Library About Earnings Whisper Numbers

Earnings Whisper® Numbers Are the Market's True Earnings Expectation

The Earnings Whisper® number has been closer to actual reported earnings than Wall Street's published consensus estimate 69.7% of the time.

That one statistic raises an obvious question: why?

For more than two decades, investors have heard about “whisper numbers.” They were often described as the market's true earnings expectation, yet almost nobody knew where they actually came from. Some believed they were leaked by company insiders. Others assumed they were rumors circulating through trading desks and message boards.

Even the first major academic study of whisper numbers couldn't pin down their origin. In 1999, professors Mark Bagnoli, Messod Beneish, and Susan Watts published “Whisper Forecasts of Quarterly Earnings per Share” in the Journal of Accounting & Economics. The researchers gathered whisper numbers from financial news services, websites, and investor message boards, and could conclude only that they appeared to come from a variety of sources. Yet despite that uncertainty, the study found that whisper numbers were more accurate than Wall Street consensus estimates — and were a better proxy for what investors actually expected. Trading strategies built on whispers earned nearly three times the returns of identical strategies built on consensus estimates.

The research proved that whisper numbers worked. It didn't explain why.

The real source of the whisper number

Long before that study was published, Earnings Whispers had already spent years chasing the same question. Like everyone else, we searched message boards, trading desks, news reports, and every place whisper numbers were rumored to appear. One thing became increasingly clear: the most reliable whisper numbers weren't rumors at all.

They were simply the analysts' current expectations.

That may seem surprising until you understand how Wall Street analysts actually work. Analysts constantly update detailed financial models as new information arrives — customer checks, supply chain data, conference presentations, macroeconomic developments, competitor earnings, and conversations with management all influence their expectations throughout the quarter. Their models may change dozens of times before a company reports.

Their published estimates usually don't. Updating a published research report requires time, approvals, distribution to clients, and updates to consensus databases. Analysts simply don't issue a new report every time their model moves by a penny. But when institutional clients call and ask, “What do you really think they're going to earn?”, analysts share their latest thinking.

Those updated expectations become the whisper number.

The Street confirmed it

That process was publicly confirmed in 1999 at a Riksbank conference in Stockholm, where UBS Warburg analyst Per Affrell explained that analysts continuously updated their internal earnings models but rarely published every revision. When clients asked for their latest expectations, analysts shared the updated numbers directly. As those expectations spread through institutional investors and trading desks, they became known as the “whisper number.”

The whisper was never secret information. It was simply the market's most current expectation.

How Earnings Whispers collects the data

Recognizing this changed everything. Rather than relying on rumors or message boards, Earnings Whispers focused on gathering analysts' latest expectations directly. In the early years, that meant calling analysts individually in the weeks leading up to earnings announcements.

As Wall Street evolved, so did our process. Today, much of the information comes through electronic research updates, institutional data sources, analyst revisions, and the same inputs analysts use to continually update their own models. The technology has changed, but the objective hasn't: measure the market's actual expectation — not what was published weeks earlier.

Why consensus estimates fall behind

Consensus estimates are snapshots. The Earnings Whisper® number is dynamic.

Consensus estimates reflect the latest published research reports, but markets don't stop processing information after a report is published. Analysts keep refining their models, investors keep gathering information, and expectations keep evolving right up until the moment earnings are released. By earnings day, the published consensus often no longer reflects what investors actually expect.

The Earnings Whisper® number does.

Two decades of evidence

Over the past two decades, Earnings Whispers has published more than 143,000 Earnings Whisper® numbers — the largest database of market expectations ever assembled. As of July 16, 2026, the Earnings Whisper® number has been closer to actual reported earnings than the published consensus estimate 69.7% of the time.

More importantly, stock prices consistently demonstrate which expectation matters.

Average close-to-close move on the day of the report, since February 23, 1999.

Since February 23, 1999, companies that beat the Earnings Whisper® number have finished the trading day higher an average of 1.9%, rising 59.8% of the time. By contrast, companies that beat Wall Street's consensus estimate but failed to meet the Earnings Whisper® number finished lower 54.8% of the time, declining an average of 0.7% on the day.

If published consensus estimates represented the market's true expectation, beating them should produce positive returns. History shows otherwise. Markets reward companies for exceeding what investors actually expect — not what analysts officially published weeks earlier.

The differences become even more pronounced over the following days and weeks through the well-documented Post-Earnings Announcement Drift (PEAD), further evidence that the Earnings Whisper® number better captures the expectations already embedded in stock prices.

Expectations drive markets

Investors often believe stock prices react to whether companies beat or miss earnings estimates. In reality, they react to whether companies beat or miss expectations. Consensus estimates measure what analysts officially published. The Earnings Whisper® number measures what the market actually expects. That distinction explains why companies can beat consensus and still see their stocks decline — or miss consensus and rally sharply.

For more than two decades, investors have called it the Earnings Whisper® number.

The market has consistently treated it as the expectation that matters.

The archive figure counts Earnings Whisper® numbers published since December 16, 1998. Accuracy is measured across events since 2003 that carried both a published Earnings Whisper® number and a consensus estimate, counting only quarters in which the published estimate did not change — a moving estimate would make “which was closer” an unfair comparison. Reaction figures average the earnings events from February 23, 1999 through July 16, 2026. Averages do not reflect transaction costs and are not a guarantee of future results.