National Culture and Stock Price Crash Risk * Tung Lam Dang † Robert Faff ‡ Luong Hoang Luong § Lily Nguyen ¶ First version: January, 2016 This version: January, 2017 * Corresponding author: Lily Nguyen. Nguyen gratefully acknowledges the financial support from the Australian Research Council (ARC) to this project. † Department of Finance, The University of Danang, Danang, Vietnam, [email protected], +84 9 1585 8458. ‡ UQ Business School, The University of Queensland, Australia, r.faff@business.uq.edu.au, +61 7 3346 8055. § UNSW Business School, The University of New South Wales, Australia, [email protected], +61 3 9466 8035. ¶ La Trobe Business School, La Trobe University, Australia, [email protected], +61 3 947 93971.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
National Culture and Stock Price Crash Risk∗
Tung Lam Dang† Robert Faff‡ Luong Hoang Luong§ Lily Nguyen¶
First version: January, 2016This version: January, 2017
∗Corresponding author: Lily Nguyen. Nguyen gratefully acknowledges the financial support from theAustralian Research Council (ARC) to this project.†Department of Finance, The University of Danang, Danang, Vietnam, [email protected], +84
9 1585 8458.‡UQ Business School, The University of Queensland, Australia, [email protected], +61 7 3346
8055.§UNSW Business School, The University of New South Wales, Australia,
where i, j, k, and t denote firm, country, industry, and year, respectively. The dependent
variable, CrashRisk, refers to NCSKEW , DUV OL, or CRASH. Control is a vector of firm-
and country-level characteristics as discussed in Section 2.4, all measured in year t− 1, except
ROA. The regression is estimated by pooled OLS with industry and year fixed effects and
robust standard errors clustered at the firm level. Our independent variable of interest is the
9
country individualism score (IDV ). If β1 is positive and statistically significant, then we find
support for a positive effect of individualism on stock price crash risk.
[Insert Table 3 about here]
Table 3 reports the regression results for equation (4). In the regression of NCSKEW
on IDV and other control variables as shown in column 1, the coefficient on IDV is positive
and statistically significant at the 1% level, suggesting a positive effect of individualism on
stock price crash risk. A coefficient estimate of 0.128 on IDV indicates that stock price crash
risk increases by 12.8% for every 1 percentage point increase in the individualism index. This
result is economically significant. Columns 2 and 3 show the regression results for DUV OL
and CRASH, respectively. Similar results obtain as the coefficients on IDV are all positively
and strongly significant at the 1% level. As in the NCSKEW regression, the results are also
of economic significance.
Turning to the firm-level control variables, we see that the coefficient estimate of BIG4
is negative and significant as shown in columns 1 and 2, suggesting that firms with greater
transparency, captured by the use of Big 4 auditing firms for their accounting statements, are
associated with lower future stock price crash risk. DeFond et al. (2015) find that the increased
transparency broadly reduces crash risk among nonfinancial firms. Consistent with Hutton
et al. (2009), the coefficient on DISACC is significantly positive in all regressions, suggesting
that firms more likely to manage reported earnings are more likely to crash. For the remaining
control variables, the results are all consistent with prior work in the crash risk literature in
terms of sign and significance.
As for the country-level control variables, we find that firms domiciled in countries with more
transparent information environments, better creditor protection rights, more well-developed
stock markets, and higher GDP growth rates are less crash-prone.
3.3 Instrumental Variable Regressions
While the baseline results support our hypothesis that culture has a positive effect on crash
risk, we cannot rule out the possibility that our culture variable is endogenous, because time-
varying country and firm unobservables omitted from the regressions can be correlated with
both individualism and crash risk. To address this important concern, we use an instrumental
variable (IV) approach. Based on prior work, we use two sets of instruments as follows.
10
First, we follow Shao et al. (2013) and use three instruments, namely, (1) genetic distance
from the United States (Genetic Distance), (2) the license to drop pronouns (Pronoun Drop),
and (3) British Rule (British Rule). Genetic Distance, from Spolaore and Wacziarg (2009),
is the FST distance1 that measures a genetic distance from the United States. This measure
aggregates differences in the distribution of gene variants across populations, and thus captures
the degree of genealogical relatedness of different populations, which has been shown to be
highly correlated with individualism. Pronoun Drop, from Kashima and Kashima (1998), is a
dummy variable related to the requirement to use pronouns in a language (the license to drop
pronouns), which equals 1 if a country’s grammatical rules license person-indexing pronoun
drop and 0 otherwise. According to Kashima and Kashima (1998) and Shao et al. (2013),
Pronoun Drop is associated with the degree of psychological differentiation between the speaker
and the social context of speech and inversely correlated with individualism. British Rule is a
dummy variable equal to 1 if a country has historically been under the British rule (Treisman,
2000) and 0 otherwise. A colony of Britain should share some similarities in cultural values.
Shao et al. (2013) find that Pronoun Drop is highly correlated with individualism when British
Rule is incorporated.
To check the relevance of the instruments, in Panel A of Table 4 we present the first-stage
regressions, where we regress IDV on the instruments and the same set of independent variables
as in the baseline regressions. Column 1 shows that all the instruments have strongly significant
coefficients. The p-value of the F -test for the joint significance of instruments, shown at the
bottom of the panel, is close to 0, thus rejecting the null hypothesis of weak instruments. The
overidentification tests with p-value for both Sargan and Basmann statistics are all larger than
0.1 for all the crash risk measures (not reported for brevity), suggesting that the instruments
for individualism are valid. In addition, it is reasonable to assume that these instruments are
uncorrelated with stock price crash risk because there is no economic mechanism under which
the instruments affect stock price crash risk other than through culture.
Columns 2–4 show that the coefficient estimates on the fitted individualism (IDVFitted) are
all positive and significant at the 1% level (except for the NCSKEW regression), suggesting
1FST distance, also known as “coancestor coefficients”, is based on indices of heterozygosity, which refers tothe probability that two alleles at a given locus selected at random from two populations will be different (see,for example, Spolaore and Wacziarg, 2009):
FST is
{= 0 if the allele distributions are equal across the two population,
> 0 if otherwise.
11
that the positive effect of individualism on stock price crash risk is unlikely to be driven by
omitted unobservables.
[Insert Table 4 about here]
The second set of instruments that we use are based on Kwok and Solomon (2006) and
Li et al. (2013): religion (Religion), demography (Ethnic Fractionalization), and geography
(Geography). Religion, from La Porta (1998), is the percentage of people in the Protestant,
Roman Catholic, and Muslim religious faiths in 1980. Ethnic Fractionalization, from Alesina
et al. (2003), is measured as the degree of ethnic heterogeneity in a given country. We use the
continent of a country as a proxy for geography. According to Li et al. (2013), these variables
are selected as potential determinants of culture based on theory and data availability. Panel B
of Table 4 reports the two-stage least squares (2SLS) regression results. The regressions control
for firm- and country-level characteristics, as well as industry and year fixed effects.
Panel B of Table 4 reports the 2SLS regression results with this set of instruments. The
first-stage results show that the instruments are highly correlated with individualism. The null
hypothesis of weak instruments is strongly rejected at the 1% level, with p-value of the F -test
for the joint significance of the instruments close to 0. As before, the coefficient estimates of the
fitted IDV as shown in columns 2–4 are all positive and strongly significant at the 1% level,
which again suggests that our results on the positive effect of individualism on stock price crash
risk still hold.
3.4 Hierarchical Linear Regressions
A further issue is that our data can be viewed as multilevel data. At the country level, we have
firms from 36 countries. At the firm level, we have 19,080 firms. We therefore follow Li et al.
(2013) and Shao et al. (2013) to employ a hierarchical linear model (HLM) regression where
the set of firms within countries form the base-level observations while the countries serve as
the higher-level observations. The power of HLMs comes from their ability to correctly pool
firm-level effects across countries while also examining country-level relations.
We closely follow Li et al. (2013) in preparing the data for the HLM regression. First, we
center each country-level independent variable by its grand mean (averaged across countries)
so that every transformed variable has a mean of zero, and we add the suffix “−ctry” to each
of these variables. Second, we center each firm-level independent variable by its grand mean
12
(averaged across firms and countries for a given fiscal year), so that every transformed variable
has a mean of zero. Third, we create country-level mean values (averaged within a country)
on those grand-mean-centered variables in step 2 and add the suffix “−ctrymean” to each of
these variables. Finally, we create within-country residuals by taking the grand-mean adjusted
variables in step 2 and subtracting the corresponding within-country means in step 3. We name
these firm-level deviations separately from their corresponding country-level means by adding
the suffix “−firmdev”.
Table 5 presents the HLM regressions, all of which include industry and year fixed effects.
The coefficient estimates of IDV are all positive and strongly significant at the 1% level, sug-
gesting a positive effect of individualism on stock price crash risk. These results provide support
for the baseline results on the relationship between individualism and crash risk.
[Insert Table 5 about here]
3.5 Additional Robustness Tests
To provide further robustness, we conduct several tests. First, we use alternative measures for
individualism because these proxies for culture can allay the concern that our baseline results
are only valid to the choice of a particular proxy. Following Chen et al. (2015), we use the
World Value Survey (WVS) on individualism as an alternative measure of individualism. An
advantage of using these data is that the measure is time-varying and thus mitigates a common
problem with time-invariability in country-level proxies for culture. The World Value Survey
covers 97 countries and is carried out in 7 waves, namely, 1981–1984, 1985–1988, 1989–1993,
1994–1998, 1999–2004, 2005–2009, and 2010–2014. We run the same regressions as in equation
(4) using the WVS individualism data for the periods 1999–2004, 2005–2009, and 2010–2014.
Since the WVS individualism index is time-varying, we are able to include country fixed effects
to control for any time-invariant country characteristics. Columns 1–3 in Panel A of Table 6
show the regressions with the alternative proxy for individualism, where we include country
fixed effects. As we can see, the coefficient estimate of the WVS individualism measure remains
positive and strongly significant at the 1% level, corroborating earlier results on the relation
between culture and crash risk.
[Insert Table 6 about here]
13
We also use an updated version of the individualism index as in Tang and Koveos (2008). To
allow for the updated components of cultural dimensions of Hofstede (2001), Tang and Koveos
(2008) establish a framework in which changes in economic conditions are the source of cultural
dynamics, while the endurance of institutional characteristics is the foundation for cultural
stability. A major advantage of Tang and Koveos (2008) is that they develop an integrated
empirical model to update Hofstede (2001)’s dimensions of culture, so that it provides a more
up-to-date and reliable reflection of each country’s true cultural predisposition. In columns
4–6 in Panel A of Table 6, we report the baseline regressions using Tang and Koveos (2008)’s
measure of individualism. Again, we find support for the positive effect of individualism on
stock price crash risk.
In the second test, we add more controls for country-level variables, such as additional
dimensions of cultures of Hofstede (2001) and country religion. Specifically, we include the power
distance scores (PDI), uncertainty avoidance (UAI), and masculinity scores (MAS). We use
the percentage of people in the Protestant (Protestantism), Roman Catholic (Catholicism), and
Muslim (Muslism) religious faiths in 1980 from La Porta (1998) as proxies for religion. Panel B
of Table 6 shows the same regressions as in equation 4 but with these additional controls. We
continue to find support for a positive effect of individualism on crash risk after controlling for
these variables.
To sum up, we show that the baseline regression results still hold under these robustness
tests.
4 Possible Mechanisms
After establishing a positive relation between culture and stock price crash risk, we now attempt
to identify possible mechanisms under which individualism positively affects crash risk. We
acknowledge that these underlying mechanisms are not necessarily mutually exclusive, and if
anything, may jointly contribute to this positive relation.
4.1 Trading by Overconfident Investors
First, traders from individualistic cultures are more overconfident because literature (e.g., Kag-
itcibasi, 1997; Markus and Kitayama, 1991) has shown that people from individualistic culture
tend to be more overconfident. For example, Chui et al. (2010) document the specific evidence
14
that individualism is directly linked to overconfident traders as it is highly correlated with
trading volume and volatility.
Second, overconfident investors’ trading activities can cause stock price crashes. The theo-
retical models of Odean (1998) and Daniel et al. (1998) show that overconfidence induces traders
to trade more aggressively even in the face of transaction costs and adverse expected payoffs.
In their models, overconfident traders place too much weight on their own views and too little
weight on other investors’ views when forming judgments about the value of a security. These
traders expect high profits from trading on their opinions, and their aggressive trading generates
excess volatility, which can lead to stock price crashes. Hong and Stein (2003) document that
investor belief heterogeneity, which is a form of trader overconfidence whereby each investor
irrationally thinks that his private signal is more precise than the others’, causes stock price
crashes.
These two streams of the literature suggest that overconfident investors’ trading activities
can be a channel through which individualism leads to stock price crashes. We thus argue that
if individualism induces overconfidence in investors whose trading activities causes stock price
crashes, then the positive effect of individualism on stock price crash risk should be stronger
when there are more overconfident traders in the market. To test this conjecture, we use
trading volume as a proxy for the presence of overconfident traders because behavioral finance
models predict higher trading volume in the presence of overconfident traders (e.g., Daniel and
Hirshleifer, 2015; Odean, 1998). We estimate the following regression model:
where i, j, k, and t denote firm, country, industry, and year, respectively. OM indicates
the presence of overconfident managers and is proxied by overinvestment (OV ERINV EST ),
capital expenditure (CAPEX), or research and development to total assets (R&D). All the
other variables are the same as in equation (4). We estimate the regressions with the pooled
OLS, where we include industry and year fixed effects and use robust standard errors clustered
17
at the firm level. Our interest is in the coefficient estimates of IDV (β1) and the interaction
term, IDV ×DTURN (β2).
[Insert Table 8 about here]
Table 8 reports the results for the regression equation (6). Columns 1–9 show that the
positive effect of individualism on stock price crash risk is more pronounced in firms with
overinvestment and risky investment because the coefficient estimates of both IDV and IDV ×
OM are positive and strongly significant at the 1% level in all model specifications, regardless
of which measure is used for OM . These results support our hypothesis that bad-news hoarding
activities by managers, whose overconfidence is fostered by their individualistic cultures, cause
stock price crashes.
5 Conclusion
We examine the effect of national culture on stock price crash risk, a currently unexplored
question in the literature. We find that individualism has a positive effect on stock price
crash risk, which suggests that firms headquartered in individualistic countries are associated
with higher stock price crash risk. We also find that this positive effect becomes stronger for
firms with high trading volume, overinvestment, and risky investments. Our results suggest
that individualistic cultures encourage overconfident managers’ bad-news hoarding activities,
and overconfident investors’ aggressive trading behavior, which eventually can lead stock price
crashes. This is the first study to document national culture as a potential factor contributing
to stock price crash risk.
18
References
Ahmed, A. S. and S. Duellman (2013). Managerial overconfidence and accounting conservatism.Journal of Accounting Research 51 (1), 1–30.
Alesina, A., A. Devleeschauwer, W. Easterly, S. Kurlat, and R. Wacziarg (2003). Fractionaliza-tion. Journal of Economic Growth 8 (2), 155–194.
An, H. and T. Zhang (2013). Stock price synchronicity, crash risk, and institutional investors.Journal of Corporate Finance 21, 1–15.
Andreou, P. C., C. Antoniou, J. Horton, and C. Louca (2016). Corporate governance andfirm-specific stock price crashes. European Financial Management 22 (5), 916–956.
Beugelsdijk, S. and B. Frijns (2010). A cultural explanation of the foreign bias in internationalasset allocation. Journal of Banking and Finance 34 (9), 2121–2131.
Callen, J. L. and X. Fang (2015). Religion and stock price crash risk. Journal of Financial andQuantitative Analysis 50 (1-2), 169–195.
Chen, J., H. Hong, and J. C. Stein (2001). Forecasting crashes: trading volume, past returns,and conditional skewness in stock prices. Journal of Financial Economics 61 (3), 345–381.
Chen, Y., P. Y. Dou, S. G. Rhee, C. Truong, and M. Veeraraghavan (2015). National cultureand corporate cash holdings around the world. Journal of Banking and Finance 50, 1–18.
Chui, A. C. W. and C. C. Y. Kwok (2007). National culture and life insurance consumption.Journal of International Business Studies 39 (1), 88–101.
Chui, A. C. W., A. E. Lloyd, and C. C. Y. Kwok (2002). The determination of capital struc-ture: Is national culture a missing piece to the puzzle? Journal of International BusinessStudies 33 (1), 99–127.
Chui, A. C. W., S. Titman, and K. C. J. Wei (2010). Individualism and momentum around theworld. Journal of Finance 65 (1), 361–392.
Daniel, K. and D. Hirshleifer (2015). Overconfident investors, predictable returns, and excessivetrading. Journal of Economic Perspectives 29 (4), 61–88.
Daniel, K., D. Hirshleifer, and A. Subrahmanyam (1998). Investor psychology and securitymarket under- and over-reactions. Journal of Finance 53 (6), 1839–1885.
Dechow, P. M. and R. G. Sloan (1995). Detecting earnings management. Accounting Re-view 70 (2), 193–225.
DeFond, M. L., H. Mingyi, L. Siqi, and L. Yinghua (2015). Does mandatory IFRS adoptionaffect crash risk? Accounting Review 90 (1), 265–299.
Dimson, E. (1979). Risk measurement when shares are subject to infrequent trading. Journalof Financial Economics 7 (2), 197–226.
Djankov, S., C. McLiesh, and A. Shleifer (2007). Private credit in 129 countries. 84 (2), 299–329.
Eun, C. S., L. Wang, and S. C. Xiao (2015). Culture and R2. 115 (2), 283–303.
Ferris, S. P., N. Jayaraman, and S. Sabherwal (2013). CEO overconfidence and internationalmerger and acquisition activity. Journal of Financial and Quantitative Analysis 48 (1), 137–164.
19
Grinblatt, M. and M. Keloharju (2009). Sensation seeking, overconfidence, and trading activity.Journal of Finance 64 (2), 549–578.
Heaton, J. B. (2002). Managerial optimism and corporate finance. Financial Management 31 (2),33–45.
Hirshleifer, D., A. Low, and S. H. Teoh (2012). Are overconfident CEOs better innovators?Journal of Finance 67 (4), 1457–1498.
Hofstede, G. (2001). Culture’s consequences: Comparing values, behaviors, institutions, andorganizations across nations. Sage Publication.
Hong, H. and J. C. Stein (2003). Differences of opinion, short-sales constraints, and marketcrashes. Review of Financial Studies 16 (2), 487–525.
Hribar, P. and H. Yang (2016). CEO overconfidence and management forecasting. ContemporaryAccounting Research, 33 (1), 204–227.
Hu, J., J.-B. Kim, and W. Zhang (2015). Insider trading and stock price crashes: Internationalevidence from a natural experiment. Available at SSRN: http://ssrn.com/abstract=2359069or http://dx.doi.org/10.2139/ssrn.2359069 .
Hutton, A. P., A. J. Marcus, and H. Tehranian (2009). Opaque financial reports, R2, and crashrisk. Journal of Financial Economics 94 (1), 67–86.
Jin, L. and S. C. Myers (2006). R2 around the world: New theory and new tests. Journal ofFinancial Economics 79 (2), 257–292.
Kagitcibasi, C. (1997). Individualism and collectivism. In J. W, B. Marshall, H. Segall, andC. Kagitcibasi (Eds.), Handbook of Cross-cultural Psychology. Boston: Allyn & Bacon.
Kashima, E. S. and Y. Kashima (1998). Culture and language: The case of cultural dimensionsand personal pronoun use. Journal of Cross-Cultural Psychology 29 (3), 461–486.
Kaufmann, D., A. Kraay, and M. Mastruzzi (2011). The worldwide governance indicators:Methodology and analytical issues. Hague Journal on the Rule of Law 3 (2), 220–246.
Kim, J.-B., Y. Li, and L. Zhang (2011a). CFOs versus CEOs: Equity incentives and crashes.Journal of Financial Economics 101 (3), 713–730.
Kim, J.-B., Y. Li, and L. Zhang (2011b). Corporate tax avoidance and stock price crash risk:Firm-level analysis. Journal of Financial Economics 100 (3), 639–662.
Kim, J.-B., Z. Wang, and L. Zhang (2016). CEO overconfidence and stock price crash risk.Contemporary Accounting Research 33 (4), 1720–1749.
Kim, J.-B. and L. Zhang (2016). Accounting conservatism and stock price crash risk: Firm-levelevidence. Contemporary Accounting Research 33 (1), 412–441.
Kothari, S. P., S. Shu, and P. D. Wysocki (2009, March). Do managers withhold bad news?Journal of Accounting Research 47 (1), 241–276.
Kwok, C. C. Y. and T. Solomon (2006). National culture and financial systems. Journal ofInternational Business Studies 37 (2), 227–247.
La Porta, R. (1998). Law and finance. Journal of Political Economy 106 (6), 1113–1155.
Li, K., D. Griffin, H. Yue, and L. Zhao (2013). How does culture influence corporate risk-taking?Journal of Corporate Finance 23, 1–22.
20
Liang Shao, Chuck CY Kwok, O. G. (2010). National culture and dividend policy. Journal ofInternational Business Studies 41 (8), 1391–1414.
Malmendier, U. and G. Tate (2005). CEO overconfidence and corporate investment. Journalof Finance 60 (6), 2661–2700.
Malmendier, U. and G. Tate (2008). Who makes acquisitions? CEO overconfidence and themarket’s reaction. Journal of Financial Economics 89 (1), 20–43.
Malmendier, U., G. Tate, and J. Yan (2011). Overconfidence and early-life experiences: Theeffect of managerial traits on corporate financial policies. Journal of Finance 66 (5), 1687–1733.
Mark, D. and J. Zhang (2014). A review of archival auditing research. Journal of Accountingand Economics 58 (23), 275–326.
Markus, H. R. and S. Kitayama (1991). Culture and the self: Implications for cognition,emotion, and motivation. Psychological Review 98 (2), 224–253.
North, D. C. (1990). Institutions, Institutional Change and Economic Performance. Cambride,UK: Cambridge University Press.
Odean, T. (1998). Volume, volatility, price, and profit when all traders are above average.Journal of Finance 53 (6), 1887–1934.
Robin, A. J. and Z. Hao (2015). Do industry-specialist auditors influence stock price crash risk?Auditing: A Journal of Practice and Theory 34 (3), 47–79.
Schrand, C. M. and S. L. C. Zechman (2012). Executive overconfidence and the slippery slopeto financial misreporting. Journal of Accounting and Economics 53 (1-2), 311–329.
Shao, L., C. C. Y. Kwok, and R. Zhang (2013). National culture and corporate investment.Journal of International Business Studies 44 (7), 745–763.
Spolaore, E. and R. Wacziarg (2009). The diffusion of development. Quarterly Journal ofEconomics 124 (2), 469–529.
Statman, M., S. Thorley, and K. Vorkink (2006). Investor overconfidence and trading volume.Review of Financial Studies 19 (4), 1531–1565.
Tang, L. and P. E. Koveos (2008). A framework to update Hofstede’s cultural value indices: Eco-nomic dynamics and institutional stability. Journal of International Business Studies 39 (6),1045–1063.
Treisman, D. (2000). The causes of corruption: A cross-national study. 76 (3), 399–457.
Williamson, O. E. (2000). The new institutional economics: Taking stock, looking ahead.Journal of Economic Literature 38 (3), 595–613.
21
App
endix
1:
Vari
able
Definit
ions
Acro
nym
Desc
ripti
on
Data
sourc
es
NC
SK
EW
The
neg
ati
ve
rati
oof
the
thir
dce
ntr
al
mom
ent
of
firm
-sp
ecifi
cw
eekly
retu
rns
over
the
sam
ple
vari
ance
rais
edto
3/2
(Chen
etal.,
2001).
Data
stre
am
DUVOL
Dow
n-t
o-u
pvola
tility
,w
hic
his
the
log
of
the
rati
oof
the
standard
dev
iati
on
on
dow
nw
eeks
toth
est
andard
dev
iati
on
on
up
wee
ks
(Chen
etal.,
2001).
Data
stre
am
CRASH
Adum
my
vari
able
equalto
1fo
ra
firm
-yea
rth
at
exp
erie
nce
sone
or
more
crash
wee
ks
duri
ng
the
fisc
al-
yea
rp
erio
d,
and
0oth
erw
ise
(Kim
etal.,
2011b).
Data
stre
am
IDV
Indiv
idualism
index
.H
ofs
tede
(2001)
BIG
4A
dum
my
vari
able
equal
to1
ifth
efirm
isaudit
edby
Big
4audit
ors
and
0oth
erw
ise.
Worl
dsc
op
eDISACC
The
lagged
thre
e-yea
rm
ovin
gsu
mof
the
abso
lute
annual
dis
cret
ionary
acc
ruals
,w
her
edis
cret
ionary
acc
ruals
are
esti
mate
dfr
om
the
modifi
edJones
model
(Dec
how
and
Slo
an,
1995).
Worl
dsc
op
e
DTURN
The
change
inav
erage
month
lysh
are
turn
over
from
yea
rt−
1to
yea
rt,
wher
em
onth
lysh
are
turn
over
isca
lcula
ted
as
the
month
lytr
adin
gvolu
me
div
ided
by
the
tota
lnum
ber
of
share
souts
tandin
gduri
ng
the
month
.
Worl
dsc
op
e
SIGM
AT
he
standard
dev
iati
on
of
firm
-sp
ecifi
cw
eekly
retu
rns
over
the
fisc
al-
yea
rp
erio
d.
Data
stre
am
RET
The
mea
nof
firm
-sp
ecifi
cw
eekly
retu
rns
over
the
fisc
al-
yea
rp
erio
d.
Data
stre
am
SIZE
The
natu
ral
log
of
the
mark
etva
lue
of
equit
ym
easu
red
at
the
end
of
the
fisc
al
yea
r.W
orl
dsc
op
eM
TB
The
rati
oof
the
mark
etva
lue
of
equit
yto
the
book
valu
eof
equit
yat
the
end
of
the
fisc
al
yea
r.W
orl
dsc
op
eLEV
The
rati
oof
long-t
erm
deb
tto
tota
lass
ets
at
the
end
of
the
fisc
al
yea
r.W
orl
dsc
op
eROA
The
rati
oof
op
erati
ng
inco
me
toto
tal
ass
ets.
Worl
dsc
op
eDisclosure
Am
easu
reofth
ele
vel
offinanci
aldis
closu
reand
availabilit
yofin
form
ati
on
toin
ves
tors
,ca
lcula
ted
usi
ng
the
surv
eyre
sult
son
the
level
and
effec
tiven
ess
of
financi
aldis
closu
refr
om
the
AnnualG
lobalC
om
pet
itiv
enes
sR
eport
sfo
r1999
and
2000.
Jin
and
Myer
s(2
006)
Anti−
SelfDea
lingIndex
The
effici
ency
of
judic
ial
syst
ems.
La
Port
a(1
998)
Creditor
Rights
Cre
dit
or
Pro
tect
ion
Index
.D
jankov
etal.
(2007)
Rule
ofLaw
The
rule
-of-
law
indic
ato
rof
Kaufm
ann
etal.
(2011),
whic
hca
ptu
res
per
cepti
ons
of
the
exte
nt
tow
hic
hagen
tshav
eco
nfiden
cein
and
abid
eby
the
rule
sof
soci
ety,
and
inpart
icula
r,th
equality
of
contr
act
enfo
rcem
ent,
pro
per
tyri
ghts
,th
ep
olice
,and
the
court
s,as
wel
las
the
likel
ihood
of
crim
eand
vio
lence
.
Kaufm
ann
etal.
(2011)
Stock
Mark
etT
he
rati
oof
stock
mark
etca
pit
aliza
tion
toG
DP
.T
he
Worl
dB
ank
GDP
The
rati
oof
GD
Pto
tota
lp
opula
tion.
The
Worl
dB
ank
GDPGrowth
Annual
GD
Pgro
wth
The
Worl
dB
ank
Con
tinen
tT
he
conti
nen
tw
her
eth
eco
untr
yis
geo
gra
phic
ally
loca
ted.
The
Worl
dB
ank
Religion
The
per
centa
ge
of
peo
ple
inth
eP
rote
stant,
Rom
an
Cath
olic,
and
Musl
imre
ligio
us
fait
hs
in1980,
from
La
Port
a(1
998).
Ale
sina
etal.
(2003)
Ethnic
Factionalization
The
deg
ree
of
ethnic
het
erogen
eity
ina
giv
enco
untr
y.A
lesi
na
etal.
(2003)
Gen
etic
Distance
To
mea
sure
agen
etic
dis
tance
from
the
Unit
edSta
tes.
Sp
ola
ore
and
Wacz
iarg
(2009)
Pronou
nDrop
Adum
my
vari
able
rela
ted
toth
ere
quir
emen
tto
use
pro
nouns
ina
language
(the
lice
nse
todro
ppro
nouns)
,w
hic
heq
uals
1if
aco
untr
ys
gra
mm
ati
cal
rule
slice
nse
per
son-i
ndex
ing
pro
noun
dro
p,
and
0oth
erw
ise.
Kash
ima
and
Kash
ima
(1998)
British
Rule
Adum
my
vari
able
,w
hic
heq
uals
1if
aco
untr
yhas
his
tori
cally
bee
nunder
Bri
tish
rule
,and
0oth
erw
ise.
Tre
ism
an
(2000)
22
Figure 1: Individualism and Crash RiskThese plots show the relationship between the country average of stock price crash risk and
individualism.
23
Table 1: Summary Statistics
This table reports summary statistics for 100,751 sample firm-year observations for 19,080 firms from 36countries over the period 2000–2009. Variables are defined in Appendix 1. The descriptive statistics arethe mean, median, standard deviation, 25th percentile, and 75th percentile.
Country ControlsDisclosure 0.932 0.103 0.873 1.000 1.000Anti− Self Dealing Index 0.607 0.188 0.483 0.651 0.651Creditor Rights 1.894 0.984 1.000 2.000 3.000Rule of Law 0.862 0.135 0.885 0.913 0.933Stock Market 1.052 0.475 0.707 1.048 1.351GDP 10.089 1.055 10.225 10.485 10.604GDPGrowth 0.070 0.087 0.027 0.064 0.115
24
Table 2: Univariate Results
This table reports the univariate results on the tests of differences in means and medians for the stockprice crash risk measures between two groups of firms formed on country-level individualism indexduring the sample period from 2000 to 2009. For each year, sample firms are assigned into either“low-individualism” or “high-individualism” groups based on the individualism index of their countries.The High-Individualism group consists of firms domiciled in countries whose individualism scores areabove the cross-sectional median score; the Low-Individualism group contains firms from countries whoseindividualism scores are below the cross-sectional median score. Variables are defined in Appendix 1.
Low- High- Difference DifferenceCrash Risk Measures Individualism Individualism in Mean in Median
This table reports the pooled OLS regressions of stock price crash risk on individualism and othercontrol variables. The dependent variable is one of NCSKEW , DUV OL, and CRASH. The variableof interest is the individualism index (IDV ). Variables are defined in Appendix 1. Robust t-statisticsin parentheses are computed using standard errors clustered at the firm level. *, **, and *** denotestatistical significance at the 10%, 5%, and 1% levels, respectively.
Industry and year fixed effects Yes Yes YesAdj. R2 0.040 0.045 0.039Obs. 100,751 100,751 100,751
26
Table 4: Instrumental Regressions
This table reports the 2SLS regressions of stock price crash risk on individualism and other controlvariables. Variables are defined in Appendix 1. All control variables are measured in year t− 1, exceptROA. Robust t-statistics use standard errors clustered at the firm level. *, **, and *** denote statisticalsignificance at the 10%, 5%, and 1% levels, respectively.
Panel A. Instruments: Genetic Distance, Pronoun Drop, and British Rule
F -test of excluded instruments (p-value) (0.00)R2 0.892 0.037 0.042 0.036Obs. 100,751 100,751 100,751 100,751
28
Tab
le5:
Hie
rarc
hic
al
Lin
ear
Regre
ssio
n
This
table
rep
ort
sth
ehie
rarc
hic
al
linea
rre
gre
ssio
ns
of
stock
pri
cecr
ash
risk
on
indiv
idualism
wit
hro
bust
standard
erro
rcl
ust
ered
at
the
firm
level
.V
ari
able
sare
defi
ned
inA
pp
endix
1.
All
contr
ol
vari
able
sare
mea
sure
din
yea
rt−
1,
exce
ptROA
.*,
**,
and
***
den
ote
stati
stic
al
signifi
cance
at
the
10%
,5%
,and
1%
level
s,re
spec
tivel
y.
NCSKEW
DUVOL
CRASH
−firmdev
−ctrymea
n−ctry
−firmdev
−ctrymea
n−ctry
−firmdev
−ctrymea
n−ctry
BIG
4–0.0
18***
0.0
28
–0.0
11***
0.0
15
–0.0
05**
0.0
27*
(–2.7
2)
(0.6
8)
(–3.2
1)
(0.7
5)
(–2.0
4)
(1.8
7)
DISACC
0.0
04**
–0.0
93***
0.0
03**
–0.0
44***
0.0
01
0.0
51***
(1.9
6)
(–3.9
0)
(2.5
4)
(–3.9
9)
(0.8
1)
(6.5
2)
DTURN
0.0
83***
0.6
60***
0.0
35**
0.4
19***
0.0
42***
0.0
74
(2.9
5)
(3.5
5)
(2.3
6)
(4.5
2)
(3.6
6)
(1.1
0)
NCSKEW
0.0
65***
0.3
58***
0.0
34***
0.1
82***
0.0
15***
0.0
24**
(8.6
4)
(7.0
0)
(10.1
8)
(7.8
2)
(8.8
9)
(2.0
0)
SIGM
A–0.2
53*
1.4
88**
–0.3
54***
0.4
63
–1.3
71***
–1.4
05***
(–1.7
7)
(2.3
7)
(–4.8
1)
(1.5
8)
(–26.3
5)
(–7.2
6)
RET
4.5
16***
23.4
23***
2.4
91***
12.9
58***
1.5
58***
0.1
00
(16.2
8)
(5.8
1)
(17.1
0)
(6.5
1)
(14.5
4)
(0.0
7)
MCAP
0.0
05**
0.0
27***
–0.0
01
0.0
11***
–0.0
08***
0.0
01
(2.4
0)
(3.3
3)
(–0.6
1)
(2.8
3)
(–10.8
6)
(0.4
7)
MB
0.0
00
0.0
33***
0.0
00
0.0
19***
0.0
00
–0.0
05**
(0.6
7)
(4.2
7)
(0.8
1)
(5.1
0)
(1.1
3)
(–1.9
9)
LEV
0.0
68***
0.5
23***
0.0
36***
0.2
45***
0.0
03
–0.1
14***
(6.1
9)
(4.7
2)
(6.1
8)
(4.4
9)
(0.8
0)
(–2.9
4)
ROA
–0.0
17***
–0.2
46***
–0.0
11***
–0.1
60***
0.0
02
–0.1
73***
(–2.9
0)
(–3.0
2)
(–3.4
4)
(–4.1
3)
(0.9
8)
(–6.3
3)
IDV
0.0
47***
0.0
11***
0.0
40***
(3.1
9)
(3.5
6)
(3.1
0)
Disclosu
re
–0.1
00**
–0.0
68***
–0.0
17
(–1.9
9)
(–2.7
4)
(–0.8
8)
Anti−
SelfDea
lingIndex
0.1
18***
0.0
76***
0.0
25**
(4.1
5)
(5.4
5)
(2.3
4)
Cred
itorRights
–0.0
15**
–0.0
08***
–0.0
11***
(–2.4
2)
(–2.8
7)
(–4.9
8)
Rule
ofLaw
–0.1
23
–0.0
13
–0.0
25
(–0.8
6)
(–1.0
2)
(–0.4
2)
Stock
Market
–0.0
33***
–0.0
20***
0.0
03
(–4.0
4)
(–4.8
0)
(0.8
0)
GDP
0.0
55***
0.0
31***
–0.0
05
(5.2
7)
(6.3
1)
(–1.6
3)
GDPGrowth
–0.1
49***
–0.1
08***
–0.0
55***
(–3.5
6)
(–5.3
1)
(–3.6
7)
Ad
j.R
20.0
43
0.0
48
0.0
43
Ob
s.100,7
51
100,7
51
100,7
51
29
Table 6: Additional Robustness Tests
This table reports robustness tests. Panel A shows the regression results using alternative proxies forindividualism. Panel B reports the regressions controlling for additional factors. Variables are definedin Appendix 1. All control variables are measured in year t − 1, except ROA. Robust t-statistics inparentheses are based on the standard errors clustered at the firm level. *, **, and *** denote statisticalsignificance at the 10%, 5%, and 1% levels, respectively.
Industry and year FE Yes Yes Yes Yes Yes YesAdj. R2 0.041 0.046 0.040 0.042 0.047 0.042Obs. 100,751 100,751 100,751 100,751 100,751 100,751
31
Table 7: Possible Mechanisms – Overconfident Traders
This table reports the regression results for the tests on whether trading by overconfident traders explainsthe effect of individualism on stock price crash risk. Variables are defined in Appendix 1. All controlvariables are measured in year t − 1, except ROA. Robust t-statistics in parentheses are based on thestandard errors clustered at the firm level. *, **, and *** denote statistical significance at the 10%, 5%,and 1% levels, respectively.