Tentative Panel Data Model Selection Plan

Based on few studies on comparison of panel data models and exploring the eveiws and stata help I have come across a tentative flow chart showing the symptoms which will lead to a particular model selection. Since it is a tentative flow chart any suggestions are welcomed.

panel data model selection- noman arshed


  1. Eberhardt, M. and Francis, T. (2010) Productivity Analysis in Global Manufacturing Production, Economics Series Working Papers 515, University of Oxford, Department of Economics
  2. Eberhardt, M. (2012). Estimating panel time-series models with heterogeneous slopes.The Stata Journal12(1), 61-71.
  3. Gengenbach, C., Palm, F. C., & Urbain, J. P. (2005).Panel cointegration testing in the presence of common factors. METEOR, Maastricht research school of Economics of Technology and Organizations.
  4. Pesaran, M. H. and Ron, P. S. (1995). Estimating long-run relationships from dynamic heterogeneous panels. Journal of Econometrics, 68(1), 79-113.
  5. Pesaran, M. H. (2006). Estimation and inference in large heterogeneous panels with a multifactor error structure. Econometrica, 74(4), 967-1012.
  6. Saikkonen, P. (1991) Asymptotically efficient estimation of cointegration regressions. Econometric theory, 7, 1-21
  7. Swamy, P. A. (1970). Efficient Inference in a Random Coefficient Regression Model. Econometrica, 38(2), 311-323.
  8. Stata Help Manual
  9. Eviews Help Manual

5 thoughts on “Tentative Panel Data Model Selection Plan

  1. nasiru inuwa says:

    my suggestions are:first the clasification should be based on static, dynamic and nonstationary panels.second, in the place where you stated hausman test you included pols but the little i knew hausman test choose the suitability between FE and RE alone, but BF LM test does between POLS and RM. I stand to be corrected if i made mistakes

    1. Noman Arshed says:

      thank you for your suggestion, yes i thought of making it in the bases of static, dynamic and non stationary panels, but instead of that I used the assumptions which these models are covering. Hopefully I will make another.
      Actually the diagram is made such a way that it is assumed that there is hetroscedasticity before selecting FE or RE that is why in such case doing LM test is not necessary. Yes if I stay within the static model system then I would have explained LM or Redundant Cross section test too.

  2. amin says:

    it was very a detail chart to know about panel data analysis but kindly explain it with small notes so that students who are not much familiar with econometrics can easily understand

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s