What Have We Learned from a Survey of Japanese Professional Forecasters? Taking Stock of Four Years of ESP Forecast Experience

This paper aims to take stock of four years of ESP forecast experience since its launch in 2004. The consensus performs well, compared to individual forecasters, which is confirmed probably for the first time in Japan. It satisfies the average form of rational expectations hypothesis defined in Pesaran and Weale (2006) at least in short forecast horizon. Remaining promising research agenda includes real-time data analysis of updating forecasters, subjective distribution, non-specialists' forecasts, changes in forecasts due to the Lehman shock.