To what extent did deviations from the Taylor rule between 2002 and 2006 help to promote price stability and maximum sustainable employment? To address that question, this paper estimates a New Keynesian model with unemployment and performs a counterfactual experiment where monetary policy strictly follows a Taylor rule over the period 2002:Q1 - 2006:Q4. The paper finds that such a policy would have generated a sizeable increase in unemployment and resulted in an undesirably low rate of inflation. Around mid-2004, when the counterfactual deviates the most from the actual series, the model indicates that the probability of an unemployment rate greater than 8 percent would have been as high as 80 percent, while the probability of an inflation rate above 1 percent would have been close to zero.