| 讲座简介： | Abstract: This paper investigates a multivariate spatial autoregressive model which consists of a finite number of equations, incorporates own-variable spatial lags and cross-variable spatial lags as explanatory variables, and allow for correlation between disturbances across equations. In addition, we extend the model to a simultaneous equations spatial autoregressive model to capture the simultaneity in different endogenous variables. We study parameter spaces, the identification of parameters, asymptotic properties of quasi-maximum likelihood estimation, and computational issues. Monte Carlo experiments illustrate the advantages of QML and FIML, broader applicability and efficiency improvement, compared to instrumental variables based estimation methods in the existing literature. |
