Shiller (1981) - Do Stock Prices Move Too Much to Be Justified by Subsequent Changes in Dividends?

Overview

Shiller (1981) challenges the efficient market hypothesis by showing stock prices are 5-13 times more volatile than can be justified by subsequent dividends. The paper launched the excess volatility literature and remains the cleanest single attack on the idea that prices are rational forecasts of fundamentals.

The Variance Bounds Test

If prices equal the present value of expected future dividends, the price series must be smoother than the realized present value of those dividends — a rational forecast cannot vary more than the thing it forecasts. Shiller's innovation was to test this inequality directly: compute the ex-post rational price (the present value of actual subsequent dividends, with a terminal value), then compare its variance to the variance of the observed price series.

The Result

The observed price index was several times more volatile than the ex-post rational price across the sample — the paper's headline factor of 5 to 13 depending on the discount rate and sample. Since the ex-post rational price already contains all the information a rational market could have used, prices that move far more than it must reflect something other than information about dividends: shifting sentiment, fads, or self-referential dynamics.

The Debate It Provoked

Econometricians immediately challenged the test's statistical foundations — small-sample bias, unit-root issues, and the choice of terminal value and discount rate. Shiller's numbers absorbed some of the criticism but the core finding survived in the robust form: aggregate stock prices move far more than dividend fundamentals can justify. The debate produced better tests (including Shiller's own later work) and better data, and the excess volatility result kept surviving.

Why It Matters

Excess volatility is the empirical cornerstone of behavioral finance: if aggregate prices are not rational forecasts, then the entire classical edifice built on rational pricing needs an alternative. For practitioners it is the theoretical license to think about sentiment and crowding in market timing, and a warning against over-interpreting price moves as news.

Key Takeaways

• Prices moved 5-13x more than ex-post fundamentals in the original sample — far beyond rational-forecast bounds.
• The test's critics were partly right and wholly unsuccessful: robustness refinements preserved the anomaly.
• Aggregate volatility is the anomaly; sector and firm-level noise cannot explain market-wide swings.