When Machines Read the News: Using Automated Text Analytics to Quantify High Frequency News Impacts
Posted: 19 Jan 2010 Last revised: 3 Mar 2011
Date Written: December 20, 2009
Abstract
We examine intraday market reactions to stock-specific news. Using pre-processed data from an automated news analytics tool based on linguistic pattern recognition we exploit information on the relevance and the direction of company-specific news. Concise news-implied reactions in returns, volatility as well as liquidity demand and supply are quantified by a high-frequency VAR model on 20 second intervals. We find significant responses in prices, volatility, spreads and trading volume to relevant news for a cross-section of stocks traded at the London Stock Exchange. The results confirm the usefulness of linguistic news classification in structuring the very noisy intraday news flow.
Keywords: firm-specific news, news sentiment, high-frequency data, volatility, liquidity, abnormal returns
JEL Classification: G14, C32
Suggested Citation: Suggested Citation