Errors occur when doing bootstrapping with xtivreg and clustered standard errors at firm level

[Haoxi Liang]

Hi, my current object is to run a IV regression with fixed effects and clustered standard errors at firm level, and to have the bootstrapped standard errors from 100 repetitions for hypothesis testing.
But I have errors occur when running commands as below:

xtset gvkey fyear
bootstrap, reps(100) cluster(gvkey): xtivreg y (x1 x2 x3 x4 = z1 z2 z3 z4) w1 w2 w3 w4 w5, fe vce(cluster gvkey)

Errors occur like

Bootstrap replications (100): repeated time values within panel
the most likely cause for this error is misspecifying the cluster(), idcluster(), or group() option
r(451);

And I have also tried similar commands like:

xtivreg y (x1 x2 x3 x4 = z1 z2 z3 z4) w1 w2 w3 w4 w5, fe vce(bootstrap, reps(100)) cluster(gvkey)

But it seems that `xtivreg` does not have the `cluster` option.

I have already checked for any duplicated observations, and things go well when I run without IV before:

xtreg y x1 x2 x3 x4 w1 w2 w3 w4 w5, fe vce(bootstrap, reps(100)) cluster(gvkey)

I wonder whether there will be other ways to do such regression with established commands, thank you!

[Andrew Musau]

Clustering in panels is a bit more complicated. See https://www.stata.com/support/faqs/s...th-panel-data/.

[Haoxi Liang]

Clustering in panels is a bit more complicated. See https://www.stata.com/support/faqs/s...th-panel-data/.

Thank you very much for your kind reply and I have successfully run the below commands by referring to your link:

gen new_gvkey = gvkey
tsset new_gvkey fyear
bootstrap, reps(100) cluster(gvkey) idcluster(new_gvkey): xtivreg y (x1 x2 x3 x4 = z1 z2 z3 z4) w1 w2 w3 w4 w5, fe vce(cluster gvkey)

And I wonder whether I get you right through the commands above, thank you again!

[George Ford]

a successful kluge.

might try ivreghdfe. it does not require you xtset/tsset the data.

[Haoxi Liang]

a successful kluge.

might try ivreghdfe. it does not require you xtset/tsset the data.

Thank you for your kind advice, but when I tried to run the `ivreghdfe`:

ivreghdfe y (x1 x2 x3 x4= z1 z2 z3 z4) w1 w2 w3 w4 w5, absorb (gvkey) vce (cluster gvkey)

It appears that

option vce() not allowed
r(198);

or

ivreghdfe y (x1 x2 x3 x4= z1 z2 z3 z4) w1 w2 w3 w4 w5, absorb (gvkey) cluster (gvkey)

It appears that

last estimates not found
r(301);

I wonder whether I understand your advice well, thank you again!

[George Ford]

make sure you install ftools, reghdfe, and ivreg2

https://github.com/sergiocorreia/ivreghdfe#installation