xtdpdsys no observations r(2000)

[Kaylin Morris]

Hello everyone,

I'm new to statistics and Stata so I'm all lost here.

My data is panel data, consisting of 98 countries, each country has a number of firms, starting from 2013 to 2022 (in short, this is World Bank Enterprise Survey data, and it's unbalanced - some countries only cover 1 year or 2 years, and some cover maximum of 3 years).

I want to run a system GMM regression to estimate how digital finance can affect firms' access to finance. My dependent variable is ATF, independent variables: NoMMA, NoMMT, VMMT, NoMIBT, VMIBT, control variables: GDP_G, CPI, Trade, Upop, Age, OpSec, Size, Ownership, ExStat. The dependent variable has every value for every firm, the independent variables have many missing values.

xtset Firm Year
panel variable: Firm (weakly balanced)
time variable: Year, 2013 to 2022
delta: 1 unit

For the SYS-GMM, I'm using this command:

xtdpdsys ATF NoMMA NoMMT VMMT NoMIBT VMIBT, noconstant lags(2) twostep artests(2)

However, when I run it I get r(2000) no observations knowing that all my variables are numeric. I even changed those 5 independent variables to other variables that have every value for each firm, but it still shows the same error.

Can anyone help me with this problem? Thank you in advance!

[Sebastian Kripfganz]

You have specified option lags(2), which adds two lags of the dependent variable as regressors. This requires at least 3 time periods for the model in levels and at least 4 time periods for the model in first differences.

More on dynamic panel data GMM estimation:

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