Probit: Heteroscedasticity, autocorrelation

[danielpark]

Hi,

I am currently working on my master thesis in finance. I have a probit model with a set of regressors that I test one-by-one (alone and with a dummy). I want to make sure that my estimates are robust. How can I test for heteroscedasticity in a probit model and how do i calculate robust estimates (as far as I understood, QML is not robust to heteroscedasticity; not sure about GLM)? Also, is there a way to calculate probit autocorrelation-consistent standard errors?

This is a bit out of the field of my studies so I apologize if my questions are stupid.

Thanks in advance.