Identifying Volatility Clusters Using the PPM: A Sensitivity Analysis

Several previous works show that, in general, financial time series are characterized by periods of large volatility followed by periods of relative quitness. In this paper we consider the product partition model (PPM) to identify changes in the volatility extending it to identify multiple change points in normal variances assuming known means. Yao’s prior cohesions and a conjugate prior distribution for the variance – which in this case is a Inverted-Gamma distribution – are assumed. The ultimate goal is to provide a sensitivity analysis to the product estimates assuming different prior specifications for the parameter which indexes the Yao’s cohesions and also for the variance. We analyze a Chilean stock market return series and conclude that the product estimates for the volatility of this series are strongly influenced by the prior specifications of both parameters. Copyright Springer Science + Business Media, Inc. 2005