[bug] esttab ommit the beta coefficient for interaction term

[fredericky123]

sysuse auto,clear
eststo:  regress price weight c.mpg##i.foreign,beta
esttab ., beta

The output of beta coefficient for the interaction are omitted

[benjann]

Option beta in esttab calls estadd beta, which has been written long before Stata had factor variables and does not work with interactions. However, since regress nowadays stores the beta coefficients in e(beta) you can simply read the values from there. Example:

. clear all
. sysuse auto
. regress price weight c.mpg##i.foreign,beta
. esttab, main(beta) nobase drop(_cons)

----------------------------
                      (1)   
                    price   
----------------------------
weight              1.216***
                   (6.36)   

mpg                 0.516*  
                   (2.38)   

1.foreign           1.754***
                   (4.08)   

1.foreign#~g       -1.244** 
                  (-2.83)   
----------------------------
N                      74   
----------------------------
beta coefficients; t statistics in parentheses
* p<0.05, ** p<0.01, *** p<0.001

or

. estout, cells(beta)

-------------------------
                        .
                     beta
-------------------------
weight           1.215683
mpg              .5162482
0.foreign               0
1.foreign        1.753747
0.foreign#~g            0
1.foreign#~g    -1.243539
-------------------------

ben

[fredericky123]

Thanks! However, I used the command reghdfe, I hope there will be one day esttab also support the interaction

[sergiocorreia]

If reghdfe is missing an e() matrix it should be pretty easy to add btw

[fredericky123]

sysuse auto,clear
eststo:  reghdfe price weight c.mpg##i.foreign,
esttab ., beta

Could you please give me an example