# Is there a specific test for this procedure? (NKPC and coefficient stability)

I'm running a GMM regression in order to estimate a New Keynesian Phillips Curve, with forecasting purpose, and i want to know if there is a testing procedure for the following problem:

(i) First i estimate the parameters with the full sample, 195 observations.

(ii) Then i estimate using 2/3 of the sample, 130 observations (not using the 1/3 final observations).

I want to know which testing procedure should i use to evaluate if the estimated coefficients had a significant change from the (ii) regression sample to the (i) regression sample, i.e., to test if they remained stable during the final 1/3. I'm using Eviews 10.

• Should Andrews-Fair wald statistic do the job? – EMSO Nov 4 '19 at 3:59