Forecasting Real Estate Business: Empirical Evidence from the Canadian Market

Abstract

In this paper, we compare the out-of-sample forecasting ability of three ARIMA family models: ARIMA, ARIMAX, and ARIMAX-GARCH. The models are tested to forecast turning points and trends in the Canadian real estate index using monthly data from April 2002 to March 2011. The results indicate that the ARIMAX model, which includes exogenous macroeconomic variables such as the gross domestic product, the consumer price index, the difference in long-term and short-term interest rates, and the exchange rate of the Canadian dollar against the US dollar and their lags, provides the best out-ofsample forecasts. Overall, the models are suitable only for short-term forecasts.