Financial incentives and risk taking in private equity funds

Private equity (PE) funds make investment decisions on behalf of their limited partners (LPs), typically pension funds, university endowments and wealthy individuals. The general partners (GPs), who manage the funds, are compensated through a mix of fixed and variable fees that creates an option-like payoff. In addition, GPs commit to co-invest a fraction of the equity in the portfolio company, along with the LPs. While stock option awards to corporate managers has been studied extensively, little is know n about the incentives provided by the complex compensation structures used in the PE industry. To address this issue, we intend to build a theretical model that generates testable predictions, which we can take to the data. The model will analyze the GP' s incentives to make risky investments as a function of the different components of his or her total compensation. We have already secured access to data on fees and co-investment percentages from several (anonymous) LPs. The information will be derived m anually from private placement memorandums for over 100 Norwegian PE funds. It will be combined with income and balance sheet data for the individual portfolio companies as well as macro-level data on the economy. We are also exploring the possibility to obtain data on the firms' bank relationships. We further intend to collect information on different personal characteristics of the GPs (education, age, etc.), which could help explain their willingness to take risk. We will empirically establish how the different GP compensation components, and in particular the co-investment percentage, affect various aspects of risk taking, for example, portfolio company leverage, bank relationship concentration, industry concentration at the fund level, and the speed with which the fund is invested. We believe that this data set will be unique in its ability to test incentive effects in the PE industry, allowing us to push the research frontier a small step forward