In almost all cases a decision maker cannot identify ex ante the true process. This observation has led researchers to introduce several sources of uncertainty in forecasting exercises. In this context, the research reported in these pages finds an increase of forecasting power of financial time series when parameter uncertainty, model uncertainty and optimal decision making are included. The research contained herein evidences that although the implementation of these techniques is not often straightforward and it depends on the exercise studied, the predictive gains are statistically and economically significant over different applications, such as stock, bond and electricity markets.

H.K. van Dijk (Herman) , M.J.C.M. Verbeek (Marno)
Erasmus University Rotterdam , Thela Thesis, Amsterdam
Tinbergen Instituut Research Series
Erasmus School of Economics

Ravazzolo, F. (2007, November 23). Forecasting Financial Time Series Using Model Averaging (No. 415). Tinbergen Instituut Research Series. Retrieved from