My dependent variable is gini coeff and explanatory variables included lagged gini variable , trade and FDI variables (which can cause endogeneity problem),GDP,GDP-square ? I thought of using System GMM since i have lagged dependent variable and endogeneity problem. Also,please guide when there is square term of GDP in my explanatory variables can still system GMM be used?
Read the following article, this might be helpful for you.
ageconsearch.umn.edu/bitstream/122704/2/sjart_st0159.pdf
Thanks Piero for really good suggestions...Since i'm new to dynamic panels....Please guide me ...in my case, GDP and gini can have casuality so square term will be problem as my variable gdp is not exogenous.. And secondly, if i take large T (i have the data) then which technique other than System GMM can be used to deal with lagged dependent variable as well as endogeneity issue... Simple OLS and fixed effect will give bias estimates ?? Thanks
This is complementary to what Prof. Piero has discussed above.
(1) The estimated coefficients of the squared independent variables should be interpreted as marginal effects. Another version of Threshold regression.
(2) I would prefer the xtabond2 version of the system GMM approach because it has some advantages like the control of cross-sectional dependence (Roodman, 2009ab).
References
Roodman, D., (2009a). “A Note on the Theme of Too Many Instruments”,
Oxford Bulletin of Economics and Statistics, 71(1), pp. 135-158.
Roodman, D., (2009b). “How to do xtabond2: An introduction to
difference and system GMM in Stata”, Stata Journal, 9(1), pp. 86-136.
- Badges
- Science topic
- Similar topics
- Geoscience
- Asia
- India