The governance debate identifies the central problem of the separation of ownership and control in the large corporation and centers on the alignment of the agent's interests with those of the principal. Key factors in monitoring senior managements' performance include the composition and independence of outside board members, issues of transparency, outside reporting, accounting standards, and shareholder composition. Empirical research on corporate governance (e.g., Klein 2000; Peasnell et al. 2002) typically investigates quantifiable relationships between measures of corporate performance and specific remedies to agency problems, including the number and independence of directors on a company board or board committees and the independence of external auditors. This large body of research identifies important aspects of the problem of minimizing the conflicts between principal and agent.
Until fairly recently, however, issues of bias in human cognition and perception, decision making under uncertainty, risk assessment, and the impact of emotion and affect on behavior received somewhat less attention in this literature. Monitors are frequently assumed to be rational actors (Fama 1980; Shleifer and Vishny 1997; Prentice 2000). (1) Findings from cognitive psychology and behavioral studies, however, indicate that judgment, decision making, and behavior are not exclusively based on logical reasoning but are also subject to numerous heuristics and cognitive biases (Tversky and Kahneman 1974; Kahneman and Tversky 1979; Fischhoff 2002), affect (Slovic et al. 2002), visceral factors (Schelling 1984; Loewenstein et al. 2001), and pressures toward conformity with the group or authority (Asch 1956; Janis 1982). Divergence from utility maximization over time adds a temporal dimension to this literature (Strotz 1955; Thaler 1981). These influences tend to steer human judgment, inference, and behavior away from the predicted outcome of expected utility theory and lead to systematic violations of the normative assumptions central to the economist's rational model.
Members of the board of directors and external auditors are, for example, expected to care about their reputation and their prospects in the job market, which is theorized to discipline their actions. There are, however, limits to reputation as a determinant to behavior. Auditors and board directors are, no less than other individuals, subject to the common human preference for immediate gratification, typically with insufficient regard for potential negative future consequences. The gratification from a bonus, re-election to a board of directors, renewal of an auditing contract, or the prospects of employment in a client's firm is certain and experienced in the present or the immediate future. In contrast, expected damage from questionable activities, including reputational damage, legal or financial sanctions, and loss of career is merely potential and in the more distant future. The magnitude of negative outcomes may be further discounted and reduced in perceived severity and probability by self-serving justifications and overoptimism. (2)
This paper examines corporate governance from an agency perspective and asks why some of the standard means of monitoring the activities of senior decision makers in large corporations are prone to failure. The focus of the paper is on issues of agent behavior, questioning some of the assumptions of the familiar neoclassical maximizing model. The issue of corporate governance touches on core assumptions in economics concerning the behavior of individual agents, the tendency toward equilibrium in financial markets, the role of the corporate firm, and the subject of regulation. Institutional economics has a long tradition of fundamental criticisms of some of the core assumptions of mainstream economics, in particular with regard to the strict assumptions about rationality and utility maximization. The unease of this school with the neoclassical assumption of perfectly optimizing behavior of agents predates the discussion of human cognition and decision making under uncertainty. More recently, behavioral economics bases its challenge to the dominating assumptions of agent behavior of the normative neoclassical tradition on people's observable behavior. The implications of these findings for the theory on financial markets and corporate governance supports long-held views of institutionalist thought. The present paper provides an analysis of human behavior which may partially bridge the gap between mainstream economic research and the institutionalist tradition.
A key question is whether independence and impartiality of monitors are reasonable assumptions of corporate governance. Arguments are introduced that departures from arm's length outcomes should be expected as the rule, rather than the exception. The second section introduces the principal-agent problem arising from the separation of decision and risk bearing or, more traditionally, the separation of ownership and control. The roles of the board of directors and the external auditors are briefly defined. The third section discusses the rational model and shows how conventional means to minimize the principal-agent problem are premised on critical assumptions underlying this standard model of economics. The fourth section identifies and illustrates a range of biases, heuristics, and social pressures that affect the judgment of individuals and groups. The next section assesses potential partial solutions to the agency problem that center on the feasibility of independence of the board of directors and external auditors. This is contrasted with findings from cognitive, social, and behavioral research that demonstrates how corporate governance mechanisms may be undermined by common human traits.
The Principal-Agent Problem
The governance debate discusses the problem of the separation of ownership and control, seminally identified by Adolph Berle and Gardiner Means (1932) for the large corporation where managers do not bear the major share of the wealth effects of their decisions. The problem of how to control the manager of a firm is hardly new. Researchers from Adam Smith (1776) to Berle and Means and, more recently, Michael C. Jensen and William C. Meckling (1976), Eugene Fama (1980), and Andrei Shleifer and Robert Vishny (1997) have suggested top management may decide expropriation of the owner/ shareholder is an optimal choice decision, and superior to depending on competition in the market for managers. This raises the issue of how to impose discipline on the actions of the manager. It would be ideal if the two parties could devise a contract that regulates what the manager can do with the firm's capital, how any surplus gets divided, and what happens if any of the covenants are violated. Since it is not possible to contract for every contingency, however, complete contracting is not feasible (Williamson 1988).
The next best solution to minimize agency problems might be perfect monitoring, but again this is impractical as monitoring costs rise very rapidly and excessive monitoring can lead to counterproductive responses in the firm (e.g., evasion, damage to motivation). Applied solutions to minimizing the agency problem have aimed at an effective design of incentive contracts and monitoring systems, realizing that, for practical reasons, the manager keeps a considerable degree of freedom within these constraints (Grossman and Hart 1986). Important elements of contracts and monitoring systems include the establishment of a board of directors and various board committees, the use of external auditing, public reporting of audited financial reports, adherence to established accounting rules and principles, supervision of areas of corporate activity by government and nongovernment agencies, and legal liability provisions.
An evaluation of the ways in which incentive contracts have been designed suggests that these may, in part, be due to motivations other than a minimization of the agency problem (Paul Krugman, "The Outrage Constraint," The New York Times, August 23, 2002). Much of the notion of stock options disciplining managers, for example, is based on the assumptions of the efficient market hypothesis (EMH). However, given numerous recent examples of rewards for managerial underperformance, why should the solution to bad management be a bigger paycheck? The managerial power view interprets some current compensation practices as evidence of the agency problem rather than as a remedy to it (Bebchuk and Fried 2003). The paper will return to these arguments.
The Board of Directors
The board of directors is a control mechanism of "professional referees" (Fama and Jensen 1983a) designated to discipline the top management of the firm and, if need be, to replace it with more effective individuals (Hermalin and Weisbach 2003). The inclusion of independent directors on the board is thought to lower the probability of collusion of management to expropriate security holders (Peasnell et al. 2001). Ideally, managers' incentives are closely aligned with those of shareholders through the use of efficient compensation contracts and close monitoring of executive performance (see, e.g., Grossman and Hart 1983 for an outline of the optimal contracting view). Recent research contradicts this ideal view of the board's monitoring function. The influence of the CEO in selecting board members can have a seriously negative impact on the board's effectiveness (Hermalin and Weisbach 2003), and board capture has been identified as a significant problem (Krugman 2002; Bebchuk and Fried 2003). While the Sarbanes-Oxley Act of 2002 provides the board's nominating committee with greater powers over the election of directors, the impact of this provision on the functional independence of directors remains to be seen.
External audits serve to evaluate an organization's...