i am working on nonperforming loans and have taken Npa as dependent variable and have taken yearly data of 25 countries for 23 years time series data.
My independent variables are ten bank specific and ten macroeconomic. My variables are i(0) and i(1).
What estimation technique will be better please suggest
Very good document: http://paneldataconference2015.ceu.hu/Program/Michael-Abrigo.pdf
I have a feeling that you might have too many independent variables. If you could select a much smaller number of independent variables in your study, I would suggest you to consider dynamic panel models. Specifically, if you are only interested in the long-run coefficients, a cross-sectionally augmented distributed lag (CS-DL) approach might be suitable. If you are also interested in the short-run coefficients, you could consider CS-ARDL approach. Both approaches take care of cross-section dependence and heterogeneous slopes. The details can be found, among the others, in Chudik, Mohaddes, Pesaran, and Raissi (2013): Debt, Inflation and Growth: Robust Estimation of Long-Run Effects in Dynamic Panel Data Models.
I would run both panel ARDL and GMM because both of the methodologies have their advantages and disadvantages. But panel ARDL is stimm more advantageous...
- Badges
- Science topic