In this paper we investigate the effect of central bank transparency on survey forecasts. Similar to Ehrmann et al. (2010), we find that greater transparency can reduce the degree of disagreement across individual forecasters and it can also improve the forecasting performance of survey respondents. However, our empirical approach is more rigorous than that of Ehrmann et al. (2010) as we test both for causality and misspecification. The analysis is carried out on a panel dataset that is much richer than those used by previous studies. This unique dataset allows us to identify the effects of various aspects of transparency separately and to assign weights to them reflecting their relative importance in reducing uncertainty. Finally, we construct a new composite measure of central bank transparency using the estimated weights.
JEL: C53, D83, E50.
Keywords: central bank transparency, survey forecast, weighted transparency index, dynamic panel model, overlapping observations.