Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
Using a large international data set, we analyze whether business cycle forecasters herd or anti-herd. In general, we find evidence for anti-herding, i.e. forecasters appear to scatter their forecasts deliberately away from the forecasts of others. Anti-herding tends to be more prevalent for the longer (next year) horizon. There is some evidence for a reduced level of anti-herding at times of increased forecast uncertainty and when the forecasts are being revised more substantially.