Lakonishok, Shleifer & Vishny (1994) - Contrarian Investment, Extrapolation, and Risk
Key Insights
- LSV show contrarian strategies outperform because investors extrapolate past growth too far, not because value strategies are riskier.
- Value stocks do NOT have higher risk in bad times.
Edit on GitHub — registry.json
Overview
Lakonishok, Shleifer and Vishny (1994) investigate why contrarian value strategies outperform, showing the premium arises from investor extrapolation bias rather than higher risk. Their answer to the value puzzle is behavioral: the premium is compensation for a mistake, not for bearing danger.
The Debate They Entered
By 1994, the value effect was well documented but contested on mechanism. Fama and French's reading treated value stocks as fundamentally riskier and HML as a risk factor. LSV attacked exactly this claim: if value is risk compensation, value stocks should earn their excess returns by performing badly in bad times — precisely when risk shows up.
Method
Using NYSE/AMEX data from 1968 to 1989, they form portfolios on price-to-earnings, price-to-book, and past sales growth, separating "value" stocks (cheap, low past growth) from "glamour" stocks (expensive, high past growth). They compare returns over the next five years and, critically, decompose performance into good times and bad times across the economic cycle.
The Findings
Contrarian portfolios beat glamour portfolios by substantial margins over the sample. The decisive result is the risk test: value stocks did not underperform glamour stocks in bad economic states — if anything, glamour stocks were the ones that disappointed. This contradicts the risk-compensation story: a factor that earns high average returns without producing losses in downturns cannot be priced as a risk premium in the CAPM sense.
Why It Matters
The paper's extrapolation explanation — investors project recent growth too far into the future, overpaying for glamour and abandoning value — became the standard behavioral account of value investing. It grounds the practitioner discipline of rebalancing into out-of-favor names and justifies treating value as a behavioral bet with a defined time horizon rather than a free lunch.
Key Takeaways
- Extrapolation bias, not risk, explains the value premium — growth expectations overshoot reality.
- Value strategies survived the bad-times test in the 1968-1989 sample, undercutting the risk story.
- Expect multi-year horizons: the premium pays off over five-year windows, which is why it persists.