annual returns on stocks and bonds or simulated thirty-year returns on the
same asset classes. The two investment choices, stocks and bonds, were un-
labeled and the subjects were asked to allocate their retirement contribu-
tions between the two funds. The results are consistent with myopic loss
aversion. Those viewing the annual returns allocated, on average, 41 per-
cent to stocks, whereas those viewing the long-term returns allocated 82
percent. The difference in the median allocations was even larger: 40 per-
cent versus 90 percent. It is quite disturbing that alternative presentation
modes of the “same” data could yield dramatic differences in investment
choices. The results raise difficult questions for employers and regulators
about the proper frequency of performance reporting.
In this appendix, we reviewed several tests of myopic loss aversion, all of
which are supportive of the concept. The evidence on myopic loss aversion
is consistent with a more general phenomenon Kahneman and Lovallo
(1993) labeled “narrow framing” (i.e., thinking about gambles or invest-
ments one at a time rather than aggregating them into a portfolio). Put dif-
ferently, aggregation could reduce aggravation. Unfortunately, there is very
little we know about the way people set their evaluation frequency, or more
generally, the way people define frames. These issues are promising avenues
for future research.
MYOPIC LOSS AVERSION 221