Abstract:

The Multivariate Spatial Durbin Model (MSDM) is a significant advance in spatial econometrics, very relevant in the context of research problems. This model extends spatial analysis by capturing the complexity and dynamism of interactions between variables in a spatial context that is often ignored by classical spatial models. Furthermore, this article aims to estimate the parameters of MSDM model applied to large and complex data sets through Monte Carlo simulations. This model was then estimated using Maximum Likelihood Estimation (MLE), and to test the accuracy of the model using the Maximum Likelihood Ratio Test (MLRT) with a computational approach. The research results show that the MSDM model para-meter estimates are accurate as indicated by an accuracy value that is smaller than the 5% significance level. The model becomes more efficient as the sample size increases..