I have a data set of 5 countries for 30 years and i want to analyse the relationship between y and x in panel setting. However, the countries are not similar in nature moreover. one country is large enough to drive the whole results. I cant drop the country because i have already fewer observations. How can i solve this problem without estimating individual time series models
You can select and try using a suitable 2nd generation panel data econometric method, since such methods are characterised by rejection of the cross-sectional independence hypothesis – eg. ECM panel co-integration, Pedorni Dynamic Ordinary Least Square (PDOLS), Fully Modified Ordinary Least Square (FMOLS), Common Correlation Effects Mean Group (CCEMG), Augmented Mean Group (AMG) and Average Correlation Coefficient (ACC).
The two approaches you can choose from and run with are the factor structure approach (Bai & Ng, 2001; Choi, 2002; Phillips & Sul, 2003; Pesaran, 2003; Moon & Perron, 2004) or imposition of either none or few restrictions on the residuals covariance matrix (Chang, 2002; 2004) by applying nonlinear instrumental variables methods or bootstrap approaches to solve the parameter problems arising from cross-sectional dependence. Such methods are robust enough to account for cross-sectoral dependence, and have the potential or power circumvent problems relating to the presence of endogeneity between dependent and independent variables and serial correlation between co-integrated panels.
You can select and try using a suitable 2nd generation panel data econometric method, since such methods are characterised by rejection of the cross-sectional independence hypothesis – eg. ECM panel co-integration, Pedorni Dynamic Ordinary Least Square (PDOLS), Fully Modified Ordinary Least Square (FMOLS), Common Correlation Effects Mean Group (CCEMG), Augmented Mean Group (AMG) and Average Correlation Coefficient (ACC).
The two approaches you can choose from and run with are the factor structure approach (Bai & Ng, 2001; Choi, 2002; Phillips & Sul, 2003; Pesaran, 2003; Moon & Perron, 2004) or imposition of either none or few restrictions on the residuals covariance matrix (Chang, 2002; 2004) by applying nonlinear instrumental variables methods or bootstrap approaches to solve the parameter problems arising from cross-sectional dependence. Such methods are robust enough to account for cross-sectoral dependence, and have the potential or power circumvent problems relating to the presence of endogeneity between dependent and independent variables and serial correlation between co-integrated panels.
Dear Mehraj.
I am sure of ECM panel cointegration test which accounts for cross sectional dependence.
I also recommend @Elvis Munyaradzi G for more information.
Regards
Dear Mehraj,
I think you are faced with two important issues. First, your your model has a large t and small n. You must adopt an estimation technique that addresses this very important issue. Second, and the less serious issue is the heterogeneity inherent in your model. There are a number of ways to deal with this. For robustness, you can apply a battery of tests. For example you can sub-group your panel based on income level, percapital income, level of development, etc.
I hope this helps.
- Badges
- Science topic
- Similar topics
- Biological Science
- Histology
- Sectioning