Maximum likelihood estimation of quantitative spatial economic models

This study presents an approach to apply the maximum likelihood estimation (MLE) method to estimate the parameters in quantitative spatial economic models. The proposed method can be applied to any model in which the unique values of the error terms can be recovered from the observed data on the endogenous variables. Through Monte Carlo experiments, I demonstrate three advantages of MLE over the generalized method of moments estimation in the context of quantitative spatial economics. First, the MLE estimator is less biased and has lower variance. Second, MLE can be applied even if the data on exogenous local variables or multiple-period data are unavailable. Third, MLE is less likely to find parameter values which lead to an unstable equilibrium. Further, I present an empirical illustration of MLE with the help of Japanese data