EViews.com

Hi everyone,

I want to estimate a single-equation GMM model with optimal weighting HAC matrix and HAC Newey-West robust standard errors with 3 lags, but I am not sure that's what Eviews 12 is giving.

In the estimation weighting matrix HAC options, I assume I don't have to select whitening options, and that Kernel is Bartlett with Newey West automatic bandwidth with 3 lags, but then there is no option to select the type of standard errors and the pretended lags, as in other software. I am not sure that I am getting it the way I want.

Here is what Capture yields:

{%equation}.gmm(instwgt=hac, instbw=neweywest, instnwlag=3, gmmiter=2) eonia=(1-c(1))*(c(2)+inflation(12)*c(3)+gap(3)*c(4))+eonia(-1)*c(1) @ inflation(-1 to -6) inflation(-9) inflation(-12) gap(-1 to -6) gap(-9) gap(-12) m3(-1 to -6) m3(-9) m3(-12) bondyield(-1 to -6) bondyield(-9) bondyield(-12)
{%equation}.gmm(instwgt=hac, instbw=neweywest, instnwlag=3, gmmiter=2, cov=hac) eonia=(1-c(1))*(c(2)+inflation(12)*c(3)+gap(3)*c(4))+eonia(-1)*c(1) @ inflation(-1 to -6) inflation(-9) inflation(-12) gap(-1 to -6) gap(-9) gap(-12) m3(-1 to -6) m3(-9) m3(-12) bondyield(-1 to -6) bondyield(-9) bondyield(-12)

Picture is what I get:

Thank you in advance

Did you check the Options tab of the estimation dialog?

Hello Gareth,

Yes, I think I can only select the options for the matrix, but I also need to select the type and lags for the standard errors. In the previous post I showed what the estimation delivers. Do I have to assume the standard errors are the ones I want by those results?

Yes, any options you set for the covariance matrix will apply to standard errors