The theory of agency, which has seen many recent applications in the social sciences and management literatures, is essentially a theory of failures: It seeks to understand the problems created when one party attempts to control another's behavior, given that control is costly and it often does not pay to extract perfect compliance from that agent. There are said to be two core problems of agency, moral hazard and adverse selection; Arrow refers to these as the problems of “hidden information” and “hidden action.” Societal failures in control and service are interesting in part because they are so plenary. Given the theoretical importance of agency in elucidating the nature of such failures and the methods devised in social systems to ameliorate them, we might expect to find the types and consequences of agency failures well-understood by now. Indeed, applications of agency theory have cut across the economics literature and entered the study of many disciplinary, not to speak of institutional settings. Thus, besides the strictly formal work, we see agency models applied in such areas as government regulation and bureaucratic behavior (and corruption), interest groups and corporate political action, corporate governance, the employment relation and compensation systems in firms, collective bargaining (and bargaining in general), centralization and privatization in economies, diplomatic behavior, auditing relationships and control in the firm, municipal bond issuance, and so on. Given their importance, it is surprising — indeed, stunning — to find a lack of consensus on exactly what the problems of agency are. This paper analyzes the variety of definitions of these problems that appear in the literature, and serves, in a way, as a survey and reconceptualization of the critical issues surrounding the problems of agency. Some of the major variables that emerge from the analysis are then employed together to identify some major types of agency relations. I conclude that the key definitional distinction relating to these concepts should be observability vs. judgment disability, rather than hidden information vs. hidden action. I then show how the elemental concepts underlying moral hazard and adverse selection can be systematically sorted into parallel processes of problems of agent and principal action, focussing on the principal side. Among the insights generated is that adverse selection is better conceived as two separate, if related problems, those of adverse claims and of adverse performance. The logic of what I call the core agency sequence generates the following problems, respectively, as the principal creates, specifies, observes, interprets, judges, and intervenes/redirects agency: Adverse Claims, Communicative Hazard, Moral Hazard, Cognitive Hazard, Adverse Performance, and Control Hazard. The problem of adverse claims suggests the need to generate what I call a theory of testaments, for which the existing literature in such areas as reputation and legitimacy would be a subset. The paper also identifies and discusses certain other problematic situations in agency, including what I call the problem of distrust in the commons.
Read full abstract