Model selection and model uncertainty in Econometrics

The choice of statistical model and uncertainty about which specification is the correct or most appropriate one are fundamental issues in any area of applied research. There is an increasing awareness of the need to account for model uncertainty in empirical economics. The few studies taking account of model uncertainty demonstrate that important insights and more robust results can be obtained.

The aim of the project is to develop tools for addressing model uncertainty within a Bayesian framework and demonstrate their viability with substantial applications in several areas of importance to applied econometrics. We adopt a Bayesian approach since this -- as opposed to a frequentist approach -- offers a coherent framework for thinking about and quantifying model uncertainty.

Publications

Jacobson, T. and S. Karlsson, (2004), "Finding Good Predictors for Inflation: A Bayesian Model Averaging Approach", Journal of Forecasting, 23, 479-496.

Eklund, J. and S. Karlsson, (2007), "Forecast combination and model averaging using predictive measures", Econometric Reviews, 26, 329-363.

Hultblad, B. and S. Karlsson, (2008) "Bayesian simultaneous determination of structural breaks and lag lengths", Studies in Nonlinear Dynamics & Econometrics, Vol. 12: No. 3, Article 4.

Andersson, M. K. and S. Karlsson, (2008), "Bayesian forecast combination for VAR models", in S. Chib, G. Koop, and B. Griffiths, eds. Advances in Econometrics, Volume 23, Bayesian Econometrics.