Standard erros vs. robust standard errors

[Guest]

Hi,

I've made a regression analysis (OLS) for stock returns and found that heteroscedasticity was present. Therefore, I clustered standard errors at the firm level. This is correct, right?
Furthermore, can someone explain what is the difference between standard errors and robust standard errors?

Thank you in advance!

Best,
Guest

[Carlo Lazzaro]

Guest:
if you detect heteroskedasticity only after -regress-, you should simply invoke the -robust- option, that actually takes heteroskedasticity into account (whereas -vce(cluster clusterid)- option was coneceived for dealing with autocorrelation).
See -help _robust- as a good first step to take to address your last question.