Lesson 7 of 8
The behavioural answer
Contrarian portfolios beat glamour ones in Lakonishok, Shleifer and Vishny's sample, and they read it as a mistake exploited rather than a risk paid for.
By Javier Sanz
Founder of ValueScreener. Built Ninety Nine, a securities services provider, and sold it to a tier-one global bank; ran operations and internationalization at Alpaca, the brokerage API.
Who writes thisEditorial processLinkedIn (opens in a new tab)X (opens in a new tab)
In one minute
- Lakonishok, Shleifer and Vishny examined New York and American Stock Exchange companies from 1968 to 1990 on price to book, price to earnings and price to cash flow.
- Their abstract states the reading plainly: the returns come from exploiting the mistakes of the typical investor rather than from bearing more fundamental risk.
- The reading has a cost, and it is the honest limit of this lesson: a premium explained by a recurring mistake lasts exactly as long as the mistake does.
The efficient-market position can accept a measured premium and call it payment for risk. The behavioural reply accepts the same measurement and gives it a different cause: investors extrapolate recent performance too far, which makes recent disappointments too cheap and recent successes too dear.
Two readings of one measurement
Payment for risk
- The cheap group carries a risk the measurement did not capture.
- The premium is a wage for bearing it, not an error to be corrected.
- The premium should persist, because the risk does.
A mistake being exploited
- Investors extrapolate recent results too far in both directions.
- The premium is the correction of that extrapolation.
- It persists only while the mistake does, which is a weaker promise.
Both readings accept that the return difference was measured. They disagree about what produced it, and the disagreement is testable in principle.
What the study did
The paper, funded by the National Bureau of Economic Research and circulated in May 1993 before publication in the Journal of Finance in 1994, examined returns from all companies listed on the New York Stock Exchange and the American Stock Exchange between 1968 and 1990. It sorted them on price to book value, price to earnings and price to cash flow.
The authors state their conclusion in the abstract: value strategies produced higher returns because they exploit the mistakes of the typical investor, and not because the strategies are fundamentally riskier.
The mechanism they propose
The mistake is extrapolation. A company that has grown fast is assumed to keep growing fast; a company that has disappointed is assumed to keep disappointing. Prices are then set on those extended paths rather than on the rate at which growth actually reverts.
When the extension proves too long in either direction, the correction runs in the opposite direction to the price that was paid. Buying the group that was extrapolated downwards collects the correction.
Why this is a weaker claim than it sounds
A risk explanation makes a strong promise: the premium persists because the risk does. A mistake explanation makes a weaker one: the premium persists while enough investors keep making the mistake.
That weakness is the honest content of this lesson. If the cause of a return difference is a repeated error, then anything that changes how many people make the error changes the return difference. Publication is one such thing. So are cheaper data, faster screens and a larger number of people running the same sort.
The instrument that came from the same instinct
Shiller's cyclically adjusted price to earnings ratio, developed with John Campbell, divides price by a ten-year average of inflation-adjusted earnings. The reason is the same one Graham gave in chapter 12: a single year's earnings carries the cycle and the accounting choices of that year, and a multiple built on it mostly measures where in the cycle you are standing.
Check yourself
What did Lakonishok, Shleifer and Vishny state as the cause of the returns they measured?
- Compensation for greater fundamental risk
- The exploitation of mistakes made by the typical investor
- Differences in trading costs between the groups
Show the answer
Answer: B.
Their abstract says the strategies yield higher returns because they exploit the mistakes of the typical investor, and not because they are fundamentally riskier.Why is a behavioural explanation a weaker promise than a risk explanation?
- Because behavioural studies use smaller samples
- Because a premium caused by a repeated mistake lasts only as long as enough investors keep making it
- Because it cannot be measured
Show the answer
Answer: B.
A risk keeps existing. A mistake can stop being made, and anything that changes how widely it is made changes the premium.Why does a cyclically adjusted multiple average ten years of earnings?
- To include more companies in the sample
- Because one year's earnings carries the cycle and that year's accounting choices, so a multiple on it partly measures where in the cycle you are
- Because ten years is the longest history available
Show the answer
Answer: B.
Averaging reduces the weight of a single year's position in the cycle and of the choices made in reporting it.Sources
- Lakonishok, Shleifer and Vishny, "Contrarian Investment, Extrapolation and Risk," National Bureau of Economic Research, May 1993, published in the Journal of Finance, 1994: NYSE and AMEX companies, 1968 to 1990, sorted on price to book, price to earnings and price to cash flow.
- Basu, Journal of Finance, June 1977, for the measurement both readings are trying to explain.
- Shiller, Irrational Exuberance, 2000, and the cyclically adjusted price to earnings ratio developed with John Campbell.
What this leaves out
- The study covers two US exchanges between 1968 and 1990. Nothing in it describes a later period, and nothing in this lesson claims it does.
- The abstract's conclusion is the authors' interpretation of their own sample. This lesson reports the interpretation and the objection to it, and settles neither.
- The worked example uses invented round numbers and is labelled illustrative.
- Nothing here is advice about any security.