The Sociology of Groups and the Economics of Incentives:

When working together, people engage in non-contractual and informal interactions that constitute the sociology of the group. We use behavioral models and a unique survey of medical groups to analyze how group sociology influences physician incentive pay and behavior. We conclude that informal interactions among group members influence pay practices and behaviors, but the relationship is complex. No single aspect of group sociology is entirely consistent with all the patterns in the data. Factors emphasized in the economic theory of agency, notably risk aversion, also shape pay policies but these factors cannot account for all the observed empirical relationships. (JEL D21, D22, J40, J41) Economic models of compensation treat pay practices as a solution to an incentive problem. High levels of performance require high levels of effort and, beyond some minimal point, providing this effort is costly. Firms desiring high levels of performance from employees should therefore link economic rewards closely to an individual’s productive contribution. Yet this tight linkage is hardly a universal characteristic of pay systems. Firms exhibit enormous variation in the degree to which compensation responds to individual performance. Explaining this diversity is one of the fundamental tasks for the economics of organizations (Baker, Jensen and Murphy, 1988; Baker, Gibbons and Murphy, 1994, Gibbons, 1996 and Prendergast 1999). This paper introduces the concept of “the sociology of groups” into a theoretical and empirical analysis of variation in pay systems. The term “group sociology” refers to the noncontractual and informal interactions that occur between members of work groups. These interactions can take the form of activities (mutual help, mutual monitoring) and psychological experiences (guilt, envy, shame, greed, peer pressure). It is widely believed among economists, sociologists, and human resource professionals that group sociology influences the design of incentive contracts, but theory and evidence regarding these influences has been scarce. We argue below that group sociology influences both the benefits and costs of incentives and therefore the kind of incentive arrangements that firms will use. Although the economic issues we discuss are quite general, our analysis focuses on pay practices in a narrowly defined setting: medical groups. We concentrate on these organizations for three reasons. First, the medical groups we study tend to be small and to have a flat organizational structure. This simplicity, combined with the fact that the key revenue generating activities (e.g. patient office visits) are regularly recorded for billing purposes, makes it feasible to link compensation to the performance of individual physicians. The second reason for our focus on medical groups is that our data allow us to observe (rather than infer) the incentive formula that prevails in the group. Our sample of medical groups, like professional groups in general, rely upon administrative rules that specify how