Dear forum members,
i am interested in some comments and opionins regarding the following matter:
I have macroeconomic and risk indicator panel data on 8 developing countries for the period 1996 to 2016, totaling to about 168 observations. Im doing regressions on 4 different dependent variables (macroeconomic) with about 10 explanatory variables (the risk indicators).. Research has led me to the conclusion that my two best options for regression here are either fixed effects xtreg regression or pooled ols.
The respective tests (hausman and etc.) were not that helpful as the results were kind of mixed. The estimates of the fixed effects model seemed to make less sense compared to the estimates of the pooled ols model. also the adj. r square values indicated a better fit of the ols model... So currently i assume that my pooled ols model with cluster for country seems to be alright: xtset country_name time
reg y x1 x2 x3 ... x10 i.time, cluster (country_name).
However... i still am uncertain whether this approach might lead to too biased estimates... or if it is okay considering to not overestimate the results and that they have no external validity.
thanks for any hints, comments and opinions.
Hello
why you are ignoring an important step before proceeding for running regressions. you could probably start with the stationarity property of your variables (before than check for panel heterogeneity and cross-sectional dependence in order to make a correct choice of unit-root tests). Having established the nature of variables you can start looking for the possibility of co-integration among them and if found, you could use dynamic panel data models (available in the literature) as per the requirements of the data.
Best.
look at the answers to the question by A Shepherd at https://www.researchgate.net/post/When_to_use_cluster-robust_standard_erros_in_panel_anlaysis?pli=1&loginT=aM8NA9s1kHaghuka25YhjFgXLmTJVPviqMO_pz-IIIH_Zx1oGyGmI9A7XQe9sLcrahUl2uBfPGnlRmd5oQ&uid=8LgpH4fSJkAlkH4Nqe8i2xxz1A26qVNVI5Xy&cp=re65_x_p2&ch=reg&utm_medium=email&utm_source=researchgate&utm_campaign=re65&utm_term=re65_x&utm_content=re65_x_p2#view=5b72c750f4d3ec881d647e6b
look at the answers to the question by A Shepherd at https://www.researchgate.net/post/When_to_use_cluster-robust_standard_erros_in_panel_anlaysis?pli=1&loginT=aM8NA9s1kHaghuka25YhjFgXLmTJVPviqMO_pz-IIIH_Zx1oGyGmI9A7XQe9sLcrahUl2uBfPGnlRmd5oQ&uid=8LgpH4fSJkAlkH4Nqe8i2xxz1A26qVNVI5Xy&cp=re65_x_p2&ch=reg&utm_medium=email&utm_source=researchgate&utm_campaign=re65&utm_term=re65_x&utm_content=re65_x_p2#view=5b72c750f4d3ec881d647e6b
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression