Estimating and Testing Dynamic Corporate Finance Models

Abstract

We assess the finite sample performance of simulation estimators that are used to estimate the parameters of dynamic corporate finance models. We formulate an out-of-sample specification test and propose a new set of statistical benchmarks that can be used to estimate and evaluate these models. These benchmarks are based on model policy functions. Our Monte Carlo simulations show that the estimators are largely unbiased with low root mean squared errors. When computed with an optimal weight matrix, the specification tests associated with the estimators are close to correctly sized. These tests have excellent power to detect misspecification.

Keywords: dynamic models of capital structure, policy function, value function, model evaluation, Monte Carlo

JEL Classification: C14, C52, C61, G31, G32