Dear all,
My main model is y = x1 + x2 + e
I would like to conduct the DWH test for endogeneity for variable x1. If I understand it correctly, when one uses the fixed-effect model, we can regress the endogenous variable on the set of exogenous variable and then use predict the residuals by using the comand:
predict uhat, resid
However, in my case, the Hausman test along with interest time-invariant variables used in the main model indicate that I would have to use the random-effect model. Then I run the regression of my endogenous variable using the xtreg, re. However, the command "predict uhat, resid" does not work after the xtreg x1 z1 z2, re
Would that it be correct in this case to use "predict uhat, e"?
Thank you
My main model is y = x1 + x2 + e
I would like to conduct the DWH test for endogeneity for variable x1. If I understand it correctly, when one uses the fixed-effect model, we can regress the endogenous variable on the set of exogenous variable and then use predict the residuals by using the comand:
predict uhat, resid
However, in my case, the Hausman test along with interest time-invariant variables used in the main model indicate that I would have to use the random-effect model. Then I run the regression of my endogenous variable using the xtreg, re. However, the command "predict uhat, resid" does not work after the xtreg x1 z1 z2, re
Would that it be correct in this case to use "predict uhat, e"?
Thank you
