Great investors tend to be avid readers, always looking for information that would give them a financial advantage. There are decades, even centuries, of examples of professionals who combined something they read (in a book, article, or regulatory filing) with their market experience to gain lucrative insight. For example, investment manager Jim Chanos’ close 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 linguistic patterns, can uncover. LLMs have gained popularity thanks to OpenAI’s ChatGPT, an advanced chatbot powered by a series of pre-trained generative transformer language models. OpenAI has released several versions of its LLM, including GPT-3.5, GPT-4 and GPT-4o among the most recent.
Nearly a decade ago, Chicago Booth Review published an article entitled “Why words are the new numbers” on a coming revolution in text analysis. The predicted revolution has occurred and it has demolished the monopoly that numbers had long held in forecasting models. Numbers are of course still important, but text analysis is taking over and everything is now potential data.
Straight talk during earnings calls? Data. The formal prose of annual declarations? Data. Press articles? Data. All Internet? Data.
LLMs are trained on large amounts of text covering a wide range of information and can apply their knowledge repositories to evaluate new information. Where a human depends on past experience and intuition, LLMs use data and models from their training.
And they operate at a scale beyond human capabilities, rapidly analyzing mountains of text and allowing traders and investors to extract information faster and more accurately than ever before. They can connect ideas from different parts of a text to better understand its overall content. LLMs can even be personalized, trained to become experts in accounting irregularities or, for example, shopping center leases or risk management.
Every asset manager with a technology team now has the opportunity to tap into – and benefit from – a massive knowledge base, and that’s exactly what many are doing. Funds use 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 the direction of a business.
From the result 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 have actively managed funds in your retirement accounts, there’s a good chance that the professionals managing the strategies are leveraging the research power of LLMs.
It’s logical to ask whether the benefits of LLM strategies will disappear as soon as everyone uses them equally. This is the result of arbitrage strategies: their returns fall when too many investors pursue limited opportunities. However, the opportunities here seem more abundant than in arbitrage scenarios. While this field is still in its infancy, researchers are still finding new ways to apply AI to extract information about investments and trading opportunities. Additionally, new data sources spanning the gamut from text to image to audio to video are making it possible to uncover insights that are not so easily priced into markets.
Researchers, like traders, strive to stay ahead of the curve. Here are 10 of their recent sightings.