high accruals, defined as earnings minus actual cash flow, and as firms with
high net issuance of equity. Firms with high accruals may become overval-
ued if investors fail to understand that earnings are overstating actual cash
flows, and Chan et al. (2001) confirm that such firms indeed earn low re-
turns. Overvalued firms may also be identified through their opportunistic
issuance of equity, and we have already discussed the evidence that such
firms earn low long-run returns. Controlling for actual investment opportu-
nities as accurately as possible, Polk and Sapienza find that the firms they
identify as overvalued appear to invest more than other firms, suggesting
that sentiment does influence investment.
Further evidence of distortion comes from Baker, Stein, and Wurgler’s
(2003) test of the cross-sectional prediction that equity-dependent firms
will be more sensitive to stock price gyrations than will non-equity depend-
ent firms. They identify equity-dependent firms on the basis of their low
cash balances, among other measures, and find that these firms have an in-
vestment sensitivity to stock prices about three times as high as that of non-
equity dependent firms. This study therefore provides initial evidence that
for some firms at least, sentiment may distort investment, and that it does
so through the equity-dependence channel.
8.2. DividendsA major open question in corporate finance asks why firms pay dividends.
Historically, dividends have been taxed at a higher rate than capital gains.
This means that stockholders who pay taxes would always prefer that the
firm repurchase shares rather than pay a dividend. Since the tax exempt
shareholders would be indifferent between the dividend payment and the
share repurchase, the share repurchase is a Pareto improving action. Why
then, do investors seem perfectly happy to accept a substantial part of their
return in the form of dividends? Or, using behavioral language, why do
firms choose to frame part of their return as an explicit payment to stock-
holders, and in so doing, apparently make some of their shareholders worse
off?
Shefrin and Statman (1984) propose a number of behavioral explana-
tions for why investors exhibit a preference for dividends. Their first idea
relies on the notion of self-control. Many people exhibit self-control prob-
lems. On the one hand, we want to deny ourselves an indulgence, but on
the other hand, we quickly give in to temptation: today, we tell ourselves
that tomorrow we will not overeat, and yet, when tomorrow arrives, we
again eat too much. To deal with self-control problems, people often set
rules, such as “bank the wife’s salary, and only spend from the husband’s
paycheck.” Another very natural rule people might create to prevent them-
selves from overconsuming their wealth is “only consume the dividend, but
don’t touch the portfolio capital.” In other words, people may like dividends
A SURVEY OF BEHAVIORAL FINANCE 59