Topic: Investment · Type: Timely · Reading time: ~6 min

The US stock market is sitting at its second-most expensive valuation in 154 years of recorded data. Not second this decade. Second in a century and a half. The only time it was pricier was December 1999 — three months before the dot-com crash. And yet most of the advice you'll read is some version of "stay the course." Both of those things can be true. Here's how to hold them together without losing your mind.

What the numbers actually say

The most widely cited long-term valuation tool is the Shiller CAPE ratio — the Cyclically Adjusted Price-to-Earnings ratio developed by Nobel laureate economist Robert Shiller. Unlike the standard P/E ratio, which measures price against a single year of earnings, CAPE smooths earnings over the past decade and adjusts them for inflation, filtering out the noise of one-off booms and recessions.

As of early 2026, the Shiller CAPE for the S&P 500 sits at approximately 41 — more than double its long-run historical median of around 16. The only period it's been higher was the dot-com mania of the late 1990s, when it peaked at 44.

The Buffett Indicator tells a similar story. This metric — named for Warren Buffett, who called it "probably the best single measure of where valuations stand at any given moment" — compares the total market capitalisation of US equities to US GDP. A reading above 100% is generally considered overvalued. The current reading is approximately 215%. Buffett himself once said that when it gets above 200%, "you are playing with fire."

The picture doesn't improve when you look elsewhere. The S&P 500's forward price-to-earnings ratio sits around 23 times expected earnings — above its 5-year average of 19.9 and its 10-year average of 18.6. Notably, 10 out of 11 S&P 500 sectors are currently trading above their 25-year average forward P/E. This isn't a tech-sector distortion. It's the whole market.

Worth knowing: At current CAPE levels, historical data suggests probability-weighted 10-year real returns of roughly 0–5% annually — compared to the long-term average of around 7%. That's not a crash forecast. It's a compressed return forecast.

Why "overvalued" doesn't mean "about to crash"

Here's the uncomfortable truth that most alarming headlines omit: valuation is a terrible short-term timing tool. The CAPE ratio spent most of the period from 2015 to 2025 well above its historical average — and the market delivered strong returns throughout. The S&P 500 returned 26% in 2023, 25% in 2024, and continued rising into 2025. Being "right" about overvaluation a decade early is indistinguishable from being wrong.

What CAPE does do well is predict 10- to 20-year returns. The relationship between starting valuation and decade-long outcomes has an R-squared of 0.78 in Invesco's analysis — meaning high starting valuations reliably predict lower long-run returns, even if they tell you nothing about what happens next year.

The critics of CAPE have a point, too. Today's S&P 500 is dominated by asset-light, high-margin technology companies whose earnings power genuinely looks different from the industrial economy the ratio was originally calibrated against. Accounting changes since the 1990s also tend to make earnings look lower during downturns, mechanically inflating the ratio. And share buybacks — now far more common than dividends — boost earnings per share without necessarily reflecting improved business results.

Wharton professor Jeremy Siegel has argued for years that the "new normal" CAPE should be around 25–30, not the historical 17. Even if he's right, a current reading of 41 is still elevated against his adjusted benchmark.

If you strip the Magnificent Seven technology companies entirely from the CAPE calculation, the ratio falls from around 41 to roughly 33. Still historically expensive — but not dot-com-bubble territory.

The concentration problem nobody's talking about enough

The headline valuation numbers obscure a more specific concern: the S&P 500 is more top-heavy than at almost any point in its history. The ten largest companies in the index now account for over 40% of its total market capitalisation — a level not seen since 1932. Seven of those companies are technology names directly tied to the AI investment cycle.

This matters because index fund investors — the vast majority of passive investors — own all of it. When you buy a standard S&P 500 tracker, roughly a third of your money goes into seven stocks. If those seven disappoint on earnings, or if AI capital expenditure slows (over $500 billion was spent on data centres in 2025 alone), the drawdown in a supposedly "diversified" index could be severe and concentrated.

If you've been quietly reassured that your globally diversified index fund protects you from this, check the weighting. MSCI World gives US equities around 65–70% of its allocation, and the US exposure is itself heavily concentrated in mega-cap tech. Global diversification helps — but only if you're actually diversified, not just holding a fund that's 65% US large-cap technology with a different label.

Where the value actually is

The US market isn't the global market. And right now, the divergence in valuations between US equities and the rest of the world is historically wide.

International equities outperformed US stocks in 2025 — non-US stocks returned roughly 30% for the year as of mid-December, significantly ahead of the S&P 500 — and the gap in valuations remains significant. The Vanguard Total International Stock ETF, which holds developed and emerging market equities, trades at a price-to-earnings ratio of around 17 — approximately 40% cheaper than the S&P 500 on that measure.

European equities in particular look cheap relative to history. Chinese and Hong Kong markets, despite a sharp rally in valuations through 2025, remain among the lowest CAPE-ratio markets globally. These aren't risk-free alternatives — geopolitical uncertainty, currency risk, and structural economic headwinds are real. But for investors who haven't looked at international allocation recently, the case for rebalancing has rarely been clearer.

Small-cap US stocks also look more attractive than large caps on a relative basis. The Russell 2000 currently trades at a nearly 10% discount to the S&P 500 — an unusual inversion compared to the historical norm, where small caps typically command a premium.

This is worth understanding if you're thinking about how to build a diversified portfolio from scratch: the composition of your equity exposure matters as much as the overall allocation.

What this means for how you invest

None of this is an argument to sell your equities and wait. Market timing is, with very few exceptions, a way to crystallise losses and miss recoveries. The investor who pulled out of US equities in 2018 — when CAPE was also elevated — missed years of strong returns.

What it is an argument for:

Temper your return expectations. If you've been mentally projecting 8–10% annual returns on your portfolio indefinitely, the data says to revise that down for the next decade. Planning around 4–5% real returns from US equities is more defensible at current valuations.

Check your geographic exposure. If your equity holdings are mostly S&P 500 trackers, you have concentrated exposure to expensive US large-cap technology. That's not inherently wrong — but it should be a deliberate choice, not an accidental one. A globally diversified fund with meaningful ex-US exposure has meaningfully lower starting valuations.

Don't stop your regular contributions. Dollar-cost averaging into an overvalued market is not a mistake. It means you'll buy fewer shares at today's prices, and more shares if prices fall. The case for dollar-cost averaging doesn't depend on the market being cheap — it depends on you not being able to predict when it will be.

Don't let this stop you from starting. The single biggest cost for most investors isn't buying into an expensive market — it's staying in cash waiting for the right moment that never arrives.

The honest verdict

Is the stock market overvalued? By the most robust long-term measures available, yes — significantly so. Does that mean a crash is imminent? No, and anyone telling you otherwise is forecasting, not analysing.

What high valuations reliably predict is lower future returns over a 10-year horizon, not a specific correction at a specific time. The practical implication isn't to exit the market. It's to enter it with realistic expectations, ensure your geographic and sector exposure is intentional rather than accidental, and understand that the next decade of US equity returns is unlikely to look like the last one.

The data doesn't tell you when to act. It tells you what you're paying for. Right now, you're paying a lot.