Boosted Return with News: News, Volatility, and Portfolio Implications

Abstract

We use XGBoost, a powerful and state-of-the-art machine learning algorithm, to predict next-day volatility jumps using financial news. We show that using over 1,400 news topics, next-day RSJ, a recent development of volatility jump measure, is reasonably predictable, and using the news-induced RSJ to form one-day portfolio achieves outstanding portfolio performance where annualized Sharpe ratio is 2.27 with only 44 stocks available. The performance is significantly higher than portfolio selection based on today’s RSJ (0.619) or today’s news sentiment (1.94). The improved Sharpe ratio is mainly from higher return. The portfolio earns a significant alpha relative to rational and behavioral pricing factors. We show that investor recognition and attention potentially explain the portfolio performance, although overall model fit is low in terms of R² and adjusted R², suggesting that the portfolio’s performance is largely unexplained by the pricing factors considered.

Keywords: Fintech, Machine Learning, Asset Pricing, Volatility Modelling, Volatility Jumps, Textual Analysis