Saving the coefficients of the time fixed effect and averaging it

[Lily Ksh]

Dear Statalists,

I would like to make out of sample predictions based on a two way fixed effect model. Since I cannot get the time fixed effect for out of sample, I would like to store the estimated time fixed effect coefficients and average it to use it for predictions. For eg: xtreg y x1 x2 x3 i.year, fe. I would like to average all of the estimated coefficients for the year dummy variables so that I can add it to to the estimates from "predict yhat, xb"

[Andrew Musau]

These coefficients are not estimated consistently if the T dimension is not very large.

I would like to average all of the estimated coefficients for the year dummy variables so that I can add it to to the estimates from "predict yhat, xb"

As you include the year effects as indicators, the prediction already includes them.