The duration analysis of structural breaks: is stability destabilizing?

This article investigates how the duration of a regime (i.e., the length of stability) affects the likelihood of a permanent structural break by devising the duration dependence in structural break method, which combines a structural break test and a duration dependence test. First, the locations of structural breaks are identified by Bai and Perron’s (1998) method. Then, it is estimated how a hazard rate changes in duration between structural breaks by parametric duration analysis using the Weibull and the log-logistic hazard functions. This study reveals the evidence of positive duration dependence, that is, stability is destabilizing, in 13 out of 27 international stock indices and the pooled data. In other words, as one regime continues over time with unchanged parameter values, a new structural break is more likely to occur. This method discloses a lower degree of duration dependence than the duration-dependence Markov-switching model (Durland and McCurdy, 1994) that considers temporary switches between a limited number of regimes. Also, some new patterns are emerged, e.g., more predominant positive duration dependence in bear markets, the secondary stock exchanges of a country and developing countries.
Publisher statement: This is an Accepted Manuscript of an article published by Taylor & Francis in Applied Economics on 28th November 2014, available online: http://www.tandfonline.com/doi/abs/10.1080/00036846.2014.985370 .