How Well Does the CAPE Ratio Predict Returns Outside the U.S.? The 90%–1% Country Scorecard
Does the valuation gauge that says U.S. stocks are expensive work anywhere else?
Yes — in most countries, but how well it works ranges from 90% to 1%. That gap, not the ratio itself, is the useful part.
The CAPE ratio — the cyclically adjusted price-to-earnings ratio, or Shiller P/E — divides a market's price by its average inflation-adjusted earnings over the past ten years. In the United States it has a strong record: a high starting CAPE has reliably meant lower returns over the next decade. But the U.S. is one market out of dozens. This briefing scores the gauge country by country, maps where the world's markets stand in October 2026, and shows the one case where it got a five-year call badly wrong. All figures are as of November 15, 2026, unless noted otherwise.
The short answer: it works abroad — about half the story
Across developed markets, the starting CAPE explains about 48% of the next 10 to 15 years of returns.
That number comes from Norbert Keimling, head of capital market research at Star Capital, who tested the ratio in 14 equity markets outside the United States between 1979 and 2013. He found the link between starting CAPE and subsequent long-term returns in every market he examined — not just the U.S. one Robert Shiller originally studied.
"Explains 48%" needs translating into plain language. If you knew nothing about a market except its starting CAPE, you could account for roughly half of the difference in future returns between markets. The other half is everything else: earnings growth, currency moves, wars, policy changes, and luck.
For scale, that 48% is strong. Joachim Klement, in his paper on whether the Shiller P/E works in emerging markets, noted that most traditional stock market prediction models explain less than 20% of the variation in future returns. On that yardstick, CAPE is one of the more reliable forecasting tools available — but "more reliable" still leaves half the story unexplained.
The country scorecard: Japan 90%, the UK 86% — and Canada 1%
The same gauge is nearly a crystal ball in Japan and almost useless in Canada — and the difference is the lesson.
Keimling's country-by-country results show how uneven the signal is:
Source: Norbert Keimling, Star Capital, "Predicting Stock Market Returns using the Shiller CAPE," as summarised by Monevator. Figures are R-squared values: the share of the variation in subsequent 10- to 15-year returns accounted for by the starting CAPE.
Two things stand out. First, the U.S. number itself changes with the window: 82% since 1970, but only 46% since 1881. The modern window reads higher, because the longer one includes eras when earnings data was thinner and the market looked very different. Second, Canada's 1% is not a rounding error. Keimling is blunt that the explanatory power varies by country and time period — and Canada is the extreme case where the gauge has told investors almost nothing about what came next.
The payoff bands behind the scorecard are just as stark. Across the 14 markets Keimling studied, a starting CAPE below 8 was followed by average real capital growth of 13.1% a year over the next 15 years — and even in the worst case of more than 4,000 observation periods, the following 15 years still delivered 5.7% a year in real terms. A starting CAPE above 32 produced almost no capital growth at all: about 0.0% a year on average, typically between −2.8% and 2.0%. Note that these are price gains only; dividends, which add to total return, are not counted.
Keimling's broader finding still holds across the uneven scorecard: in all countries a relationship between valuation and subsequent long-term returns could be observed, and with the exception of Denmark, a low CAPE below 15 was always followed by greater returns than a high CAPE.
Where the world's markets stand now: the October 2026 map
The most expensive market in the world is not the United States — it is Taiwan, at 43.4.
Siblis Research's October 2026 country CAPE table, covering 28 markets, shows a spread of 7.2 times from the cheapest market to the dearest. Only 4 of the 17 markets with more than 20 years of history sit in the upper half of their own historical range — most markets are mid-range or lower by their own standards.
Source: Siblis Research, "CAPE Ratios by Country," October 2026. Price divided by the average of the previous ten years of inflation-adjusted earnings, aggregated across each market's main index.
Japan is the trap in this table. At 33.2 it looks expensive — the third-highest reading in the world. But Japan's own historical median is 31.1, because the 1980s bubble sits inside its history. Against its own past, Japan is barely above normal. The United States is the opposite case: Research Affiliates put the U.S. reading at 39.3 on September 30, 2025, against a historical median of 16.5. The same headline number means something very different in each country.
One more wrinkle: there is no single U.S. CAPE. Shiller's own S&P 500 series stood at 40.09 at the close on November 12, 2026, and Legal & General put it at 40.5 in early November — the 98.9th percentile of its near-150-year history. Research Affiliates' version reads 39.3. Siblis, using each market's main index and its own earnings definitions, reads 35.1. The providers disagree on the decimal, not the message: the U.S. is far above its own history, and most of the rest of the world is not.
Why the same number means different things in different countries
A CAPE of 20 in one country is not the same signal as a CAPE of 20 in another — compare each market with its own past.
Siblis gives three reasons why cross-market ranking is the weakest use of CAPE:
- Inflation histories differ: the ratio restates ten years of earnings in today's money. Two markets with very different inflation over that decade are not adjusting the same quantity, and the adjustment is larger where inflation was higher.
- Growth rates differ: the gap between current earnings and the ten-year average is the growth delivered over that decade. A market that has grown quickly carries a structurally higher CAPE than a slow-growing one at the same price.
- Index composition differs: a market heavy in banks and miners will read below one heavy in software, whatever either is actually worth — and CAPE does nothing to correct for that.
China shows why the range matters more than the level. Its CAPE sits low in its own historical range — but that range is stretched by the bubbles of the past, so a low reading is mostly a comparison with a peak nobody expects to see again. The comparison the measure genuinely supports is a market against its own past, not against its neighbours.
When CAPE gets it wrong: the 2012 call that backfired
In 2012 the gauge said buy Britain and beware America. The next five years did the opposite.
Klement used the CAPE ratio to forecast cumulative five-year real returns from July 2012 to 2017. The UK, approximately fairly valued on its historical CAPE, was forecast to return 43.8%. The U.S., significantly overvalued, was forecast to return 24.5%. An investor who overweighted the UK against the U.S. on that signal would have regretted it: the "overvalued" S&P 500 beat the FTSE All-Share over the next five years — and over the next ten.
The miss is not a mystery. Five years sits inside CAPE's blind spot. Shiller himself has said the ratio does not predict what happens next year very well, but predicts what happens over the next five to ten years with remarkable consistency — and even that understates it: the strongest evidence is at 10 to 15 years. A five-year bet is a coin flip dressed up as a forecast.
The counterweight is what happens when the window is long enough. Mebane Faber showed that allocating to the cheapest 25% of countries by CAPE between 1993 and 2018 would have earned 3,052%, against 962% for the S&P 500. The signal works — but on its own timetable, not the investor's.
How to actually use it
Use CAPE to set expectations, not to time trades.
1. Check the market against its own history
Look up the market's CAPE and its own historical median — Siblis and Research Affiliates both publish country tables. A reading far above its own median is the warning; a reading near it is not, whatever the absolute number.
2. Translate the reading into a range, not a point forecast
History gives bands, not targets. When the U.S. CAPE has started above 30, subsequent 10-year real returns have averaged just 2.9%; when it has started below 10, they have averaged 10.5% — but individual decades have landed far from both averages. Our 100-year crash scorecard shows the same pattern: the gauge warns, it does not schedule.
3. Remember the number is not unique
The U.S. CAPE is 40.09, 39.3, or 35.1 depending on whose table you read. The providers use different indices and different earnings definitions. Treat the second decimal as noise and the message — far above its own history — as the signal.
4. Treat it as a 10- to 15-year lens
Nothing in the evidence says anything about next month or next year. If you want the "buy the cheap countries" route rather than the measurement itself, the two ETFs that try it — and the gap between the famous 14.7% backtest and the 8.3% live return — are covered in our CAPE-based ETF briefing.
FAQ
Why is the U.S. CAPE 40 on one site and 35 on another?
Different providers use different indices and different earnings definitions. Shiller's series uses the S&P 500 with his own earnings data; Research Affiliates derives its U.S. CAPE from the S&P 500 with a different method; Siblis uses each market's main index with normalised earnings. The readings differ by a few points, but all three agree the U.S. sits far above its own historical median of about 16.5.
Can CAPE tell me which country to invest in?
It can rank expected returns over 10 to 15 years, which is not the same as a buy signal. The cheapest-quartile strategy beat the S&P 500 by a wide margin from 1993 to 2018, but the 2012 UK-versus-U.S. call shows the same gauge pointing the wrong way for five years at a time. It is an expectations tool, not a timing tool — and it is not personalised advice.
Does CAPE work in emerging markets?
Klement's conclusion was that the Shiller P/E is a reliable indicator for future real stock market returns not only in the United States but also in developed and emerging markets in general. The same caveats apply: compare each market with its own history, expect a wide range of outcomes, and give it a decade or more.
Key Takeaway
The CAPE ratio travels well, but it does not travel evenly. Across developed markets it explains about half of the next decade's returns — 90% in Japan, 86% in the UK, 82% in the post-1970 U.S., and 1% in Canada. October 2026's map puts Taiwan at the top at 43.4, the U.S. far above its own history, and most other markets mid-range or lower by their own standards. And the 2012 miss is the discipline: five years is too short for this gauge, and a single country's reading is only meaningful against that country's own past.
Use CAPE to set your expectations for the next decade — and never to call next year.
Sources:
• Siblis Research, "CAPE Ratios by Country," October 2026 — siblisresearch.com/data/cape-ratios-by-country/
• Monevator, "CAPE ratio by country" (Research Affiliates data, September 30, 2025) — monevator.com/cape-ratio-by-country/
• Norbert Keimling, Star Capital, "Predicting Stock Market Returns using the Shiller CAPE" — via ValueWalk and Monevator
• Joachim Klement, "Does the Shiller-PE Work in Emerging Markets" — via Monevator
• AAII, "Navigating Market Cycles With Shiller's CAPE Ratio" — aaii.com/journal/article/281759
• Legal & General, "CAPE fear?" (November 3, 2026) — blog.landg.com/categories/markets-and-economics/cape-fear/

Comments
Post a Comment