Abstract

Market participants expend considerable resources on forecasting services to improve their decision-making processes. One of the more popular sources of these services is that of Bloomberg. In recent years Bloomberg has regularly polled reputable experts on macroeconomic forecasts in many countries. This study aims to test the accuracy of Bloomberg's forecasts. Specifically we test whether the surveyed experts' median opinions are statistically similar to the announced figures from U.S.A., Japan, Canada, United Kingdom, Germany, France and Italy. The results provide an indirect test of the Rational expectation hypothesis. The forecasts and announced figures are first analyzed to discern unit roots and cointegration before inferences are made from the accuracy regression. This research shows that for the most part Bloomberg's forecasts surveys are unbiased and rational predictors of macro economic indicators for most countries. A somewhat longer time series of forecasts is available for the U.S. The earlier part of the sample indicates that forecasters of most of the macroeconomic variables appeared to be biased/not consistent with rational expectations. However over the latter part of the sample, forecasts accuracy for the U.S. seems to have improved significantly and compares favorably with forecasts produced for the other G7 countries.