We explore how the ECB sets interest rates in the context of policy reaction functions. Using both real-time and revised information, we consider linear and nonlinear policy functions in inflation, output and a measure of financial
conditions. We find that amongst Taylor rule models, linear and nonlinear models are empirically indistinguishable within sample and that model specifications with real-time data provide the best description of in-sample
ECB interest rate setting behavior. The 2007-2009 financial crisis witnesses a shift from inflation targeting to output stabilisation and a shift, from an asymmetric policy response to financial conditions at high inflation rates, to a more symmetric response irrespectively of the state of inflation. Finally, without imposing an a priori choice of parametric functional form, semiparametric models forecast out-of-sample better than linear and nonlinear Taylor rule models.