Esttab and vce

[Ariel Karlinsky]

Esttab/Estout is a user add-on, can be found here: http://repec.sowi.unibe.ch/stata/estout/

Is there anyway to change the labeling of vces per what I want? for example, this code:

clear all
sysuse auto
eststo: reg price mpg
eststo: reg price mpg, vce(cluster rep78)
esttab, label stats(vce)

will output this:

----------------------------------------------------
                              (1)             (2)   
                            Price           Price   
----------------------------------------------------
Mileage (mpg)              -238.9***       -226.4   
                          (-4.50)         (-2.74)   

Constant                  11253.1***      10965.2** 
                           (9.61)          (6.89)   
----------------------------------------------------
vce                           ols         cluster   
----------------------------------------------------
t statistics in parentheses
* p<0.05, ** p<0.01, *** p<0.001

I wish to output something like this:

----------------------------------------------------
                              (1)             (2)   
                            Price           Price   
----------------------------------------------------
Mileage (mpg)              -238.9***       -226.4   
                          (-4.50)         (-2.74)   

Constant                  11253.1***      10965.2** 
                           (9.61)          (6.89)   
----------------------------------------------------
Standard Errors          Regular     Repair Record Clusters   
----------------------------------------------------
t statistics in parentheses
* p<0.05, ** p<0.01, *** p<0.001

Can this be done?

[Ariel Karlinsky]

I found a solution with estout's "substitute()" command. just use:

 esttab ,label stats(vce) substitute(ols Regular cluster "Repair Record Clusters")

[Max Piper]

I believe I have a similar question. I want to display different models in one table that use the "vce" option, but that cluster standard errors by different variables in the table.
For example, the dataset used below contains weight data on 48 pigs over time, which I, for the sake of this example, divide into 6 pig families.

Now, I want to compare the same regression, once with standard errors clustered at the -pig- level (each with a different -id-), once with them clustered only at the -pigfamily- level (each family has a separate -pigfamily- value).

This means that the estout-output should indicate in the "vce" row not only "cluster", but what cluster. How can this be done?
Below is the stata code, and the output.

webuse pig.dta, clear

gen pigfamily = round(id/10)+1
estimates clear
eststo: reg weight week, vce (cluster id)
eststo: reg weight week, vce (cluster pigfamily)

esttab * using "thatlldo.tex", replace stats(N r2_a vce, labels("observations" "adjusted \(R^{2}\)" "VCE estimation")) ///
    substitute (cluster "pig-level SE" "pig-family-level SE")

{

\def\sym#1{\ifmmode^{#1}\else\(^{#1}\)\fi}

\begin{tabular}{l*{2}{c}}

\hline\hline

&\multicolumn{1}{c}{(1)}&\multicolumn{1}{c}{(2)}\\

&\multicolumn{1}{c}{weight}&\multicolumn{1}{c}{weight}\\

\hline

week & 6.210\sym{***}& 6.210\sym{***}\\

& (67.39) & (65.19) \\

[1em]

_cons & 19.36\sym{***}& 19.36\sym{***}\\

& (47.87) & (25.86) \\

\hline

observations& 432 & 432 \\

adjusted \(R^{2}\)& 0.930 & 0.930 \\

VCE estimation& pig-level SE & pig-level SE \\ [this should say "pig-family-level SE" or similar!]

\hline\hline

\multicolumn{3}{l}{\footnotesize \textit{t} statistics in parentheses}\\

\multicolumn{3}{l}{\footnotesize \sym{*} \(p<0.05\), \sym{**} \(p<0.01\), \sym{***} \(p<0.001\)}\\

\end{tabular}

}

[Ariel Karlinsky]

You can use clustvar for this:

sysuse auto
eststo: reg price mpg
eststo: reg price mpg, vce(cluster rep78)
esttab, stats(vce clustvar)

[Max Piper]

Thank you. That works.
But do you also know how the variable label, instead of the variable name of clustvar, can be displayed? using the label option of estout only displays labels of dependent and independent variables:

sysuse auto,clear
estimates clear
eststo: reg price mpg
eststo: reg price mpg, vce(cluster rep78)
esttab, label stats(vce clustvar)

Best,
Max

[Ariel Karlinsky]

Just use substitute:

esttab, label stats(vce clustvar) substitute(rep78 "Repair Record")