but a higher actual penalty. See Kaplow & Shavell, supra, at 584-85.
89. The term originates with Arlen, supra note 5, at 833. Note too that we define traditional strict liability as strict liability that imposes a fixed sanction on wrongdoers which does not depend on the probability of detection. See id. at 842. This may be contrasted with "sanction-adjusted" strict liability, under which actual sanction levels rise or fall. See infra Part II.A.2.
90. In addition, the fine that enables strict vicarious liability to induce optimal monitoring (when it is capable of doing so) is very complicated. Thus, the standard argument that strict liability places low information demands on courts does not apply to strict vicarious liability employed to induce monitoring or investigation of misconduct. See Arlen, supra note 5, at 847, 856-57.
91. Our analysis does not require that the government always get the information, just that there is a positive risk it will. The greater the risk, the worse the liability enhancement effect. The assumption that the government may obtain information about wrongdoing that the firm detects is reasonable for several reasons. First, even when a corporation does not report its discovered crimes to the government, corporate enforcement efforts may increase the firm's expected liability. Corporations recognize that the government often discovers evidence of possible corporate wrongdoing on its own. The government may well respond to evidence of possible wrongdoing by subpoenaing corporate records. These records will include documentary evidence resulting from corporate enforcement efforts -records that may contain evidence of wrongdoing which prosecutors may use to prove their case against the corporation. See id. Neither firms nor their managers can shield such records by asserting a Fifth Amendment right against self-incrimination. See Harry First, Business Crime: Cases and Materials 382-401 (1990) (discussing scope of Fifth Amendment protection for corporate records). The government also may induce the firm to reveal detected wrongdoing through criminal liability rules that heavily penalize firms that do not report misconduct. See infra Part II.B. (discussing such a regime). Second, in some cases the market may penalize firms that detect wrongdoing but do not report it, such as when customers are the victims of the wrong. Third, managers and employees may face strong pressures to reveal enforcement information. Some statutes impose personal liability on managers who fail to report certain violations to the proper authorities. See, e.g., California Corporate Criminal Liability Act of 1989, Cal. Penal Code § 387 (West 1988 & Supp. 1997). Other statutes provide cash bounties to those who report corporate wrongdoing. See, e.g., Securities Exchange Act of 1934 § 21A(e), 15 U.S.C. § 78u-1(e) (1994) (bounty provision for information on insider trading); False Claims Act, 31 U.S.C. § 3730(d) (1994) (qui tam provisions). Finally, the threat of higher corporate sanctions should the firm not report may induce innocent corporate officials whose compensation is tied to firm profits to report. See Arlen, supra note 5, at 858-60 (noting that penalizing nonreporting may induce innocent corporate managers to report crimes they discover).
92. See Arlen, supra note 5, at 842-43.
93. For example, the government used Princeton Newport's own trading records to determine that an employee might have engaged in illegal trading, and then later obtained evidence of alleged wrongdoing by the firm from the firm's own documents and taped conversations of its traders' telephone calls. See James B. Stewart, Den of Thieves 348-52 (1991). Other securities firms also faced liability based on their own records.
94. This assumes that the wrong does not benefit the firm. If the wrong does benefit the firm, the firm will be even less likely to monitor because monitoring would impose an additional cost of 2B, which is the benefit to the firm of the two wrongs deterred.
95. For a full mathematical proof of this claim, see Arlen, supra note 5, at 850-58.
96. See Arlen, supra note 5, at 850-57 (providing a mathematical proof of this point). An elaboration of our earlier example can illustrate the point. Previously we supposed that, without monitoring, misconduct was detected with a probability of 1/5 and five agents would commit the wrong. Assume now that optimal monitoring increases the probability of detection to 1/3 and reduces the number of offending agents to two. In this case, the social marginal benefit of monitoring is 3h, which is three times the social cost of wrongdoing to others. The private marginal benefit, however, is 1F - (2/3)F = (1/3)F, which is the expected decrease in the firm's expected liability, taking into account the liability enhancement effect. Thus, although three wrongs are deterred, the firm's expected liability only falls by (1/3)F. If F = h/p*, the firm's marginal benefit of policing - and thereby deterring three wrongs - would be only h, which is less than the benefit to society of deterring those wrongs. Thus, if, as the definition of optimal monitoring implies, the marginal cost of undertaking optimal rather than nonoptimal monitoring equals the social marginal benefit, 3h, then this cost will exceed the firm's private marginal benefit if F = h/p*. To counteract the depressing effect of the liability enhancement effect, the sanction must exceed this amount. See id.
97. Composite duty-based regimes - which reduce but do not eliminate liability if the firm meets its monitoring, investigating, or reporting duties - can also induce optimal enforcement measures, provided that the implementation of such measures reduces the firm's penalty enough to warrant the investment. In addition, as we discuss below, composite regimes that mitigate (rather than eliminate) liability when firms implement optimal policing measures can induce such measures while simultaneously meeting all other liability aims, including the optimal regulation of activity levels and inducement of preventive measures.
98. See Craswell & Calfee, supra note 42, at 298; see also Mark F. Grady, Proximate Cause and the Law of Negligence, 69 Iowa L. Rev. 363, 403 (1984); Marcel Kahan, Causation and Incentives to Take Care Under the Negligence Rule, 18 J. Legal Stud. 427, 437 (1989).
99. See Craswell & Calfee, supra note 42, at 298; see also Mark F. Grady, Proximate Cause and the Law of Negligence, 69 Iowa L. Rev. 363, 403 (1984); Marcel Kahan, Causation and Incentives to Take Care Under the Negligence Rule, 18 J. Legal Stud. 427, 437 (1989).
100. See Craswell & Calfee, supra note 42, at 298; see also Mark F. Grady, Proximate Cause and the Law of Negligence, 69 Iowa L. Rev. 363, 403 (1984); Marcel Kahan, Causation and Incentives to Take Care Under the Negligence Rule, 18 J. Legal Stud. 427, 437 (1989).
101. The problem of uncertain legal standards is likely to be particularly acute when optimal monitoring is firm-specific because there is no standard monitoring technology for all firms in the industry. Similarly, courts are likely to have much more difficulty assessing monitoring measures that involve "nondurable" precautions, such as the human effort involved in detecting securities fraud, than they are in evaluating durable monitoring technologies such as video cameras or tape recorders. See supra note 44.
102. See Arlen, supra note 5, at 847; see also infra Part II (comparing a duty-based regime to sanction-adjusted strict liability).
103. See infra Part II.
104. By "commit" we mean the ability of the firm to establish monitoring programs ex ante which are sufficiently fixed that the firm cannot reduce its monitoring efforts once agents adjust their behavior to reflect the threatened level of monitoring. Note that this monitoring must also be observable for agents to believe a firm's threats to monitor at a specific level. The credibility problem has been previously noted, generally in models involving government enforcement efforts. See, e.g., Debra J. Aron & Pau Olivella, Bonus and Penalty Schemes as Equilibrium Incentive Devices, with Application to Manufacturing Systems, 10 J.L. Econ. & Org. 1, 14-19 (1994); Nahum D. Melumad & Dilip Mookherjee, Delegation as Commitment: The Case of Income Tax Audits, 20 Rand J. Econ. 139, 142-44 (1989); Jennifer F. Reinganum & Louis L. Wilde, Equilibrium Verification and Reporting Policies in a Model of Tax Compliance, 27 Int'l Econ. Rev. 739, 740 (1986). We are, to our knowledge, the first to consider the impact of strict versus duty-based liability on this problem.
105. By "commit" we mean the ability of the firm to establish monitoring programs ex ante which are sufficiently fixed that the firm cannot reduce its monitoring efforts once agents adjust their behavior to reflect the threatened level of monitoring. Note that this monitoring must also be observable for agents to believe a firm's threats to monitor at a specific level. The credibility problem has been previously noted, generally in models involving government enforcement efforts. See, e.g., Debra J. Aron & Pau Olivella, Bonus and Penalty Schemes as Equilibrium Incentive Devices, with Application to Manufacturing Systems, 10 J.L. Econ. & Org. 1, 14-19 (1994); Nahum D. Melumad & Dilip Mookherjee, Delegation as Commitment: The Case of Income Tax Audits, 20 Rand J. Econ. 139, 142-44 (1989); Jennifer F. Reinganum & Louis L. Wilde, Equilibrium Verification and Reporting Policies in a Model of Tax Compliance, 27 Int'l Econ. Rev. 739, 740 (1986). We are, to our knowledge, the first to consider the impact of strict versus duty-based liability on this problem.
106. By "commit" we mean the ability of the firm to establish monitoring programs ex ante which are sufficiently fixed that the firm cannot reduce its monitoring efforts once agents adjust their behavior to reflect the threatened level of monitoring. Note that this monitoring must also be observable for agents to believe a firm's threats to monitor at a specific level. The credibility problem has been previously noted, generally in models involving government enforcement efforts. See, e.g., Debra J. Aron & Pau Olivella, Bonus and Penalty Schemes as Equilibrium Incentive Devices, with Application to Manufacturing Systems, 10 J.L. Econ. & Org. 1, 14-19 (1994); Nahum D. Melumad & Dilip Mookherjee, Delegation as Commitment: The Case of Income Tax Audits, 20 Rand J. Econ. 139, 142-44 (1989); Jennifer F. Reinganum & Louis L. Wilde, Equilibrium Verification and Reporting Policies in a Model of Tax Compliance, 27 Int'l Econ. Rev. 739, 740 (1986). We are, to our knowledge, the first to consider the impact of strict versus duty-based liability on this problem.
107. Even under strict liability, firms will not face credibility problems if they have adequate incentives to establish a reputation for making credible threats, or if they can use third parties to implement their policing measures. Reputation is most likely to be effective in situations where deviations by either party are quickly observed and the future costs of losing one's credibility are high. See generally Benjamin Klein & Keith Leffler, The Role of Market Forces in Assuring Contractual Performance, 89 J. Pol. Econ. 615, 624-25 (1981) (even when consumers can perfectly verify quality of a good after the fact, high reputation firms will have incentive to "cheat" and supply low-quality goods unless these firms are earning a continual stream of rental income from producing the high quality goods, the discounted value of which exceeds the one-time wealth increase obtained from low quality production). But reputation can only solve a credibility problem in some circumstances. For example, a firm can credibly develop a reputation for policing only if it is properly viewed as being in a potentially infinitely-lived relationship with its agents. If instead the firm and its agents are in a finite relationship with a fixed time horizon, the agents' knowledge that the firm has an incentive to cheat in the last period will eliminate the reputational benefit to the firm of policing in the second-to-last period, which, as agents will understand this, in turn eliminates the firm's incentive to police in the third-to-last period, and so forth. Thus, in such a situation, reputation will not solve the credibility problem. See generally Drew Fudenberg & Jean Tirole, Game Theory 166 (1996). Even if the firm may be potentially infinitely-lived, reputation will not be sufficient to induce optimal policing if the probability the firm will exist in future periods is very small, because then the expected benefit of developing a reputation also will be small. See generally Klein & Leffler, supra; cf. Alexander & Cohen, Corporate Crime, supra note 40 (providing empirical evidence that poor prior performance tends to precede environmental crime is consistent with view that reduced likelihood of repeat dealing increases likelihood that employees will commit crime). Moreover, reputational effects can solve the credibility problem only if agents can verify the firm's policing efforts once it has implemented them. This often will be difficult for them to do. Thus, when a firm threatens to sample or monitor probabilistically, an agent's observation of ex post monitoring will not enable the agent to determine, for certain, whether the firm truthfully announced its monitoring strategy. See Reinganum & Wilde, supra note 60, at 742, 754-55 (discussing this point). Similarly, firms may be unable to establish reputations if agents can determine ex post whether the firm has monitored but cannot determine how diligent its monitors are. Finally, firms will not be able to establish accurate reputations for reporting wrongdoing if agents do not know when, and how much, wrongdoing has occurred; in this case, agents who observe a firm reporting wrongs will not be able to determine whether the firm is reporting all the wrongs it detects - and thus they should assume it will report any wrongdoing they do - or only some portion of the wrongs it detects - in which case it may not report their misconduct. Thus, firms will have difficulty establishing policing reputations for wrongs likely to be committed by employees and middle level managers, but may be able to establish effective reputations for implementing policing measures aimed at wrongdoing committed by senior officers who are privy to information about both the firm's monitoring policies and about whether the firm has detected possible wrongdoing.
108. Even under strict liability, firms will not face credibility problems if they have adequate incentives to establish a reputation for making credible threats, or if they can use third parties to implement their policing measures. Reputation is most likely to be effective in situations where deviations by either party are quickly observed and the future costs of losing one's credibility are high. See generally Benjamin Klein & Keith Leffler, The Role of Market Forces in Assuring Contractual Performance, 89 J. Pol. Econ. 615, 624-25 (1981) (even when consumers can perfectly verify quality of a good after the fact, high reputation firms will have incentive to "cheat" and supply low-quality goods unless these firms are earning a continual stream of rental income from producing the high quality goods, the discounted value of which exceeds the one-time wealth increase obtained from low quality production). But reputation can only solve a credibility problem in some circumstances. For example, a firm can credibly develop a reputation for policing only if it is properly viewed as being in a potentially infinitely-lived relationship with its agents. If instead the firm and its agents are in a finite relationship with a fixed time horizon, the agents' knowledge that the firm has an incentive to cheat in the last period will eliminate the reputational benefit to the firm of policing in the second-to-last period, which, as agents will understand this, in turn eliminates the firm's incentive to police in the third-to-last period, and so forth. Thus, in such a situation, reputation will not solve the credibility problem. See generally Drew Fudenberg & Jean Tirole, Game Theory 166 (1996). Even if the firm may be potentially infinitely-lived, reputation will not be sufficient to induce optimal policing if the probability the firm will exist in future periods is very small, because then the expected benefit of developing a reputation also will be small. See generally Klein & Leffler, supra; cf. Alexander & Cohen, Corporate Crime, supra note 40 (providing empirical evidence that poor prior performance tends to precede environmental crime is consistent with view that reduced likelihood of repeat dealing increases likelihood that employees will commit crime). Moreover, reputational effects can solve the credibility problem only if agents can verify the firm's policing efforts once it has implemented them. This often will be difficult for them to do. Thus, when a firm threatens to sample or monitor probabilistically, an agent's observation of ex post monitoring will not enable the agent to determine, for certain, whether the firm truthfully announced its monitoring strategy. See Reinganum & Wilde, supra note 60, at 742, 754-55 (discussing this point). Similarly, firms may be unable to establish reputations if agents can determine ex post whether the firm has monitored but cannot determine how diligent its monitors are. Finally, firms will not be able to establish accurate reputations for reporting wrongdoing if agents do not know when, and how much, wrongdoing has occurred; in this case, agents who observe a firm reporting wrongs will not be able to determine whether the firm is reporting all the wrongs it detects - and thus they should assume it will report any wrongdoing they do - or only some portion of the wrongs it detects - in which case it may not report their misconduct. Thus, firms will have difficulty establishing policing reputations for wrongs likely to be committed by employees and middle level managers, but may be able to establish effective reputations for implementing policing measures aimed at wrongdoing committed by senior officers who are privy to information about both the firm's monitoring policies and about whether the firm has detected possible wrongdoing.
109. Even under strict liability, firms will not face credibility problems if they have adequate incentives to establish a reputation for making credible threats, or if they can use third parties to implement their policing measures. Reputation is most likely to be effective in situations where deviations by either party are quickly observed and the future costs of losing one's credibility are high. See generally Benjamin Klein & Keith Leffler, The Role of Market Forces in Assuring Contractual Performance, 89 J. Pol. Econ. 615, 624-25 (1981) (even when consumers can perfectly verify quality of a good after the fact, high reputation firms will have incentive to "cheat" and supply low-quality goods unless these firms are earning a continual stream of rental income from producing the high quality goods, the discounted value of which exceeds the one-time wealth increase obtained from low quality production). But reputation can only solve a credibility problem in some circumstances. For example, a firm can credibly develop a reputation for policing only if it is properly viewed as being in a potentially infinitely-lived relationship with its agents. If instead the firm and its agents are in a finite relationship with a fixed time horizon, the agents' knowledge that the firm has an incentive to cheat in the last period will eliminate the reputational benefit to the firm of policing in the second-to-last period, which, as agents will understand this, in turn eliminates the firm's incentive to police in the third-to-last period, and so forth. Thus, in such a situation, reputation will not solve the credibility problem. See generally Drew Fudenberg & Jean Tirole, Game Theory 166 (1996). Even if the firm may be potentially infinitely-lived, reputation will not be sufficient to induce optimal policing if the probability the firm will exist in future periods is very small, because then the expected benefit of developing a reputation also will be small. See generally Klein & Leffler, supra; cf. Alexander & Cohen, Corporate Crime, supra note 40 (providing empirical evidence that poor prior performance tends to precede environmental crime is consistent with view that reduced likelihood of repeat dealing increases likelihood that employees will commit crime). Moreover, reputational effects can solve the credibility problem only if agents can verify the firm's policing efforts once it has implemented them. This often will be difficult for them to do. Thus, when a firm threatens to sample or monitor probabilistically, an agent's observation of ex post monitoring will not enable the agent to determine, for certain, whether the firm truthfully announced its monitoring strategy. See Reinganum & Wilde, supra note 60, at 742, 754-55 (discussing this point). Similarly, firms may be unable to establish reputations if agents can determine ex post whether the firm has monitored but cannot determine how diligent its monitors are. Finally, firms will not be able to establish accurate reputations for reporting wrongdoing if agents do not know when, and how much, wrongdoing has occurred; in this case, agents who observe a firm reporting wrongs will not be able to determine whether the firm is reporting all the wrongs it detects - and thus they should assume it will report any wrongdoing they do - or only some portion of the wrongs it detects - in which case it may not report their misconduct. Thus, firms will have difficulty establishing policing reputations for wrongs likely to be committed by employees and middle level managers, but may be able to establish effective reputations for implementing policing measures aimed at wrongdoing committed by senior officers who are privy to information about both the firm's monitoring policies and about whether the firm has detected possible wrongdoing.
110. Even under strict liability, firms will not face credibility problems if they have adequate incentives to establish a reputation for making credible threats, or if they can use third parties to implement their policing measures. Reputation is most likely to be effective in situations where deviations by either party are quickly observed and the future costs of losing one's credibility are high. See generally Benjamin Klein & Keith Leffler, The Role of Market Forces in Assuring Contractual Performance, 89 J. Pol. Econ. 615, 624-25 (1981) (even when consumers can perfectly verify quality of a good after the fact, high reputation firms will have incentive to "cheat" and supply low-quality goods unless these firms are earning a continual stream of rental income from producing the high quality goods, the discounted value of which exceeds the one-time wealth increase obtained from low quality production). But reputation can only solve a credibility problem in some circumstances. For example, a firm can credibly develop a reputation for policing only if it is properly viewed as being in a potentially infinitely-lived relationship with its agents. If instead the firm and its agents are in a finite relationship with a fixed time horizon, the agents' knowledge that the firm has an incentive to cheat in the last period will eliminate the reputational benefit to the firm of policing in the second-to-last period, which, as agents will understand this, in turn eliminates the firm's incentive to police in the third-to-last period, and so forth. Thus, in such a situation, reputation will not solve the credibility problem. See generally Drew Fudenberg & Jean Tirole, Game Theory 166 (1996). Even if the firm may be potentially infinitely-lived, reputation will not be sufficient to induce optimal policing if the probability the firm will exist in future periods is very small, because then the expected benefit of developing a reputation also will be small. See generally Klein & Leffler, supra; cf. Alexander & Cohen, Corporate Crime, supra note 40 (providing empirical evidence that poor prior performance tends to precede environmental crime is consistent with view that reduced likelihood of repeat dealing increases likelihood that employees will commit crime). Moreover, reputational effects can solve the credibility problem only if agents can verify the firm's policing efforts once it has implemented them. This often will be difficult for them to do. Thus, when a firm threatens to sample or monitor probabilistically, an agent's observation of ex post monitoring will not enable the agent to determine, for certain, whether the firm truthfully announced its monitoring strategy. See Reinganum & Wilde, supra note 60, at 742, 754-55 (discussing this point). Similarly, firms may be unable to establish reputations if agents can determine ex post whether the firm has monitored but cannot determine how diligent its monitors are. Finally, firms will not be able to establish accurate reputations for reporting wrongdoing if agents do not know when, and how much, wrongdoing has occurred; in this case, agents who observe a firm reporting wrongs will not be able to determine whether the firm is reporting all the wrongs it detects - and thus they should assume it will report any wrongdoing they do - or only some portion of the wrongs it detects - in which case it may not report their misconduct. Thus, firms will have difficulty establishing policing reputations for wrongs likely to be committed by employees and middle level managers, but may be able to establish effective reputations for implementing policing measures aimed at wrongdoing committed by senior officers who are privy to information about both the firm's monitoring policies and about whether the firm has detected possible wrongdoing.
111. In some cases firms can solve their credibility problems by hiring third parties to monitor, investigate, and report. This solution, however, will not solve the credibility problem in all cases and imposes its own costs. First, insider investigations often will be more effective than those conducted by outsiders because the insiders will have better information. This is particularly likely because firms will often withhold crucial information from outsiders; for example, they are likely to be reluctant to provide outsiders with the regular reports on production costs, pricing, customers, and negotiations with suppliers that are necessary to monitor for antitrust violations. In addition, third party enforcers are effective only if they have an incentive to investigate and report misconduct even when the firm does not want them to do so. Contractual arrangements alone cannot necessarily provide this incentive because, even if the contract rewards a third party for reporting wrongdoing, the firm can secretly negotiate with its third party monitors to get them to monitor ineffectively or to cover up detected wrongdoing. Cf. Avinash Dixit & Barry Nalebuff, Making Strategies Credible, in Strategy and Choice 161, 166-68 (Richard Zeckhauser ed., 1991) (one problem with use of third party enforcers to solve credibility problem is that, ex post, it often will be profitable for principal to renegotiate contract, unless third party reputation prevents renegotiation, or firm and third party cannot communicate in the future); Jerry Green, The Strategic Use of Contacts with Third Parties, in Strategy and Choice, supra, at 241, 241-42 (same). Finally, the use of third party enforcers will not induce firms to report those wrongs that they themselves detect, as is necessary in order to minimize enforcement costs. See supra note 46.
112. To be precise, the credibility problem occurs when the firm cannot commit to such efforts ex ante and agents cannot verify its policing efforts ex ante.
113. When policing measures are also preventive measures, the firm may have a credible incentive to police even under a strict liability regime. For example, if detecting fraud also prevents the wayward agent from engaging in additional fraud, the firm's threat to monitor may or may not be credible, according to the relative magnitude of its future expected liability and the cost of monitoring. Even in this case, however, monitoring may be suboptimal if the benefit to the firm of preventing future wrongs is not large enough to induce optimal expenditures on monitoring.
114. To see this, consider our example of a firm where five agents commit the wrong if there is no policing and only three commit it if the firm polices optimally. See text preceding note 52. Assume that the firm is strictly liable for all wrongs and announces that it will engage in optimal monitoring. The question arises: if agents believe that the firm will police, does it have an incentive to do so? In our example, if agents believe the firm will police, only three commit a wrong. If it monitors optimally, the firm faces a 50% probability that each wrong will be detected. Thus, under traditional strict liability, the firm's expected costs if it actually does police are: M* + 3(1/2)F. However, the firm's expected costs if it does not police are: 3(1/5)F. Thus, it is better off if it does not undertake the threatened policing measures. Agents, knowing this to be the case, thus will not believe the firm's threats.
115. This analysis implicitly assumes that firms and agents only pursue pure strategies, that firms either undertake policing measures or do not, and that agents either commit a wrong or do not. In reality, either or both could pursue mixed strategies. Firms could pursue probabilistic policing, under which agents face only a probability of being subject to policing. Agents could respond with a mixed strategy, adopting a probability of engaging in misconduct. The possibility of mixed strategies does not eliminate the credibility problem, however. Consider the question of whether firms will report wrongdoing. If ex post the firm has no incentive to report because reporting only increases its expected liability, then even if it announces a positive probability of reporting, it still will have no incentive to report if it actually detects a wrong. It will only implement a positive probability of reporting if it has reason to report which is independent of any deterrent effect of the threat of reporting - for example, if reporting terminates the wrong more quickly, thereby reducing its severity. Similarly, if a firm cannot commit to a monitoring policy, and monitoring is unobservable, it will not implement a probabilistic monitoring program unless it has an incentive to monitor which is independent of the deterrent effect of the threat of monitoring - for example, if rapid detection enables it to reduce the severity of the wrongs it is liable for. Thus, any probabilistic monitoring it does do will be based only on this standalone benefit; when credibility problems exist, it will not consider the deterrent effect of any additional monitoring. See Reinganum & Wilde, supra note 60, at 754-55 (if enforcement is unobservable and the enforcer cannot commit to its enforcement measures, mixed strategies will not yield positive enforcement efforts unless there is a stand-alone benefit to policing).
116. The only potential benefit from reporting is signaling the firm's willingness to report in the future. This benefit is likely to be significant, however, only if firms can develop a reputation for reporting. Firms can develop such a reputation only if agents and firms are in a relationship without a fixed final termination point and agents can verify the firm's reporting behavior. Agents must thus be able to determine whether a firm fails to report because it chooses not to or because it fails to detect misconduct. See supra note 61 (discussing reputation); supra note 62 (discussing third parties).
117. See Aron & Olivella, supra note 60, at 14-19 (discussing problem of inducing monitoring when monitoring is unobservable by agents); cf. Reinier Kraakman et al., When Are Shareholder Suits in Shareholder Interests?, 82 Geo. L. Rev. 1733, 1741 (1994) (noting prospect of suit that is counter to firm's ex post interests will not deter managerial misconduct unless firm makes credible threat to sue).
118. See Aron & Olivella, supra note 60, at 14-19 (discussing problem of inducing monitoring when monitoring is unobservable by agents); cf. Reinier Kraakman et al., When Are Shareholder Suits in Shareholder Interests?, 82 Geo. L. Rev. 1733, 1741 (1994) (noting prospect of suit that is counter to firm's ex post interests will not deter managerial misconduct unless firm makes credible threat to sue).
119. See supra notes 61-62 (discussing when credibility problems exist).
120. See supra note 62 (discussing third-party policing).
121. This is particularly likely to hold when the level of monitoring is difficult for employees to determine ex post, but can be established in court through introducing documents, testimony of monitors, and other evidence. In other words, on a day-to-day basis it often will be difficult to determine the amount of monitoring, while it will not be difficult to do so in a courtroom where the firm has an incentive to introduce evidence unavailable to its workers. Cf. supra note 61 (discussing reputation). Nevertheless, ex ante monitoring may sometimes be difficult for courts to verify ex post if it is a random process. See Reinganum & Wilde, supra note 60, at 740.
122. See infra Part U.C. (discussing the informational requirements of duty-based liability in more detail).
123. For example, in some cases payout policies might be such that a wrongdoer's benefit from misconduct is directly proportional to the firm's benefit net of any expected criminal liability. This is likely, for example, when shareholder-managers of firms with highly concentrated ownership commit a wrong. In this situation, corporate criminal liability can optimally deter wrongdoing even when agents are insolvent by holding firms strictly liable for their agents' crimes subject to an expected sanction equal to the social cost of wrongdo-ing to others. This regime would ensure that agents do not profit from misconduct by ensuring that the firm does not profit. See supra text accompanying note 40.
124. In order for this to be true, the ex ante sanction imposed by the market must equal the social cost of wrongdoing, which means that either the market must always detect wrongs or must impose a reputational sanction which exceeds the actual cost to victims of the wrong. Cf. Karpoff & Lott, supra note 19, at 760-66. This market penalty does not obviate the need for corporate liability because the fact that the firm bears the full social cost of the harm does not mean agents will necessarily be optimally deterred. Deterring agents generally will require that firms implement policing measures. Market forces alone cannot induce optimal policing because the market essentially effects a regime of traditional strict liability. See supra Parts I.C. & D. Thus, to induce optimal policing, market forces generally must be supplemented by a duty-based regime.
125. See Arlen, supra note 5, at 865-66.
126. Probability-fixed regimes created by use immunity or evidentiary privileges have also been criticized on the ground that they will induce firms to shift resources from prevention into auditing, which, it is argued, is less effective at deterring wrongdoing. See Dana, supra note 5. We agree with Professor Dana that these regimes lead to a relative increase in the amount of auditing and other policing measures. Yet this is a reason to adopt such regimes, not to reject them. Under the present regime, firms are undertaking too little policing. Increasing policing, therefore, promotes social welfare. And, provided that the expected sanction equals the social cost of wrongdoing, in theory firms nevertheless will undertake optimal prevention. The problem with probability-fixed regimes, as we noted, is that the sanction needed to satisfy the condition that the expected sanction equals the social cost of misconduct will often be so high as to exceed the firm's assets, causing these regimes to be unable to induce optimal prevention or policing. For a discussion of additional problems with privileging the information that firms obtain through policing measures, see infra Part IV.A.
127. See Arlen, supra note 5. As firm insolvency already may be a substantial problem, it often will be impossible to implement the huge fines required by probability-fixed regimes, in which case neither prevention nor policing nor activity levels will be optimal. Cf. Alexander & Cohen, Corporate Crime, supra note 40 (providing empirical evidence that poor prior performance tends to precede environmental crime); Arlen & Carney, supra note 36 (a substantial number of firms which committed fraud-on-the-market securities fraud had a net worth less than shareholders' total harm); Mark Cohen, Theories of Punishment and Empirical Trends in Corporate Criminal Sanctions, 17 Managerial & Decision Econ. 399, 403 (1996) (finding that 35.7% of organizations convicted of federal crimes between 1984 and 1990 could not afford to compensate for harm caused by offense). For a discussion of ex ante composite liability as a partial solution to the problem of firm insolvency, see infra note 89.
128. See Arlen, supra note 5. As firm insolvency already may be a substantial problem, it often will be impossible to implement the huge fines required by probability-fixed regimes, in which case neither prevention nor policing nor activity levels will be optimal. Cf. Alexander & Cohen, Corporate Crime, supra note 40 (providing empirical evidence that poor prior performance tends to precede environmental crime); Arlen & Carney, supra note 36 (a substantial number of firms which committed fraud-on-the-market securities fraud had a net worth less than shareholders' total harm); Mark Cohen, Theories of Punishment and Empirical Trends in Corporate Criminal Sanctions, 17 Managerial & Decision Econ. 399, 403 (1996) (finding that 35.7% of organizations convicted of federal crimes between 1984 and 1990 could not afford to compensate for harm caused by offense). For a discussion of ex ante composite liability as a partial solution to the problem of firm insolvency, see infra note 89.
129. See Arlen, supra note 5. As firm insolvency already may be a substantial problem, it often will be impossible to implement the huge fines required by probability-fixed regimes, in which case neither prevention nor policing nor activity levels will be optimal. Cf. Alexander & Cohen, Corporate Crime, supra note 40 (providing empirical evidence that poor prior performance tends to precede environmental crime); Arlen & Carney, supra note 36 (a substantial number of firms which committed fraud-on-the-market securities fraud had a net worth less than shareholders' total harm); Mark Cohen, Theories of Punishment and Empirical Trends in Corporate Criminal Sanctions, 17 Managerial & Decision Econ. 399, 403 (1996) (finding that 35.7% of organizations convicted of federal crimes between 1984 and 1990 could not afford to compensate for harm caused by offense). For a discussion of ex ante composite liability as a partial solution to the problem of firm insolvency, see infra note 89.
130. See Arlen, supra note 5. As firm insolvency already may be a substantial problem, it often will be impossible to implement the huge fines required by probability-fixed regimes, in which case neither prevention nor policing nor activity levels will be optimal. Cf. Alexander & Cohen, Corporate Crime, supra note 40 (providing empirical evidence that poor prior performance tends to precede environmental crime); Arlen & Carney, supra note 36 (a substantial number of firms which committed fraud-on-the-market securities fraud had a net worth less than shareholders' total harm); Mark Cohen, Theories of Punishment and Empirical Trends in Corporate Criminal Sanctions, 17 Managerial & Decision Econ. 399, 403 (1996) (finding that 35.7% of organizations convicted of federal crimes between 1984 and 1990 could not afford to compensate for harm caused by offense). For a discussion of ex ante composite liability as a partial solution to the problem of firm insolvency, see infra note 89.
131. Cf. Orts & Murray, supra note 5, at 7 (noting that even if firm's internal assessments are privileged, strict liability will not provide optimal incentives to audit if government can use underlying facts contained in audit against firm).
132. See, e.g., Jennifer Arlen, Shielding Audits Will Aggravate Pollution Problems, 17 Nat'l L.J. A23 (Oct. 3, 1994).
133. See Arlen, supra note 5, at 857-58 (discussing a version of this regime). An adjustable regime that reduces sanctions discretely - rather than continuously - with an increasing probability of detection would induce suboptimal policing unless a reduction in sanctions occurred at the precise point that a firm adopted optimal policing measures. But such a regime would require the court to determine whether the firm had adopted optimal policing measures, and would thus be a duty-based regime in effect.
134. See supra Parts I.A. & B.
135. See supra text accompanying notes 52-53.
136. Reporting reduces enforcement costs by saving the government the resources needed to uncover the firm's information independently. It also reduces the risk that firms can escape sanctions through insolvency by raising the probability of detection and lowering the sanction necessary to induce wrongdoers to internalize the cost of misconduct. See supra note 46.
137. The sanctions must meet these requirements in order to ensure that ex ante the firm's expected liability equals h, as is necessary if this regime is to induce optimal prevention, sanctioning, activity levels, and policing.
138. Note that if g = p*°, then g(h/p*°) = h < h/p**. See infra Appendix note 206. 86 Quasi-strict liability is similar to the regime proposed by Professor Dana, in that the firm's sanction is mitigated if a firm reports wrongdoing. See Dana, supra note 5. Dana's regime differs from the two adjusted strict liability regimes we describe, however, in that he does not advocate adjusting the sanction for changes in the probability of detection resulting from policing practices. Rather, Dana appears to be combining his reporting mitigation provision with a traditional strict liability regime. The reporting mitigation provision is consistent with our conclusions. Unlike Dana, however, we conclude that the residual strict liability should be adjusted because otherwise firms will undertake insufficient monitoring and investigation as a result of perverse effects and credibility problems. See supra Parts I.C. & D.
139. In addition, this regime will not induce optimal monitoring if, in addition to increasing the overall probability of detecting misconduct, monitoring also increases the probability that firms will detect wrongdoing before the government does. In this case, firms will receive a private benefit from monitoring: the increased likelihood of detecting the wrong first, thereby rendering a firm eligible for sanction mitigation. This private benefit can induce excessive monitoring under a quasi-strict liability regime, as firms increase monitoring over optimal levels solely to detect first, and thus qualify for sanction mitigation. See infra Appendix, Part III.C.3 (discussing excessive monitoring).
140. In some cases market penalties may substitute for legally imposed strict liability in a composite regime. Recall that firms must face expected residual liability equal to the social cost of misconduct in order to regulate activity levels and induce preventive measures -including the obvious measure of declining to reward agents for engaging in misconduct. Firms must often pay the price for misconduct in market settings such as securities fraud. In this case, residual liability is unnecessary if the market forces firms to internalize the costs of wrongdoing. See supra note 74 (sanction imposed by market must equal harm divided by probability of detection).
141. A composite regime could sanction breaches of an ex ante monitoring duty independently of underlying misconduct, as where the government searches for and sanctions any shortcoming in a monitoring program proactively before finding evidence of misconduct. Although administrative economies will ordinarily dictate investigating a firm's monitoring program ex post, in tandem with its misconduct, insolvency concerns may lead to severing this connection in order to inspect monitoring efforts more frequently. Put differently, if liability for underlying misconduct is likely to exhaust a firm's assets, the prospect of facing additional liability for breach of a monitoring duty will have little effect if it can only be imposed when the firm is also liable for the underlying misconduct. In this situation, to induce optimal monitoring the government may need to periodically evaluate a firm's ex ante policing measures and impose sanctions on firms whose efforts are suboptimal. Note, however, that ex post policing duties such as investigating and reporting misconduct generally can only be enforced in conjunction with prosecuting the misconduct itself.
142. See infra Appendix, Part III.A. (describing such regime).
143. This Part focuses on monitoring and reporting. The analysis could easily be extended to include either a multi-tiered regime based on ex ante monitoring on the one hand and ex post investigating and reporting on the other, or, probably better, a multitiered regime with separate mitigation provisions for monitoring, investigating, and reporting. We remain open to the possibility that a multi-tiered regime may be optimal.
144. For a discussion of the relative merits of civil versus criminal corporate liability, see the sources cited supra note 12.
145. Alternatively, the sanction could be simply the net social cost of crime to others divided by the actual probability of detection. The advantage of basing the residual fine on the optimal probability of detection is that if there are many similarly situated firms, the court could apply the same optimal probability of detection to many firms. This approach only works if mitigation induces optimal policing, however. In addition to inducing optimal prevention, activity levels, and sanctioning, this residual liability is consistent with optimal policing. Specifically, a firm that adheres to its legal duty to police will not be induced to undertake excessive policing by the threat of strict residual liability. See infra Appendix, Part III.A.2.
146. See supra note 46 (corporate liability should ensure that firms always report wrongdoing).
147. This fine structure will induce optimal policing but is not the minimum default sanction that will do so. For a discussion of the minimum mitigation amount, see infra Appendix, Part III.A.3. It might seem that all composite regimes suffer in comparison to traditional strict liability because they require larger penalties and thus are more vulnerable to failure as a result of firm insolvency. In fact, however, the default sanction under our composite regimes is not necessarily higher than the optimal sanction under strict vicarious liability. The minimum default sanction under our composite regimes equals the residual liability h/p* plus an amount calculated to induce optimal monitoring. The sanction h/p° certainly will suffice to induce optimal policing, but a lower sanction will also achieve the desired result. See infra Appendix, Part III.A.3. By contrast, under traditional strict vicarious liability the sanction simply equals h/p. However, the optimal measure of h/p under traditional strict liability is the harm divided by the expected probability of detection based on the policing measures the firm is likely to implement. Because firms subject to a traditional strict liability rule often will not undertake efficient policing - and indeed in some cases may not implement any policing measures, see supra Parts LC. & D. - the probability of detection under strict liability may be substantially lower than under our composite regimes. Thus, if firms do not implement any policing measures, the optimal sanction under traditional strict liability will equal h/p°, an amount that exceeds the minimum optimal default sanction under a composite regime.
148. See infra Appendix, Part III.A.3. If the firm faces a credibility problem with respect to monitoring, this sanction must be adjusted so that agents' expectations that the firm will monitor optimally are correct. See infra Appendix, Part III.A.4.
149. See supra note 46 (corporate liability should ensure that firms always report wrongdoing).
150. This specification of the sanction assumes that reporting only affects the firm's expected liability for this particular wrong. In other words, reporting one wrong does not deter other wrongs. This assumption is justified if employees are fully informed about the costs and benefits to the firm of reporting any given wrong - since in this case reporting one wrong does not provide employees with any information about the firm's willingness to report other wrongs. If reporting does help deter other wrongs, then a lower default sanction than described here will be capable of inducing optimal reporting. See infra Appendix note 188.
151. See infra Appendix, Equation (19).
152. H, is M* + 3h.
153. See supra Part I.D. 102 Thus, courts could employ a rough rule of thumb to determine the default sanction, provided the standard for optimal policing is set correctly.
154. See Craswell & Calfee, supra note 42, at 280, 288-89. The conclusion that liability may induce excessive policing holds for government imposed civil and criminal sanctions provided that, as is generally the case, the government need not show that the firm's failure to comply with its legal duties "caused" the harm. Cf. Kahan, supra note 56, at 437-39 (showing that uncertainty leads to suboptimal caretaking when private plaintiffs must show "but-for" causation).
155. That is, the court's administrative and information costs. Firms require the same information under all regimes.
156. Courts probably need to know the probability of detection to determine whether the firm took due care. To determine whether due care was taken, courts must know the net social cost of wrongdoing deterred - i.e., the net social cost of the marginal wrong. See infra Appendix, Equation (6). This equals the social cost of wrongdoing to others plus the cost of committing the wrong minus the benefit of the wrong to the marginal wrongdoer and to the firm. As it often will be difficult to determine the benefit of wrongdoing to the marginal wrongdoer, courts often will need to determine that benefit by determining the expected individual sanction, pf, which will equal the net benefit of wrongdoing to the marginal wrongdoer.
157. Under composite liability, but not under sanction-adjusted strict liability, the net social cost of wrongdoing (item 2 in Table 3) must be net of the benefit of wrongdoing to the firm.
158. This assumes that the cost function is well-behaved in that marginal costs are constant or increasing and marginal benefits are constant or decreasing, so that if the court finds that a particular level of monitoring is a local maximum it also is a global maximum.
159. Cf. Mark Grady, A New Positive Economic Theory of Negligence, 92 Yale L.J. 799 (1983) (negligence analysis involves a specific analysis of whether there exist any precautions that defendant should have taken but did not, not a global analysis of what is due care); Mark Grady, Untaken Precautions, 18 J. Legal Stud. 139 (1989) (same).
160. Cf. Mark Grady, A New Positive Economic Theory of Negligence, 92 Yale L.J. 799 (1983) (negligence analysis involves a specific analysis of whether there exist any precautions that defendant should have taken but did not, not a global analysis of what is due care); Mark Grady, Untaken Precautions, 18 J. Legal Stud. 139 (1989) (same).
161. r(M)W). Unlike in the case of sanction-adjusted strict liability, courts need not necessarily determine the cost to the wrongdoer of committing the wrong because, although the net benefit of the wrong to an agent equals his direct benefit, b, minus his cost of committing the wrong, the direct benefit should equal the marginal cost of doing the wrong, which equals his expected fine, pW, plus the cost of committing the wrong. So the net benefit simply equals the expected individual sanction.
162. We expect that courts often will be able to determine whether a firm has detected misconduct. They will generally have access to the information produced both by the firm's monitoring programs and audits and by its internal investigations. See supra note 49 (discussing Fifth Amendment). The government also may be able to reduce the likelihood that detected wrongs will go unreported by providing properly designed bounties to individuals who report wrongs. See Arlen & Kraakman, supra note 11 (discussing bounty provisions).
163. Under sanction-adjusted strict liability and a composite regime, firms will need to be able to calculate optimal monitoring and investigation and determine whether to report wrongdoing.
164. For a complete discussion of this regime, see infra Appendix, Part III.B.
165. As before, the present analysis assumes that the firm cannot commit to reporting wrongdoing it detects. Therefore, the aggravation provision will induce optimal reporting only if reporting lowers the firm's expected liability for the wrong it has detected. See supra text accompanying notes 97-99 and infra text accompanying Appendix note 188.
166. A/g(M°), where g(M°) is the probability the government will detect the wrong if the firm does not monitor optimally and does not report the wrong.
167. A/g°. This condition ensures that even if the firm cannot guarantee that it will report wrongdoing (for example, because of agency problems), it nevertheless will have an incentive to undertake efficient monitoring.
168. This sanction will induce optimal monitoring because in this situation, the firm's expected costs if it monitors optimally and reports detected wrongdoing equal social costs when monitoring is optimal. The firm's expected costs if it does not monitor optimally (but does report) equal social costs when monitoring is suboptimal. Notice that, under this regime, there is no liability enhancement effect because the firm dramatically reduces its expected liability by talcing due care. See infra Appendix, Part III.A.2 (showing that this regime will not induce excessive monitoring).
169. See supra Table 4.
170. See infra Appendix, Part III.C.1.
171. m. This amount must equal the social cost of wrongdoing to others.
172. See infra Appendix, Part III.C.
173. hh, weighted according to the probability that the firm or the government will detect wrongdoing first. See supra note 119.
174. R are variable and depend on the firm's actual monitoring efforts (provided the firm at least took due care). Under such a regime, the firm will not have an incentive to engage in excessive monitoring because implementing additional monitoring above the required level will not change the firm's expected sanction, which equals the social cost of wrongdoing to others.
175. R the firm reduces its expected liability by monitoring excessively and it may be profitable to do so.
176. R, because the firm will always detect first and thus can ensure itself complete mitigation by reporting. As a result, the optimal sanctions under this regime will be identical to those under the mitigation-aggravation regime, because there will be no risk that the government will get there first.
177. Mitigation-mitigation also may be superior when the central concern is that courts will not set the monitoring standard high enough. The mitigation-mitigation provision provides firms with an extra incentive to monitor - that is, the hope that by doing so they will be able to detect and report misconduct before the government discovers it. See supra text accompanying notes 122-23 and note 123. By contrast, under the mitigation-aggravation provision, a firm that will be deemed to have monitored optimally has little to no incentive to incur additional monitoring costs to ensure that it detects wrongdoing.
178. See Orts & Murray, supra note 5, at 22-24. Privilege statutes thus differ from use immunity in that a privilege protects the material from discovery for any purpose, whereas use immunity allows the government to obtain the material for use against others, but not for use against the firm.
179. See id. (describing exceptions to standard privilege).
180. See Dana, supra note 5, at 971 (stating that 18 states have recently adopted environmental audit privilege or immunity statutes).
181. See Arlen, supra note 5, at 865-66 (discussing evidentiary privileges); see also supra Part II.A.
182. Particularly when the firm must report the wrongdoing in order to keep the audit privilege, auditing could increase the firm's expected liability unless firms that report are granted mitigation. Cf. Orts & Murray, supra note 5, at 45-69 (proposing evidentiary privilege which excludes information about underlying facts but noting that privilege must be accompanied by mitigation provision).
183. See supra Part II.A.3.
184. See Arlen, supra note 79; cf. Arlen, supra note 5, at 865-66 (considering modified privilege, akin to use immunity, under which information cannot be used against firm but can be used against individual wrongdoers).
185. See Arlen, supra note 79; cf. Arlen, supra note 5, at 865-66 (considering modified privilege, akin to use immunity, under which information cannot be used against firm but can be used against individual wrongdoers).
186. See Dana, supra note 5, at 993-1000 (discussing impact of privilege laws on relative expenditures on audits and other enforcement measures).
187. Perhaps for this reason, some states with environmental audit privileges also grant immunity from prosecution to firms that conduct environmental audits, report detected wrongdoing, and take prompt action to correct the wrong. See id. at 971 n.7.
188. In addition, the optimal default sanction imposed on firms that do not report detected wrongdoing must also be higher under an audit privilege regime.
189. Cf. Arlen, supra note 5, at 865-66 (risk of insolvency also is greater if state employs modified use immunity rule under which information cannot be used against firm).
190. See Sentencing Guidelines, supra note 6, at § 8. Environmental crimes are governed by all provisions of the Sentencing Guidelines except those relating to fine amounts. Courts may apply the Sentencing Guidelines by analogy.
191. Under the Sentencing Guidelines, the base fine equals the greater of (i) the amount determined from a fine table (which is based on offense level); (ii) the pecuniary gain to the organization from the offense; or (iii) the pecuniary loss from the offense caused by the organization, to the extent the loss was caused intentionally, knowingly, or recklessly. See id. at § 8C2.4. Often, the pecuniary loss will exceed the other measures. But see infra text accompanying notes 151-57 (discussing base fine).
192. See Sentencing Guidelines, supra note 6, at § 8C2.5(f).
193. See id. at § 8C2.5(g). Should the firm satisfy the last one or two requirements it receives partial mitigation.
194. See EPA Guidelines, supra note 8; Antitrust Division Guidelines, supra note 8. See generally Laurence A. Urgenson, Voluntary Disclosure: Opportunities and Issues for the Mid-1980s, 943 PLI/Corp 225 (June 1996). This description of the programs focuses on their common features. For a more detailed discussion of the programs, see Urgenson, supra. In addition, there is evidence that in other areas some prosecutors have decided not to prosecute when firms with an ongoing compliance program report wrongdoing to the government and take any necessary steps to correct it. Firms are still subject to substantial civil penalties, however, which may include treble damages in the case of antitrust violations and double to treble damages in the case of government procurement fraud, as well as a possible risk of a qui tam action. See Gruner, supra note 9, at § 8.5.2 (discussing prosecutorial trends under the Sentencing Guidelines and mitigation where potential defendants adopt voluntary reforms).
195. For example, under our simple composite regime the amount of mitigation is h/p**-h/ P°.
196. Indeed, careful examination of the Sentencing Guidelines reveals that they particularly disadvantage large firms. Under the Sentencing Guidelines, the sanction equals the base fine (which may equal the harm caused) multiplied by a culpability score (or "multiplier") which reflects the firm's culpability. Mitigation provisions affect the size of the multiplier. Under the Guidelines, all firms are eligible for 5 points of mitigation for reporting, investigation, and accepting responsibility for the wrong. This in turn results in a reduction by 2 in the maximum multiplier - e.g., from 4 to 2 - and a reduction by 1 in the minimum multiplier - say from 2 to 1. This would appear to grant equal mitigation to all firms. But this is misleading. Under the Sentencing Guidelines, larger firms generally can expect to start with a higher culpability score than smaller firms, say a maximum multiplier of 4 instead of a maximum of 2.4. The 2-point decrease in the multiplier granted to a larger firm - from 4 to 2 - only decreases its expected sanction by 50%. By contrast, a 2-point reduction in the multiplier imposed on a firm facing a multiplier of 2.4 would reduce its multiplier to 0.4 - resulting in a six-fold reduction in its expected sanction.
197. r, only if 2g>1. This implies that the firm will report only if the probability the government will detect a wrong the firm has detected equals or exceeds 50%. This amount of mitigation, therefore, will not always be sufficient to induce reporting.
198. See Sentencing Guidelines, supra note 6, at § 8C2.5(f). In addition, there is a rebuttable presumption against mitigation if an individual within substantial authority personnel participated in the offense. See id. The Commentary to the Sentencing Guidelines provides as follows: "High-level personnel of the organization" means individuals who have substantial control over the organization or who have a substantial role in the making of policy within the organization. The term includes: a director; an executive officer; an individual in charge of a major business or functional unit of the organization, such as sales, administration, or finance; and an individual with a substantial ownership interest. Id. at § 8A1.2 cmt. 3(b). "Substantial authority personnel" means individuals who within the scope of their authority exercise a substantial measure of discretion in acting on behalf of an organization. The term includes high-level personnel, individuals who exercise substantial supervisory authority (e.g., a plant manager, a sales manager), and any other individuals who, although not a part of an organization's management, nevertheless exercise substantial discretion when acting within the scope of their authority (e.g., an individual with authority in an organization to negotiate or set price levels or an individual authorized to negotiate or approve significant contracts). Whether an individual falls within this category must be determined on a case-by-case basis. Id. at § 8A1.2 cmt. 3(c).
199. See id. (defining high level personnel).
200. See Sentencing Guidelines, supra note 6, at § 8C2.5 cmt. 13.
201. See infra Appendix, Part III.C.
202. See supra Part III. This rule should apply even if the firm's benefit exceeds the harm, provided that this private benefit is one that society counts as a social benefit. The social cost of the wrong should include the dynamic costs of wrongs, including victims' expenditures to prevent such wrongs. See Fred S. McChesney, Boxed In: Economists and Benefits from Crime, 13 Int'l Rev. L. & Econ. 225, 229 (1993) (discussing this point); Fred S. McChesney, Desperately Shunning Science?, 71 B.U. L. Rev. 281, 285 (1991) (same); Gordon Tullock, The Welfare Costs of Tariffs, Monopolies and Theft, 5 W. Econ. J. 224, 228-31 (1967) (same).
203. See supra Part III. This rule should apply even if the firm's benefit exceeds the harm, provided that this private benefit is one that society counts as a social benefit. The social cost of the wrong should include the dynamic costs of wrongs, including victims' expenditures to prevent such wrongs. See Fred S. McChesney, Boxed In: Economists and Benefits from Crime, 13 Int'l Rev. L. & Econ. 225, 229 (1993) (discussing this point); Fred S. McChesney, Desperately Shunning Science?, 71 B.U. L. Rev. 281, 285 (1991) (same); Gordon Tullock, The Welfare Costs of Tariffs, Monopolies and Theft, 5 W. Econ. J. 224, 228-31 (1967) (same).
204. See supra Part III. This rule should apply even if the firm's benefit exceeds the harm, provided that this private benefit is one that society counts as a social benefit. The social cost of the wrong should include the dynamic costs of wrongs, including victims' expenditures to prevent such wrongs. See Fred S. McChesney, Boxed In: Economists and Benefits from Crime, 13 Int'l Rev. L. & Econ. 225, 229 (1993) (discussing this point); Fred S. McChesney, Desperately Shunning Science?, 71 B.U. L. Rev. 281, 285 (1991) (same); Gordon Tullock, The Welfare Costs of Tariffs, Monopolies and Theft, 5 W. Econ. J. 224, 228-31 (1967) (same).
205. As previously noted, the precise statement of the optimal residual liability is more complicated to the extent that the Sentencing Guidelines' regime is a mitigation-mitigation regime.
206. Specifically, under the Sentencing Guidelines, the fine is based on the pecuniary losses caused by the firm's misconduct only if these losses were caused intentionally, knowingly, or recklessly - and even then, only if these losses exceed both the firm's pecuniary gain and an arbitrary amount set forth in the Offense Level Fine Table. See Sentencing Guidelines, supra note 6, at § 8C2.4.
207. See id. (detailing arbitrary threshold amounts for each offense level).
208. See id. at § 8C2.6 (detailing minimum and maximum multipliers for different culpability scores).
209. The comment about prevention measures depends in part on whether courts con-sider prevention measures in assessing the firm's eligibility for mitigation for "prevention and detection." Some of the measures we consider to be prevention measures - in particular salary structure - probably will not be examined under § 8C2.8(f) of the Sentencing Guidelines, in which case the strict liability residual is needed to provide adequate incentives to undertake these measures.
210. See supra note 8. The Environmental Protection Agency refuses to adopt a blanket immunity proposal, but has said it generally would not seek gravity-based penalties and would refrain from recommending firms for prosecution if the firm detects the wrong as a result of a comprehensive environmental audit, takes prompt steps to correct the wrong, cooperates with the EPA, and outlines a program to prevent future violations. See Urgenson, supra note 141, at 235-36.
211. See Gruner, supra note 9, at § 8.5.2 (discussing prosecutorial trends under Sentencing Guidelines and mitigation where potential defendants adopt voluntary reforms).
212. The Sentencing Guidelines determine the basic parameters of the firm's liability without regard to other forms of liability or to whether the market will force the firm to bear some or all of the cost of wrongdoing to others. Judges can take into account collateral consequences of conviction, including civil obligations, in determining what fine to impose within the range of fines set by the Sentencing Guidelines, but such considerations do not affect the base fine amount. See Sentencing Guidelines, supra note 6, at § 8C2.8. For a discussion of the need to consider market forces in determining the appropriate sanction for fraud, see Jonathan Lott, Jr., The Level of Optimal Fines to Prevent Fraud When Reputations Exist and Penalty Clauses Are Unenforceable, 17 Managerial & Decision Econ. 363 (1996).
213. As previously noted, the Sentencing Guidelines may induce efficient policing measures in some cases, but the mitigation provisions will not induce efficient policing measures in other circumstances. To the extent that policing is suboptimal, more wrongs will be committed than is socially desirable.
214. Allowing for variations in wealth only complicates the analysis without changing any fundamental conclusions. For a discussion of strict corporate liability when wealth (and thus insolvency points) vary, see Jennifer Arlen, Controlling Corporate Torts: Strict Liability Versus Negligence Reconsidered (June, 1996) (unpublished manuscript, on file with author).
215. within the realm of misconduct that we are particularly concerned with - socially undesirable misconduct - the wage per worker will not include any expected individual liability because the misconduct is an affirmative wrong (which the worker need not take care to avoid), the marginal wrong is socially undesirable and, at the margin, the firm does not want to encourage such wrongs because its expected liability equals the social cost of wrongdoing to others. Nor does the wage reflect the net benefit of wrongdoing (k-(b-pf-C)), because it is assumed that b is not observable ex ante, although it is observable ex post. See infra note 162. Infra-marginally, when harms are socially desirable, firms might try to encourage a worker to commit a wrong. This can be accommodated in this model by having B equal the firm's net benefit and b equal the worker's benefit including any payments from the firm. Finally, the firm will not penalize the worker for committing a wrong because the government sanctions the worker to the full extent of the worker's wealth. See infra notes 165 & 171 and accompanying text. The model could easily be expanded to allow the firm to penalize workers without changing the central results of the paper. For a discussion of supra-compensatory wages, see supra note 36 (explaining why firms will not eliminate this problem by paying supra-compensatory wages) and Shavell, supra note 36.
216. These benefits include both direct monetary benefit (including impact on employees' stockholdings), and the benefit to the employee of any increase in salary and job security resulting from the wrong. The latter will be particularly hard for the firm to determine because it will depend on the employee's level of risk aversion, preference for intangibles (such as loyalty and community), and the degree to which the employee's human capital is firm-specific.
217. Each employee's benefit of wrongdoing may be unknown to the firm because it is unobservable or it is prohibitively expensive for the firm to determine the benefit of wrongdoing with respect to each employee and all possible wrongs ex ante. See supra note 161.
218. For simplicity, it is assumed that the agent's benefit is independent of the corporation's benefit. The possibility that the firm can share its benefit with the agent is discussed supra Part I; see also supra note 160.
219. See Becker, supra note 20; see also supra text accompanying notes 20-23 (discussing the limits of pure agent liability).
220. See supra note 21 (explaining why the problem of agent insolvency cannot always be solved with nonmonetary sanctions). Where nonmonetary sanctions are appropriate, W can be reinterpreted as the maximum amount of monetary and nonmonetary sanctions that the government can optimally impose. It is optimal to set the sanction equal to the maximum possible sanction - rather than employing a lower individual sanction and higher firm sanction - because it is assumed that imposing a fine is costless, therefore this instrument should be employed to its fullest extent. By contrast, corporate sanctions, while costless to impose, produce costly reactions -e.g., expenditures on monitoring and prevention. Thus, social welfare is maximized by finding the optimal corporate sanctions assuming that society has employed the low cost solution - individual sanctions - to their fullest extent.
221. See supra Part I (defining these terms).
222. See infra note 171 (the firm and its agents face same probability of detection). Throughout this analysis, it is assumed that a firm that reports also cooperates with the government.
223. The equation governing the optimal level of government enforcement is given in Arlen, supra note 5. Directly incorporating government prevention and enforcement expenditures into the model would not significantly change our results. In such a model one would simply calculate the optimal relative levels of government and private expenditures on enforcement, assume that the social planner selects the optimal level of public enforcement, and then proceed as follows to determine what liability rule will induce optimal private enforcement expenditures. See id. (employing such model).
224. Because the firm does not sanction wrongdoers (and so detection by the firm is irrelevant if the firm does not report) p°(M) is effectively the probability the government will detect. See infra text accompanying note 171.
225. See infra note 188 (discussing what happens if state cannot correctly determine whether firm detected wrong).
226. 1(M). This would change the magnitude of the optimal sanction but would not alter our essential results.
227. The present analysis assumes that enforcement measures do not eliminate the problem of agent insolvency. This insolvency constraint is intended to capture the fact that in many circumstances, even if government and firm behavior employ optimal enforcement measures, some nonoptimal wrongs will be committed.
228. The present analysis assumes, and is limited to the circumstances where, Equation (6) is concave.
229. See supra note 46.
230. r(M)W - B.
231. See supra Part I.
232. See supra Part IV.B.
233. For a discussion of the relative merits of the composite regimes see supra Part III.
234. See discussion in supra Part IV.B.
235. See supra Part I.
236. In determining the optimal residual liability, it is assumed that the firm monitors optimally and reports all wrongdoing it detects because the firm is eligible for mitigation only if it engages in such behavior. Appendix, Part III.A.3. determines the amount of mitigation which induces both optimal monitoring and reporting. This equation implicitly assumes that the firm is a perfectly price discriminating monopolist. The results are not qualitatively different if the firm is in a perfectly competitive market.
237. This term is positive provided that the net social cost to others of misconduct with average expected benefit is positive.
238. This analysis assumes that Equations (6) and (11) are concave.
239. r is set optimally, and thus the firm sets P = P*.
240. Equation (16) assumes that a firm that does not monitor will not report because under this regime a firm which does not monitor optimally has no reason to report. Reporting occurs ex post and thus only serves to increase the firm's expected liability for the detected wrong; absent the possibility of fine mitigation, the firm does not benefit. See infra note 188; see also supra note 61 (discussing reputation).
241. H at which strict liability induces the firm to take due care. If the perverse effects of strict liability are severe enough, however, M° will be zero. See supra Part LC. (discussing perverse effects).
242. See supra Part II.B. (discussing composite regimes) and note 61 (discussing reputation).
243. r] where q is the conditional probability that the state having detected the wrong will correctly determine that the firm knew of the wrong and did not report it.
244. See supra Part I.D. What we refer to as the "credibility problem" is often referred to in the economics literature as the "time-consistency problem" or the "commitment problem."
245. See supra note 66 (discussing pure versus mixed strategies). This is a condition for monitoring to be at least locally dominant. In some cases, global dominance may require that it be in the firm's best interests to monitor even if none of the employees believes it will monitor.
246. Equation (20) is simply Equation (17) adjusted to reflect the assumption that all employees believe the firm will monitor and thus monitoring does not affect the number of wrongs that are committed.
247. See Part II.C. and Table 3 (discussing relative information costs in more detail).
248. See supra Table 4.
249. A/g(M°)
250. See supra text accompanying Equation (18).
251. See supra Part I.C. & D.
252. In addition, the firm is given credit for reporting only if it reports before the government detects the wrong on its own.
253. r(M*) is the probability that the government detects before the firm detects and reports, whereas p°(M*) is the probability that the government eventually detects wrongdoing when firms do not report detected wrongs.
254. If the government never detects first then the conditions for optimal residual liability and mitigation for reporting are identical to those under the mitigation-aggravation regime.
255. hh.
256. This assumes that Equation (26) is concave.
257. This assumes that the firm is liable for reported wrongs only if a wrong occurred.
258. One partial solution to this problem might be to alter the residual fine to the level that induces optimal monitoring, and then adopt a duty-based regime to govern prevention (since a duty-based regime will be less sensitive to the fine level). Cf. Cooter, supra note 37. Activity levels would not be optimal, however. Alternatively, the duty-based regime governing monitoring could be set so that the firm is entitled to mitigation only if it engages in the precise optimal level of care, no more no less.
259. See supra Part I.D. (discussing credibility problem in detail).
260. see supra note 61 (discussing reputation as solution to credibility problem); supra note 62 (discussing use of third-party policers).
261. r(M) = p*; p°(M) = p° and g(M) = g. Note that the probability that the firm detects a wrong will be independent of its reporting strategy. The firm will not report if h(P,b)/p* > gh(P,b)/p°. This implies that the firm will not report if p° > gp*. We know the following, where R means firm reports and Pr = Probability: p* = Pr [Caught/firm reports] = Pr [Caught/Firm Detect; R] Pr [Firm Detect; R] + Pr [Caught/Firm Not Detect; R] Pr [Firm Not Detect] = Pr [Firm Detect] + Pr [Caught/Firm Not Detect; R] Pr [Firm Not Detect] p° = Pr [Caught (by gov't)/Firm not report] = Pr [Caught/Firm detect; not report] Pr [firm detect; not report] + Pr [Caught/Firm Not Detect; not report] Pr [Firm Not Detect; not report] = gPr [firm detect; not report] + Pr [Caught/Firm Not Detect; not report] Pr [Firm Not Detect; not report] The fact that the probability the firm detects or not is independent of reporting implies p* = Pr [Firm Detect] + X p° = gPr [Firm Detect] + X where X = Pr [Caught/Firm Not Detect] Pr [Firm Not Detect] Thus gp* = gPr [Firm Detect] + gX < p° = gPr [Firm Detect] + X The firm will not report.
262. See supra Part I.D.
263. See supra Part II.A. (discussing this regime in more detail).
264. See supra Part II.C.
265. Information costs are discussed in detail in supra Part II.C.