Predicted values from spatial error models

[Frank Porell]

Hello,
I have little experience in estimating spatial regression models. After some work I have been able to estimate a spatial error model and get the same coefficient estimates using spatreg, spreg, and the R program spdep. The issue I have is that when I test the residual errors for spatial autocorrelation I get an insignificant result with R spdep and significant results using spatreg and spreg. It appears that the reason for the discrepancy is that the residuals differ because the predicted values differ in the R program and the other Stata programs. I have not been able to find anything readable that explains how predicted values are generated in these programs. Any advice is appreciated.

[Marcos Almeida]

Welcome to the Stata Forum / Statalist,

Your query is really interesting. It will demand expertise in spatial analysis. What is more, in two statistical packages.

With regards to Stata, I guess the odds of getting insights will increase if you share command , output and data (full, abridged or mock). By the way, there are members of this forum who are authors of spatial analysis user-written programs in Stata.

Hopefully you will get the message information you wish.