Which markets are cheap?

CAPE research scenarios

CAPE observations: September 30, 2026. This is the separate CAPE research model, not the current benchmark equity models used in Assumptions.

A CAPE ratio on its own is a fact about price. The useful question is what it implies for the next decade, which depends on where the multiple sits relative to its own history, what earnings are expected to do, and what the market pays out along the way. Running each market through those three inputs over a 10-year horizon gives the figures below.

The spread is the point. US Large sits at the 99th percentile of its own history and implies 0.72% a year after inflation. Hong Kong implies 6.92%. The difference reflects this model's valuation, growth and income assumptions.

Annualized over 10 years. Real returns are after inflation; nominal adds US expected inflation so the figures sit on the same footing as the optimizer. Percentile is where each market’s CAPE falls within its own history rather than against other markets, so a high number means expensive for that market specifically. Starting valuation has historically explained roughly half the variation in ten-year returns, which leaves the other half to everything else. These are expectations over a decade, not a forecast for any single year, and a market can stay expensive for a very long time. Rows marked * use the closest available index rather than an exact match.