Relatore
Mauro Costantini (Brunel University, Department of Economics and Finance)
Abstract We present a regime-switch lognormal frailty model that accounts for both unobserved factors and regime changes to estimate within-sector dependence in a multivariate frame- work. We employ the penalised partial likelihood of McGilchrist and Aisbett (1991) and its extension by Ripatti and Palmgren (2000). We use the data set of public listed firms grouped into sectors on the London Stock Exchange over the period 1985-2012. We show that the within-sector dependencies are highly significant providing evidence of firm failure clustering, and neglecting this effect may likely lead to inappropriate estimates of hazard rates.
Joint work with Francis Atsu
Contact person
Alessandra Luati