تنظیمات مدیران برای عوامل با ترجیحات اجتماعی

Abstract This study explores a nested representation of ethical, moral, social identity, motivated, opportunistic and reciprocal agent preferences to characterize screening contracts in a principal–agent model under adverse selection. This leads to a ranking of the type of social preferences that principals should seek in agents, based upon the information rents associated with each agent type. When moral hazard is introduced the ranking further depends upon the interaction of limited liability with self-selection. These results are interpreted in light of the 2010 Dodd-Frank Act and principal–agent experiments.

. Introduction The decade 2000–2009 has come to be regarded as ‘The Noughties;’ primarily for the economic and financial crises brought on by a breakdown in corporate ethics, laissez-faire regulation and limited liability in leveraged securitization. A rogue's gallery of corporate misbehavior in this era would likely include executives at Adelphia, Enron, Fannie Mae, Freddy Mac, Goldman Sachs, Madoff Investments, Tyco and WorldCom. Table 1 shows that those who were convicted of corporate wrongdoing received sentences that were often far longer than the average number of years given for a murder conviction in the US. Personal consequences aside, the crises of the early and late Noughties brought on the destruction of untold wealth via two major recessions. Table 1. Penalties for unprincipled behavior during the ‘Noughties’. Violator Firm Sentence Average murder sentence in state courts Bernie Ebbers WorldCom 25 years Full-size image (<1 K) Dennis Kozlowski Tyco 8–24 years Kenneth Lay Enron 20–30 yearsa Bernie Madoff Madoff Investment Securities 150 years Timothy Rigas Adelphia 20 years Richard Scrushy HealthSouth 6.8 years Jeffrey Skilling Enron 24 years a Lay passed away before appealing or serving his sentence. Table options The canonical approach to the study of corporate governance in financial economics – agency theory – was created in recognition of the potential for opportunistic behavior in organizations characterized by principal–agent relationships. Yet (rightly or wrongly) the conventional wisdom is that high-powered compensation packages based on this theory were at the root cause of unprincipled behavior such as revenue smoothing; backdating options; mark-to-magic accounting; the lend-to-securitize criterion for subprime mortgages; and the warehousing of collateralized debt obligations (owing to limited liability). Indeed, with the context of the early Noughties clearly in mind, a founding father of the agency-theoretic approach to corporate governance, Jensen (2007), has identified agent integrity as a key element for successfully operationalizing the agency framework. For example, Jensen points out that without integrity agents with stock options that are in-the-money would have little reason to report the truth when their firm's stock is currently overvalued. For Jensen, integrity is a positive rather than normative concept and is itself not inconsistent with the agency theoretic approach to corporate governance. Consequently, if agent behavior reflects a preference for integrity this can potentially mitigate some of the unintended consequences associated with pay-for-performance. 1 This then raises the greater issue of what types of agent preferences serve to enhance a firm's value? As firms are social organizations they are a natural environment for social preferences to be expected to arise. The positive examination of the implications of an agent's social preferences for contract design is currently somewhat of a growth industry. An incomplete list of social preferences that have been considered includes ethical, inequity averse, intrinsic, mission-oriented, moral, motivated, reciprocal, social identity, social norm and virtuous. What all of these preferences have in common is that they are examples of what Jensen (2008) calls the positive analysis of normative values. To wit, just as economics has traditionally focused on the positive analysis of alternative institutional structures, it can also shed light on how normative values reflected in standards of behavior affect the management of conflicts of interest. In this way, the present study steps away from a situation of de gustibus non est disputandum with respect to agents’ preferences and evaluates a subset of the aforementioned preferences within a screening environment where the principal is uninformed about an agent's social preferences or lack thereof (adverse selection). As Delfgaauw and Dur (2007) note, what separates this environment from one in which the principal is uninformed about an agent's ability/productivity is that intrinsic motivation affects an agent's willingness to work even in the absence of productivity differences. The aforementioned social preferences reflect intrinsic concerns about opportunistic behavior.2 Examples of such intrinsic concerns include an agent's own behavior with reference to a social standard or the type of firm with whom they are employed. In particular, consider an exogenously determined standard of effort, e*. Some agents may experience intrinsic disutility if e* is not met. Depending upon the functional form of this disutility, the agent can be said to possess moral preferences ( Casadesus-Masanell, 2004); social identity preferences ( Akerlof and Kranton, 2008); or ethical preferences ( Stevens and Thevaranjan, 2010). In a certain respect there is nothing preventing any firm from expressing a standard for effort, yet opportunistic agents and the firms that employ them are rarely modeled within this context. Alternatively, opportunistic agent behavior can be subject to a social constraint, as is the case for the virtuous agents in Carlin and Gervais (2009). Instead of facing an incentive compatibility constraint, virtuous agents set their level of effort equal to the maximum allowed. This is consistent with viewing ethical behavior as socially constrained behavior (e.g., Arce, 2004). Another type of social preference that falls into this category are agents who prefer to work for certain firms because either the work itself achieves a social purpose or the firm's output fulfills a social purpose. These are known as (intrinsically) motivated agents or mission-oriented firms. In particular, Besley and Ghatak (2005) and Makris (2009) consider the case in which agents are motivated by how the firm's mission serves a social purpose; e.g., public hospitals, public universities and nonprofits. Alternatively, Delfgaauw and Dur (2007) consider agents who are socially motivated by the work (effort) itself. Perhaps the work is fun or the agent lives by the adage, any job worth doing is worth doing well. For example, workers at the internet shoe giant Zappos.com purportedly adhere to a “work hard, play hard” mentality. A commonly invoked defense of the assumption of agent opportunism is that not all agents are opportunistic, but one has to account for those agents who are. Such an accounting therefore connotes adverse selection in agent types. By contrast, the studies cited above primarily focus on the contractual implications of social preferences under moral hazard, even though adverse selection is a natural context for accounting for agents with (non-)opportunistic preferences. Indeed, uncertainty over an agent's type introduces contractual distortions even in the absence of moral hazard. Moreover, the fixed component of a contract reduces the opportunity cost of ethical/socially concerted behavior (Osterloh and Frey, 2004 and Arce, 2011); and what is observed is an emphasis on the fixed component of pay as compared to the contract offered to opportunistic agents. Within the context of the Noughties, our results are consistent with the way in which Treasury Department pay czar Kenneth Feinberg has unilaterally cut incentive compensation in favor of raising the fixed (salary) component of pay for C-level executives in firms receiving bailouts (Enrich and Solomon, 2009). Implicit in this approach is the argument that increased fixed pay and decreased incentive pay reduce the incentive for misbehavior when performance criteria are out-of-the-money and also reduce the opportunity cost of ethical behavior that may adversely affect performance criteria. Hence, the focus is on the effect of agents’ social preferences on the fixed and variable components of pay as compared to opportunistic agents. Consideration of screening for social preferences under adverse selection represents an intuitive yet previously unexplored alternative to the existing literature, where it is either assumed that an agent's social preferences are observable or an exogenous (efficient) matching process exists for pairing principals with agents. Endogenizing the sorting process through screening contracts leads to several novel results. For example, the principal cannot screen between the different social preference types considered in this study. Consequently, we present a series of results where screening takes place between opportunistic agents and each particular form of social preferences. Under pure adverse selection agents with social preferences exert more effort and require less pay-for-performance (lower bonuses) than do their opportunistic counterparts. Yet once moral hazard (in the form of limited liability) comes into play some agents with social preferences earn informational rents whereas others do not. This is because agents with social preferences differ in the extent to which they substitute fixed pay for bonuses. Hence, it becomes possible to rank social preference types in terms of the informational rents they require for screening to successfully occur. The overall message is that once the possibility of agents with social preferences is recognized, screening requires differences in the composition of pay (fixed versus bonuses). The analysis proceeds as follows. Section 2 introduces a nested characterization of ethical, moral, social identity, motivated, opportunistic and reciprocal agent preferences. Screening contracts for each type of agent (versus an opportunistic one) are derived in Section 3 for the case of pure adverse selection. In Section 4 the fixed component of pay is discussed in a mixed setting involving both moral hazard and adverse selection. This is followed by a discussion of the results in terms of the Noughties, the Dodd-Frank Act and the experimental literature on social preferences.

Discussion and conclusion Corporate misbehavior during The Noughties (2000–2009) has focused attention on the role of high-powered incentives that were originally introduced to align the interests of principals and agents. Indeed, the Congressional summary to the 2010 Dodd-Frank Act (DFA) states that the purpose of DFA is to, “create a sound economic foundation to grow jobs, rein in Wall Street and big bonuses, end bailouts and too big to fail, and prevent another financial crisis.” In operationalizing DFA Treasury Department pay czar Kenneth Feinberg has altered the composition of executive compensation to emphasize fixed pay over bonuses. The underlying rationale is increased fixed pay (salary) will (i) reduce the opportunity cost of ethical behavior that may potentially reduce performance measures; and (ii) reduce the incentive for unprincipled behavior to achieve incentive pay when performance measures are out-of-the money. Affected corporations have almost unilaterally reacted that restrictions and regulations on compensation will make them less attractive for employees and therefore less profitable. This begs the question as to what types of agents will be attracted to accept this new pay structure? This study shows that it is possible to successfully screen for non-opportunistic agents who will exert more effort and accept less incentivized compensation as compared to opportunistic agents. This is accomplished by considering the current literature on agency and social preferences, which is concerned with the well-accepted concept that behavior within firms reflects both market and social considerations. In particular, various alternative social preferences have been proposed for agents; yet little is understood about the relationship between them. Hence, a first contribution of this paper is to compare ethical, social-identity, mission-oriented, and moral agent preferences within an adverse selection environment. Once agents are differentiated according to preferences (type) the issue of adverse selection requires the design of a menu of contracts that screen for agents with social versus opportunistic preferences. Moreover, screening contracts under adverse selection naturally address the issue of executive compensation and recruitment under the DFA regime. In particular, these preferences can be categorized into two types: norm-based and motivated preferences (N- and M-types, respectively). It is shown that M-types are a nested subset of N-types. When ranking agent types according to information rents that are paid by the principal, N-types do not earn information rents but opportunistic agents (O-types) do. By contrast, O-types do not earn information rents when the principal instead screens for M-types. The difference is due to the additional compensation N-types potentially receive when they do not achieve the social norm/standard for effort set by the principal. When the model is extended to allow for moral hazard (via limited liability) in addition to adverse selection, agents with social preferences accept a contract that has a higher fixed and lower (or no) incentive (bonus) component as compared with opportunistic agents. This finding is consistent with a McKinsey report by Day et al. (2002) on pay for performance in the aftermath of the dot.com/new economy scandals.11 They suggest that weakening incentive structures and fostering an inclusive corporate culture results in a proper distinction between motivation and alignment that is beneficial to firm and employee alike. The present paper therefore provides a formal analysis in which behavior within the firm is simultaneously market- and socially-driven. The results are also consistent with the Dodd-Frank Act's recommendation for altering the composition of executive compensation and the conventional wisdom that pay-for-performance incentivized unprincipled behavior during the Noughties. At the same time, it must be acknowledged that unprincipled behavior in the form of revenue smoothing, mark-to-magic accounting, etc., is not directly addressed. This is a subject for future research. Nevertheless, the model is consistent with the vast majority of agency-theoretic models of corporate governance that take the incongruence between principal and agent to be adequately represented through an agent's effort aversion. In all cases the agents with social preferences exert more effort than opportunistic agents. Hence, there is less opportunistic behavior, suggesting that the composition of pay may be an effective screening device for addressing concerns stemming from the corporate malfeasance associated with the Noughties. In addition, norm-based agents exert more effort than motivated agents and are preferred to motivated agents in terms of information rents. The results can also be considered in light of recent experiments on agency and social preferences. For example, Fehr et al. (2007) consider a moral hazard experiment and find for a wide variety of treatments that the use of verification technologies, bonus agreements and agent shirking substantially differ from what is theoretically predicted. Both principal and agent tend to be “fairer” than would be expected from opportunistic agents and principals who assume they are dealing with opportunistic agents. That is, principals pay more and agents exert more effort than is theoretically predicted for opportunistic agents. The authors appeal to social preferences to explain the difference and employ inequity aversion due to its tractability. Interestingly enough, in each of their treatments the principal specifies a standard for effort, e*, which is consistent with inducing preferences for norm-based agents; yet the contractual forms they consider are never a function of e*, as e* does not enter into the social preferences they consider (inequity-aversion with respect to payoffs). Within this context Result 2 identifies the dependence of the contractual forms on e* and suggests the need to consider a combination of fixed and incentive components of pay in order to screen for agents with social preferences. In adverse selection experiments with agents possessing differential skills sets, Cabrales and Charness (2011) find complete separation among skill types – using fixed contracts – and a rejection rate of contracts that is higher than theoretically predicted. This suggests that incentive-compatible contract menus can be very effective, so long as they are not unfair. Whereas in Cabrales and Charness hidden information stems from unknown skill levels, in the present study the principal does not know the agents’ social preferences. Here the separation between opportunistic and N- or M-types involves a bonus menu instead of a fixed contract menu.