Event Trading Using Market Response

RavenPack | July 22, 2011

This paper shows how to enhance event-driven trading strategies.

By considering the prevailing market environment and market response to news events like layoffs, executive appointments, or analyst recommendations.

I find that a market response technique significantly improves both the daily Hit Ratio and the risk-adjusted performance of the benchmark strategies that take either long or short position throughout the entire backtesting period.

Generally, I find that the market response approach works best for the most frequent event categories, since this allows for a more local extraction of the market environment around the time of the event.

For events with at least 10,000 instances during the backtest period, the market response approach improves the benchmark strategy in 76% of the cases, while as much as 95% when focusing on raw returns rather than excess returns.

Introduction

Understanding how news events impact individual stock prices is a challenging exercise. With the emergence of News Analytics, quantitative researchers have better tools and data to test and understand the complex relationships between public information and stock prices.

Even though it may be possible to classify certain news events as being either positive or negative to a company, how the market responds or interprets these events may depend on the prevailing market environment. In contrast to previous research that relies on aggregated news scores to arrive at a market sentiment proxy [Hafez, 2010], this paper considers how financial markets tend to respond to given types of events dynamically over time, thereby capturing the current interpretation of such events.

This study is structured as follows. Section 2 presents basic statistics around events as detected by RavenPack, while examples of the median stock price development 60 days prior and post event are shown in Section 3. In Section 4, I introduce the concept of market response and provide some examples based on a one day investment horizon. Finally, in Section 5, I present the conclusions of the study.

Event Frequency and Timing

Data & News Analytics

To evaluate the market reaction to company-specific events, I use news analytics data from RavenPack going back to the year 2000. The data set includes tens of thousands of records per day, each representing the mention of a company in a financial news story.

Currently, RavenPack tracks around 28,000 companies globally, which represent more than 98% of the investable global market. Furthermore, they currently identify 278 market moving events in the news including earnings announcements, lawsuits, layoffs, and product recalls. Each record is tagged with a timestamp to the millisecond and contains data for sentiment, novelty, relevance, event categories, among other news analytics. Furthermore, each company is identified systematically using its respective point-in-time ticker symbols and/or other security identifiers, and both ”dead” and ”survivor” companies are included in the data set.




We will process your personal data with the purpose of managing your personal account on RavenPack and offering our services. You can exercise your rights of access, rectification, erasure, restriction of processing, data portability and objection by emailing us at privacy@ravenpack.com. For more information, you can check out our Privacy Policy.