French paper evaluates forecasting strategies

A working paper published by the Banque de France last week analyses the predictive power of two alternative forecasting strategies, one that combines forecasts and another that combines information sets.

The paper, Nowcasting French GDP in Real-Time from Survey Opinions: Information or Forecast Combinations? by Frederique Bec and Matteo Mogliani, applies a range of different combination techniques to the quarterly growth rate of French GDP.

It finds that when a predictive model is obtained by combining information sets, it outperforms even "the most accurate forecast combination strategy".

Moreover, the authors find that a simple average of individual forecasts outperforms individual forecasts, and also outperforms (or is at least statistically equivalent to) more sophisticated forecast combination schemes.