Arellano-Bond Estimator for AR(2) model

[Chinmay Sharma]

Hi All,

My dataset resembles the following:

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(individual time y x lagy lag2y)
1 1999 23  3  .  .
1 2000  2  2 23  .
1 2001  2 32  2 23
1 2002  3  2  2  2
2 1999  2  3  .  .
2 2000  2  2  2  .
2 2001  2  2  2  2
2 2002  2  2  2  2
2 2003  3  3  2  2
3  200  2  2  .  .
3 1999  2  .  .  .
3 2000  3  2  2  .
end

My model is basically an AR(2) model, with the dependent variable y being determined by its 2 lags, and another variable, say x. Now, I wish to apply the Arellano-Bond procedure to obtain consistent estimates of the AR terms. I am familiar with estimation of the AR(1) model, but I am not sure how to use the AB estimator with 2 lags. I currently am doing the following:

xtabond y x , lags(2) artests(2)

I am not certain if this takes care of the issue of the second lagged depedent variable in the moment conditions automatically (in other words, does the option lags(2) ensure that appropriate lags in levels and differences are chosen as instruments?)

Many thanks,
CS

[Sebastian Kripfganz]

The xtabond default is to choose all available valid lags as instruments. The number of instruments hence does not change when you add further lags of the dependent variable as regressors.

It is always a bit dangerous to rely on default settings and I would recommend to explicitly specify the instruments as you want to include them.

There are also more flexible community-contributed commands for GMM estimation of dynamic panel models; see for example:

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

[Chinmay Sharma]

Thank you Sebastian Kripfganz