Third-Party Research & Methodology Only

This section shares summaries of third-party academic research and descriptions of quantitative models. The content represents the findings of the original researchers, not the opinions or recommendations of Foxholm Financial. Foxholm Financial does not publish hypothetical or backtested performance metrics on its quantitative research pages. All content is restricted to methodology, signal construction, factor logic, and risk architecture. SEC rules require that investment advisers not present misleading performance data, and our methodology-only approach reflects that standard and the firm's fiduciary obligations.

Contrarian Investment, Extrapolation, and Risk (1994)

Academic Research Review Value Investing Behavioral Finance

Robert Stowe, AAMS® Investment Advisor

This page reviews "Contrarian Investment, Extrapolation, and Risk," a 1994 paper by Josef Lakonishok, Andrei Shleifer, and Robert Vishny. The researchers found that "value stocks" (stocks that appear cheap based on measures like price-to-earnings or price-to-book) significantly outperformed "glamour stocks" (stocks that appear expensive on those same measures) over holding periods of 1–5 years.

Published in The Journal of Finance, 49(5), 1541–1578, the paper's key argument is that this outperformance comes from investor mistakes (extrapolating past growth too far into the future), not from value stocks being riskier. This was a direct challenge to the efficient market view that cheap stocks are cheap because they deserve to be. The paper remains one of the most cited works in the debate over why value investing works.

Key Findings

The paper's central contribution is a detailed, multi-measure demonstration that cheap stocks beat expensive stocks by a wide margin, and that this gap is better explained by investor psychology than by differences in risk. The researchers used several different ways to measure "cheapness" and found the same pattern every time.

Value Beats Glamour

The researchers sorted stocks into five equal groups (quintiles) based on four different valuation measures: price-to-book, price-to-earnings, price-to-cash-flow, and past sales growth. In every case, the cheapest quintile significantly outperformed the most expensive. The difference was roughly 10–11 percentage points per year over 5-year holding periods.

This is not a small edge. An annual gap of that size, compounded over several years, produces enormous differences in total wealth. The consistency across four independent measures of cheapness makes it difficult to dismiss as a statistical accident. When multiple ways of identifying "cheap" stocks all point to the same conclusion, the pattern is more likely to be real.

The Extrapolation Error

Glamour stocks had strong past earnings growth that investors expected to continue. Value stocks had poor past growth that investors expected to persist. In reality, the growth rates converged: glamour stock growth slowed and value stock growth recovered. Investors who extrapolated past trends into the future systematically overpaid for glamour and underpaid for value.

The researchers documented this directly by tracking what happened to earnings growth after the stocks were sorted into groups. Glamour stocks, which had grown earnings quickly over the prior five years, saw their growth rates decline toward the market average. Value stocks, which had seen weak or negative earnings growth, saw their growth rates improve. The market priced these stocks as if their recent trends would continue indefinitely, but reversion to the mean was the more common outcome.

Multi-Dimensional Value

Using multiple value measures together produced even stronger results than any single measure alone. For example, combining low price-to-book with low past sales growth identified a group of stocks that outperformed by an even wider margin than sorting on either measure individually.

The researchers found that using multiple value measures identified a group of stocks that outperformed by a wider margin than sorting on either measure individually. A stock that is cheap on price-to-book but has strong recent sales growth might be cheap for a temporary reason. A stock that is cheap on multiple measures is more likely to be genuinely undervalued by the market.

Is Value Riskier?

The most contested part of the paper is its direct test and rejection of the risk-based explanation for value outperformance. At the time, the prevailing academic view held that if cheap stocks earned higher returns, it must be because they carried higher risk. The researchers challenged this interpretation head-on.

Performance During Bad Times

If value stocks are riskier, they should perform especially poorly during recessions or market downturns, when investors need their money most. The researchers tested this directly and found that value stocks did not perform worse during the 25 worst months for the overall stock market. In fact, value stocks held up at least as well as glamour stocks during these periods.

This finding undercuts the core logic of the risk explanation. If value stocks do not let investors down during the worst economic environments, it is hard to argue that their higher returns are compensation for bearing that kind of risk. The researchers also examined whether value stocks were more volatile (had bigger price swings) than glamour stocks. They were not.

A Behavioral Anomaly

The researchers argue that the value premium is a behavioral anomaly driven by investor mistakes rather than fair compensation for risk. The mechanism is straightforward: investors look at a company's recent track record and assume it will continue. Companies with impressive past growth get bid up to prices that are hard to justify unless that growth persists. Companies with poor past growth get sold down to prices that assume things will never improve.

This was controversial because it challenged the interpretation offered by Eugene Fama and Kenneth French, who had documented similar value premiums but attributed them to risk. The debate between "value works because of risk" and "value works because of investor mistakes" remains one of the central questions in academic finance. Lakonishok, Shleifer, and Vishny landed firmly on the side of mistakes.

Practical Implications

Why This Matters

If value outperformance comes from behavioral mistakes rather than risk, it may persist as long as investors keep making those mistakes. Risk-based premiums can shrink as more investors discover them and bid up the prices of risky assets. But behavioral premiums can survive discovery because they are rooted in human psychology: the tendency to chase winners, avoid losers, and extrapolate recent trends.

This distinction also affects how investors should think about value strategies during long stretches of underperformance. If the premium is compensation for risk, a prolonged drawdown might mean the risk has increased and the strategy is working as expected. If the premium is behavioral, a prolonged drawdown likely means the mispricing has gotten larger, and the expected future payoff has actually increased.

Contrarian Discipline

Successfully investing in value stocks requires buying what other investors are selling and avoiding popular "story stocks." This is psychologically difficult. The stocks that score well on value measures tend to be companies with bad headlines, slowing businesses, or unfashionable industries. Buying them means acting against the prevailing market narrative.

The paper's findings imply that this psychological difficulty is precisely what allows the premium to persist. If buying cheap, unpopular stocks were emotionally easy, more investors would do it, and the mispricing would shrink. The discomfort is, in a sense, the price of admission to the extra returns.

Time Horizon Matters

The researchers found that the value premium was larger over longer holding periods. Over one year, the difference between value and glamour was meaningful but moderate. Over three to five years, it was substantial. Value strategies can underperform over short periods, and impatient investors who abandon the strategy before it pays off miss the premium entirely.

The researchers observed that the value premium was more pronounced over longer 3-to-5-year horizons. Evaluating a value approach over a 6-month or 12-month window is likely to produce misleading conclusions. The original research measured results over 1, 2, 3, and 5 year horizons, and the case for value strengthened with each extension.

Growth Rate Reversion

The finding that earnings growth rates revert to the mean is one of the most practically useful takeaways from the paper. Companies growing at 30% per year rarely sustain that pace; investors who pay for perpetual high growth tend to be disappointed. Conversely, companies with declining earnings often stabilize and recover, rewarding investors who bought at depressed prices.

This pattern of mean reversion in corporate fundamentals has been confirmed by subsequent research across many markets and time periods. It provides a simple but powerful mental model: be skeptical when the market prices a company as if its recent trend will last forever, whether that trend is positive or negative.

How the Researchers Tested This

Data and Time Period

The study used all stocks listed on the New York Stock Exchange (NYSE) and American Stock Exchange (AMEX) from 1963 to 1990. This provided a broad cross-section of U.S. companies across nearly three decades, covering diverse market environments including the inflationary 1970s, the bull market of the 1980s, and the 1987 crash.

How the Portfolios Were Built

The researchers sorted stocks into quintiles (five equal groups) based on each fundamental ratio. The top quintile represented the cheapest stocks (value), and the bottom quintile represented the most expensive (glamour). They then tracked the performance of each quintile over 1, 2, 3, and 5 year holding periods following the sorting date.

The researchers also tested two-dimensional sorts, grouping stocks by combinations of value measures (for example, price-to-book and past sales growth together). This allowed them to identify whether using multiple signals improved the reliability of the value effect. They specifically tested whether value stocks underperformed during bad economic times by examining returns during the worst market months and during NBER-designated recession periods.

Limitations and Caveats

Limitations to Consider

U.S.-only data: The study covers only NYSE and AMEX stocks. Later research by Fama and French (1998) and others confirmed the value effect internationally, but the original paper's evidence is limited to one country.
Trading costs not included: The paper reports returns before transaction costs. Value stocks tend to be smaller and less liquid, which means buying and selling them in large quantities can be expensive. Real-world returns would be lower than the paper reports.
Liquidity constraints: Value stocks are often smaller companies with lower trading volumes. Institutional investors managing large portfolios may find it difficult to build meaningful positions without moving the stock price against them.
The value trap problem: Some cheap stocks are cheap for good reasons. A company with a low price-to-book ratio might be facing permanent business decline rather than temporary undervaluation. The paper's quintile approach averages across many stocks, masking the losses from individual value traps.
Period-specific results: The study covers 1963–1990. The post-2020 environment has shown extended periods of value underperformance, raising questions about whether the premium's magnitude has changed as more investors have become aware of it.