The relevance of real-time data in estimating reaction functions for the euro area

This paper tackles the issue of the incompleteness of information available to the central bank, when taking its monetary policy decisions. It is focused on euro area data and based on the simplistic assumption that the central bank follows a simple monetary policy rule à la Taylor. Along the lines of the work by Orphanides, our aim is to assess whether estimates of reaction functions, which are carried out using revised data for the euro area, can be misleading. In essence, the analysis yields indications which are consistent with the findings by Orphanides for the United States. First, the results suggest that it would be preferable for a central bank not to attach too much weight to output-gap measures in policy analysis, given that such measures are subject to large revisions over time. Second, the coefficients of a simple Taylor rule estimated in real time differ quite considerably from those related to the same rule estimated on the basis of ex post revised data. More precisely, a coefficient for inflation larger than one (which is a requirement for a unique equilibrium in many theoretical models) in real time is found only when a forward-looking specification based on the Survey of Professional Forecasters is estimated. On the contrary, when revised data are used, the same result is obtained when the Taylor rule includes the current inflation rate. This shows how misleading inferences can occur when appropriate available data are not taken into account.