期刊名称：Documents de Travail du Centre d'Economie de la Sorbonne
印刷版ISSN：1955-611X
出版年度：2015
出版社：Centre d'Economie de la Sorbonne
摘要：The official institutions (NBER, OECD, CEPR and others) provide business cycle chronology with a lag from 3 months up to several years. Markov-Switching Dynamic Factor Model (MS-DFM) allows to produce the turning points more timely. The Kalman filter estimates of the model can be obtained in one step with limited number of series or in two steps on a much richer dataset. While the choice of correct series is a challenge for the one-step method, the problem of the two-step method is the potential misspecification. In this paper we apply one-step and two-step approaches to the French data and compare their performance. Both methods give qualitatively similar results and prove to reproduce the OECSD business cycle chronology on the 1993-2014 monthly sample well. We find that the two-step method is more precise in determining the beginnings and the ends of recessions. Also, both methods produce extra signals corresponding to downturns which were too short to belong to OECD chronology of recessions.
关键词：Dynamic factor models; Markov switching models; business cycle turning points
Loading...
