Missing SE in xtlogit Model with Random Effects

[Mike Zhang]

Dear Statalist,

I am using STATA 17 and have run into an issue with the xtlogit regression model I am working on. The model includes random effects and clustered standard errors at the state level. Each observation is a firm-year. I have also included year and city dummies in the model. I have noticed that the standard error for one of my control variables ("Control1"), a continuous firm-level variable, is missing in the output, but the issue is resolved when I switch to a normal logit model. Can anyone explain why this is happening and what steps I should take? Additionally, would it be acceptable to simply use the normal logit model instead? I have attached my code and output for your reference.

xtlogit DV control1 control2 control3 i.year i.cbsacode1 i.industry, re vce(cluster state)
-------------------------------------------------------------------------------------------
| Robust
DV | Coefficient std. err. z P>|z| [95% conf. interval]
--------------------------+----------------------------------------------------------------
Control1 | -5.914205 . . . . .
Control2 | .6409042 .078724 8.14 0.000 .486608 .7952003
Control3 | 1.61525 .1153349 14.00 0.000 1.389197 1.841302

logit DV control1 control2 control3 i.year i.cbsacode1 i.industry, vce(cluster state)

-------------------------------------------------------------------------------------------
| Robust
DV | Coefficient std. err. z P>|z| [95% conf. interval]
--------------------------+----------------------------------------------------------------
Control1 | -5.914214 .4270465 -13.85 0.000 -6.751209 -5.077218
Control2 | .6409045 .0795968 8.05 0.000 .4848976 .7969114
Control3 | 1.615251 .1135764 14.22 0.000 1.392645 1.837856

Thank you in advance!

[Mike Zhang]

Any help you can provide would be greatly appreciated

[Carlo Lazzaro]

MIke:
it's difficult to say without an excerpt/example of your dataset on which to test your code.
It is also weird that Stata did not throw any note before the output table.
In addition, what do you mean by "normal logistic regression"? A -logit- with default standard errors?

[Mike Zhang]

MIke:
it's difficult to say without an excerpt/example of your dataset on which to test your code.
It is also weird that Stata did not throw any note before the output table.
In addition, what do you mean by "normal logistic regression"? A -logit- with default standard errors?

Thanks for your reply. Oddly, Stata didn't give any note on this. And sorry for the confusion. "Normal logistic regression" refers to logit with clustered SE.