Predictive gains from forecast combinations using time-varying model weights
Several frequentist and Bayesian model averaging schemes, including a new one that simultaneously allows for parameter uncertainty, model uncertainty and time varying model weights, are compared in terms of forecast accuracy over a set of simulation experiments. Artificial data are generated, characterized by low predictability, structural instability, and fat tails, which is typical for many financial-economic time series. Sensitivity of results with respect to misspecification of the number of included predictors and the number of included models is explored. Given the set up of our experiments, time varying model weight schemes outperform other averaging schemes in terms of predictive gains both when the correlation among individual forecasts is low and the underlying data generating process is subject to structural locations shifts. In an empirical application using returns on the S&P 500 index, time varying model weights provide improved forecasts with substantial economic gains in an investment strategy including transaction costs.
|Date of creation:||26 Jul 2007|
|Contact details of provider:|| Postal: Postbus 1738, 3000 DR Rotterdam|
Phone: 31 10 4081111
Web page: http://www.eur.nl/ese
More information through EDIRC
Please report citation or reference errors to , or , if you are the registered author of the cited work, log in to your RePEc Author Service profile, click on "citations" and make appropriate adjustments.:
- Rodney Strachan & Herman K. van Dijk, "undated".
"Bayesian Model Averaging in Vector Autoregressive Processes with an Investigation of Stability of the US Great Ratios and Risk of a Liquidity Trap in the USA, UK and Japan,"
MRG Discussion Paper Series
1407, School of Economics, University of Queensland, Australia.
- Strachan, R.W. & van Dijk, H.K., 2007. "Bayesian model averaging in vector autoregressive processes with an investigation of stability of the US great ratios and risk of a liquidity trap in the USA, UK and Japan," Econometric Institute Research Papers EI 2007-11, Erasmus University Rotterdam, Erasmus School of Economics (ESE), Econometric Institute.
- Ravazzolo, F. & van Dijk, D.J.C. & Paap, R. & Franses, Ph.H.B.F., 2006. "Bayesian Model Averaging in the Presence of Structural Breaks," Econometric Institute Research Papers EI 2006-33, Erasmus University Rotterdam, Erasmus School of Economics (ESE), Econometric Institute.
- Jana Eklund & Sune Karlsson, 2007. "Forecast Combination and Model Averaging Using Predictive Measures," Econometric Reviews, Taylor & Francis Journals, vol. 26(2-4), pages 329-363.
- Eklund, Jana & Karlsson, Sune, 2005. "Forecast Combination and Model Averaging using Predictive Measures," Working Paper Series 191, Sveriges Riksbank (Central Bank of Sweden).
- Eklund, Jana & Karlsson, Sune, 2005. "Forecast Combination and Model Averaging Using Predictive Measures," CEPR Discussion Papers 5268, C.E.P.R. Discussion Papers.
- Guidolin, Massimo & Timmermann, Allan, 2009. "Forecasts of US short-term interest rates: A flexible forecast combination approach," Journal of Econometrics, Elsevier, vol. 150(2), pages 297-311, June.
- Massimo Guidolin & Allan Timmerman, 2007. "Forecasts of U.S. short-term interest rates: a flexible forecast combination approach," Working Papers 2005-059, Federal Reserve Bank of St. Louis.
- Guidolin, Massimo & Timmermann, Allan G, 2007. "Forecasts of US Short-term Interest Rates: A Flexible Forecast Combination Approach," CEPR Discussion Papers 6188, C.E.P.R. Discussion Papers.
- Harvey, Andrew C. & Trimbur, Thomas M. & Van Dijk, Herman K., 2007. "Trends and cycles in economic time series: A Bayesian approach," Journal of Econometrics, Elsevier, vol. 140(2), pages 618-649, October.
- Harvey, A.C. & Trimbur, T.M. & van Dijk, H.K., 2005. "Trends and cycles in economic time series: A Bayesian approach," Econometric Institute Research Papers EI 2005-27, Erasmus University Rotterdam, Erasmus School of Economics (ESE), Econometric Institute.