Published online by Cambridge University Press: 29 December 2016
This paper provides a structural approach to testing investment equations based on the log-likelihood function of a nonlinear investment rule. The analysis integrates the predictions of the q-theory for the commonly studied active region of investment and provides new inferences on how real and financing frictions affect the probability that a firm invests. The empirical findings are consistent with the macro-finance literature suggesting that q-theory models with nonconvex investment frictions better explain the data. I also find that both real and financing costs of investment are related to the capital intensity of the industry in which firms operate.