policy. Based on these interviews Lintner proposed what we would now call
a behavioral model. In his model, firms first establish a target dividend pay-
out rate based on notions of fairness, in other words, on what portion of
the earnings it is fair to return to shareholders. Then, as earnings increase
and the dividend payout ratio falls below the target level, firms increase
dividends only when they are confident that they will not have to reduce
them in the future.
There are several behavioral aspects to this model. First, the firm is not
setting the dividend to maximize firm value or shareholder after-tax wealth.
Second, perceptions of fairness are used to set the target payout rate. Third,
the asymmetry between an increase in dividends and a decrease is explicitly
considered. Although fewer firms now decide to start paying dividends, for
those that do Lintner’s model appears to be valid to this day (Benartzi,
Michaely, and Thaler 1997; Fama and French 2001).
Baker and Wurgler (2004) argue that changes in dividend policy may
also reflect changing investor sentiment about dividend-paying firms rela-
tive to their sentiment about nonpaying firms. They argue that for some
investors, dividend-paying firms and nonpaying firms represent salient cat-
egories and that these investors exhibit changing sentiment about the cate-
gories. For instance, when investors become more risk averse, they may
prefer dividend-paying stocks because of a confused notion that these firms
are less risky (the well-known “bird in the hand” fallacy). If managers are
interested in maximizing short-run value, perhaps because it is linked to
their compensation, they may be tempted to change their dividend policy in
the direction favored by investors.
Baker and Wurgler find some supportive evidence for their theory. They
measure relative investor sentiment about dividend-paying firms as the log
B/M ratio of paying firms minus the log B/M ratio of nonpaying firms, and
find that in the time series, a high value of this measure one year predicts
that in the following year, a higher fraction of nonpaying firms initiate a
dividend and a larger fraction of newly-listed firms choose to pay one. Sim-
ilar results obtain for other measures of sentiment about dividend-paying
firms.
8.3. Models of Managerial IrrationalityThe theories we have discussed so far interpret the data as reflecting actions
taken by rational managers in response to irrationality on the part of in-
vestors. Other papers have argued that some aspects of managerial behav-
ior are the result of irrationality on the part of managers themselves.
Much of section 2 was devoted to thinking about whether rational
agents might be able to correct dislocations caused by irrational traders.
Analogously, before we consider models of irrational managers, we should
ask to what extent rational agents can undo their effects.
A SURVEY OF BEHAVIORAL FINANCE 61