xtabond2 with unbalanced panels and missing values in between

[Guest]

Hi everyone,

I am conducting an empirical work where I am estimating the effect of financial development on economic growth. I have data from 1960-2015 for around 130 countries but for many countries data are observed in different time periods and in some countries some data are missing in between (e.g data for 1970-1980 available but 1980-1985 missing and then 1986-2015 available). I know that GMM can accommodate unbalanced panels, however, I am not sure if it can accommodate missing values in between. Can somebody help ?

Thank you !

[Sebastian Kripfganz]

Missing values (at random) in between are not generally a problem for GMM estimation. For such a situation, using forward-orthogonal deviations instead of first differences is preferred, because the former retains more information when the panel data has gaps. See slides 66 to 71 of my 2019 London Stata Conference presentation:

• Kripfganz, S. (2019). Generalized method of moments estimation of linear dynamic panel data models. Proceedings of the 2019 London Stata Conference.