Saddled with Attention: Overreaction to Bankruptcy Filings

Published date01 December 2019
DOIhttp://doi.org/10.1111/irfi.12199
Date01 December 2019
AuthorNicolas Waissbluth,Tomas Reyes
Saddled with Attention:
Overreaction to Bankruptcy
Filings*
TOMAS REYES AND NICOLAS WAISSBLUTH
Department of Industrial and Systems Engineering, Ponticia Universidad Católica de
Chile, Santiago, Chile
ABSTRACT
We analyze the effect of investor attention on stock prices around Chapter 11
bankruptcy lings. We measure investor attention as abnormal search volume
from Google, and nd that attention-grabbing companies have more negative
abnormal stock returns in the days before and during bankruptcy lings and
more positive abnormal returns immediately thereafter. That is, for companies
receiving high attention, investors overreactto a bankruptcy ling; for compa-
nies receiving low attention, they underreact. This pattern is more pro-
nounced for companies with low institutional ownership and holds after
controlling for standard predictors of stock performance during bankruptcy.
JEL Codes: G14; G33; D82
Accepted: 22 April 2018
I. INTRODUCTION
Limited attention plays an important role in investor behavior and is associated
with multiple market pricing anomalies. Since Kahneman (1973) characterized
attention as a scarce cognitive resource, several authors have argued that lim-
ited attention from economic agents explains numerous nancial phenomena
(Hirshleifer and Teoh 2003; Peng and Xiong 2006; Barber and Odean 2008; Del-
laVigna and Pollet 2009; Hirshleifer et al. 2009). Whether investors are paying
attention to specic stock market events affects the way in which they process
information and react to it. This can ultimately alter the effect of both positive
and negative information on asset prices.
There are two standard mechanisms through which attention can affect
stock prices. First, in conjunction with new information, attention can lead to
* We acknowledge the support of FONDECYT (Grants 1160048 and 1171894) and Núcleo Milenio
Research Center for Entrepreneurial Strategy under Uncertainty (Grant NS130028). We thank the
editors and the anonymous referee for their constructive feedback. We also appreciate helpful com-
ments and suggestions from Manuel
Alvarez, Robin Greenwood, Katerina Manoff, Santiago Mingo,
Julio Pertuzé, and Stephen Zhang. We also thank Diego Martínez for excellent research assistance.
Part of the work was completed while Reyes was at UC Berkeley.
© 2018 International Review of Finance Ltd. 2018
International Review of Finance, 19:4, 2019: pp. 787819
DOI: 10.1111/ir.12199
faster information diffusion and therefore faster price discovery (DellaVigna
and Pollet 2009; Hirshleifer et al. 2009; Drake et al. 2012). Second, attention
itself can articially drive stock demand and induce noise trading behavior that
ultimately leads to price pressure from retail investors (Barber and Odean 2008;
Da et al. 2011; Joseph et al. 2011). This second theory traditionally applies to
attention driven by both positive and negative information.
In this paper, we posit the existence of a third mechanism through which
attention can exacerbate mispricing that stems from biased beliefs. We look for
evidence for this third mechanism around bankruptcy lings, which are
unusual, associated with large price changes, and yield extremely negative news
stories that are generally not open to interpretation. This empirical setup pro-
vides a powerful natural setting that allows us to diminish any upward price
pressure stemming from a neutral or positive interpretation of a news event. As
a robustness check, we also examine a sample of earnings announcements with
negative earnings surprises and nd analogous patterns.
There is an ample body of literature related to bankruptcy issues. However,
most of these studies deal with the prediction of nancial distress and the work-
ings of corporate failure mechanisms. In contrast, few studies focus on the mar-
ket performance of rms during bankruptcy (Clark and Weinstein 1983; Datta
and Iskandar-Datta 1995; Dawkins et al. 2007; Coelho 2013). According to
these studies, the average effect of a bankruptcy ling on a rms stock is a con-
sistent price decrease over the weeks prior to the ling and a steep price drop at
the time of the ling.
Our primary hypothesis is that, before and during their bankruptcy lings,
rms receiving higher levels of investor attention experience more negative
abnormal returns (ARs) than their low-attention counterparts; in the period
after their lings; however, these companies experience a signicant reversal
marked by higher ARs. That is, a surge in retail investorsattention around
bankruptcy lings leads to an overreaction to the bankruptcy news.
This overreaction is not fully consistent with the two previously mentioned
standard mechanisms through which attention is known to affect stock prices.
If attention were a facilitator of information diffusion (DellaVigna and Pollet
2009; Hirshleifer et al. 2009), we would not expect an overreaction. On the
other hand, if attention were a trigger of retail net buying (Barber and Odean
2008), we would expect a price pressure effect in the opposite direction.
Specically, in this paper, we test three hypotheses: (i) attention is negatively
associated with abnormal stock returns before and during bankruptcy lings;
(ii) attention is associated with a return reversal after bankruptcy lings; and
(iii) these effects are more signicant for companies with low levels of institu-
tional ownership.
To test our hypotheses, we conduct a short-term event study of cumulative
abnormal returns (CARs) around bankruptcy ling dates and analyze these
returns as a function of attention and institutional ownership levels. Our sam-
ple consists of 155 Chapter 11 bankruptcy cases led between 2004 and 2014
that involve active trading of common stock after the ling date. We measure
© 2018 International Review of Finance Ltd. 2018788
International Review of Finance
attention as abnormal internet search volume for company-specic search
terms from Google Trends. We consider cases after 2004 because data from this
source is available from that time onward.
We start by examining the cross-section of bankruptcy returns for the pre-,
during, and postling periods and identify systematic patterns of returns
around the ling date. We nd patterns consistent with prior research and the
efcient market hypothesis (Fama 1970); the market is able to anticipate the
bankruptcy ling to some extent, the ling reveals a signicant amount of neg-
ative information, and there is no signicant drift in the postling period given
that no new information is revealed.
However, consistent with our hypotheses, we identify a different pattern of
CARs through time when considering the attention paid to each company.
Companies receiving high levels of attention exhibit evidence of
overreactionfaster stock price declines before and during their bankruptcy l-
ings, and more positive CARs afterward. In contrast, companies receiving lower
levels of attention show evidence of underreactionslower decreases in stock
price before and during their lings, and more negative CARs thereafter.
More specically, for the preling and ling periods, we nd a statistically
signicant difference between the CARs of two groups of companies receiving
high and low levels of attention, respectively. Our results show that companies
receiving higher attention experience signicantly larger price drops. In the pre-
ling period, we nd a 14.2% CAR for the high-attention group, compared to
5.16% for the low-attention group. On the ling days, we nd a 47.6% CAR
for the high-attention group, and 31.2% for the low-attention group. This
conrms that attention is related to negative CARs before and during bank-
ruptcy lings and supports our rst hypothesis.
In contrast, for the postling period, we nd that companies receiving high
levels of attention have a positive average CAR of 14.8%, 22.2% higher than
the low-attention group, which exhibits a negative average CAR of 7.43%.
This suggests an attention-driven reversal after the ling and supports our sec-
ond hypothesis (i.e., positive short-term postling CARs are related to higher
levels of attention).
We then turn to regression models to explain CARs for each period in terms
of attention. Our results reveal a relevant effect of attention that holds when
controlling for standard predictors of bankruptcy stock performance (rm size,
nancial distress level, debtor-in-possession nancing, and prepackaged bank-
ruptcies). We nd economically and statistically signicant effects of abnormal
search volume on stock performance, conrming that attention drives prices
down before and during bankruptcy lings, and up thereafter. This differs sub-
stantially from traditional ndings regarding attention and stock returns, which
suggest that attention is a facilitator of either information diffusion or retail net
buying.
We also analyze whether the effect of attention interacts with the level of
institutional ownership of bankrupt rms. A higher concentration of individual
investors tends to exacerbate market anomalies, and therefore we expect the
© 2018 International Review of Finance Ltd. 2018 789
Attention to Bankruptcy Filings

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT