Communicational Bias in Monetary Policy: Can Words Forecast Deeds?

Communication with the public is an ever-growing practice among central banks and complements their decisions of interest rate setting. In this paper we examine one feature of the communicational practice of the Central Bank of Chile which summarizes the assessment of the Board about the most likely future of monetary policy. We show that this assessment, which is called communicational bias or simply c-bias, contains valuable information regarding the future stance of monetary policy. We do this by comparing, against several benchmarks, the c-bias ability to correctly forecast the direction of monetary policy rates. Our results are consistent with the hypothesis that the Central Bank of Chile has (in our sample period) matched words and deeds. In fact, the c-bias is a more accurate predictor than two versions of random walks and than a uniformly-distributed random variable. It also outperforms, at some horizons, the predictive ability of a discrete Taylor-rule-type model that uses persistence, output gap and inflation-deviation-from-target as arguments. Furthermore, the c-bias is more accurate than survey-based forecasts at several forecasting horizons. We also show that the c-bias can provide information to improve monetary policy rate forecasts based on the forward rate curve.