Does agency theory have universal relevance? A reply to Lubatkin, Lane, Collin, and Very

摘要

We are fully sympathetic to the view held by Lubatkin, Lane, Collin, and Very (this issue) that social context is an important and often over-looked influence on principal-agent (P/A) relations. Clearly, P/A relations are embedded within a social context. This context plays an important role in (a) defining self-interest for both agents and principals, (b) determining the degree of potential incongruence between the objectives of agents and those of principals, and (c) designing efficient mechanisms for aligning agent interests with those of principals. Agency theory has been used and interpreted in many ways depending on the author's disciplinary tradition and the problem at hand (see for instance, Eisenhardt, 1989; and Gomez-Mejia, Welbourne, & Wiseman, 2000 for reviews). One of the most popular targets for critics of agency theory (including some of our own work, e.g., Gomez-Mejia & Wiseman, 1998; Wiseman, Gomez-Mejia, & Fugate, 2000) is the P/A model by Jensen and Meckling (1976). Their model has driven a considerable amount of research on P/A relations (according to the Web of Science (2006), it has been cited 3414 times as of this writing) and is thus often used as representative of agency theory. In developing their version of agency theory, Jensen and Meckling incorporated assumptions borrowed from managerial economics. For example, relying on subjective expected utility notions of individual behavior, Jensen and Meckling assumed agents were inherently risk averse. While this may reflect settings where agents have undiversifiable risks, this clearly represents a special case of agency theory, which is mostly applicable to a market economy with dispersed shareholdings and a competitive labor market. As our past work has suggested, agents may or may not be risk averse and thus risk aversion represents a special case of goal conflict (Wiseman & Gomez-Mejia, 1998). Though we certainly do not want to become apologists for a model that we ourselves have criticized, it does seem to us a bit harsh to criticize an entire theory based on one application of the theory.