Does the ECB Rely on a Taylor Rule During the Financial Crisis? Comparing Ex-post and Real Time Data with Real Time Forecasts

We assess the differences that emerge in Taylor rule estimations for the ECB when using ex-post data instead of real time forecasts and vice versa. We argue that previous comparative studies in this field risk mixing up two separate effects. First, the differences resulting from the use of ex-post and real time data per se and, second, the differences emerging from the use of non-modified real time data instead of real-time data based forecasted values (and vice versa). Since both effects can influence the ECB reaction to inflation and the output gap either way, we use a more clear-cut approach to disentangle the partial effects. However, “good” forecasts have to be as close as possible to the forecasts the ECB governing council had at hand when taking its interest rate decision. Therefore we use two approaches to generate the forecasts: First, forecasts generated relying on a pure AR process and, second, explicit ECB staff projections which are available only at a quarterly frequency. So we found it indispensable to estimate all variants of the reaction function using also quarterly data. Our estimation results indicate that using real time instead of ex post data leads to higher estimated inflation and output gap coefficients. If real time forecasts are used (since actual data become available with a lag), the output response is reduced while the inflation response depends crucially on the inclusion of an interest rate smoothing term, the data frequency and forecast type.