Recent developments allow a nonparametric separation of the continuous sample path component and the jump component of realized volatility. The jump component has very different time series properties than the continuous component, and accounting for this allows improved forecasting of future...

This paper provides empirical evidence that volatility markets are integrated through the time-varying term structure of variance risk premia. These risk premia predict the returns from selling volatility for different horizons, maturities, and products, including variance swaps, straddles, and...

Recent modeling developments have created tradeoffs between attribution-based models, models that rely on causal relationships, and “pure prediction models†such as neural networks. While forecasters have historically favored one technology or the other based on comfort or loyalty to a...

This study provides new insights on the nexus between Tweet sentiments and stock price in China. Based on machine learning, we classify the Tweets from Weibo, a Twitter's variant in China into five sentiments of anger, disgust, joyful, sadness, and fear. Using wavelet analysis, we find close...