Is the Bank of Canada concerned about inflation or the state of the economy?

This paper examines the historical behaviour of the Bank of Canada (BoC) from 1991, when the BoC adopted inflation targeting, until 2015. We use a newly released dataset that contains quarterly vintages of real-time historical data and BoC staff forecasts, and we present three novel empirical findings. First, over the full sample period we find that the BoC responds both to inflation and the real economy. The long-run coefficient on inflation exceeds unity, satisfying the Taylor principle. Second, we fi nd that there is considerable variation in the monetary policy coefficients. During the early period of our sample, we fi nd a strong response to inflation and no response to the real economy. But over time we fi nd that the response to inflation weakens, the Taylor principle disappears, and the response to the real economy rises substantially. At the later part of our sample, the BoC appears to respond to the real economy, but not to inflation. Third, we investigate if the BoC is responding to an alternative inflation measure that captures persistent inflation deviations, that is inflation that remains away from its target for an extended period. We augment our monetary policy rule with this new inflation measure and find that the BoC responds asymmetrically to positive and negative persistent inflation deviations, suggesting that persistent inflation overshooting and undershooting elicit different responses.