S-GMM interpretation

[Philippe Caeyers]

Hello everyone!

I'm confused by my Stata-output. I'm doing a system-GMM estimation with bank risk as a dependent variable. My main variable of interest is supervisory stress, for which I use 2 different proxies (1 and 2 use proxy 1; 3 and 4 use proxy 2). I notice that the significance of some variables completely disappears when switching to the second proxy, look at Bank Size for example. I'm struggling with how to explain this, as there seems to be no relationship between Bank Size and the component that differs across proxies. Could this have to do with my instrumentation?

Here is my code.

xtabond2 risk l.risk size L2GDP L2INFL L2INT L2LI CAP MON SUP DEP LIQ ENT ACT STAFFASSETS2 SKILL y* if inrange(year, 2014, 2019), gmm(l.risk L2GDP L2INFL) iv(L1FREE size y*, equation(level)) robust orthogonal small

Kind regards,

Philippe

[Sebastian Kripfganz]

Note that xtabond2 has a few serious bugs:

1. It is likely to produce incorrect coefficient estimates when used with option orthogonal.
2. When used with time dummies y*, the Sargan-Hansen overidentification test results might be incorrect.

To avoid these problems, I recommend to use the xtdpdgmm command instead:

• Kripfganz, S. (2019). Generalized method of moments estimation of linear dynamic panel data models. Proceedings of the 2019 London Stata Conference.
• XTDPDGMM: new Stata command for GMM estimation of linear (dynamic) panel data models