Coordination (Games) in Organizations

23 October 2006 at 12:46 pm Nicolai Foss Leave a comment

| Nicolai Foss |

The basic thrust of new institutional economics and contract-theory approaches to organizations and contracts is to conceptualize virtually any issue related to economic organization in terms of solving incentive conflicts. The motivation for this presumably is an underlying argument that in the absence of such conflicts, the first-best can be reached without major obstacles.

The view that all, or most, organizational phenomena are reducible to problems of aligning incentives is one that is implicitly (and sometimes explicitly) contradicted by contributions to organization studies (e.g., March & Simon, Thompson), the executive and leadership literature (Barnard; Carlsson; Kotter; Selznick), political science (Calvert), and (rational choice) sociology (Coleman). For example, in Foundations of Social Theory, Coleman observes that charismatic authority may be a response to coordination problems that don’t necessarily turn on misaligned incentives. Implicitly, such contributions suggest that the prisoners’ dilemma isn’t the only kind of interaction structure that is helpful for understanding organization; coordination games may also hold potential in this respect.

Economists have typically not been much interested in coordination games in the context of (economic) organization (however, see this paper; and here is my own modest contribution on this). Economists arguably have tended to think of coordination games as trivial; except in the case of multiple identical outcomes, the problem of how to coordinate on an optimum seems trivial. However, accumulating experimental and theoretical work on learning in games has revealed that coordination games aren’t necessarily trivial. The increased importance of complementarities has further directed attention to coordination games.

A recent experimental paper by Jordi Brandts and David Cooper, “A Change Would Do You Good … An Experimental Study on How to Overcome Coordination Failure in Organizations,” published in the June issue of the American Economic Review, takes ideas on coordination games in an exciting direction. Here is the abstract:

We study how financial incentives can be used to overcome a history of coordination failure using controlled laboratory experiments. Subjects’ payoffs depend on coordinating at high effort levels. In an initial phase, the benefits of coordination are low and play typically converges to an inefficient outcome. We then explore varying financial incentives to coordinate at a higher effort level. An increase in the benefits of coordination leads to improved coordination, but large increases have no more impact than small increases. Once subjects have coordinated on a higher effort level, reductions in the incentives to coordinate have little effect on behavior.

This is clearly a highly puzzling finding. Brandts and Cooper try to explain by arguing that incentives may function as a kind of focal point for coordination rather than as devices that call forward more effort per se:

Breaking the trap of coordination failure requires a coordinated move to higher effort levels. An increase in the bonus rate provides the necessary coordinating device. If we look at the individual level data, virtually all employees respond to an increase in the bonus rate by bumping up their effort levels regardless of the size of the increase. If enough employees make a large increase to their effort level, other employees will generally follow their lead and the firm will overcome its history of coordination failure. The bonus increase is more important as a trigger starting the process of change rather than as a determinant of the eventual quantitative outcome of this process. This also explains why cutting the bonus rate generally doesn’t undermine successful coordination – the primary effect of increasing the bonus, coordinating a mutually desirable shift to higher effort levels, isn’t relevant for firms that are already coordinated and face a decreased bonus rate.