III. Empirical Studies of Overreaction
and UnderreactionThe third section contains four important empirical papers. Chapter 8, by
Josef Lakonishok, Andrei Shleifer, and Robert Vishny (LSV), offer, the de-
finitive study of value investing, a theme of behavioral finance since Ben-
jamin Graham. De Bondt and Thaler (1985, 1987) had shown that simple
value strategies, such as buying stocks that had greatly underperformed in
the market for 3–5 years, or had extremely low ratios of price to book
value, earn excess returns. We hypothesized that these excess returns
could be attributed to investor overreaction. Efficient market adherents,
such as Eugene Fama and Kenneth French conducted a series of studies
that confirmed the basic facts of the De Bondt and Thaler findings, but
offered a different interpretation, namely that the apparent excess returns
to value stocks were attributable to risk. In this chapter LSV respond to
the Fama-French challenge in various ways. Specifically, they show that
value stocks do not appear to be riskier than growth stocks (e.g., do not
underperform in down markets or recessions). They also find that value
stocks do particularly well in the days around earnings announcements,
suggesting that investors had biased expectations of earnings, as the be-
havioral theory predicts.
The question of whether the returns to value strategies can be attributed
to risk is also addressed by Kent Daniel and Sheridan Titman in Chapter 9.
The heart of their paper is an investigation of whether there is a “distress
factor” shared by value firms that can explain their excess returns. Daniel
and Titman reject this interpretation. Instead, they find that value stocks
tend to move together because stocks with similar factor loadings tend to
become distressed (i.e., cheap) at the same time.
Value strategies are long-term phenomena. Stocks that have done well or
poorly over a period of years subsequently show mean-reverting returns.
But over shorter horizons, such as six months or a year, just the opposite is
true. That is, the best performers over the past year tend to keep outper-
forming over the subsequent year. Over the past decade much attention has
been given to this phenomenon, that has come to be known as momentum.
Narasimhan Jegadeesh and Sheridan Titman summarize the research in this
area in Chapter 10. They find the evidence supporting the existence of ex-
cess returns to be quite strong, and judge that it is very difficult to claim
that these excess returns can be attributed to risk.
The final empirical paper in this section addresses a topic of growing con-
cern, namely the role of brokerage stock recommendations. Roni Michaely
and Kent Womack survey this important field in Chapter 11. Specifically
they cover two important topics: first, do recommendations have informa-
tion content? (Yes); second, is there evidence of biases stemming from either
cognitive biases or from conflicts of interests? (Yes and yes).
xiv THALER