Inference on filtered and smoothed probabilities in Markov Switching Factor Models


In this paper we show how to compute asymptotic variances of filtered and smoothed probabilities in a Dynamic Factor Markov-Switching autoregressive model. Using real time data of four economic indicators, we estimate these probabilities and their variances for the US economy. Our results show that the confidence intervals based on the asymptotic distribution of filtered probabilities provide more accurate information to identify the business cycles turning points.

Información adicional

  • Presentador: Rocío Álvarez-Aranda
  • Proveniente: Universidad Central
  • Fecha: Miércoles, 28 Marzo 2018
  • Hora: 12:00
  • Lugar: Sala R1, FAE