System and Difference GMM -syntax

[Emma Gal]

Hi, I am researching how conflict moderates the impact of remittances on inequality in Latin America. I ran a difference GMM and though I got the results I wanted, it did not pass the Sargan test and so now I am trying to run a system GMM but am not sure if my syntax is correct.

My difference GMM command was
xtabond Gini l.RemittancesGDP conflict_index remitlag_conflict_now GDPpercapitaconstant2010US Inflation GovernmentexpenditureofGDP NetODAreceivedofGNI TradeopennessTradeofGDP Unemploymentrate, lags(1) vce(robust) artests(2) nocons

but this failed the Sargan test:

chi2(76) = 152.1699
Prob > chi2 = 0.0000

How would I run the system GMM? Currently this is my command but it is giving me very not statistically significant results which i would not expect
xtabond2 Ginipre l.Ginipre RemittancesGDP conflict_index remit_conflict_now GDPpercapitaconstant2010US Inflation GovernmentexpenditureofGDP NetODAreceivedofGNI TradeopennessTradeofGDP Unemploymentrate, gmm(l.Ginipre RemittancesGDP remit_conflict_now, lag (1 2) collapse) iv(l.conflict_index GDPpercapitaconstant2010US Inflation GovernmentexpenditureofGDP NetODAreceivedofGNI TradeopennessTradeofGDP Unemploymentrate) nocons twostep robust

Thank you for reading!

[Sebastian Kripfganz]

Just two observations:

• Do not use the nocons option with system GMM. This is almost never a good idea.
• All variables in the iv() option are implicitly assumed to be uncorrelated with the unobserved country-specific effects, which is almost never a reasonable assumption for these types of variables in such an application. [Also, the default settings for the iv() option may not produce what you want it to produce. Carefully check the command's documentation. If in doubt, always explicitly specify the eq() suboption.]

More on dynamic panel data GMM estimation in Stata:

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