estat sbknown with panel data

[Lorenzo Capasso]

Hello everyone
I would run the structural break for a panel data, but I have the following problem. Can you help me?

. xtset CompanyID Year

Panel variable: CompanyID (unbalanced)
Time variable: Year, 2018 to 2020, but with gaps
Delta: 1 unit
. reg logdissent Board_women lnCEO_pay Institutional_Ownership Board_Size REMCOM_size Board_Independence R
> EMCOM_Independence CEO_duality lnCEO_tenure lnTotal_Asset Stock_Return ROA MBV Leverage Volatility Board
> _minority REMCOM_minority Chairman_min i.Industry_dummies, vce(robust)

Linear regression Number of obs = 232
F(27, 204) = 5.40
Prob > F = 0.0000
R-squared = 0.3371
Root MSE = 1.6593

-----------------------------------------------------------------------------------------
| Robust
logdissent | Coefficient std. err. t P>|t| [95% conf. interval]
------------------------+----------------------------------------------------------------
Board_women | -1.653851 1.742134 -0.95 0.344 -5.088747 1.781046
lnCEO_pay | .0422516 .162857 0.26 0.796 -.2788471 .3633503
Institutional_Ownership | 2.420419 1.258245 1.92 0.056 -.0604144 4.901252
Board_Size | .0438092 .056622 0.77 0.440 -.0678301 .1554486
REMCOM_size | -.2619893 .1973738 -1.33 0.186 -.6511434 .1271649
Board_Independence | .7951045 1.019648 0.78 0.436 -1.215295 2.805504
REMCOM_Independence | -1.737793 .7230597 -2.40 0.017 -3.163421 -.3121643
CEO_duality | .3249947 .3446475 0.94 0.347 -.3545332 1.004523
lnCEO_tenure | .0719688 .1705765 0.42 0.674 -.2643503 .4082879
lnTotal_Asset | .3814851 .1635779 2.33 0.021 .0589649 .7040052
Stock_Return | .4271535 .3219049 1.33 0.186 -.2075338 1.061841
ROA | -1.724825 1.665783 -1.04 0.302 -5.009184 1.559535
MBV | .0163252 .0200782 0.81 0.417 -.0232623 .0559127
Leverage | -1.596521 1.012267 -1.58 0.116 -3.592368 .3993265
Volatility | -3.378435 1.672556 -2.02 0.045 -6.676147 -.0807223
Board_minority | -1.456112 1.626032 -0.90 0.372 -4.662096 1.749871
REMCOM_minority | 1.006785 .2510135 4.01 0.000 .5118712 1.501698
Chairman_min | -.316853 .4091401 -0.77 0.440 -1.123539 .4898326
|
Industry_dummies |
2 | 3.061067 .8387441 3.65 0.000 1.407348 4.714786
3 | 2.450273 1.046696 2.34 0.020 .3865438 4.514002
4 | 1.09852 .9455269 1.16 0.247 -.7657381 2.962779
5 | 4.080751 .8843371 4.61 0.000 2.337139 5.824364
6 | 2.639974 .8178243 3.23 0.001 1.027501 4.252446
7 | 2.444533 1.22498 2.00 0.047 .0292873 4.85978
8 | 3.186229 .9021271 3.53 0.001 1.40754 4.964918
9 | 3.263872 1.093146 2.99 0.003 1.108559 5.419184
10 | 2.3831 .9516223 2.50 0.013 .5068233 4.259376
|
_cons | -11.47272 2.585971 -4.44 0.000 -16.57137 -6.374058
-----------------------------------------------------------------------------------------

. estat sbknown, break(2020)
sample may not include multiple panels
r(459);

Thank you for your help

[Clyde Schechter]

-estat sbknown- is designed to test for structural breaks in a single time series; it does not work with panel data. And for that reason, it also is usable only after -regress- or -ivregress 2sls-, not after -xtreg-.

[Lorenzo Capasso]

Thank you Clyde !!!

[Lorenzo Capasso]

Clyde Schechter Does an alternative exist for panel data ?

[Clyde Schechter]

I have not used it myself, so can't evaluate it nor advise on specifics of its use, I believe -xtbreak-, from SSC can do this.

[JanDitzen]

Clyde Schechter is correct. xtbreak can estimate the number of and occurrence for breaks and supports three different tests for structural breaks. We have a GitHub page describing the method and command and there is a working paper describing it as well.