Multivariate Nonlinear Models for Turning Points Detection: Direct Versus Indirect Indicators

• Activity Number: 588
• Type: Topic Contributed
• Date/Time: Wednesday, August 7, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): Riccardo Gatto*+ and Monica Billio and Gian Luigi Mazzi and Laurent Ferrara
• Companies: Eurostat - European Commission and Università di Venezia and Eurostat - European Commission and Universitè Paris 10
• Keywords: Business Cycle Analysis ; Turning points ; Markov switching models

Abstract

In a period of high economic uncertainty it is crucial to deliver timely and not contradictory signals on the occurrence of turning points avoiding misleading messages. The paper compares two alternative strategies to construct real-time turning points indicators for the growth and business cycle of the Euro area. The first one is based on the construction of a multivariate Markov switching model for both growth and business cycle using Euro area data, while the second one is based on modelling the cycles of member states and then aggregating the returned recession signals to derive Euro area turning points. While the first approach is easy to compute and transparent, the second one delivers as by product turning points indicators at country level, providing then a more complete information. Both approaches are simulated in real-time to compare their recession signals. The main finding is that both approaches deliver very similar messages over almost all past simulation periods. However, for the end of the period they agree on the existence of a euro area slowdown phase, with some uncertainty concerning whether or not the Euro area is currently experiencing a recession.