Forecasting Canadian Inflation Rate Using CPI and Interest Rate Data: A Time Series Analysis

Forecasting short-term inflation in Canada requires accounting for both price trends and market volatility. To address this challenge, this study applies a time series model that combines monthly Consumer Price Index (CPI) data with the 3-month interbank interest rate a signal often associated with monetary policy adjustments. The model specification, ARIMA(0,1,2) coupled with a GARCH(1,1) process, is estimated under a t-distribution to accommodate heavy-tailed residuals. With monthly observations from January 2002 to August 2024 for training the model and withholding the following seven months as a track record, the analysis finds that CPI changes have a mild but stable positive trend. Moreover, interest rate appears as a significant factor, which increases its predictive power in inflation-linked models. The approach improves short-run accuracy, while retaining economic feasibility. This design could have useful application for policymakers who try to predict pressure in prices. Extensible cases include the model of nonlinear feedback or the inclusion of external drivers energy shocks and global financial volatility.