A new study has discovered that the brain activity of professional investors can predict future stock market performance more accurately than traditional stock metrics or the investors’ own predictions.
The study, conducted by researchers from Erasmus University in the Netherlands, was published in the journal Proceedings of the National Academy of Sciences (PNAS) on April 9, 2024.
It challenges the long-held belief that it is impossible for investors to reliably forecast the stock market.
This belief, known as the efficient market hypothesis, assumes that stock prices reflect all available information and that no one can consistently beat the market.
But the researchers found that the amount of activity in a specific part of the brain called the nucleus accumbens (NAcc) was higher when investors were evaluating stocks that ended up overperforming in the market one year later.
Remarkably, this effect remained significant even when taking into account traditional stock metrics and the investors’ own predictions.
The nucleus accumbens is a small area deep within the brain that is associated with reward processing and anticipation.
And the study suggests that when investors are presented with information about a company, their nucleus accumbens activity reflects an intuitive, anticipatory response that can predict the company’s future stock performance.
Interestingly, the nucleus accumbens activity was most predictive during the initial presentation of company information, such as the company profile and stock price graph.
This suggests that investors’ initial, gut reactions to a company may be more valuable for forecasting stock performance than their later, more deliberate analyses.
The Rise of Neuroforecasting
The study is part of a growing field called “neuroforecasting”, which uses brain activity from small samples of people to forecast market-level outcomes.
Previous neuroforecasting studies have predicted things like album sales, movie box office results, and crowdfunding success.
This is the first study to demonstrate neuroforecasting’s potential in the complex, high-stakes world of professional stock market investing.
Methodology
The study involved 34 professional investors from leading Dutch investment companies.
These participants had an average of 19 years of experience in the finance industry, 12 years of experience in asset management, and 15 years of experience in equity analysis.
Two participants were excluded from the analysis due to excessive head movement during the fMRI scanning, resulting in the final sample of 34 participants, 33 of whom were men.
The ranged in age from 29 to 66, with an average age of 48.
The fMRI scanning session lasted approximately 85 minutes.
During this time, the participants completed the prediction task while their brain activity was being measured.
In the task, participants were presented with 45 anonymized investment cases in randomized order, each consisting of five information screens: company profile, price graph, fundamentals, relative valuation, and a news item.
They saw each screen for between 7 and 20 seconds.
After viewing all five screens for a given case, participants were asked to predict whether the company’s stock would overperform or underperform its market segment one year in the future.
The cases were based on real-world companies from various sectors, with stock data sampled from the period of 2000-2011.
The period of 11 years was meant to ensure that the cases were not affected by a single economic trend.
Half of the cases were chosen based on their overperformance one year after the selected data period, while the other half underperformed.
The researchers used objective, historical stock market data to determine the over- or underperformance of the stocks relative to their market segment benchmarks.
To prevent any recognition of the cases, participants were not informed about the identities of the investment stocks, or the period from which they were sampled.
Motivating the Participants
The researchers used a performance-based incentive to ensure that the participants were motivated to make accurate predictions.
Specifically, the participants were informed that whoever made the most accurate predictions of the stock outcomes would be awarded a prize of €500.
In the end, two participants tied for the best performance, so each took home €250 (these prizes were awarded based on the accuracy of the participants’ conscious predictions, not the predictions derived from their brain activity).
This type of performance-based compensation helps to ensure that participants are engaged and motivated.
In this case, it is unlikely that the prize money would significantly bias the results, as the brain activity measured by fMRI reflects subconscious, intuitive processes that are difficult to consciously control or manipulate.
Results
The researchers analyzed the fMRI data by focusing on predefined brain regions that have been implicated in neuroforecasting studies, including the NAcc.
They then used statistical models to test whether brain activity in these regions was related to future stock performance, controlling for traditional stock metrics and investors’ own predictions.
The results were striking: while the investors’ predictions were no better than chance at forecasting stock performance, their NAcc activity levels predicted with 68% accuracy whether a stock would overperform or underperform.
In other words, the subconscious brain activity of these professional investors, specifically in the region associated with reward anticipation, was significantly better at predicting stock market outcomes than the conventional methods they use in their daily work.
Specifically, their NAcc activity was most predictive during the initial presentation of company information, specifically the “company profile,” “price graph,” and “fundamentals” screens, whereas the activity levels measured during the presentation of the “relative valuation” and “news item” screens were significantly less predictive.
The researchers caution that more evidence is needed before financial institutions should start collecting neural data as an integral part of their investment process.
Yet the study opens up exciting new avenues for research and challenges traditional assumptions about the predictability of the stock market.
The immediacy of MRI readings in this case also hints at the possibility that trader intuition might be more emotionally driven (and thus less introspective) than previously thought.
“In conclusion,” the researchers write, “our study suggests that future stock market performance can be predicted by brain measures of professional investors.”
And that raises the question, they write, “whether financial institutions should invest in collecting such information.”
Study Details:
- Title: “Brain activity of professional investors signals future stock performance”
- Authors: Leonard D. van Brussel, Maarten A.S. Boksem, Roeland C. Dietvorst, Ale Smidts
- Publication Date: April 9, 2024
- Journal: Proceedings of the National Academy of Sciences (PNAS)
- DOI: https://doi.org/10.1073/pnas.2307982121