Clustered Std Errs vs Unclustered

[TheBestMilke]

Greetings,

I recently ran two regressions, one clustering around the states from my data and the other without clustering. Both have fixed time effects. I've noticed that the standard errors for my main coefficients under observation are much larger than the standard errors from my unclustered regression. Is there a reason for this difference (the clustered std errors are about twice as big) and could it be an effect of serial correlation?

Thanks!

 

[Stephen Tashi]

Greetings,

I recently ran two regressions, one clustering around the states from my data and the other without clustering. Both have fixed time effects. I've noticed that the standard errors for my main coefficients under observation are much larger than the standard errors from my unclustered regression. Is there a reason for this difference (the clustered std errors are about twice as big) and could it be an effect of serial correlation?

You should have another go at explaining your question. "Clustering" is general idea, but it isn't clear what specific clustering technique you used. Your statement indicates that the data may have a time stamp to it, but not much else.