Essays on internal capital markets

My dissertation investigates the internal capital allocation decision of a multidivision firm by building a model and testing the theoretical predictions using segment level investment data. It contains three chapters. The first chapter studies how distortions in internal capital allocations can arise due to asymmetric information and limited commitment. Division managers have private information and may leave firms' to seek venture capital (VC) financing if their ideas are not funded. Headquarters designs a mechanism to induce truth-telling and to retain managers with good quality projects. This chapter shows that, in contrast to the results established in the literature, the optimal mechanism involves both under- and over-investment in good quality projects. The second chapter extends the analysis to include product market competition. If managers leave to form a new firm using VC financing, depending on the structure of the product market competition, a new entrant may reduce the profits of exiting firms. This chapter shows that, as a result, firms may be more willing to retain managers by bearing the cost of information rent to managers of low quality projects, further exacerbating the inefficiency in internal capital markets. The paper also derives empirical implications linking product market characteristics and the levels of VC efficiency to investment inefficiency in internal capital markets. The third chapter tests these implications using segment level investment data from Compustat. In this chapter, IPO activity at the SIC two-digit level is used as a proxy for managers' outside options. The segment data is grouped into smallest segments that operate in industries with low IPO activity and largest segments that operate in industries with high IPO activity. This chapter documents that investment in smallest segments increases with the IPO activity of the largest segments and that smallest segments tend to over-invest compared to the stand-alone firms. These results are consistent with the theoretical predictions of the model: managers of low quality projects receives information rent in the form of higher capital allocation, and the rent increases as the outside offers of the managers with high quality projects increases.