Is growing investor/trader use of large language models (LLM) extinguishing known stock return anomalies? In their March 2026 paper entitled “Do LLMs Make Markets More Efficient?”, Runjing Lu, Yongxin Xu and Luka Vulicevic examine how use of LLMs is affecting reactions of individual stocks to recent newsworthy events with and without outages of LLMs from three major providers (ChatGPT, Claude and Gemini). Together, these three account for nearly 80% of LLM usage. They classify outages as (1) any, (2) single-provider severe or (3) multi-provider, as documented by each provider. They focus on outages that coincide with news releases and persist beyond the NYSE close. They use RavenPack Event Sentiment Scores for articles from the Dow Jones Newswire that have ticker-specific relevance scores above 75. They control for time-varying stock/firm characteristics, past returns, new type and calendar effects. They measure daily abnormal stock returns relative to those of a characteristic-matched benchmark portfolio. Using daily outage, stock/firm and news/sentiment data during March 2023 through November 2025, they find that:Keep Reading
Recent studies, based on the distribution of reported in-sample test statistics, find that publication bias in finance is modest and that most published factors are true discoveries. What about unreported testing performed during the factor discovery process? In the April 2026 revision of their paper entitled “The False Discovery Rate in Finance: Identification Failure and Search-Adjusted Estimation”, Marcos Lopez de Prado and Frank Fabozzi argue that in-sample statistics alone do not reveal the false discovery rate (FDR) when there is an underlying anomaly search-and-selection (data snooping) process. Based on their experience in empirical finance and modeling of unreported testing, they conclude that:
Will investors in passive index funds overwhelm the ability of active investors to keep prices near fundamental value? If so, what happens? In their March 2026 paper entitled “A Model for Passive That Breaks the Market”, Michael Green, Hari Krishnan and Stephan Sturm model the impact of passive share on equity market behavior. Their model has the following assumptions:
Passive fund managers ignore fundamental value.
Equity index volatility tends to be higher when prices are low.
The fundamental value of the broad stock market tends to increase over the long run.
Active investors historically tend to push prices toward some notion of fair value. However, they may stop resisting above some passive share threshold, shorten their investment horizons and make little use of fundamentals.
Strength of reversion to fair value decreases as passive share increases.
The model considers cases for which active investors either do or do not change their behavior when faced with increased passive share. Using the above modeling assumptions and data for the S&P 500 during 1926-1994 as a baseline for the U.S. equity market without passive investing, they conclude that:
Does lack of critical assessment of cause and effect undermine reliability of findings in scientific research? In his December 2025 presentation package entitled “Investment Lessons from Cosmology: Draw Your Assumptions Before Your Conclusions”, Marcos Lopez de Prado describes how recent developments in cosmology demonstrate the potential power of causal inference to prevent false discoveries in financial research. Based on this demonstration, he concludes that:
Does lack of liquidity among stocks in anomaly portfolios effectively block exploitation? In their November 2025 paper entitled “Liquidity Constraints and the Illusion of Anomaly Profitability”, Álvaro Cartea, Mihai Cucuringu, Qi Jin and Jiexiu Zhu assess exploitability of anomaly trading strategies after accounting for individual stock liquidities. They define liquidity of a stock as its capacity to absorb incremental volume relative to recently observed average daily volume without material price impact. They estimate anomaly portfolio profitability based on liquidity-constrained dollar trade sizes/profit for each anomaly portfolio stock. They apply this approach to 128 U.S. stock return anomalies, with both in-sample (same as originally published) and out-of-sample results. They initially assume zero trading costs to isolate the impact of liquidity constraints. They then estimate trading costs (either half the bid-ask spread or price impact estimates), exclude trades expected to be unprofitable and generate the combined effects of liquidity constraints and trading costs. Using data for stocks per the 128 anomalies during January 1930 through December 2023, they find that:
How is the increasing role of interacting algorithms changing financial markets? In his November 2025 paper entitled “Algorithmic Exuberance”, Marc Schmitt presents an Algorithmic Exuberance model, which automatically stimulates market volatility from two coupled feedback channels (see the figure below):
Market-algorithmic reflexivity (trading systems learning from one another).
Information-algorithmic reflexivity (algorithmic amplification of news, narratives and sentiment).
The model derives a Reflexivity Index (RI) that quantifies the strength and persistence of market volatility from these feedback channels, and measurable Reflexivity Share of Variance (RSV) and Implied Reflexivity (IR) components. Using broad U.S. stock market data from 1980 through 2024, he finds that:Keep Reading
A subscriber suggested review of The Book of Alternative Data: A Guide for Investors, Traders, and Risk Managers , a 2020 book by Alexander Denev and Saeed Amen. In this book, the authors address “the ever growing importance of data, and in particular, alternative data. We live in a world, which is rich with data, where many datasets are accessible and available at a relatively low cost. …This book is aimed at investors who are in search of superior returns through nontraditional approaches. These methods are different from fundamental analysis or quantitative methods that rely solely on data widely available in financial markets. It is also aimed at risk managers who want to identify early signals of events that could have a negative impact, using information that is not present yet in any standard and broadly used datasets.” Based on prior research and their experience, they conclude that:
Different risk metrics capture different aspects of risk, and the relative importance of different aspects of risk varies across investors. Widely used risk metrics do not serve the interests of long-term investors because they destroy price series history. Is there a better risk metric for such investors? In the October 2025 draft of their paper entitled “Submergence Intensity: A Contextualized Risk Metric for Long-Term Investing”, Dane Rook and Ashby Monk introduce submergence intensity as a risk metric for long-term investors. They designed this metric to overcome the following shortcomings of existing risk metrics:
Many risk metrics reflect either typical/average risk or extreme (tail) risk, but not both.
Many risk metrics are insensitive to the order in which returns occur.
Many risk metrics are insensitive to asymmetries in returns (positive or negative, or part of a drawdown or a recovery.
Most risk metrics critically depend upon a small number of parameters, but these parameters are often not explicit, or not adaptable.
Many risk metrics effectively penalize liquid assets and therefore distort relative riskiness of assets.
The authors compare submergence intensity with other risk metrics, and discuss how investors can adapt it to specific preferences. Based on theoretical considerations, they conclude that:Keep Reading
The Sharpe ratio is the most widely used measure of investment efficiency. Is it truly reliable? In their September 2025 paper entitled “How to Use the Sharpe Ratio”, Marcos Lopez de Prado, Alexander Lipton and Vincent Zoonekynd review the shortcomings of conventional Sharpe ratio analysis and tackle these issues via several corrections. They also address the problems of statistical error and false discovery rates. Using theoretical analysis and Monte Carlo simulations, they conclude that:
What are life lessons from one of the leading researchers in finance? In the August 2025 transcript of his interview entitled “My Life in Finance in 12 Questions”, Campbell Harvey offers the following notable points relevant investors regarding (1) most important findings and (2) interpretation of academic research:
