other, they perform poorly.^25 A challenge to both behavioral and rational
approaches is to explain why extending the formation period switches the
result in this way.
There is some evidence that tax-loss selling creates seasonal variation in the
momentum effect. Stocks with poor performance during the year may later be
subject to selling by investors keen to realize losses that can offset capital
gains elsewhere. This selling pressure means that prior losers continue to lose,
enhancing the momentum effect. At the turn of the year, though, the selling
pressure eases off, allowing prior losers to rebound and weakening the mo-
mentum effect. A careful analysis by Grinblatt and Moskowitz (1999) finds
that on net, tax-loss selling may explain part of the momentum effect, but by
no means all of it. In any case, while selling a stock for tax purposes is ra-
tional, a model of predictable price movements based on such behavior is not.
Roll (1983) calls such explanations “stupid” since investors would have to be
stupid not to buy in December if prices were going to increase in January.
A number of studies have examined stock returns following important cor-
porate announcements, a type of analysis known as an event study.
Event Studies of Earnings Announcements. Every quarter from 1974 to
1986, Bernard and Thomas (1989) group all stocks traded on the NYSE and
AMEX into deciles based on the size of the surprise in their most recent
earnings announcement. “Surprise” is measured relative to a simple random
walk model of earnings. They find that on average, over the sixty days after
the earnings announcement, the decile of stocks with surprisingly good news
outperforms the decile with surprisingly bad news by an average of about
4 percent, a phenomenon known as post-earnings announcement drift. Once
again, this difference in returns is not explained by differences in beta be-
tween the two portfolios. A later study by Chan, Jegadeesh, and Lakonishok
(1996) measures surprise in other ways—relative to analyst expectations, and
by the stock price reaction to the news—and obtains similar results.^26
Event Studies of Dividend Initiations and Omissions. Michaely, Thaler,
and Womack (1995) study firms that announced initiation or omission of a
dividend payment between 1964 and 1988. They find that on average, the
shares of firms initiating (omitting) dividends significantly outperform (un-
derperform) the market portfolio over the year after the announcement.
A SURVEY OF BEHAVIORAL FINANCE 37(^25) In fact, De Bondt and Thaler (1985) also report that one-year big winners outperform
one-year big losers over the following year, but do not make much of this finding.
(^26) Vuolteenaho (2002) combines a clean-surplus accounting version of the present value
formula with Campbell’s (1991) log-linear decomposion of returns to estimate a measure of
cash-flow news that is potentially more accurate than earnings announcements. Analogous to
the post-earnings announcement studies, he finds that stocks with good cash-flow news subse-
quently have higher average returns than stocks with disappointing cash-flow news.