| |
[Wikipedia]
According to the peak-end rule, we judge our past
experiences almost entirely on how they were at their peak
(pleasant or unpleasant) and how they ended. Virtually
all other information appears to be discarded, including net pleasantness
or unpleasantness and how long the experience lasted. Evaluations
are based primarily on the most extreme
and the final moments of an episode with
all other moments having essentially no influence on judgment. This
heuristic was first suggested by Daniel Kahneman and others. He
argues that because people seem to perceive not the sum of an experience
but its average, it may be an instance of the representativeness
heuristic. |
[Odean]
"Investors who insist on hunting for the next brilliant stock
would be well advised to remember what California prospectors discovered
ages ago: All that glitters is not gold". |
All
that Glitters: The Effect of Attention and News on the Buying Behavior
of Individual and Institutional Investors, Barber and Odean,
2003
"We test the hypothesis that individual investors are more
likely to be net buyers of attention-grabbing stocks than are institutional
investors. We speculate that attention-based buying is a result
of the difficulty that individual investors have searching the thousands
of stocks they can potentially buy. We look at three indications
of how likely stocks are to catch investors' attention: daily abnormal
trading volume, daily returns, and daily news. Consistent with our
predictions, we find that individual investors
display attention based buying behavior. They are net buyers on
high volume days, net buyers following both extremely negative and
extremely positive one-day returns, and net buyers when stocks are
in the news. Professional investors are less prone
to indulge in attention-based purchases. With more time and resources,
professionals are able to continuously monitor a wider range of
stocks. Consistent with the predictions of our model, we find that
stocks bought by individual investors on high-attention days tend
to subsequently under perform stocks sold by those
investors". |
The
Cross-Section of Analyst Recommendations, Sorescu and Subrahmanyam,
2004
We analyze the relation between analyst attributes (years of experience,
reputation of the analysts' brokerage houses) and the short- and
long-term price reactions to recommendations made by the analysts.
We find that in the long-term, the recommendation changes of highly
experienced analysts outperform those of low-experience ones. In
addition, investors appear to overreact to dramatic upgrades of
low-ability analysts, and underreact to small upgrades by high-ability
analysts. These results are consistent with the Griffin and Tversky
(1992) argument that agents place too much emphasis on the
strength of the signal (the dramatic nature of the event) and insufficient
emphasis on the weight (the ability of the analyst making the recommendation).
Agents are prone to attaching undue importance to the enthusiasm
in a recommendation letter, and not enough importance to the credibility
of the recommendation writer. Since the investor bias is stronger
for more extreme (high-strength) signals, the market overreacts
to such signals. At the same time, the market underreacts to the
weight (quality) of the signal. The net result is that prices experience
reversals (overreaction) following high-strength, low-weight signals
and drift (underreaction) following high-weight, low-strength signals.
|
Internet
Stock Message Boards and Stock Returns, Antweiler and Frank,
2002
"This paper examines whether stocks with high posting levels
also have unusual subsequent returns and/or risk. They do. We find
that portfolios with particularly high message posting have abnormally
poor returns. The poor returns in the high message posting
portfolios are accompanied by high volatility.
Consideration is given to market manipulation, differences of opinion,
and anxiety reduction as possible explanations for the observed
patterns". |
Is
the Market Mad? Evidence from Mad Money, Engelberg, Sasseville
and Williams, 2006
"We documented statistically signifcant abnormal returns for
the stocks recommended by Jim Cramer on the popular television show,
Mad Money. The average cumulative abnormal overnight return for
the smallest quartile of recommended stocks is 5.19%, and these
returns completely disappear within 12 trading days. After documenting
this market inefficiency, we analyzed the trading activity following
Cramer's recommendations. Our findings that trading volume and buy-sell
imbalance are unusually high on the day following Cramer's recommendation
suggests that uninformed traders buy the stocks recommended by Cramer
on the previous night. The uninformed traders do this despite the
fact that these stocks became overpriced overnight and earn negative
cumulative abnormal returns over the next two weeks. Our Finding
that short sales volume is unusually high on the day following Cramer's
recommendations suggest that some arbitraguers are aware of Cramer's
effect on security prices. Taken together, our results suggest that
the aggregate losers in our event study are the Mad Money viewers
who decide to buy the recommended securities when the markets open
the following day, and that the winners are the market makers and
arbitraguers who sell the overpriced recommended stocks on day 1,
as well as the traders who sell the recommended stocks on days 2
through 12. Individual investors who watch Mad Money would be wise
to wait before purchasing the small stocks Cramer recommends, as
these stocks tend to fall to their original levels following the
overnight price spike caused by his recommendation". |
Extrapolation
Bias: Insider-Trading Improvement Signal, Fuller & Thaler
Research Library
Extrapolation Bias exploits overreaction to past, negative information.
This report deals with the behavioral biases that cause investors
to overreact and describes one aspect of our Extrapolation Bias
strategy which exploits overreaction. Naive extrapolation is largely
the result of behavioral biases associated with two heuristics --
the representativeness heuristic and the saliency heuristic. Representativeness
involves the tendency of humans to generalize about a population
of future outcomes after observing a small sample -- for example,
after observing only one or two outcomes, humans will frequently
conclude that these two outcomes are representative of future outcomes.
Saliency involves the tendency of humans to assign too high a probability
to low frequency events after observing a recent, vivid example
of the event. For example, immediately after a plane crash has been
reported in the news, people will greatly overestimate the probability
of future plane crashes. The first step in our Extrapolation Bias
strategy is too identify conditions under which investors are most
likely to naively extrapolate recent, negative information into
the future. The second step is to determine whether the negative
information investors are extrapolating into the future is, most
likely, temporary. |
|
