RavenPack | January 10, 2021
Evidence that news analytics and sentiment data provides original sources of alpha not already factored in by existing indicators and models.
In this selection of research studies published by our in-house data science team as well as independent researchers, you will find evidence that our news analytics provide original sources of alpha not already factored in by existing indicators and models.
When conducted on a portfolio of 141 stocks, a strategy based on the observation generated an in-sample annual excess return of 4.78%, and 9.87% out-of-sample.
RavenPack research found that high co-mentions of the virus in the news with high beta stocks such as those in the airline or pharma sector, for example, could be successfully used as the basis for modeling future returns.
Researchers at Duke University and the University of Luxembourg found that economic news could be used in a 3-step process, first to estimate higher frequency econometric variables which could then be aggregated to estimate GDP, then finally to use GDP forecasts to estimate future crude oil prices.
The addition of news sentiment delivered an annualized return of 210bps with an information ratio of 0.70 – a result that was 3x better than the benchmark strategy.
Using an automated news-based strategies with almost all research showing, in the case of news-based approaches, Small Caps substantially outperform their Large-Cap brethren
Our research concluded that news sentiment-based strategies worked, especially in the short-to-medium term with the region producing information ratios close to 4.0, and 20% annualized returns.
A strategy based on insider trading data achieved annualized returns of 8.6% and 13.9% with a global Mid/Large-Cap and Small-Cap strategy respectively, and information ratios of 1.82 and 2.64.
Researchers at Monash University found that overreactions tend to be short-lived, with prices mean-reverting over time, providing contrarians with the perfect ‘buy the dip’ trading strategy.
An analysis of the recommendations produced by 150+ brokers, as well as a model that used RavenPack sentiment data, ranked how well we performed during the height of the pandemic. The RavenPack signals ranked 4th and generated a 16.82% hypothetical return.
RavenPack's news analytics-based election monitor successfully forecasted the winner of the U.S. presidential election, and has now accurately forecasted the winner in five out of the last six U.S. presidential elections
Please use your business email. If you don't have one, please email us at info@ravenpack.com.
By submitting this form, you agree to RavenPack's terms of service and privacy policy.
Your request has been recorded and a team member will be in touch soon.
Schedule a personalized trial to see how our actionable insights can boost your investment strategies.