Jensen and Meckling (1976) first discussed importance of agency costs in determining the optimal capital structure, arguing such costs arising from conflicts between different groups within a firm increase as the firm employs more debt financing. Galai and Masulis (1976) formalized their idea in the classical option pricing framework of Black and Scholes (1973), regarding equity as call option on firm value. Adopting the project of the highest volatility will maximize equity value possibly at the sacrifice of debt value. Assuming managers act in the interests of equityholders, managers are able to transfer wealth to equityholders from bondholders by taking excessive risk. Such an incentive problem is referred to as risk-shifting or asset substitution.
The fairly large literature on risk management and capital structure theories are developed under the widely accepted assumption that an increase in leverage increases equityholders risk-shifting incentive and associated agency cost of debt. However, a little thought will reveal such a naive point of view lacks theoretical justification. Since standard call option premium is a monotone increasing function of volatility, the optimal asset volatility for equityholders is infinite irrespective of leverage level. The firms in reality do not seem to seek for projects with infinite volatility.
Gavish and Kalay (1983), in their pioneering study, demonstrates equityholders' gains from an unexpected increase in variance of the investment do not increase monotonically with the firm's leverage. The gains are expected to decrease when leverage is high. Their analysis raises a serious question whether the agency costs of asset substitution really increase monotonically in leverage. In a similar setting, Green and Talmor (1986) confirm Gavish and Kalay's result and further formalize the optimal risk policy of the firm.
They show as the promised debt payment increases, equityholders' risk-shifting incentive increases monotonically, although equityholders' gains do not. Although these studies provide profound insights regarding the nature of agency costs due to asset substitution, it should be pointed out they are based on the restrictive models in which all agents are risk neutral and the risk-free interest rate is zero. Their assumption on the future cash flows or asset returns is fairly general, however, it precludes important probability distributions such as log-normal, which we will adopt in the following analysis.
Download
Risk Shifting Incentive of the Firm when its Value is Correlated with Interest Rates
