Causality analysis of the Canadian city house price indices: A cross-sample validation approach

Abstract

We examine pair-wise causality between 11 Canadian city house price indices. Monthly data that span from 1990:7 to 2015:9 are employed. The traditional Granger-causality framework is relaxed by following the cross-sample validation approach of Ashley and Tsang (2014). This allows us to overcome the ad hoc partition of the sample and examine predictability both “in” and “out-of-sample”. Toronto emerges as the driving force of the Canadian Housing Market.