Despite the key influence of news coverage on the financial markets, there has been no large-scale examination of this across different countries. In this project, we use semantic fingerprinting, a novel language processing technology, to translate qualitative descriptions in news articles into quantitative measures. We develop a “news intensity” measure, study its impact on the volatilities of stock markets across 24 countries, and examine its relationship with economic policy uncertainty and investors’ attention.
In today’s information age, we are bombarded with news from all directions, thanks to the widespread use of computers and smartphones connected to the internet. Still, our limited attention spans have significant implications for the financial markets.
When investors decide which stocks or funds to buy, they first select a set of options to consider and then choose one of them. Attention is a scarce resource for individuals and when we’re overloaded with information, the options that grab our attention are more likely to be considered and chosen. Individual investors are more likely to buy stocks that catch their attention, such as those in the news. This means news coverage has important consequences for the behaviour of investors and, by extension, the stability of the financial markets, which is essential for economic growth.
The scope of almost all existing studies on news and stock markets is limited in three important ways.
Firstly, the news stories examined often include information that directly references the asset studied, which does not allow market reactions to the news coverage per se to be disentangled from reactions to the underlying information.
Secondly, there is a lack of large-scale evidence showing the importance of news coverage, with most existing studies being anecdotal. The few large-scale studies that feature media coverage use information that directly references an asset.
Thirdly, with very few exceptions, the assets studied in the literature are US-based, raising significant questions about whether the findings can be generalised globally.
Our research moves beyond these limitations by conducting a comprehensive study of the impact of indirect news coverage across a large group of countries.
We adopt a new measure called ‘news intensity,’ which estimates the quantity of indirectly relevant news coverage. It is calculated using semantic fingerprinting, a novel language processing technology implemented by Cortical.io.
This technology translates qualitative descriptions into quantitative measures. The semantic fingerprinting represents text as a sparse binary 128×128 matrix of 16,384 topics. The semantic fingerprint of a word is represented in the matrix by topics with which the word is associated. The semantic fingerprint of a text is the aggregation of the semantic fingerprints of keywords in that text.
In this project, we take the semantic fingerprints of the countries in our sample and use them as filters for the semantic fingerprints of news stories to focus on the news most relevant to each country. This approach attributes to each country the news stories likely to be relevant to it, even if the country itself is not mentioned.
This research provides valuable insights into the role of news coverage in financial markets and offers a robust framework for policymakers and organisations to understand and mitigate the impacts of news on economic stability.
Organisations interested in our research can partner with us with confidence backed by an external and independent benchmark: The Knowledge Exchange Framework. Read more.
Project last modified 29/08/2025