Estimating the significance of fixed effects with clustered standard errors

[Jane Wilson]

Hi all,

I am estimating the following equation.

 xtreg lemp lunion lunemployment lfertility lyouth i.year, fe vce(cluster id)

.

I am interested in calculating the significance of the fixed effects. However, since I am clustering my data u_i = 0 no longer appears.

I have tried estimating the equation as:

 reg lemp lunion lunemployment lfertility lyouth i.year i.id, vce(cluster id)

.

Then conducting a Wald test:

 testparm i.id

However, I am obtaining some very high estimates, some over 40,000, so I am not sure I am calculating this correctly.

I would really appreciate any help with this - whether I am computing this incorrectly, or if there is some other issue causing my F-statistics to be so large.

Thank you.