vai al contenuto della pagina vai al menu di navigazione

Statistics seminar: "Chasing volatility: a multiplicative error model with jumps"

Seminario di Statistica

14/05/2013 dalle 12:00 alle 14:00

Dove Dipartimento di Scienze Statistiche - Via Belle Arti 41 - Aula II

Aggiungi l'evento al calendario


Massimiliano Caporin,  Università di Padova 



The volatility of financial returns is characterized by rapid and large increments. Recent empirical studies indicate that diffusive stochastic volatility and jumps in returns are incapable of capturing the empirical features of equity index returns. Instead, it has been stressed that jumps in volatility can improve the overall fitting of stochastic volatility models. In this paper, we introduce the Multiplicative Error Model with jumps (MEM-J). This is an extension of the MEM by Engle and Gallo (2006) for estimating the presence of jumps in volatility, using a realized volatility measure, as volatility proxy, depurated by the effect of price jumps. The moments and the likelihood function are obtained in closed form. A Monte Carlo simulation experiment shows that properties of the model and the finite-sample features of maximum likelihood estimation. The MEM-J model with time-varying jump intensity is able to capture in the volatility process. The maximum likelihood estimates of the jump component seem to be reliable, even when jumps are rare events.The empirical application focuses on a set of stock indexes. The estimation results highlights a positive probability of jumps in volatility, which is consistent with the findings of previous studies on the topic.