Title: Stochastic volatility model with time-varying leverage effect
Authors: Toshiaki Watanabe - Hitotsubashi University (Japan)
Jouchi Nakajima - Hitotsubashi University (Japan) [presenting]
Abstract: A stochastic volatility model with a time-varying leverage effect is proposed. The leverage effect, which is captured by a correlation coefficient between innovations to today's return and tomorrow's volatility in a standard stochastic volatility model, is assumed to evolve according to an autoregressive process. An efficient Bayesian method via Markov chain Monte Carlo is developed for the estimation of the proposed model. An empirical analysis using daily stock returns, foreign exchange rate, and cryptocurrency provides evidence that the leverage effect considerably changes over time.