Agency conflicts in financial contracting with applications to venture capital and CDO markets

Abstract

In these papers I examine efficient financial contracting when incentive problems play a significant role. In the first chapter (joint with Z. Fluck and S. Myers) we focus on the venture capital industry. We build a two-stage model capturing moral hazard, effort provision, and hold-up problems between entrepreneurs and investors. Across multiple financing scenarios we solve numerically for optimal decision policies and NPV, finding significant value losses from first-best. A commitment to competitive syndicate financing increases effort and NPV and benefits all parties. However, syndicate financing raises potential information problems, and the fixed-fraction participation rule of Admati-Pfleiderer (1994) fails with endogenous effort. We find that debt financing is often less efficient than equity financing, for while it improves effort incentives it worsens hold-up and debt overhang problems in later-stage financing. In the next chapter I turn to the collateralized debt obligation or "CDO" market. CDOs are closed-end, actively-managed, highly leveraged bond funds whose managers typically receive subordinated compensation packages. I develop a model of manager trading behavior and quantify under-investment and asset substitution problems, calibrating to market parameters.

(cont.) Compared to prior studies, I find similar value losses to senior investors and significantly higher increases in debt default risk and spread costs. However, for even extremely conservative effort assumptions, the ex-ante benefit of greater effort incentives outweighs risk-shifting costs, rationalizing observed contracts. I also analyze the ability of various payout policies and trading covenants to curtail risk-shifting. Excess interest diversions, contingent trading limits, and coverage test "haircuts" of lower-priced assets are effective measures and increase allowable leverage and equity returns. In the final chapter I examine the empirical relationship between CDO trading, manager compensation, and fund performance from 2001-2004. Using a large panel data set, I find a statistically significant relationship between trades which add volatility to the portfolio and the level of subordinated manager compensation. Worse deal performance increases risk-shifting behavior so long as subordinate investors are still in-the-money. Tendencies to group trades and the effect of managerial reputation are also considered.