June 2024


The importance of narratives: Past, present and future

Our Markets and Financing Research Retreat offers a wide range of academic expertise and timely market insights.

Market sentiment and media narratives have long been considered important factors for understanding financial markets and predicting portfolio returns. But increased adoption of new technologies like artificial intelligence (AI) in financial services is transforming how these factors are measured and interpreted, according to Ronnie Sadka, professor of finance, senior associate dean for faculty in the finance department and chairperson at Boston College's Carroll School of Management and State Street Associates academic partner.

Narrative-driven exposure

During State Street’s annual Markets and Financing Research Retreat in Boston, Sadka explained the impact of media-derived narratives on markets and portfolios, and demonstrated why he and his team at MKT MediaStats are leaning into AI technology to derive insights about investing.

“When there is more discussion about an asset, the asset price moves. Some assets are positively exposed to a narrative while others are negatively exposed to a narrative,” he said. “That knowledge can allow you to build narrative-mimicking portfolios.”

With access to over 150,000 primary sources, including trade publications and corporate communications, MKT MediaStats analyzes as many as five million media articles per week related to the news coverage of specific topics or particular narratives.

Quantifying narratives

Using large language models (LLMs), his team studies numerous narrative indicators, including intensity (the attention a given narrative receives), sentiment (attitudes or opinions toward a topic) and disagreement (the level of dispersion of sentiment toward a particular narrative). This granular analysis enables them to retrieve predictive data, gain insights on macro narratives and quickly assess their impact on investment strategies.

Narrative indicators help measure exposures and stock returns in a meaningful way and quantifying those narratives can assist in understanding market price dynamics, according to Sadka. “We see that sectors that have more positive discussion tend to outperform in the future,” he said.

Relying on LLM technology to pore over large reservoirs of media content, Sadka is able to measure market returns beyond traditional indicators. His research finds that LLMs significantly improve the ability to determine relevancy and predict future changes in yield.

AI-powered investment decisions

When it comes to the future of investing, Sadka sees LLMs and AI playing a prominent role. He believes there are benefits to applying the technology to analyze narratives, identify sub-narratives, and leverage that data to build asset baskets that are tailored to those narratives.

“In general, we think people don’t trade in factors, people trade in narratives,” said Sadka. “Once you can quantify narratives that otherwise seem intangible you can run betas, build baskets, look at associations and factors, and get all those things that can help investors improve their performance.”

For this reason, Sadka notes it’s important that investors understand not only the influence of economic media narratives on asset prices over time but how to efficiently leverage technology to identify fundamentals and corresponding investments.

Stay updated

Please send me State Street’s latest Insights.