Recognising Bubbles — Grantham's Framework
Jeremy Grantham has predicted every major asset bubble of the past 50 years with unusual accuracy — including the 1989 Japanese market, the 2000 technology bubble, the 2008 housing crisis, and others. His framework for bubble recognition is not magic. It is a structured approach to identifying when markets have moved so far from historical norms that mean reversion is not merely probable but historically inevitable.
Why This Matters
Jeremy Grantham co-founded GMO (Grantham Mayo van Otterloo) in 1977 and has spent decades documenting, studying, and eventually predicting major asset bubbles. His quarterly letters are among the most intellectually rigorous and historically informed pieces of market commentary produced anywhere, combining detailed historical valuation data with psychological analysis of investor behaviour. Grantham's central thesis is simple and empirically supported: asset prices mean-revert to their long-run historical average valuations. This is not a controversial claim — it is a statistical property of any mean-reverting series. What is remarkable is how consistently this property is ignored during periods of extreme overvaluation, and how long the overvaluation can persist before reverting.
The Core Idea
Grantham defines a bubble as a statistical event: a price move that is two standard deviations or more above the long-run mean of a valuation metric. He uses this definition deliberately — not as a description of a market condition he dislikes, but as a measurable, observable, historically consistent marker that has preceded reversions with remarkable reliability. His historical analysis of major market and asset classes from the early 20th century onward found that every two-standard-deviation departure from long-run mean valuations eventually reverted — not to the mean, but often to below it (overshooting on the downside as it had overshot on the upside). The only question was timing. Grantham identifies several specific characteristics that distinguish a genuine bubble from mere overvaluation. First, narrative displacement: in a bubble, price appreciation is justified by a novel narrative that argues "this time is different" — that the historical mean is no longer relevant because of a permanent structural change. The narrative is always plausible (genuine technological change, demographic shift, or geopolitical development often underlies it) but the price implications are always more extreme than the fundamental change justifies. Second, broad participation: bubbles require widespread investor engagement, not just a specialist sector. When retail investors, institutional investors, and media commentary are uniformly optimistic about an asset class, the breadth of the participation is itself a warning sign. Third, leverage-driven demand: asset prices that require continued credit expansion to be maintained are inherently fragile — the credit, not the fundamental, is the support. The investment implication of bubble recognition is not simply to sell and go to cash. Grantham has famously noted that bubbles can persist for longer than any sceptic can remain solvent. His approach involves reducing exposure as valuations become extreme, shifting toward less overvalued assets globally, and maintaining enough defensiveness that the portfolio survives the eventual reversion.
Jeremy Grantham's Perspective
“Grantham: "All bubbles break. All overvaluations correct. Sometimes it takes years — and sometimes, near the very end, they spike even higher and make the bears look foolish. But history has no examples of a genuine two-standard-deviation overvaluation that did not eventually revert. The investor's job is not to predict when — nobody can do that reliably — but to position appropriately for the fact that it will.”
Jeremy Grantham
A Real Example
The US equity market in 1999–2000 provides Grantham's cleanest recent case study. By multiple measures — CAPE ratio, price-to-book, price-to-revenue — the US equity market had moved to two or more standard deviations above its long-run average valuation. The narrative ("the internet changes everything — traditional valuation metrics don't apply") had widespread acceptance. Retail investor participation was at historical highs. GMO moved its equity allocation heavily to international and value stocks that were not at extreme valuations, accepting substantial short-term underperformance relative to the bubble market. When the bubble broke in 2000–2002, the NASDAQ fell 78% and GMO's relative performance was exceptional — not because of a precise timing call, but because appropriate positioning for the certainty that extreme overvaluation would revert.
The Common Mistake
The most common mistake in bubble recognition is arguing from the narrative rather than from the valuation data. Every bubble has a compelling narrative — one that makes the extreme valuation seem reasonable in context. Investors who evaluate the narrative on its own terms (rather than asking whether the narrative justifies the valuation) will find it difficult to identify when narrative and valuation have diverged beyond what the fundamental story can support. Grantham's framework begins with the valuation data — is this statistically extreme relative to long-run history? — and then considers the narrative as a secondary question. The narrative can justify moderate departures from historical norms. It cannot justify two-standard-deviation departures.
Key Takeaways
What to Read Next
The final foundational lesson explores Positioning Through the Cycle — how Ray Dalio's All-Weather framework and the principles of cycle-aware investing translate into practical portfolio construction that can survive and prosper across different economic environments.
Continue →
Positioning Through the Cycle — The All-Weather Approach