Contents lists available at ScienceDirect Finance Research Letters journal homepage: www.elsevier.com/locate/frl Estimating stochastic volatility with jumps and asymmetry in Asian markets K. Saranya, P. Krishna Prasanna* Department of Management Studies, IIT Madras, India ARTICLE INFO Keywords: Stochastic volatility Monte Carlo Markov Chain Asymmetry Jumps Bayesian estimation ABSTRACT This study investigates the impact of stock market cycles on the volatility of Asian markets. It specifically addresses the combined effect of jumps, asymmetry and stochasticity while pre- dicting the market volatility. Our results indicate that the stochastic volatility process is highly persistent across the countries. Leverage effect, size and frequency of jumps are found to be significant and play a prominent role in computing market volatility. The empirical results imply that the stochastic volatility model embedded with the jump and asymmetric component sig- nificantly helps in measuring volatility especially during the turbulent periods. Our results have major implications for policy makers, regulators, mutual funds, hedge funds as well for other institutional investors. 1. Introduction Financial markets have undergone several economic drifts and turns over the last decade. Markets, the world over, witnessed abrupt changes in economic and fundamental factors leading to stock market cycles and fluctuations in volatility. Prior studies widely used GARCH family models to forecast volatility and found evidence in favor of in-sample estimates. However, their out of sample forecasts are poor due to their rigid linear structure (Tripathy and Gil-Alana, 2015). Financial markets exhibit asymmetric conditional volatilities (Hafner and Franses, 2009) especially, in emerging markets. Also, the presence of asymmetry is most apparent during stock market crashes (Wu, 2001). Engle (2004) postulated that ignoring asymmetry in volatility leads to a significant under/ over estimation of the risk. Corsi et al. (2010) indicated that jumps, when included in the model, increase the predicting power for volatility. Yet, all these contemporary studies modeled only one property of volatility i.e., either jumps or clustering or asymmetry or stochasticity. However, accurate volatility estimation requires the integrating all the properties. Moreover, forecasting stochastic volatility during the crisis periods requires contemporary investigation. Hence, this study investigates the individual and combined effect of these properties in volatility estimation. The study not only tests the empirical validity of the model but also explores out of sample validation in both turbulent and tranquil periods. This study contributes to the literature in several aspects: First, the impact of stochasticity, jumps and asymmetry on volatility estimation was investigated for the period 2005–2014. This period was characterized by the occurrence of successive crises and turbulent episodes. These events had significant effect on volatility. Second, volatility was modeled for both emerging and developed countries in Asia. Empirical studies on stochastic volatility modeling predominantly in the context of emerging markets are minimal. Finally, the conditional volatilities were also forecasted, to establish the predictive ability of alternate stochastic volatility models. Volatility of five Asian market indices (NIFTY, KOPSI, TWSE, FSSTI and Nikkei) over the period from 1 January 2005 to 31 http://dx.doi.org/10.1016/j.frl.2017.10.021 Received 4 October 2017; Accepted 23 October 2017 * Corresponding author. E-mail address: pkp@iitm.ac.in (P.K. Prasanna). Finance Research Letters xxx (xxxx) xxx–xxx 1544-6123/ © 2017 Elsevier Inc. All rights reserved. Please cite this article as: Kshatriya, S., Finance Research Letters (2017), http://dx.doi.org/10.1016/j.frl.2017.10.021