Behavioral all articles
Articles week 2:
Ackert & Deaves, Chapter 1, 3 & 10.4
Frydman, C., & Wang, B. (2020). The impact of salience on investor behavior: Evidence from a
natural experiment. The Journal of Finance, 75(1), 229-276:
Abstract: We test whether the display of information causally affects investor behavior in a highstakes
trading environment. Using investor-level brokerage data from China and a natural experiment, we
estimate the impact of a shock that increased the salience of a stock’s purchase price but did not
change the investor’s information set. We employ a difference-in-differences approach and find that
the salience shock causally increased the disposition effect by 17%. We use microdata to document
substantial heterogeneity across investors in the treatment effect. A previously documented trading
pattern, the “rank effect,” explains heterogeneity in the change in the disposition effect.
Het dispositie-effect verwijst naar onze neiging om voortijdig activa te verkopen die financiële winst
hebben opgeleverd, terwijl we vasthouden aan activa die geld verliezen.
In this paper, we use a combination of microdata from a Chinese brokerage house and a natural
experiment to estimate the causal effect of a change in information display on individual investor
trading decisions. Specifically, we obtain data from a brokerage house that increased the salience of a
stock’s capital gain by making it more visually prominent on the investor’s online trading screen
(Figure 1). To assess the impact on investor behavior, we measure the disposition effect, which is the
greater tendency to sell stocks with capital gains compared to capital losses (Shefrin and Statman
(1985), Odean (1998)).1 We hypothesize that increasing the salience of a stock’s capital gain will
increase the disposition effect.
The main result in our paper is that after the salience shock, there is a significant increase in the
disposition effect among internet investors (treatment group), compared to noninternet investors
(control group). We show that the two groups have parallel trends before the salience shock and thus
we can interpret the change in display as causally affecting trading behavior. Moreover, the change
among the treatment group is sizable: The disposition effect increases by 17%, relative to the control
group. This represents both an economically and a statistically meaningful change in trading behavior
that takes place in a highstakes environment. The average portfolio size among internet investors in
our sample is 63,916 RMB, and annual disposable income per capita of citizens in China during our
sample period is approximately 10,000 RMB.
Empirically, we find that variation in the change in the disposition effect can be explained by a
previously documented trading bias called the “rank effect” (Hartzmark (2015)). This trading pattern
refers to an investor’s tendency to sell the extreme ranked stocks in his or her portfolio.
IV. Potential Mechanisms
While our data enable us to establish a causal effect on investor behavior, we are unable to identify
the mechanism that generates our results. In this section, we consider several mechanisms that can
potentially explain the observed increase in the disposition effect. It is important to emphasize that
each of the mechanisms we consider is insufficient to explain the change in trading behavior, and
needs to be combined with a theory of the disposition effect. In other words, standard models of
investor behavior do not predict a greater willingness to sell winning stocks compared to losing
,stocks, and thus explaining an amplification of this effect must rely on a theory that can generate this
asymmetry in the first place.
With this in mind, a first potential mechanism that can generate an increase in the disposition effect
is exogenous attention allocation. Under this mechanism, attention is exogenously directed toward
salient attributes, which then receive more weight in the decision-making process (Fehr and Rangel
(2011), Bhatia (2013)). In our setting, the capital gain became more salient, in the sense that it
became more visually prominent, which may have led to an exogenous (or “bottom-up”) shift in
investor attention toward this attribute on the trading screen. The greater weight that investors place
on the capital gain during decision making can then trigger a stronger asymmetry in selling winning
and losing stocks.
Second, investors may perceive the change in information display as advice from the brokerage house
(Benartzi (2001), Madrian and Shea (2001)). This mechanism is likely to be more relevant for
unsophisticated investors, who may be looking to the brokerage house for information on how best
to manage a portfolio. Interestingly, this channel may also be useful in explaining our cross-sectional
results. Hartzmark (2015) reports evidence that unsophisticated investors from a large discount
brokerage in the United States exhibit a stronger rank effect compared to sophisticated investors. In
our data, we find that investors with stronger rank effects are the same investors who exhibit a larger
trading response after the capital gain is added to the trading screen.
Third, our results may be generated by a bounded rationality or search cost mechanism (Caplin and
Dean (2015), Gabaix (2019)). Although information on the capital gain was available prior to the
change in information display, to obtain this information investors needed to combine their
transaction history and current stock prices. This process of searching for and integrating multiple
sources of information can be costly to some investors, who may choose to engage in this process
only when the cost is lower than the benefit of using the capital gain in their trading decision.
Because the change in information display arguably reduced the search cost, it may have increased
the frequency with which investors chose to acquire and trade on the capital gain.
Finally, as shown in Figure 1, when the brokerage company added the capital gain variable to the
trading screen, it also changed the font color for winning stocks from blue to red. Similar to the
exogenous attention mechanism discussed above, this may have increased the salience of the sign of
the capital gain, which investors then assigned a higher weight in their decision process (Bazley,
Cronqvist, and Mormann (2018)). We note, however, that font color cannot explain our cross-
sectional results because we measure the rank effect in the period before the display change when
there was no variation in font color.
In summary, our results are likely driven by some combination of the mechanisms discussed above.
Our data cannot distinguish between these different sources of behavior, but the connection with the
rank effect provides preliminary evidence that may be helpful in future empirical tests of the
mechanism.
VI. Conclusion
In this paper, we provide evidence that information display has a causal effect on investor behavior in
a high-stakes and natural trading environment. The magnitude of the change in trading behavior is
economically large, as we estimate a 17% increase in the size of the disposition effect after the
change in information display. Our paper thus provides evidence from the field that investor behavior
is affected by information display. While our sample period corresponds to the early 2000s, our
,results are likely to be more applicable today as there has been a large migration toward making
investment decisions on a digital platform (Benartzi and Lehrer (2015)).
In addition, our microdata allow us to document substantial heterogeneity across investors in
sensitivity to the change in display. We find that investors who exhibit a stronger trading response to
the salience shock also exhibit a larger rank effect (Hartzmark (2015)). These results contribute to a
recent empirical literature in behavioral economics and finance that seeks to document correlations
between behavioral biases (Barber and Odean (2013), Frydman and Camerer (2016), Stango, Yoong,
and Zinman (2017), Chapman et al. (2018), Dean and Ortoleva (2019)). Measuring these correlations
is useful because it can distill a long list of empirical facts into a smaller number of principal
psychological components. This, in turn, can provide guidance for constructing a more unified model
of investor behavior (Barberis (2018)). Finally, our results provide evidence consistent with an
important assumption in several recent models of attention and economic choice (Bordalo, Gennaioli,
Shleifer (2012, 2013a, 2013b), Koszegi and Szeidl (2013)). In these models, attention is endogenously
allocated toward the attribute in which the alternative is most different relative to other alternatives.
An important assumption that links this endogenous allocation of attention to economic behavior is
that those attributes that receive more attention also receive more weight in the decision process.
Most of the evidence that supports this assumption has come from laboratory experiments. The
results reported here show that this link between attention and economic behavior also holds in a
higher-stakes environment outside the laboratory.
Tutorial:
The Impact of Salience on Investor Behavior: Evidence from a Natural Experiment
Motive of research: • In most models of investor behavior, information display has little impact on
investor behavior (the way information is framed should impact the way an investor accts)
• Laboratory experiments demonstrate the opposite • Little or contrasting evidence outside of
laboratory Need for evidence from a higher-stake (laboratory its maybe for money and people care
less) and more natural environment (here look at investors that behave rationally with own money).
Relevant info: • Hypothesis: increasing the salience of a stock’s capital gain will increase the
disposition effect • From a brokerage house in China [so, no tax effects] • 7 years of data: 01/2003 -
12/2009 • Approx. 3,500 retail investors
Capital gains became more salient, easier to visualize. The important thing is that the variables could
be computer by the investors before. didn’t change the content, only the way its framed.
Treatment group are the internet investors. Non internet investors not influenced by the salience
shock.
Look at the change in the disposition effect due to salience shock. Selling gain over the post
treatment.
What did the authors do to tackle (endogeneity) concerns?
Inclusion of controls known to affect selling propensities
Inclusion of FE
Estimation of a multiple placebo test › Use of different control groups: • Data on investors from a
second branch of the same brokerage firm • Data on investors from a third branch of another
brokerage firm • Propensity Score Matching
, Changing the control groups, results are the same. Results are robust.
Findings: heterogeneity in disposition effect, most are positive and ….
Mechanisms behind results
mechanisms ? 1. Attention constraints (attention directed to salient attributes as gains and that’s why
they start training on these attributes) 2. Endorsement effect (change of the fond could be seen as
advice) 3. Bounded rationality (inability to carry out full cost benefit analysis, lack of cognitive
capabilities).
What have we learnt?
Retail investors respond to a salience shock in a real live setting
• Their propensity to sell a winner stock [=the Disposition Effect] is approx. 17% larger after the
salience shock.
• This result fits well within the literature on framing [the way of presenting] and salience [drawing
attention to a specific attribute that then receives a larger weight in the decision]
• There is substantial heterogeneity between investors in terms of size of the DE and the impact of
the salience shock
• The fact that there appears to be a link with another behavioral bias [the Rank Effect], supports a
behavioral [and not a rational] explanation.
ChatGPT:
investigates how salience affects investor behavior, using a natural experiment in which the Securities
and Exchange Commission (SEC) announced a new regulation that increased the salience of mutual
fund fees.
The authors find that the increased salience of mutual fund fees led to a significant decrease in fees
and an increase in flows to lower-cost funds. However, the study also finds that the impact of salience
varies across investor types, with more financially sophisticated investors being less affected by
salience.
The paper also examines the mechanisms behind the impact of salience on investor behavior. The
authors find that the effect of salience is driven by both increased attention and increased
information processing. Specifically, the increased salience of fees led to increased attention to fees
and increased comparison shopping among investors.
Overall, the paper provides evidence that salience has a significant impact on investor behavior,
leading to changes in investment decisions and greater attention to fees. The study also suggests that
the impact of salience may depend on investor characteristics, such as financial sophistication.
