Expectations and Central Banks' Forecasts: The Experience of Chile, Colombia, Mexico, Peru and the United Kingdom, 2004 – 2014

The paper tests the hypotheses associated to whether or not the publication of central banks’ forecasts (and subsequent media-diffusion efforts) affects the professional forecasters’ expectations in terms of both their cross-section dispersion and the distance between their median and the central banks’ forecasts. The study considers the monetary authorities of Chile, Colombia, Mexico, Peru and United Kingdom. The focus is on forecasts for inflation and real growth and the common sample of monthly fixed-event forecasts goes from 2004 to 2014. This sample of forecasts allows highly specific tests by splitting it according to the forecasting horizon (short and medium terms) and the level of macroeconomic uncertainty (high- or low-uncertainty months). With a significance level of 10%, the general findings are that (i) the dispersion and the distance can significantly increase or decrease as a result of the publication of the official forecasts and the media-diffusion efforts, the number of increases in the distance being low with respect to the number of increases in the dispersion, though; and (ii) the number of decreases in the dispersion and distance is low for all inflation-targeting central banks considered. These findings point out the expectation management is still an elusive goal in the majority of countries considered.