Behavioral Finance Theory Building. In the past few years there has been
a burst of theoretical work modelling financial markets with less than fully
rational agents. These papers relax the assumption of individual rationality
either through the belief formation process or through the decision-making
process. Like the work of psychologists discussed above, these papers are
important existence proofs, showing that it is possible to think coherently
about asset pricing while incorporating salient aspects of human behavior.
Investor Behavior. We have now begun the important job of trying to
document and understand how investors, both amateurs and professionals,
make their portfolio choices. Until recently such research was notably ab-
sent from the repertoire of financial economists.
This is a lot of accomplishment in a short period of time, but we are still
much closer to the beginning of the research agenda than we are to the end.
We know enough about the perils of forecasting to realize that most of the
future progress of the field is unpredictable. Still, we cannot resist venturing
a few observations on what may be coming next.
First, much of the work we have summarized is narrow. Models typically
capture something about investors’ beliefs, or their preferences, or the lim-
its to arbitrage, but not all three. This comment applies to most research in
economics, and is a natural implication of the fact that researchers are
boundedly rational too. Still, as progress is made, we expect theorists to
begin to incorporate more than one strand into their models.
An example can, perhaps, illustrate the point. The empirical literature re-
peatedly finds that the asset pricing anomalies are more pronounced in
small- and mid-cap stocks than in the large-cap sector. It seems likely that
this finding reflects limits to arbitrage: the costs of trading smaller stocks
are higher, keeping many potential arbitrageurs uninterested. While this ob-
servation may be an obvious one, it has not found its way into formal mod-
els. We expect investigation of the interplay between limits to arbitrage and
cognitive biases to be an important research area in the coming years.
Second, there are obviously competing behavioral explanations for some
of the empirical facts. Some critics view this as a weakness of the field. It is
sometimes said that the long list of cognitive biases summarized in section 3
offer behavioral modelers so many degrees of freedom that anything can be
explained. We concede that there are numerous degrees of freedom, but
note that rational modelers have just as many options to choose from. As
Arrow (1986) has forcefully argued, rationality per se does not yield many
predictions. The predictions in rational models often come from auxiliary
assumptions.
There is really only one scientific way to compare alternative theories,
behavioral or rational, and that is with empirical tests. One kind of test
64 BARBERIS AND THALER