Tests for Spatial Correlation of Dependent Variables in Spatial Dynamic Panel Data Models

摘要

This paper considers a spatial dynamic panel data regression model with fixed effects, spatial correlation of dependent variables and error serial among cross sectional units. When the number of individuals n, the number of time periods T are large, and T is asymptotically large relative to n, the paper derives various Lagrange multiplier tests and likelihood ratio test statistics for this panel data regression model including tests for spatial correlation of dependent variables in jointly, marginally or conditionally (one -dimensional and two -dimensional). Limiting null distributions of the tests are derived.