Revealed Preferences for Corporate Leverage

I consider a simple model of dynamic capital structure with adjustment costs, and I show how financial decisions reveal firms' preferences for leverage. I show that the most commonly used empirical models of capital structure (the partial adjustment model, and the financial deficit model) are mis-specified. Estimating a correctly specified model yields new insights and solves spurious puzzles: The expected future financial deficit is a significant determinant of the current choice of debt versus equity, and profitable firms have a higher target leverage. However, while the value functions that I estimate display qualitatively sensible properties, they also appear extremely flat for moderate values of leverage, which is inconsistent with tax-based theories. The methodology that I propose can be readily extended to test the importance of other determinants of capital structure decisions. As an application, I show that CEO tenure flattens the left part of the estimated value function, which is consistent with agency theories.