Hi,

I am running a multivariate probit model to assess the music consumption habits of a sample of individuals (about 230 people). I have 4 equations with mainly dummy explanatory variables. When I run the multivariate model with all 4 equations using the cmp function, the output is fine. However, I removed one of my equations because it was not significant overall when run as an individual regression. When I run the model with only three equations, the standard error for the constant term in the regression for one of my dependent variables (PIRACY) is now missing.

Why might this happen? Is it a big issue?

Many thanks