Evidence-based investors have long debated the efficient market hypothesis (EMH), popularized by Gene Fama. In the new age of social media echo chambers, meme stocks and information overload, it has become fashionable to argue that markets are growing less rational. BlackRock’s William Ezratty, Gerald Garvey, Timothy McDade and Andrew Robinson, authors of the study “The impressive markets hypothesis: The Basics of Price Forecasting (Still),” published in the April 2026 issue of The Journal of Portfolio Management, push back against the narrative of declining market efficiency, arguing that markets remain far more efficient—and thus harder to beat—as stock prices remain impressive predictors of future business performance.
What problem were the authors trying to solve?
AQR’s Cliff Asness published an influential 2024 paper, “The less efficient market hypothesis”, arguing that markets are made in a measurable wayless efficient— that prices have been disconnected from fundamentals, evidenced in part by historically wide “value spreads” (the gap between the cheapest and most expensive stocks by valuation metrics like book-to-price).
The BlackRock team took this hypothesis seriously, but wanted to test it directly. Instead of looking at what prices are doing relative to current book values, they asked a more pointed question: Do stock prices today actually predict future cash flows? And has this predictive power eroded over time? They examined over 3,000 U.S. stocks each year from 2004 to 2024—a dataset that spans the Global Financial Crisis, the rise of passive investing, the COVID hit, the alternative data explosion, and the early days of big language models.
Methodology: How did they test it?
The core approach was elegant: use a company’s current valuation ratio (specifically, the book-to-price ratio—the same metric underlying the famous Fama-French “value” factor) to predictoperating cash flowa year later. They controlled for each company’s actual profitability, so they could isolate what information prices add beyond what the accounting statements already tell us.
They then applied three tests.
- Does predictive power change when value differences are wide—the condition Asness highlights as a sign of potential mispricing?
- Has power strengthened or weakened concretely in the 2020s?
- Does market predictive power differ between revenue generation and profit margins?
Key findings
The results are organized around three core findings, each with significant implications:
Prices still predict fundamentals – consistently
Higher value companies (those trading at a premium to book value) continue to generate significantly stronger operating cash flows a year later. The effect is economically meaningful: a one standard deviation change in valuation shifts the expected future return from approximately 10% to nearly 14% of assets. This holds for different industry definitions, time period fixed effects and cOLLectiON approaches.
Wide value spreads do not damage the signal
This is perhaps the most direct challenge to Asness’s thesis. When the authors interact value spreads with their predictive measure, the coefficient is actually slightly negative—meaning that prices are if anythingbetterpredictors when spreads are wide, no worse. The coefficient is statistically insignificant, so the more cautious statement is that there is simply no evidence that wide differences reflect reduced information content in prices.
Prices have improved on revenue, worse on margins
The authors divide the benefit into two components:
- How much income does the business generate from its assets?
- How much profit does he keep from each dollar of revenue?
The first is the “asset turnover” part – sales broken down by assets.
The latter is the margin part – profit divided by sales.
Multiply the two and you get the total benefit: OPCF/Assets (operating cash flow divided by total assets)
A striking divergence emerged between the two components. Market prices have risen significantlymoreaccurate in predicting future revenue generation over time – consistent with the growth of alternative data sources such as credit card transactions, web traffic and location data that are heavily focused on revenue. However, the predictive power for profit margins has declined. The cost side of the income statement appears to be increasingly undervalued by the market.
To illustrate the revenue skew in alt data, the authors refer to an Eagle Alpha report on the 20 most popular alt data products in 2021. Of these, 12 were clearly classified as revenue-focused (consumer transactions, app usage, web traffic, location). Only a single vendor — Revelio Labs, which collects employment data from LinkedIn — was classified as cost-focused.
Their findings led the authors to conclude:
“Markets may not be perfectly efficient, but they contain an impressive amount of forward-looking information. This is as true today as it was in the past.” They added: “Prices are no less impressive predictors of future fundamentals when spreads are wide, nor are their information content captured by features such as accruals, leverage, volatility and external financing.”
Key Investor Relations
- Do not abandon market prices as a source of information.The notion that markets have become informationally hollow is not supported by the data. Estimation-based strategies preserve the true content of the prediction.
- Wide value differences are not necessarily a sign of irrational markets.The gap between cheap and expensive stocks reflects a mix of risk, mispricing,ANDlegitimate forward-looking information. Investors shorting “overpriced” stocks simply on the basis of wide spreads should be careful—those lofty valuations may well be justified by the future fundamentals the market may see.
- The cost side of the income statement is the new frontier.If market prices have become adept at incorporating revenue signals (via alternative data) but are weakening on margins, then active managers who build proprietary insights into cost structures—supply chain efficiencies, labor cost trends, input pricing—may have a real advantage that markets are not yet capturing.
- LLMs can close the gap—eventually.The authors note that their data captures the alternative burst of data, but only the early beginnings of large language patterns. LLMs, which can analyze transcripts of earnings calls, regulatory filings and supplier disclosures, may be better suited than consumer transaction data to illuminate the cost side. This is a space worth watching.
- Valuation is a better tool than spreads for assessing market efficiency.Rather than using the level of value spreads as a barometer of market rationality, investors are better served by asking whether prices retain their underlying predictive content—and to that extent, the evidence is reassuring.
CONCLUSION
This paper offers a measured and empirically grounded counterpoint to the narrative of market quality deterioration. Its title—”The Impressive Markets Hypothesis”—is a deliberate nod to (and gentle rebuke of) the “Less Efficient Markets Hypothesis” against which it argues. Markets are not perfect. Anomalies exist. But the idea that social media and information overload have fundamentally broken the relationship between prices and business fundamentals doesn’t survive contact with 20 years of data.
What is perhaps most valuable to practitioners is not the certainty itself, but the nuance. The informational advantage of the market is not uniform. It looks sharper on revenue than earnings. And this asymmetry—fueled by the one-sided nature of the alternative data industry—shows where disciplined, fundamental-minded active investors can still find a consistent signal.
In the authors’ own words: “There appears to be a great opportunity for active managers to gather new information and model the cost side.” In line with Andrew Lo’s Adaptive markets hypothesiswhich states that markets become more efficient as investors compete to exploit inefficiencies, in doing so they will make the market even more efficient.
Larry Swedroe is the author or co-author of 18 books on investing, including his latestEnrich your future. He is also a consultant to RIAs as an educator on investment strategies. This article is for informational and educational purposes only and should not be construed as specific investment, accounting, legal or tax advice.
Impressive Markets Hypothesis: Prices still know the future originally published in Alpha Architect. Please read the Alpha Architect FINDINGS at your convenience.
