Why do valuations matter more for long-term returns than short-term?

The short answer

A common argument against valuation-based reasoning is that valuation measures have weak predictive power. The argument typically cites short-window studies: the R-squared between starting CAPE and one-year forward returns is around 0.05, indicating valuation explains only 5% of return variation over twelve months.

The complication is that this is the wrong horizon. Valuation reasoning is not designed to predict next year — it is designed to anchor multi-year return expectations. When the same regression is run at ten-year horizons, the R-squared rises to approximately 0.50: starting valuations explain roughly half of the variation in subsequent ten-year real returns.

The systematic increase in predictive power with horizon is itself the central result. It means that valuation matters more, not less, for long-term planning — exactly the opposite of how the framework is sometimes characterised.

→ New to long-term return forecasting? Financial education framework

What the data shows

Robert Shiller’s online database provides the canonical evidence. Regressing subsequent ten-year real returns on starting CAPE for the post-1880 period produces an R-squared of approximately 0.40-0.50 depending on subperiod and methodology. Comparable analyses on shorter windows produce much lower R-squared values, declining toward zero at sub-annual horizons.

Three regime-specific data points illustrate the mechanism. After 1929, with CAPE at 32, subsequent ten-year real returns were significantly negative — the market did not recover its pre-crash real level until the 1950s. After 1982, with CAPE at 7, subsequent ten-year real returns were approximately 14% annualised — among the strongest decadal returns in the historical record. After 2000, with CAPE at 44, subsequent ten-year real returns were approximately -3% annualised — and the S&P 500 did not durably exceed its 2000 nominal peak until 2013.

These are not isolated cases. The pattern repeats across the historical record: high starting valuations are followed by below-average returns, and low starting valuations are followed by above-average returns, over decade-length horizons. Variability around this central tendency is real, but the tendency itself is robust.

→ Shiller CAPE: monthly history (dataset)

Why it happens — the macro mechanism

The mathematics is unavoidable. Total return on equities can be decomposed into three components: dividend yield, earnings growth, and change in valuation multiple. Over long horizons, these components determine the return identity. When starting CAPE is at 40, the dividend yield component is mechanically low (around 1.5%). For total returns to match historical averages of 7% real, either earnings growth must accelerate to multi-percent above trend, or the multiple must remain elevated indefinitely.

Robert Shiller, in Irrational Exuberance, made the case that this decomposition is the core of long-term return analysis. Aswath Damodaran has formalised the relationship in his implied equity risk premium work, showing that current prices can be solved for the implied long-term return given assumed cash flow growth.

Three regime patterns illustrate the framework. In the post-1929 regime, multiple compression dominated returns for two decades — the price recovery preceded the multiple recovery. In the post-1982 regime, multiple expansion contributed roughly half of total returns, with the rest from earnings growth and dividends. In the post-2000 regime, multiple compression offset earnings growth for the first decade, producing flat real returns despite earnings expansion.

The signature of the mechanism is that horizon length determines whether the components can balance. Over one year, multiple compression can be entirely overwhelmed by sentiment shifts. Over ten years, the underlying mathematics dominates.

What it means for different economic actors

For institutional asset managers, the long-horizon predictive power of valuation is the foundation for capital market assumptions. Major firms produce ten-year return projections that embed valuation starting points; current US equity assumptions cluster in the 4-6% nominal range, well below historical averages.

For retail investors, the practical consequence is that savings rate decisions matter more from elevated valuations. If forward ten-year returns are likely to be lower than historical averages, accumulating the same target wealth requires either higher savings or longer horizons.

For pension funds and endowments, low expected returns from current valuations create real solvency challenges. Funded ratios assume return assumptions that may overstate forward returns.

For households accumulating wealth, the relevant observation is that the valuation question is about expectations, not timing. This page is itself the core summary.

Frequently asked questions

If valuation has weak short-term power, why does it matter for long-term?

The mathematics of return decomposition explains this. Over short horizons, sentiment shifts and flow dynamics dominate the price path — these can move multiples sharply in either direction regardless of starting valuation. Over long horizons, the components of total return must balance: cash flows are bounded by economic reality, and starting prices set the implicit assumption about how those cash flows will be valued. The longer the horizon, the more the mathematics asserts itself over the noise.

What about the “this time is different” argument?

This argument has merit in some periods. Structural changes — technology, demographics, monetary regimes — can shift the level of equilibrium valuations. But the historical record shows that “this time is different” arguments have been advanced in every late-cycle period, and most have eventually proved partially correct (some structural shift was real) and partially incorrect (the magnitude was overstated). The evidence does not support either complete continuity or complete discontinuity with historical patterns.

Should I sell stocks because valuations are elevated?

This page does not provide investment advice. The valuation framework provides information about expected returns over multi-year horizons, not about market timing. Many investors who have remained allocated to equities through elevated valuation periods have done well over very long horizons; others have benefited from valuation-conscious allocation adjustments. The right answer depends on individual circumstances, time horizon, and tolerance for the variability around central tendencies.

Last updated — 12 June 2026