Market Cap to GDP (Buffett Indicator)
Warren Buffett once called market capitalisation divided by GDP "probably the best single measure of where valuations stand at any given moment". This is that ratio, rebuilt from public Federal Reserve data.
History
What the ratio measures
["The numerator is the market value of equities issued by nonfinancial U.S. corporations, from the Federal Reserve's Z.1 Financial Accounts. The denominator is nominal gross domestic product from the BEA. Dividing one by the other gives the price the stock market puts on American business relative to the income the economy produces in a year.
", "Buffett's point was not that the ratio forecasts a crash. It was that a ratio far above its own history means you are paying a lot for a dollar of economic output, and that the arithmetic of long-run returns gets worse from there.
"]Why our number can differ from other sites
["Several versions circulate. The most common divides the Wilshire 5000 total market index by GDP; others use the total value of all U.S. equities including foreign-domiciled issuers listed here. We use the Federal Reserve's nonfinancial corporate equity series because it is public domain, internally consistent back to 1947 and not subject to index-vendor licensing.
", 'That choice shifts the level — our ratio runs below the Wilshire-based version because financial-sector equity is excluded to avoid double counting — but the shape over time is nearly identical, which is what the percentile reading depends on. Compare percentiles across sources, not levels.
']Reading the percentile, not the headline
['A ratio of 250% sounds alarming until you notice it has been above 100% since the mid-1990s. The useful question is where today sits against the whole record: the 90th percentile means the market is more expensive than nine tenths of all quarters since 1947.
', 'It is a slow variable. The ratio moved from roughly 50% in the 1950s to over 100% by the late 1990s as margins, sector mix and intangible capital changed — a structural drift that makes long-horizon comparisons tricky and short-horizon ones nearly useless.
']How we calculate it
['We fetch NCBEILQ027S (market value of nonfinancial corporate equities, millions of dollars) and GDP (billions) from FRED, align them by quarter starting in 1947, and divide after unit conversion. Quarters where either series is unavailable are dropped rather than interpolated.
', "The series is quarterly because the Z.1 accounts are quarterly; there is no official monthly or daily equivalent. The percentile is the share of quarterly observations since 1947 below today's ratio.
"]Distribution since 1947
Each row is the share of historical observations at or below that level. A reading in the 90th percentile is higher than nine tenths of the record.
Important
Not investment advice. This page is published for information and education only. It is not a recommendation, solicitation, or offer to buy or sell any security. Market data may be delayed, incomplete, or revised. Indicators describe the past; they do not forecast returns. Past performance does not guarantee future results. Do your own research or consult a licensed financial professional before acting on anything you read here.
Questions
Is the Buffett indicator a sell signal?
No. It has been elevated for most of the last three decades and gave no useful timing signal during the 2010s. It describes how much you are paying for a dollar of GDP, not when the market will fall.
Why is your reading different from the Wilshire 5000 version?
Different numerator. We use Federal Reserve data on nonfinancial corporate equity to keep the series public-domain and consistent back to 1947. Levels differ by design; percentiles are comparable.
How often does it update?
Quarterly, after the Z.1 release and the BEA's GDP revision. Expect a lag of several weeks after quarter end.
What is a normal reading?
There is no stable normal. The historical median is far below today's level because the U.S. economy has shifted toward asset-light, higher-margin businesses. That drift is why we publish the percentile rather than a fixed threshold.