Discussion of “CEO Compensation, Regulation, and Risk in Banks: Theory and Evidence from the Financial Crisis”

The Cerasi and Oliviero contribution in this issue explores an important question from the banking regulation perspective: how does bank CEO compensation affect risk taking? It first lays out a model that illustrates how using variable compensation for bank CEOs (an exogenous free parameter in the model) may have an ambiguous effect on the CEO’s risk-taking behavior. The key feature of the model is that both the bank’s CEO and its shareholders can monitor the project. In the presence of bank leverage constraints and a deposit insurance, more variable compensation may lead to the substitution of shareholder monitoring with monitoring by the CEO. This substitution implies that the direction of the effect of variable compensation on risk taking cannot be signed in general. This is an interesting point because it highlights that theory does not provide clear guidance on whether—and, if so, how—to regulate bank CEO pay to reduce risk taking. Whether capping bank CEO bonus pay curtails bank risk taking and, if so, by how much is a fundamentally empirical question. The fundamental challenge in addressing this empirical question is one of providing the appropriate counterfactual. Suppose one observed in the data that CEOs with a higher proportion of variable compensation take more risk. Does this imply that capping variable compensation leads to less risk taking? Unfortunately, the answer is no. A correlation between variable compensation and risk taking may arise even if risk taking by the bank is unaffected by compensation. Cheng, Hong, and Scheinkman (2015) make this point in a model where banks with heterogeneous productivity and riskiness set the optimal variable component of the compensation contract. As in the present paper, the resulting correlation between bank risk and