Volatility in the stock market: ANN versus parametric models.

Relatore: Rita L. D’Ecclesia, Daniele Clementi - Sapienza University of Rome

  • Data: 30 gennaio 2020 dalle 16:00 alle 18:00

  • Luogo: Dipartimento di Scienze Statistiche - via delle Belle Arti 41 - Aula seminari - I° piano

Abstract
Forecasting and adequately measuring equity returns volatility is crucial for portfolio selection and trading strategies. Implied volatility is often considered to be informationally superior to the realized volatility. When available, implied volatility is largely used by practitioners and investors to forecast future volatility. To this extent we want to identify the best approach to track equity return implied volatility. In this paper we use parametric and ANN approaches to track stock returns implied volatility. Using daily prices for regularly traded stocks on major international Exchanges, together with data of stock market indices we estimate time varying volatility using the E-GARCH approach, the HESTON model and introducing a novel ANN framework to replicate the corresponding implied volatility. Overall the E-GARCH and the ANN approaches result accurate to track the equity return implied volatility.

Organizzatore
Maria Letizia Guerra