Identifying turning points in the business cycle

Enrique M. Quilis
D.G. de Análisis Macroeconómico y Economía Internacional
Ministerio de Economía y Competitividad
Esta dirección de correo electrónico está protegida contra los robots de spam, necesita tener Javascript activado para poder verla

  • Abstract
    One of the critical tasks in business cycle measurement is dating its turning points. There are two broad alternative approaches: non-parametric (or empiricist) and parametric (or model-based). The first one combines band-pass linear filtering and algorithmic rules to identify turning points. The second one specifies and estimate a non-linear model (a Markov-Switching Autoregression, e.g.) and derives the turning points according to the probabilities to switch from one regime to another. Both approaches are examined in this paper and are applied to the Spanish quarterly Gross Domestic Product (GDP) time series in order to asses its performance from a practical perspective.
  • Keywords: Bussines cycle, turning points, band-pass filters, Markov-switching autoregressions, Spanish quarterly GDP.
  • AMS Subject classifications: 62M10, 60G35, 62M05, 68T10.