TY - JOUR

T1 - Combining Economic Forecasts

AU - Peel, David

PY - 1988/11

Y1 - 1988/11

N2 - It is well known that no particular forecasting agency dominates when the accuracy of economic forecasts of the UK is investigated. There are good reasons for believing that if forecasts differ, some combination of them will be an improvement over the individual forecasts. The problem is to determine what weights to attach to each forecast. Various methods have been suggested in the literature, including equal weights (averaging), optimal weights (linear regression), varying weights based on past performance, and the Bayesian approach. We review these methods and examine their performance for important macro-economic variables.

AB - It is well known that no particular forecasting agency dominates when the accuracy of economic forecasts of the UK is investigated. There are good reasons for believing that if forecasts differ, some combination of them will be an improvement over the individual forecasts. The problem is to determine what weights to attach to each forecast. Various methods have been suggested in the literature, including equal weights (averaging), optimal weights (linear regression), varying weights based on past performance, and the Bayesian approach. We review these methods and examine their performance for important macro-economic variables.

KW - economics

KW - forecasting

KW - time series

U2 - 10.1057/jors.1988.170

DO - 10.1057/jors.1988.170

M3 - Journal article

VL - 39

SP - 1005

EP - 1010

JO - Journal of the Operational Research Society

JF - Journal of the Operational Research Society

SN - 0160-5682

IS - 11

ER -