Empirical analysis of inflation targeting in an open economy: An application to Canada

In late 1980\u27s and the 1990\u27s, many industrial countries adopted inflation targeting policy to keep the inflation under control and eliminate its high costs, as well as one source of uncertainty. This step came after the failure of the money supply policies. Canada, the country of application, was the second country to adopted inflation targeting in February 1991. We derived mathematically the model and the optimal feedback rule. Our econometric analysis compares two methods of estimation—the 2SLS and the maximum likelihood estimation (MLE). The 2SLS estimates serve as starting point for the MLE. Analyses in the literature cited in this study are based on calibration of hypothetical coefficients, estimation by others or estimation methods do not go far enough. The MLE, in which all coefficients adjust to changes in policy parameter, displayed an estimation improvement over the 2SLS method in terms of a huge reduction in the standard error of the feedback rule. We tested for structural change using three tests—Chow test, average F test and CUSUM. The three tests, returning the same conclusion for the inflation and the exchange rate equations but not for the real output, show no structural change in the model coefficients. Testing for inflation-output variability trade-off indicates no variability trade-off based on MLE, while tests based on 2SLS estimates reveals otherwise. Hence, the existence of variability trade-off depends on the econometric method of estimation. Finally, the impulse response functions for the inflation rate, the output gap, the exchange rate to their shocks and a change in the foreign interest rate show that the response takes more than sixteen quarters before it dies completely