An evaluation of the leading indicators for the Canadian economy using time series analysis

Abstract

This paper conducts an in-depth statistical evaluation of the ability of Statistics Canada's leading indicators (including the new composite index introduced in February, 1993) to predict changes in GDP. Empirical transfer function models are built to investigate the historical dynamic relationships between the component and composite indicator series and GDP. Although stable dynamic relations were found the average lead times are too short for most of the indicators. The new composite index designed specifically to improve promptness gave an average lead time of only 1 month. In addition, the indicator models were evaluated by conducting ex-ante forecasts of GDP during the recent 1991–1992 recession. Although the new composite index gave better forecasts than the univariate seasonal autoregressive integrated moving average (SARIMA) model for GDP, most of the component indicators gave better forecasts than the new composite index.