This paper estimates forward-looking and forecast-based Taylor rules for France, Germany, Italy, as well as the euro area, using both final revised data and real-time data. We are particularly interested in the impact of adding asset prices to the standard Taylor rule specification. Since forward-looking Taylor rules are usually estimated via GMM, we perform extensive tests for over-identifying restrictions and instrument relevance, the latter generally eschewed in previous work. We find that asset prices can be highly relevant as instruments in policy rules. Forecast-based rules perform best using the root mean squared error metric. However, forecast-based rules are best estimated in difference form. Encompassing tests are used to select the 'best' policy rule. Finally, we perform a series of counterfactual experiments and ask whether the ECB?s monetary policy resembles that of any particular euro area core country, as well as asking what interest rates would have been like in the core countries had the ECB conducted monetary policy prior to 1999. Finally, while estimation using real-time data results in a deterioration in the forecasting performance of standard policy rules we do not find that they provided seriously misleading advice about the appropriate stance of monetary policy.
Monetary policy reaction functions asset prices instruments European Central Bank