Abstract
This study presents empirical evidence on the relationship between the forecast errors and the number of individual forecasts used in averages of model forecasts. The investigation is based on forecasts published by Her Majesty's Treasury in the monthly report ‘Forecasts for the UK economy: A comparison of independent forecasts’. The results suggest that averages of model forecasts tend to improve forecasting performance when relatively large numbers of forecasts are used in the averages (usually in excess of 40); however, results are not always consistent and averages should be used with caution.
Additional information
Notes on contributors
Antonis A. Michis
Antonis A. Michis is a research Economist at the Central Bank of Cyprus. His research interests include forecasting, market response modelling and empirical industrial organization.