Risk and the role of collateral in debt renegotiation

In his basic model of debt renegotiation, BESTER [1994] argues that collateral is more effective if high risk projects are financed. This result, however, crucially depends on the definition of risk. Using the second-order stochastic dominance criterion introduced by ROTHSCHILD AND STIGLITZ [1970], we show that it is not a project's high risk, induced by a high probability of default, that makes collateral more effective. Instead it turns out that, given the expected return, the probability of default has no impact on the collateral's effectiveness. Moreover, a higher risk of the project caused by a higher loss given default makes the use of collateral even less effective.