The Buffett Indicator Manual

Research note: every current number, historical comparison and outcome table is computed at render from the maintained Fed/BEA ratio dataset, profits cross-check, VTI closes and the Shiller real total-return inputs. The official and estimated readings are never blended. How we use AI.

The official Buffett Indicator is 250%; the market-close estimate is 289%. Both are historically extreme. Neither tells you when the market must fall—and the completed high-valuation sample is much thinner than the headline implies.

What the Buffett Indicator actually measures

The Buffett Indicator divides the market value of publicly traded US corporate equities by the annualized dollar value of US economic output. Its intuition is deliberately broad: stockholders own claims on businesses, and those claims cannot compound faster than the economy forever without either profit share, valuation, or the geographic scope of those businesses changing.

The name comes from Warren Buffett's 2001 Fortune discussion of market value relative to gross national product. He called the relationship a strong single measure of valuation and warned that a ratio near 200% meant investors were playing with fire. The quote is memorable; the implementation is not standardized. This page uses the Fed's all-domestic-sectors public-equity liability series and BEA nominal GDP because both are primary, reproducible and maintained.

The official reading and the market-close estimate

The latest official observation covers Q1 2026, was released on Jun 11, 2026, and reads 250%: $79.67 trillion of public-equity value divided by $31.87 trillion of annualized nominal GDP. It fell 14.6 percentage points from the prior quarter because equity value changed -4.2% while GDP changed +1.4%.

The official number is stale by design, so the gauge page also carries a market-close estimate: 289% as of Aug 21, 2026. We scale the last official market value by VTI's price change since that quarter end, then divide by the latest published GDP annual rate (Q2 2026, released Jul 30, 2026). This is more current, but it does not acquire official status merely because it is precise to one decimal place.

The complete 1947+ history

The full-record median is 84%. The late-1960s peak was 102%, the dot-com peak 172%, and the 2007 peak 131%. The series moved onto a visibly higher plateau after the financial crisis and first exceeded the dot-com benchmark in 2020. A fixed threshold therefore answers only one question—how far the current economy sits from its own past—not whether the old threshold is still a stable estimate of fair value.

Why today's 250% is not the same valuation as 2000

Market cap divided by GDP can be rearranged into two economically distinct pieces: market cap divided by profits, multiplied by profits divided by GDP. The first is an economy-wide price-to-earnings multiple. The second is the share of output captured as after-tax corporate profit. This identity turns the standard “the indicator is obsolete” argument into numbers we can inspect.

This does not make 250% cheap. It makes the diagnosis narrower: the market is extraordinarily large relative to domestic output, while its multiple of current economy-wide profits is elevated but below 2000. If profit share mean-reverts, the apparent denominator support disappears. If high margins persist, the raw Buffett Indicator overstates how far valuation alone moved.

Starting valuation and future real returns

To test the claim rather than repeat it, we start each official observation with the first monthly S&P observation after its initial publication date and measure subsequent inflation-adjusted total return with dividends reinvested. The dots below are completed ten-year windows; the table adds 1-, 3- and 5-year horizons. Overlapping starting quarters remain visible, so every cell discloses both sample size and distinct episode count.

The cheap end is the clearest result: readings under 70% preceded a median +10.7% annualized real return over the next decade across 5 distinct eras. The expensive end is not yet statistically mature. Only 1 quarter has a completed ten-year window, from 1 era. The current high plateau cannot grade itself until the 2030s.

Does a high Buffett Indicator predict a crash?

No. The benchmark table shows why the folklore survives: the dot-com and 2007 peaks were followed by deep drawdowns and weak five-year real returns. But it also shows why the rule is unusable as a timer. The ratio crossed old “danger” levels years before some market tops, the 130–169% band produced positive median returns at every horizon here, and the modern 170%+ cluster kept rising long after exceeding the 2000 record.

A valuation ratio can compress through falling prices, rising GDP, rising profits, or time. It does not specify which mechanism comes next. Use it to set long-horizon expectations and stress-test assumptions—not to choose tomorrow's position size.

Why different websites show different Buffett Indicator numbers

There is no official series called “the Buffett Indicator.” The Fed publishes market values; the BEA publishes GDP; every website makes implementation choices. A Wilshire-based daily proxy, the Fed's quarterly public-equity value, total corporate equities including closely held companies, GDP versus GNP, and end-of-quarter versus interpolated dates all produce different levels. A globalization adjustment adds another modeling layer rather than discovering a hidden official denominator.

Consistency matters more than choosing the number with the most recent timestamp. This page's official history always uses BOGZ1LM883164115Q ÷ GDP. The live estimate is displayed separately, uses VTI only for the post-release market move, and is replaced—not spliced into history—when the next Fed observation arrives.

Where the Buffett Indicator will mislead you

• Global companies, domestic denominator. US-listed firms earn abroad, while US GDP measures production inside the country. That mismatch grew over time.
• Profit share is not constant. Asset-light businesses, globalization, tax changes and market concentration lifted corporate profits relative to GDP. The decomposition above measures the consequence.
• Interest rates are absent. Market cap/GDP never asks what bonds yield. Pair it with the Shiller CAPE and excess CAPE yield.
• A stock is divided by a flow. Market value is measured at quarter end; GDP is an annualized flow during the quarter. That convention is valid, but abrupt quarter-end price moves matter.
• Source history can revise. The Fed warns that Z.1 releases may include major data and structural revisions. This page shows today's maintained vintage on historical release dates, not an unrevised real-time vintage archive.
• Valuation is not timing. An expensive market can become more expensive; a cheap market can remain impaired. The incomplete high-zone study is evidence, not a footnote.

How we checked it

Frequently asked questions

Corrections: none. The official series can revise with future Z.1 releases; the page recomputes when its maintained data changes. This is research, not investment advice.

Open the live Buffett Indicator gauge →