Hello folks!!
I am researching on the determinants of growth of NBFI assets for N=11 countries and T=17 years (maximum) (unbalanced panel), case of macro panels.In order to estimate these models, I have used static methods using xtgls,xtpcse,xtregar and xtscc user based command in Stata 13.I am not able to decide which of these estimates I should rely upon? Also,the coefficients of these estimates are not consistent across different models (dropping or adding variables).
Further,in order to apply dynamic panels,I checked for stationarity using panel unit root tests.My variables are of I(0) and I(1) order.I(0) variables are in growth rate terms such as Real GDP growth,Year on Year Bank assets,Year on Year Pension funds,Year on Year Insurance assets. I(1) variables include stock turnover ratio,capital adequacy ratio,etc.
Therefore,I applied panel ARDL model and estimate it using xtpmg. But my estimates are not consistent across models?
Is the model feasible for this panel?