Dear Statalist,
I am using Stata 16 on a windows device.
I want to perform the same regression model in various subsamples, each of which is created by splitting the full sample in two parts with a similar number of observations based on different variables.
To give an example, the interest rate a company has to pay shall be explained by it's Return on Assets (RoA) and the Equity ratio (Eq_ratio). Moreover, we have two dummy variables assigning the companies to high- and low risk groups, based on the country of their destination (H_risk_count) and their industry (H_risk_ind):
Now let's assume, I want to perform the regression separately for the high- and low risk group, as defined by both variables. Then, I want to generate a result table that shows only the estimators for RoA. The table should have two columns, one containing all estimators for the different high-risk samples, the other one for the low-risk samples and one row for every risk-defining variable.
Up to now, I've tried the following:
This leaves me with the following result:
However, as described I would like to have only two columns with the results of (3) in (1) and (4) in (2), so I could rename the columns as "high risk" and "low risk".
Is there a way to achieve this without to many manual changes of the output table?
Kind regards
Ingo
I am using Stata 16 on a windows device.
I want to perform the same regression model in various subsamples, each of which is created by splitting the full sample in two parts with a similar number of observations based on different variables.
To give an example, the interest rate a company has to pay shall be explained by it's Return on Assets (RoA) and the Equity ratio (Eq_ratio). Moreover, we have two dummy variables assigning the companies to high- and low risk groups, based on the country of their destination (H_risk_count) and their industry (H_risk_ind):
Code:
* Example generated by -dataex-. For more info, type help dataex clear input float(Interest_rate RoA Eq_ratio H_risk_count H_risk_ind) .04 .1 .4 1 0 .035 .2 .5 1 0 .06 .08 .25 1 1 .05 .4 .5 0 0 .027 .12 .7 1 1 .017 .1 .55 1 1 .03 .08 .3 0 1 .08 .05 .45 0 0 .04 .15 .4 0 1 .032 .1 .6 0 1 .053 .08 .15 0 1 .026 .1 .7 0 1 .091 .05 .4 0 1 .043 .04 .35 0 0 .027 .08 .8 0 1 .038 .15 .5 0 0 .045 .1 .3 1 0 .05 .03 .6 1 0 .065 -.05 .5 1 1 .072 .08 .8 1 0 end
Up to now, I've tried the following:
Code:
preserve keep if H_risk_count == 1 rename RoA country regress Interest_rate country Eq_ratio estimates store a restore preserve keep if H_risk_count == 0 rename RoA country regress Interest_rate country Eq_ratio estimates store b restore preserve keep if H_risk_ind == 1 rename RoA industry regress Interest_rate industry Eq_ratio estimates store c restore preserve keep if H_risk_ind == 0 rename RoA industry regress Interest_rate industry Eq_ratio estimates store d restore esttab a b c d, keep(country industry)
Code:
----------------------------------------------------------------------------
(1) (2) (3) (4)
Interest_r~e Interest_r~e Interest_r~e Interest_r~e
----------------------------------------------------------------------------
country -0.150 -0.0175
(-1.65) (-0.25)
industry -0.223 -0.0352
(-2.04) (-0.75)
----------------------------------------------------------------------------
N 9 11 11 9
----------------------------------------------------------------------------
t statistics in parentheses
* p<0.05, ** p<0.01, *** p<0.001
Is there a way to achieve this without to many manual changes of the output table?
Kind regards
Ingo
Comment