Estimation of the Taylor Rule for Canada Under Multiple Structural Changes

The Taylor rule is estimated under the period 1963Q2 to 1999Q4 using Canadian data and the methodology proposed by Bai and Perron (1998) to estimate regression models with multiple endogenous breaks. Although monetary rules are notorious for suffering from structural instability, recent attempts at modeling it are only considering exogenous breaks which are imposed on the data generating process (e.g. Judd and Rudebusch, 1998; and Clarida, Galí and Gertler, 2000). We show that the monetary rule cannot be evaluated over this period without taking into account parameter instability and structural changes, re- flecting changes in monetary policy preferences. Infation is modeled as a Markov-Switching (MS) process to extract expectations, which we treat as the implicit infation target. To extract the potential level of output, we also use a MS process. Modeling the rule when allowance for two breaks is made (1978Q3 and 1988Q2) illustrates well the changing policy preferences of the Bank of Canada.