Michael D. Ryall and Olav Sorenson

Theoretical research on organizations generally presumes that their leaders have the ability to direct the organization towards a set of goals. But that presumption depends crucially on the ability of leaders to understand how particular actions or directives would influence organizational outcomes, a problem of causal inference. We develop a formal model of this problem. Our model reveals that hidden (unobserved) influences can stymie inference. However, these hidden influences only become problematic under a specific set of local conditions. That fact further suggests that organizational design features can help to mitigate this problem. We introduce three types of solutions to the problem of inference in the presence of hidden influences — experimentation, illumination, and substitution — and discuss how these solutions relate to a variety of organizational design features.