Autori: Banerjee, Anindya, Lazarova, Stepana, Urga, Giovanni
Titolo: Bootstrapping sequential tests for multiple structural breaks
Periodico: European University Institute of Badia Fiesolana (Fi). Department of Economics - Working papers
Anno: 1998 - Fascicolo: 24 - Pagina iniziale: 1 - Pagina finale: 21

We show how finite sample bootstrapping methods can help to detect multiple breaks in systems of equations with long time series. The method of Banerjee and Urga (1995, 1996), where single breaks in the marginal models are imposed in the conditional model and then the conditional model estimated, is extended to cover the case of multiple (> 2) breaks in marginal and conditional models by using the technique of dominant break dating. An empirical investigation of a small monetary system for the United Kingdom establishes the viability of our method in developing congruent dynamic regression models.





Esportazione dati in Refworks (solo per utenti abilitati)