Hello everybody!
In carryong out my spatial econometrics estimates I came into something strange. I have a panel data with one dependent variable on 4 independent variables.
Now, for a specific indep. variable, the Moran'sI statistics for all the years in questions (20) appears to be non significant. However, when I perform a Durbin SDM (fixed effect) regression, the same variable appears non significant in the non spatial coefficient, but significant in the spatial coefficients (as well as strongly significant in the indirect effects spatial estimation).
The contrary for another regressor, which provided strongly significance degree of global spatial autocorrelation (in the Moran's statistics) but which does not have significant indirect spacial effects..
How is this possible? how can this be interpreted? Is there maybe something wrong? I performed all the diagnostics tests which lead me to accept the two-way durbin model as the correct spatial model, so i really do not know how to conciliate the DSM spatial results with the Moran's results.
Thank you!
Kodi
In carryong out my spatial econometrics estimates I came into something strange. I have a panel data with one dependent variable on 4 independent variables.
Now, for a specific indep. variable, the Moran'sI statistics for all the years in questions (20) appears to be non significant. However, when I perform a Durbin SDM (fixed effect) regression, the same variable appears non significant in the non spatial coefficient, but significant in the spatial coefficients (as well as strongly significant in the indirect effects spatial estimation).
The contrary for another regressor, which provided strongly significance degree of global spatial autocorrelation (in the Moran's statistics) but which does not have significant indirect spacial effects..
How is this possible? how can this be interpreted? Is there maybe something wrong? I performed all the diagnostics tests which lead me to accept the two-way durbin model as the correct spatial model, so i really do not know how to conciliate the DSM spatial results with the Moran's results.
Thank you!
Kodi
