What Makes a Market's CAPE Ratio Very High or Very Low? Taiwan's 43.4 vs. Turkey's 6.0
Why does one stock market trade at 43 times its smoothed earnings while another trades at 6?
It is not one thing. Six forces — expected growth, real interest rates, sector mix, accounting rules, buybacks, and inflation history — set a market's CAPE, and they can pull the same number in opposite directions.
In October 2026, Taiwan's CAPE stood at 43.4 and Turkey's at 6.0, a 7.2-to-1 spread across the 28 markets Siblis Research tracks. The United States sat at 35.1. The gap looks like a simple ranking of cheap versus expensive. It is not. A CAPE is a price for two things at once — the growth investors expect and the safety they are giving up — and both are priced into the number before you ever compare two countries.
The short answer: a CAPE is a price for growth and safety
Flip the ratio over and the logic shows itself.
CAPE divides a market's price by its average inflation-adjusted earnings over ten years. Its inverse, 1 divided by CAPE, is the cyclically adjusted earnings yield — the smoothed earnings an investor gets for each dollar paid. On Robert Shiller's measure, the U.S. CAPE was 40.7 in October 2026, which works out to an earnings yield of about 2.46%.
Now compare that with the safe alternative. The 10-year Treasury Inflation-Protected Security yielded 2.91% in real terms at the close on November 13, 2026. An investor could earn more, after inflation, from a government bond than from the smoothed earnings yield of the entire U.S. stock market.
That comparison is the whole mechanism. A CAPE rises when investors accept a lower earnings yield, and they do that for two reasons: they expect earnings to grow fast enough to justify it, or the safe alternative pays so little that stocks look attractive anyway. A CAPE falls when growth expectations fade or when real rates rise and the safe alternative starts paying properly. Bradford Cornell, writing on global CAPE data, puts it as two forces: the return investors demand for holding a country's equities, and how optimistic they are about that country's future earnings growth.
Everything below is a version of that trade-off.
The six forces that set a market's CAPE
1. Expected earnings growth
Growth expectations are the largest single driver of a high CAPE.
A market priced for fast future earnings can carry a high CAPE even while today's earnings look ordinary, because the price already contains the growth investors expect. Taiwan is the clearest current case: its CAPE rose 13.3 points in a single year to 43.4, the biggest move in the Siblis table, as investors priced in the semiconductor earnings behind the AI buildout. The U.S. CAPE, by contrast, was flat over the same twelve months at 35.1.
The trap is that the growth is already in the price. A CAPE of 43 says expectations are high. It says nothing about whether they will be met — and when a growth story disappoints, the CAPE falls from both sides at once: the price drops and the expected growth disappears with it.
For the other side of this question — how well CAPE actually predicts returns in different countries — see How Well Does the CAPE Ratio Predict Returns Outside the U.S.?
2. Real interest rates
Real rates set the price of the safe alternative, and the safe alternative sets the floor for everything else.
When the real yield on government bonds rises, investors demand more from stocks too, and multiples compress. The 10-year TIPS yield climbed from 1.72% in late February 2026 to 2.91% by November 13 — a rise of nearly 1.2 percentage points in under nine months, even while the Federal Reserve was cutting its policy rate. The long end of the market moved the other way.
Put the two numbers side by side and the change is concrete. At a CAPE of 40.7, the U.S. market's earnings yield is about 2.46%. In February, with the real yield near 1.7%, that earnings yield still beat the safe alternative. By mid-November it did not. Same market, same CAPE, different verdict — because the alternative moved.
Japan in 1989 is the extreme version. Stocks there traded at a dividend yield of about 0.4% while 10-year government bonds paid roughly 6%. Investors accepted almost nothing from equities because they assumed prices would keep rising. The Nikkei peaked at 38,915.87 on December 29, 1989, then lost about 60% within two and a half years — and did not pass that peak again until 2024.
3. Sector mix
What an index holds matters as much as what it is worth.
Software and chip companies usually trade at higher multiples than banks, miners, and utilities. So a tech-heavy index reads high and a bank-heavy index reads low, whatever either is "worth" in absolute terms.
The U.S. market is now the extreme case of this. Information technology made up 39.7% of the S&P 500 at the end of September 2026, against 12.2% for financials, and the ten largest companies made up 39.1% of the index. Counting the big tech names classified elsewhere — Alphabet and Meta in communication services, Amazon and Tesla in consumer discretionary — broad technology exposure reaches roughly 55% of the index. In 1990, technology was about 7%.
Compare that with the MSCI EAFE index of developed markets outside North America: financials 25.4%, industrials 18.4%, and information technology just 10.2% at the end of October 2026. A large part of the valuation gap between the U.S. and Europe is composition, not judgment — reweighting one index to match the other's sector mix narrows the gap sharply. That is why Siblis warns that comparing CAPE across markets is the weakest use of the measure: "a market heavy in banks and miners will read below one heavy in software whatever either is worth."
4. Accounting rules
The denominator is reported earnings, and the rules for reporting them have changed.
Two American accounting changes reshaped the number. SFAS 142, from 2001, stopped companies from amortizing goodwill and instead forced large write-downs when acquisitions soured — which concentrates enormous losses into downturn years. FAS 123R, from 2004, required companies to expense stock-based compensation. Both push reported earnings down relative to the cash the businesses actually generate.
Jeremy Siegel's research argues the effect is large: post-2001 GAAP earnings understate cash earnings by roughly 20–30%, and on operating earnings the U.S. CAPE would read materially lower — one widely cited application of his adjustment brings a Shiller reading in the low 40s down to about 32.
The practical point is about history, not just today. Comparing the current CAPE with its long-run average of about 17 means comparing two different accounting regimes. And across countries the problem compounds: U.S. GAAP and international IFRS rules treat items differently, so part of any cross-border gap is bookkeeping.
5. Buybacks
Buybacks lift earnings per share without lifting earnings.
CAPE divides price by earnings per share, and a company that retires its own stock spreads the same earnings over fewer shares. S&P 500 companies spent a record $1.02 trillion buying back their own stock in the twelve months to September 2025, and 17.1% of index companies cut their share count by at least 4% over that period.
The effect builds slowly — a decade of buybacks compounds into the denominator — but it means part of today's earnings-per-share growth is financial structure, not business growth. A market whose companies return cash as dividends instead shows the same total return differently, and its CAPE reads higher for the same economics. When you see a country's CAPE rising, it is worth asking how much of the earnings growth came from shrinking share counts.
6. Inflation history — and the ten-year window itself
High inflation distorts the measure, and the window it uses has a memory.
CAPE restates ten years of earnings in today's money. Where inflation has been high and volatile, that restatement is large and noisy — and investors demand a much higher return for holding equities in such a market, which pushes the price down and the CAPE with it. Turkey's 6.0 is the example at the bottom of the table: a very low CAPE that reflects very high required returns, not a bargain on sale.
The window itself also leaves fingerprints. The COVID earnings collapse of 2020 is still inside the ten-year average, depressing the denominator and lifting today's CAPE by an estimated 3 to 5 points; it rolls out of the window in 2030. A single bad year, six years ago, is still moving the number you read today.
And a market can look cheap against its own history for the wrong reason. Siblis notes that a very wide historical range usually means the market once had a bubble: China reached a CAPE of 94.4 at the end of 2007, so today's 16.0 looks low partly against a peak nobody expects to see again.
The same market, two different CAPEs
The number itself is a measurement choice.
Shiller's U.S. CAPE for October 2026 was 40.7. Siblis's U.S. CAPE for the same month was 35.1. Same market, same month, a gap of 5.6 points — because Siblis normalizes earnings by removing major purely accounting gains and losses, while Shiller uses reported earnings. Neither is wrong; they answer slightly different questions, and the gap between them is a reminder to check which version of CAPE any chart is showing before comparing it with another. For the context of that 40.7 reading, see S&P 500 CAPE Ratio Above 39: What the Index Returned the Last Time It Was Here.
Japan shows the same lesson across time. Its CAPE approached 100 in late 1989, the highest of any major market on record. Siblis puts Japan at 33.2 in October 2026 — still high by global standards, but a different world from the bubble peak, because the earnings, the rates, and the expectations behind the number all changed.
The October 2026 scorecard
Source: Siblis Research, CAPE Ratios by Country, October 2026. The pattern in the table follows the six forces: the top is dominated by markets priced for growth, the bottom by markets where investors demand high returns for risk, or where inflation distorts the measure. The U.S. is the interesting middle case — high, but flat over the year, while Taiwan climbed 13.3 points and India fell 7.2 points from 35.1 to 27.9.
How to compare two markets' CAPEs without fooling yourself
Three checks, in order.
- Check the sector mix first: If one index is 40% technology and the other is 25% banks, you are comparing two different businesses, not two prices. Compare sector by sector, or use a sector-adjusted measure.
- Check what is in the earnings: Reported or operating? Has a crash year — 2020, 2008 — sat inside the ten-year window? Has the country changed accounting rules? Any of these moves the denominator without a single business changing.
- Check the growth and the rate together: A high CAPE is a forecast, not a verdict. Ask what growth rate the price implies, and what the local real interest rate pays instead. If the safe asset pays more than the market's earnings yield, the market is asking you to believe in the growth.
For the crash-prediction question — whether a high CAPE actually warns of a fall — the record is mixed: How Accurate Is the CAPE Ratio at Predicting Stock Market Crashes? And for whether cheap-looking markets abroad are actually buyable through ETFs: Can You Pick Undervalued Markets With the CAPE Ratio?
Key Takeaway
A CAPE is not a price tag on a market. It is a price for expected growth minus the return investors demand for risk, filtered through whatever the index holds, whatever the accountants count, and whatever the last ten years did to earnings.
Taiwan at 43.4 and Turkey at 6.0 are not two ends of a cheap-to-expensive scale — they are two different answers to the same question: what growth, and what risk, is already in the price?
Sources:
- Siblis Research, "CAPE Ratios by Country," October 2026
- Robert Shiller, Yale University — U.S. stock market data (CAPE 40.7, October 2026; long-run average about 17; December 1999 peak 44.2)
- Federal Reserve, FRED series DFII10 — 10-year TIPS yield (2.91% at the November 13, 2026 close)
- S&P Dow Jones Indices, "S&P 500 Q3 2025 Buybacks," December 18, 2025
- Charles Schwab / S&P Dow Jones Indices — S&P 500 sector weights, September 30, 2026
- MSCI — MSCI EAFE Index Factsheet, October 30, 2026
- Bradford Cornell, "Around the World with CAPE"
- The Good Investors, "Understanding the Epic Bubble in Japanese Stocks" (Japan CAPE near 100, 1989)
- Jeremy Siegel's research on CAPE and accounting changes, as discussed in recent market commentary
- Dombey Capital, "What Is the CAPE Ratio?" (accounting and buyback effects)

Comments
Post a Comment