Bootstrap standard error

[kamalesh pahurkar]

I want to bootstrap the standard errors with 1000 replications. In my regression specification standard errors are clustered. How can I do this?

[Joseph Coveney]

Try the cluster() option of bootstrap.

For example, using the auto dataset

regress mpg c.displacement, vce(cluster rep78)

would become

regress mpg c.displacement, vce(bootstrap, reps(1000) cluster(rep78))

Don't forget to set the seed beforehand.

[kamalesh pahurkar]

Thanks Joseph, what is seed? Why we have to set it?

[Carlo Lazzaro]

Kamalesh:
setting the -seed- makes your data being the same every time you run the code:

. use "C:\Program Files\Stata17\ado\base\a\auto.dta"
(1978 automobile data)

. regress price mpg, vce(bootstrap, reps(200) seed(12345) dots(1))
(running regress on estimation sample)

Bootstrap replications (200)
----+--- 1 ---+--- 2 ---+--- 3 ---+--- 4 ---+--- 5 
..................................................    50
..................................................   100
..................................................   150
..................................................   200

Linear regression                                    Number of obs =        74
                                                     Replications  =       200
                                                     Wald chi2(1)  =     18.91
                                                     Prob > chi2   =    0.0000
                                                     R-squared     =    0.2196
                                                     Adj R-squared =    0.2087
                                                     Root MSE      = 2623.6529

------------------------------------------------------------------------------
             |   Observed   Bootstrap                         Normal-based
       price | coefficient  std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
         mpg |  -238.8943    54.9405    -4.35   0.000    -346.5758   -131.2129
       _cons |   11253.06   1319.023     8.53   0.000     8667.824     13838.3
------------------------------------------------------------------------------

. regress price mpg, vce(bootstrap, reps(200) seed(12345) dots(1))
(running regress on estimation sample)

Bootstrap replications (200)
----+--- 1 ---+--- 2 ---+--- 3 ---+--- 4 ---+--- 5 
..................................................    50
..................................................   100
..................................................   150
..................................................   200

Linear regression                                    Number of obs =        74
                                                     Replications  =       200
                                                     Wald chi2(1)  =     18.91
                                                     Prob > chi2   =    0.0000
                                                     R-squared     =    0.2196
                                                     Adj R-squared =    0.2087
                                                     Root MSE      = 2623.6529

------------------------------------------------------------------------------
             |   Observed   Bootstrap                         Normal-based
       price | coefficient  std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
         mpg |  -238.8943    54.9405    -4.35   0.000    -346.5758   -131.2129
       _cons |   11253.06   1319.023     8.53   0.000     8667.824     13838.3
------------------------------------------------------------------------------

. regress price mpg, vce(bootstrap, reps(200)  dots(1))
(running regress on estimation sample)

Bootstrap replications (200)
----+--- 1 ---+--- 2 ---+--- 3 ---+--- 4 ---+--- 5 
..................................................    50
..................................................   100
..................................................   150
..................................................   200

Linear regression                                    Number of obs =        74
                                                     Replications  =       200
                                                     Wald chi2(1)  =     18.65
                                                     Prob > chi2   =    0.0000
                                                     R-squared     =    0.2196
                                                     Adj R-squared =    0.2087
                                                     Root MSE      = 2623.6529

------------------------------------------------------------------------------
             |   Observed   Bootstrap                         Normal-based
       price | coefficient  std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
         mpg |  -238.8943   55.31386    -4.32   0.000    -347.3075   -130.4812
       _cons |   11253.06   1316.147     8.55   0.000      8673.46    13832.66
------------------------------------------------------------------------------

.

[kamalesh pahurkar]

Thank you for clarifying, Dr. Carlo

[Carlo Lazzaro]

Kamalesh:
Carlo (who just exploited Joseph's enlightening nitty gritty on this topic) is enough. Thanks.

[kamalesh pahurkar]

I'm using ivreghdfe, facing an errror - option vce() not allowed