Dear Statalisters,
I'm a longtime occasional lurker and a firsttime poster.
I'm using 15.1. I have a very unbalanced panel of N 160 and T 49, but for the Westerlund cointegration tests, I have extracted various strongly balanced N x T panels out of the larger set.
But when I try xtwest on different pairs of variables, about 70 percent of the time I get this error message--

. xtwest wgi aidirc, lags(0 3) leads(0 2) lrwindow(2)

Calculating Westerlund ECM panel cointegration tests.likelihood information not found
> in last estimation results
r(321);

I can't seem to determine the pattern. I have tried doing a simple xtreg on a pair before doing xtwest, no dice.
Thanks in advance for your support.