Significance test for the difference of two variables from regressions

[Orifjon Kurbanov]

Hello, everyone!

I'm trying to replicate a study for my master's thesis and have an issue with calculating the difference of two variables from the regression results.

1) Firstly, I add a link to a screenshot of what I want from the prior study (Source: Kucheev, Y.O., Ruiz, F. & Sorensson, T. Do Stars Shine? Comparing the Performance Persistence of Star Sell-Side Analysts Listed by Institutional Investor, the Wall Street Journal, and StarMine. J Financ Serv Res 52, 277–305 (2017))

2) Below is the example of my data. I am using a

regress

command where alpha (portfolio return) is on the left side and the market factors are on the right side of the equation. I am repeating the regression for 2 times based on the type of the portfolio (variable 'status'). And storing the results after running

regress

command.

* Example generated by -dataex-. For more info, type help dataex
clear
input byte status double(alpha MktRF SMB HML Mom RF)
1 .017695295640000053                 .0113                .0072                .0003  .0089 0
1 .017695295640000053 -.0007000000000000001                .0097                .0015  .0049 0
1 .023521795424509273                 .0175                -.003                .0103  .0089 0
1 .018342684506806428                 .0049                .0012                .0009  .0001 0
1 .020716443677130325                 .0062                 .001                    0  .0012 0
1  .02524816573129507  .0028000000000000004                .0005  .004699999999999999 -.0022 0
1 .021363832539977423                -.0005               -.0004                .0059  .0013 0
1 .018990073370174443                -.0124 .0028000000000000004               -.0091 -.0103 0
1  .03949072076292093   .009399999999999999               -.0036                .0053  .0052 0
1   .0392749244747177                -.0013                .0023 .0028000000000000004  .0048 0
end

3) I need to calculate the difference between _cons of each regression _cons[regression1] - _cons[regression2] and see the t-statistics to understand whether the difference gives significant results.

I tried to use

test

and

 suest

commands, but they are not giving me the results I want.

Thank you for any advice!

[Carlo Lazzaro]

Orifjon:
see https://stats.idre.ucla.edu/stata/co...s-using-suest/

[Orifjon Kurbanov]

Orifjon:
see https://stats.idre.ucla.edu/stata/co...s-using-suest/

Carlo, thank you!
Maybe I am just confused with statistics, but I would like get the t-stat significance for the difference. test and suest commands give me the chi2, which does not fit my whole regression table. Can you, please, advise me something about the right use of t-test?
Thank you beforehand!