Shiller P/E (CAPE) Explained: The 10-Year Valuation Lens

The S&P 500 — anchored by Microsoft (MSFT), Apple (AAPL), and Nvidia (NVDA) — trades at roughly ~22x trailing earnings but at ~36x Robert Shiller's cyclically adjusted P/E (CAPE), a gap sitting near levels last seen in 1929 and 1999.

What Is the Shiller P/E and Why Does It Exist?

A price-to-earnings ratio that uses 10 years of inflation-adjusted earnings instead of just one. Economist Robert Shiller (who won the Nobel Prize in 2013) introduced the metric in his 1996 work with John Campbell, arguing that one-year earnings are too noisy to capture true valuation.

The intuition is simple. A single year of earnings is whatever happens to fall inside that fiscal window — a recession, a buyback boom, a one-time tax write-off, or an AI capex cycle. Ten years smooths all of that out. What you are left with is a closer approximation of "earnings power" through the cycle.

Shiller's original paper showed that the standard one-year P/E has almost no predictive power for the next decade of stock returns. CAPE has historically explained roughly 40% to 50% of the variance — still imperfect, but the best single ratio finance has found for index-level forecasting.

For investors building a long-term valuation framework, our fundamental analysis section covers how CAPE complements (rather than replaces) DCF and free cash flow yield.

How Do You Calculate CAPE?

Take real earnings per share for the last ten years, average them, and divide today's price by that average. The formula is:

CAPE = Current S&P 500 Price ÷ 10-Year Average of Inflation-Adjusted EPS

Each year of historical EPS is restated into today's dollars using CPI. That removes the inflation distortion that makes nominal earnings look bigger over time. Shiller publishes monthly CAPE data on his Yale website, going back to 1881 — a roughly 145-year series, which makes the metric one of the most studied in finance.

Worked example. If the S&P 500 sits at ~5,900 and the 10-year average of real EPS is ~$165, then CAPE = ~5,900 ÷ ~165 = ~35.8. The current reading sits very close to this level.

What CAPE Reading Counts as Expensive?

Anything above ~30 is historically expensive; anything below ~16 is historically cheap. The long-run median sits near ~17 — meaning today's reading near ~36 is roughly twice the historical norm.

The relationship is not deterministic. Markets have stayed above ~30 CAPE for years at a time — 1997 through 2001, 2017 through 2021, and most of 2024-2026. But the math eventually mean-reverts; the question is whether reversion happens through falling prices or rising earnings.

Why Does CAPE Disagree with Trailing P/E?

Because corporate earnings in 2024-2026 are running ~30%+ above their 10-year average. The trailing P/E uses peak earnings; CAPE uses cycle-average earnings. When earnings are near the top of the cycle, trailing P/E understates valuation and CAPE captures it more honestly.

Look at the S&P 500 today. Trailing P/E sits around ~22x, which most strategists call "fair to slightly expensive". CAPE sits near ~36x, which puts the market in the most expensive ~5% of all readings since 1881. That gap is not noise — it is two different views of "earnings".

Which view is right depends on what you believe about the next decade. If you think AI and pricing power push margins permanently higher, trailing P/E is the right lens. If you think margins mean-revert to history, CAPE is closer to truth. Most evidence sits in the middle — margins are structurally higher than 1990s averages, but not by enough to justify a CAPE near 36.

Real Examples: When CAPE Has Worked and Failed

Three episodes show both sides. CAPE called the dot-com top correctly — the metric hit ~44 in December 1999, and the subsequent ten years delivered roughly -1% annualized for the S&P 500. Anyone using CAPE in 1999 would have de-risked before the crash.

CAPE called the 2009 bottom correctly. The reading dropped to ~13 in March 2009, and the next ten years delivered roughly 13% annualized — exactly what the historical relationship predicted.

The metric was wrong in 2014-2021. The reading sat near ~25-30 for most of that period, suggesting modest forward returns. The actual S&P 500 delivered ~14% annualized — well above the historical mean. The reason was a structural drop in interest rates that lifted multiples. Critics argue this is exactly when CAPE breaks: regime-shift periods.

For investor strategies that use CAPE-like long-cycle thinking, see Howard Marks — his memos repeatedly stress thinking in cycles rather than current quarters.

What Are the Common Mistakes With CAPE?

Three errors keep tripping up retail investors. First, treating CAPE as a market-timing tool — it is not. CAPE tells you the probability distribution of long-run returns, not the timing of the next 12 months. A reading near ~36 means lower expected returns, not an imminent crash.

Second, applying CAPE to individual stocks. The metric was built for the S&P 500 index, where 500 companies smooth idiosyncratic noise. Applying CAPE to Nvidia (NVDA) or Tesla (TSLA) — high-growth names whose earnings 10 years ago bear no relationship to 2026 economics — produces garbage. Use DCF for those.

Third, ignoring accounting changes. Earnings rules shifted significantly in the late 1990s and again in 2001 (FAS 142 goodwill changes) and 2017 (revenue recognition). Jeremy Siegel and others argue CAPE comparisons across decades are not apples-to-apples. The fix is to look at CAPE relative to a more recent baseline (last 30 years) rather than the full 145-year series.

For more on valuation pitfalls, our investment strategies hub covers when to use CAPE versus DCF versus dividend discount models depending on the question.

When Should You NOT Use CAPE?

When you are evaluating individual high-growth stocks, when accounting standards have shifted, and when you need a timing call. CAPE is a long-cycle index metric. It will tell you that the next decade of S&P 500 returns is likely below average; it will not tell you whether to buy Microsoft (MSFT) or Apple (AAPL) today.

For single-stock decisions, ROIC, free cash flow yield, and forward DCF are far more informative. CAPE on Costco (COST) tells you almost nothing; ROIC on COST tells you whether the moat is intact.

The honest summary: CAPE is the best long-run valuation barometer the academy has produced. It is not a precision instrument. Use it to set expectations, not to set entry points. Pair it with bond yields, cyclically adjusted profit margins, and a clear-eyed look at what could drive a regime change.

Ready to analyze these stocks yourself? Search any ticker on MainRatios to see valuations from 6 legendary investors - free.