I am writing a paper about the contribution of foreign banks to economic growth (real GDP per capita), using a panel dataset for $N=6$ countries and $T=40$ quarters.

I was planning to use GMM, but I guess my panel structure isn't suitable for GMM, so I changed my methodology:

  1. Unit root of panel data (I was planning to use a dynamic model but I found that it won't be fine with my model as $N<T$: am I right?). I used IPS because my panel is UNBALANCED.

  2. If series are not stationary I will run a cointegration test to check for the long term relation.

  3. Granger causality test.

  4. At this stage I will add the dynamic term ($y_{t-1}$ as explanatory variable) and run the GMM estimates (level, difference and 2 stages) to compare between different estimates.

What do you think? Would that be correct? One of my teachers said that when the panel is dynamic there is no need for unit root test. Can I still run unit root test for a dynamic model and is the dynamic model suitable for my data structure?


Since you have the long-narrow panel data ($T > N$), the easiest way is to fit OLS for all 6 countries separately. If you suspect there are across dependence between $N$, you can run sureg (seemingly unrelated regression) in Stata to control cross relationships in OLS residuals.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.