Abstract

In this talk, I will give an overview of an exciting new field of research: creating news sentiment measures in financial markets (the so-called news analytics) and studying how various aspects of news influence asset prices. My main focus will be on commodity markets, but I will also touch upon equity-related news. I will illuminate how commodity and equity markets respond to news, in particular, to positive and negative news sentiment. I will utilize event studies, Granger causality tests and news-augmented volatility models to assess the effect of news on the returns, price jumps, volatilities, correlations and liquidity.

I will show that there are significant statistical and economic news effects on asset prices, leading to potentially profitable trading strategies. For energy commodities, I will show a great asymmetry in the market’s behavior: negative events are accompanied by much greater losses than the gains surrounding positive events.

I propose a Local News Sentiment Level model for constructing a running series of news sentiment. Among other effects, I find strong evidence that news sentiment causes price jumps and conclude that market participants trade as some function of aggregated news.

I augment volatility models (GARCH as well as high-frequency volatility models (HEAVY)) with news sentiment and show that including news sentiment measures in volatility models results in superior volatility forecasts and can improve the assessment of risk.

Finally, I will discuss how news sentiments for various equities or commodities can be combined into a “news sentiment index” and how such index relates to well-know price indices.