Great investors tend to be avid readers, always looking for information that will give them a financial edge. There are decades, if not centuries, of examples of professionals who combined what they read—in a book, an article, or a regulatory filing—with their market experience to gain lucrative insight. For example, investment manager Jim Chanos’s careful reading of Enron’s regulatory filings and his past experience detecting fraud led him to suspect accounting irregularities at the company. He made $500 million when Enron filed for bankruptcy in 2001.
These days, however, even the most avid readers would struggle to compete with the volume of financial information that artificial intelligence, in the form of large language models, can uncover. LLMs have gained mainstream popularity thanks to OpenAI’s ChatGPT, an advanced chatbot powered by a series of pre-trained generative transformative language models. OpenAI has released several versions of its LLM, with GPT-3.5, GPT-4, and GPT-4o being among the most recent.
Almost ten years ago, Chicago Stand Review published an article titled “Why Words Are the New Numbers” This is a coming revolution in text analytics. This announced revolution has arrived and has demolished the monopoly that numbers have long held in forecasting models. Numbers are still important, of course, but text analytics is booming and everything is now potential data.
The straight talk on earnings conference calls? The data. The formal prose of annual statements? The data. The news articles? The data. The entire Internet? The data.
LLMs are trained on vast volumes of text covering a wide range of information and can apply their knowledge repositories to evaluate new information. While a human relies on past experience and intuition, LLMs use data and models from their training.
And they operate at a scale that exceeds human capabilities, rapidly analyzing mountains of text and allowing traders and investors to extract information from it faster and more accurately than ever before. They can connect ideas from different parts of a text to create a better understanding of its overall content. LLMs can even be customized, trained to become experts in accounting irregularities, or, say, shopping mall leases or risk management.
Every asset manager with a technology team now has the opportunity to tap into and leverage a massive knowledge base, and many are doing so. Funds are using LLMs to read and glean information from earnings call transcripts, 10-K regulatory filings, annual reports, social media and streaming news headlines, looking for clues about a company’s direction.
From the results of this text mining, LLMs can create direct trading signals (buy or sell instructions) or develop new predictive variables for their forecasting models. If you hold actively managed funds in your retirement accounts, chances are the professionals executing the strategies are leveraging the research power of LLMs.
It’s logical to wonder whether the benefits of LLM strategies will disappear once everyone uses them. This is the case with arbitrage strategies: their returns drop when too many investors chase limited opportunities. However, there seem to be more opportunities in this area than in arbitrage scenarios. While the field is in its infancy, researchers continue to find new ways to apply AI to extract investment insights and trading opportunities. Additionally, new data sources ranging from text to images to audio and video are helping to uncover insights that are not so easily captured by markets.
Researchers, like traders, are trying to stay ahead of the curve. Here are 10 of their recent observations.