Tracking economic activity in the euro area: Multivariate direct filter approach

Abstract

The paper applies the multivariate direct filter approach on selected business and consumer confidence indicators, and share price data to construct a real-time indicator tracking the medium-to-long-run component of the quarterly growth of the euro area gross domestic product. Results show that the created indicator behaves similarly to another established indicator, Eurocoin, but slightly leads it after mid 2009. The new indicator is also compared to the Markit Euro Area Composite Purchasing Managers Index and is found to be leading it by about one month as well as being smoother. Overall, the multivariate direct filter approach is found to have merit in tracking business cycle developments, however, the increasing number of free coefficients is an issue for the filter to be applied to detailed datasets. Keywords: real-time signal extraction, business cycles, multivariate time series JEL classification: C13, C32, E32