Do central bank forecasts matter for professional forecasters?

This paper examines to what extent public information provided by the central bank affects the forecasts formulated by professional forecasters. We investigated empirically whether disclosing GDP and inflation forecasts by Narodowy Bank Polski (the central bank of Poland) reduced the disagreement in professional forecasters' expectations. We also checked whether the strength of the forecasters’ reaction to the release of the central bank’s projection depends on the phase of the business cycle. Finally we identified the determinants of the dispersion among the forecasters.In the first step we used single equation models estimated separately for inflation and GDP forecasts. While the results confirm that by publishing its projection of future GDP growth, the central bank was reducing the dispersion of GDP forecasts, we extended the linear model and introduced asymmetry in the response of individual GDP forecasts to the release of NBP projection. Therefore in the second step we used the Smooth Transition Regression (STR) models with both logistic function and exponential transition function.The results only partially support the hypothesis on the coordinating role of the central bank existing in the literature. The main finding is that by publishing its projection of future GDP growth, the central bank was reducing the dispersion of one-year-ahead GDP forecasts. Our study indicates that the role of the central bank in reducing the forecasts dispersion was strengthening over time. We also found using non-linear STR models that the extent to which the projection release affected the dispersion of GDP forecasts varied over the business cycle. By disclosing its own projection the central bank reduced the disagreement among the forecasters the most in the periods when the economy moved from one phase of the business cycle to another.