| Is the investment behavior of multi-divisional firms driven in part by influence activities in the internal capital market? Using Compustat Segment data, I find that models of influence activities may have considerable value in predicting capital allocation decisions of multi-divisional firms. Headquarters decides how much to invest in small divisions based on the investment committee's recommendation (private signal) and observable characteristics (public signal) of the small division. Using a standard moral hazard model, I show that large, established division managers have an incentive to bias private information about investment opportunities in small divisions and skew capital budgets in their favor. Headquarters can reduce these incentives to engage in costly influence activities, by designing contracts which alter the sensitivity of investment to private and public signals. In evaluating optimal contracts, firms face a trade-off between the value of an accurate signal and the cost of mitigating influence problems. This trade-off and, the resulting implications for investment sensitivity to both private and public signals depends on the severity of the firm's influence problem and the quality of the public signal. Using division profits as a proxy for the public signal, firm attributes as proxies for the severity of the influence problem, and the informativeness of profits in predicting firm value as a proxy for the quality of the public signal, I show that manufacturing firms allocate capital to smaller divisions in a way that suggests that influence problems lead to inefficient capital allocation. Moreover, firms with operations in related or less predictable businesses, flatter organizational structures, and tighter financial constraints appear to be more vulnerable to these problems and, as a result, suffer from greater inefficiencies in their internal capital markets. |
