Do company towns solve free rider problems? A sensitivity analysis of a rational-choice explanation

Coleman (1979; 1994) has argued that sociological rational choice theory is an appropriate tool to analyze organizations. He emphasizes that the theory combines the analytical power of formal decision analysis with sociology’s attention to informal incentives. Our paper aims to caution against a potential danger in Coleman’s research agenda. Substantively important implications of the theory, namely, propositions about the interplay of formal and informal mechanisms, may be highly sensitive to specifications of the central assumption of goal-directed decision making. To demonstrate this potential sensitivity, we provide a stylized analysis of the effect of company towns on workers’ compliance with production norms, an example discussed by Coleman. Our study starts out with a forward-looking analysis, employing the orthodox rational choice assumption of a “forward-looking” rational actor. For comparison, we then apply a “backward-looking” learning model. It turns out that results are highly sensitive to variation in the underlying decision model. The forwardlooking model suggests that company towns may help workers to solve free rider problems by informal social control. By contrast, the backward-looking model implies that company towns may foster strong bilateral friendships that undermine workers’ willingness to invest in social control. We argue that such sensitivity is of relevance beyond the scope of this particular example. We therefore propose that rational choice analyses of organizations carefully assess model sensitivity to deviations from the assumption of perfect rationality.