Data & Analytics Insights

Predicting US stock volatility in unpredictable news cycles

Amit Das

Proposition, News Feeds & APIs
  • Examine how quarterly earnings reports, released outside regular trading hours, lead to significant price adjustments.
  • Discover the surge in news coverage following earnings reports and its relevance to stock valuations.
  • Explore how news sentiment can be used to predict post-earnings price trends, offering valuable trading opportunities.

Understanding and managing volatility is a crucial aspect of trading and portfolio management. It constitutes one half of the formula for the most widely used measure of performance – Sharpe ratios. Indeed, concerns around volatility have only increased in recent years. The VIX is one of the leading indexes for tracking market volatility. As can be seen, 2020 marked a clear breaking point where volatility in US markets took an upward shift. Between 2020 and present day, the VIX has had an average close of ~$21.7. This is over a 40% increase over the average close between 2015-2020. This is not only being driven by the handful of what we hope once-in-a-lifetime events that have occurred over the past few years. Looking at a more granular level, the VIX closed above the 2015-2020 average on 79% of the market days since the start of 2020.

The below chart shows the predictive power of news over short-term volatility. To focus on predicting volatility rather than direction of price movements, we group all historical daily return z-scores based on their absolute value and then plot the average news z-score over the previous 72 hours. LSEG’s news score shows a significant spike when looking at the top 10% of the most volatile trading days over the past five years. 

vix daily closing price

A multitude of approaches exist to help manage portfolio or company level exposure to broader market volatility. These range from quantitative methods, such as volatility scaling and consideration of cross-strategy correlation metrics, to more fundamental approaches that aim to maintain various forms of neutrality, be that across sectors, long/short allocations etc. These approaches are certainly useful tools for handling the challenges that have been faced in recent years. However, each approach attempts to address the issue of volatility at the stage of portfolio construction. In this article, we look at how LSEG’s News data can be used to manage volatility when selecting trades to be taken. 

News as a predictor

News signals are among the best to ascertain the market’s outlook on a stock at any point in time. News is not strictly backwards looking like many signals derived from price data are. Nor do they necessarily concern the long-term fundamentals of a company, which in turn can require long-term evaluation periods. For our current purpose, we are most interested in determining short-term, stock-specific volatility which is more than broader market volatility.

The best way to achieve this is to get a sense of interest and stability relating to the stock. The MRN (Machine Readable News) data archive contains historical records of news articles. With each record, the full body of the article as well as the companies the author believed the article was relevant to are stored. Using these records, we generate historical counts for each stock that summarise the number of high-polarity articles which are being published over a 3-day lookback window. To assess the effectiveness of our signals, we look at the relationship between the scale of historical stock returns against the news signals at the time. We lag the news signals by 24 hours to ensure that any insights they give can be effectively acted on.

One challenge that arises in doing this is that both news coverage and volatility varies from stock-to-stock. To account for this and allow for effective grouping across the entire US universe, we look at rolling z-scores of both the news coverage and the stock’s returns. This lets us determine how far away from “standard” behaviour a stock is behaving on a given day.

The above shows the general usefulness for predicting the higher end of daily moves. In the following, we demonstrate how our news polarity signals can predict single instances of extreme price re-adjustments.

price candles and news score for meta

Chart  shows Around mid-October in 2022, Facebook faced a series of headwinds, centred around their US, Canadian and Brazilian markets.

Source: FTSE Russell, data to Feb 28, 2025. Based on Kim and Wright (2005). Past performance is no guarantee of future results.

Around mid-October in 2022, Facebook faced a series of headwinds, centred around their US, Canadian and Brazilian markets. These three events coincided and led to significant increases in news coverage leading up to the 26 Oct 2022. By the time the 26th came around, Meta saw a single day decline of just over 26%. News coverage calmed down following this until it spiked up again on 7 Nov. In the week following this news spike, the stock would go on a 20% run, recouping approximately half of the losses it had taken over the previous two weeks. In this example, we have seen how abnormally high news coverage a predictor of both sides of extreme tail risk was.

Traders in US markets are facing periods of significantly heightened volatility. A suite of tools and techniques are available to help address this at the portfolio level. LSEG’s Machine Readable News data presents a potent tool for screening out high-risk stocks, empowering traders to act on their risk tolerances at the trade selection phase. Furthermore, LSEG’s News Archives have global coverage, allowing for the benefits outlined above to be lifted-and-shifted to whichever markets are most salient to investors.

Our MRN data feed provides real-time coverage of some of the leading financial news providers globally. LSEG research papers have shown the value of this feed to generate directional predictions on both short and long-term stock movements. In this instance, we instead focus on predicting the volatility of a stock. In order to do this, on a daily basis, we scan for all news that is highly relevant to a given ticker. We then assess the polarity of a company’s news coverage, without looking at the directional sentiment. In this way, we can come to a conclusion on the intensity of discussion surrounding a company at any given point in time, over any given look back period.

Read more about

Stay updated

Subscribe to an email recap from:

Legal Disclaimer

Republication or redistribution of LSE Group content is prohibited without our prior written consent. 

The content of this publication is for informational purposes only and has no legal effect, does not form part of any contract, does not, and does not seek to constitute advice of any nature and no reliance should be placed upon statements contained herein. Whilst reasonable efforts have been taken to ensure that the contents of this publication are accurate and reliable, LSE Group does not guarantee that this document is free from errors or omissions; therefore, you may not rely upon the content of this document under any circumstances and you should seek your own independent legal, investment, tax and other advice. Neither We nor our affiliates shall be liable for any errors, inaccuracies or delays in the publication or any other content, or for any actions taken by you in reliance thereon.

Copyright © 2024 London Stock Exchange Group. All rights reserved.