Culture vs Corp. Risk-taking

8/4/2019 Culture vs Corp. Risk-taking

1/47

How Does Culture Influence Corporate Risk-Taking?*

Kai LiSauder School of Business

University of British Columbia2053 Main Mall, Vancouver, BC V6T 1Z2

[email protected]

Dale GriffinSauder School of Business

University of British Columbia2053 Main Mall, Vancouver, BC V6T 1Z2

[email protected]

Heng YueGuanghua School of Management

Peking UniversityBeijing 100871 P.R. China

[email protected]

Longkai ZhaoGuanghua School of Management

Peking UniversityBeijing 100871 P.R. China

[email protected]

This version: July, 2011

Abstract

We investigate the role of national culture in corporate risk-taking. First, we postulate thatmanagerial risk-taking is influenced directly by culture through its effect on individual decision-making, and indirectly through its effect on a countrys formal institutions and a firmsorganizational structure. Second, we postulate that the influence of culture is conditioned oncertain firm environments. Using firm-level data from 35 countries and employing a hierarchicallinear modeling approach to isolate the effects of firm-level and country-level variables, we showthat individualism has positive and significant direct effects, whereas uncertainty avoidance hasnegative and significant direct effects on corporate risk-taking. The economic significance

analysis of the total effects suggests that culture has significant explanatory power in corporaterisk-taking. Greater earnings smoothing exacerbates the effect of individualism on corporate risk-taking, and larger firm size weakens the effects of individualism and uncertainty avoidance oncorporate risk-taking.

Keywords: decision-making; formal institutions; individualism; national culture; risk-taking;uncertainty avoidanceJEL Classification: G32; G11; G18

* We thank Jan Bena, Craig Doidge, Huasheng Gao, Lubo Litov, and Pedro Matos for kindly sharing theirdata. We thank Andy Chui, Amir Licht, Alexander Ljungqvist, Maria-Teresa Marchica, Michael Meloche,Fatma Sonmez Saryal, Yangru Wu, seminar participants at Fudan University, Peking University, Shanghai

Jiaotong University, Sun Yat-Sen University, and Tsinghua University, and conference participants at theFive Star Finance Forum (Beijing), the China International Conference in Finance (Guangzhou), theSecond Finance Research Summer Camp of Cheung Kong GSB (Guilin), the UBC Finance SummerResearch Conference (Kelowna), and the Northern Finance Association Meetings (Niagara-on-the-Lake)for their helpful comments, and Huasheng Gao for research assistance. This paper is the recipient of theBest Paper Award on Capital Market Research sponsored by the Toronto CFA Society at the NorthernFinance Association Meetings. Li and Griffin acknowledge financial support from the Social Sciences andHumanities Research Council of Canada. Zhao acknowledges financial support from the National NaturalScience Foundation of China (Approval numbers 70873002 and 70932002). We are responsible for allerrors.


2/47

0

How Does Culture Influence Corporate Risk-Taking?

Abstract

We investigate the role of national culture in corporate risk-taking. First, we postulate that managerialrisk-taking is influenced directly by culture through its effect on individual decision-making, andindirectly through its effect on a countrys formal institutions and a firms organizational structure.Second, we postulate that the influence of culture is conditioned on certain firm environments. Usingfirm-level data from 35 countries and employing a hierarchical linear modeling approach to isolate theeffects of firm-level and country-level variables, we show that individualism has positive and significantdirect effects, whereas uncertainty avoidance has negative and significant direct effects on corporate risk-taking. The economic significance analysis of the total effects suggests that culture has significantexplanatory power in corporate risk-taking. Greater earnings smoothing exacerbates the effect ofindividualism on corporate risk-taking, and larger firm size weakens the effects of individualism anduncertainty avoidance on corporate risk-taking.

Keywords: decision-making; formal institutions; individualism; national culture; risk-taking; uncertaintyavoidanceJEL Classification: G32; G11; G18


3/47

1

1. Introduction

Is corporate risk-taking affected by a firms national environment? Cultural theorists suggest that cultural

background may have an inescapable influence on decisions of every kind including decisions about risk

(e.g., Hofstede (2001), and House, Hanges, Javidan, Dorfman, and Gupta (2004)). Consequently, we

expect that a firms culture of origin will affect its appetite for risk, and that firms operating in different

countries will differ systematically and predictably on their level of corporate risk-taking.

In this paper, we examine the influence of national culture on corporate risk-taking. Our paper

makes the following important contribution to the management literature. First, we postulate specific

economic and psychological channels through which culture exerts its influence on risky corporate

decision-making. Second, by including some of the above channels in our empirical specification, we

capture both the direct and indirect influences of culture on risky corporate decision-making. Finally, we

establish boundary conditions on the influence of culture in corporate decisions by testing whether

cultural effects are most apparent in certain firm environments. Our paper reinforces the growing

awareness among management scholars that informal institutions such as culture matter in corporate

decisions even when those decisions are made by sophisticated professional managers (e.g., Weber,

Shenkar, and Raveh (1996), Weber and Camerer (2003),and Shao, Kwok, and Guedhami (2010)).

Using cultural values developed by Hofstede (1980, 2001), we examine how between-country

differences in adherence to the cultural values of individualism (versus collectivism), and uncertainty

avoidance affect corporate risk-taking. Cultures high on individualism emphasize individual freedom and

achievement, whereas cultures low on individualism emphasize strong group cohesion. Cultures high on

uncertainty avoidance shun ambiguous situations and prefer clear rules of conduct, whereas cultures low

on uncertainty avoidance enjoy novel events and value innovation. We hypothesize that there is a positive

relation between national individualism and corporate risk-taking, and a negative relation between

national uncertainty avoidance and corporate risk-taking. Furthermore, we expect culture to affectcorporate risk-taking both directly through its effect on individual decision-making, e.g., managerial

attitude to risky outcomes, and indirectly through its effect on a countrys formal institutions, e.g.,

investor protection and rule of law, and a firms organizational structure, e.g., equity-based managerial


4/47

2

compensation. Finally, we expect that the influence of culture is conditioned on certain firm

environments, e.g., earnings smoothing and firm size.

We empirically examine these hypotheses using industrial firms from 35 countries covered by the

Compustat Global Vantage database over the period 1997-2006. Using the standard deviation of operating

income as our primary measure, and R&D expenditures and the level of long-term debtboth normalized

by total assetsas our two other measures of corporate risk-taking, we show that there is a positive

association between individualism and risk-taking, and a negative association between uncertainty

avoidance and risk-taking, controlling for firm-level and country-level characteristics, including measures

of formal institutions. Furthermore, we demonstrate that culture influences formal institutional

development which in turn influences risky corporate decision-making. These indirect effects sometimes

are consistent in sign and reinforcing, and sometimes are opposite in sign and offsetting the direct effects

of culture. The total effects of each cultural value on corporate risk-taking are in the hypothesized

direction and of economic importance. Finally, we show that the influence of culture is conditioned on

certain firm environments: Greater earnings smoothing exacerbates the effect of individualism on

corporate risk-taking, and larger firm size weakens the effects of individualism and uncertainty avoidance

on corporate risk-taking.

We implement additional investigations to ensure our results are robust to different measures of

culture, model specifications, and samples. First, using Schwartzs (1994) culture value of harmony, we

show that there is a negative and significant association between harmony and corporate risk-taking.

Second, we add a set of religion measures, and show that the relation between cultural values and

corporate risk-taking is unaffected. Third, we employ a subsample with a minimum number of firm

observations in each country, and our main results remain. Fourth, we remove the two countries with the

largest number of sample firms, Japan and the US, and conclude that our key findings are not driven by

firms from these two countries. Finally, we remove multinational or cross-listed firms from our main

sample and find that our key findings remain unchanged.

The remainder of the paper is structured as follows. We review the related literature and develop

our hypotheses in the next section. Section 3 describes key variable construction and our sample. Section


5/47

3

4 presents the empirical specification. Section 5 presents our main results and provides some

interpretation. Section 6 implements various robustness checks on our main results. Section 7 concludes.

2. Literature Review and Hypothesis Development

Prior Literature

Our study builds upon two strands of literature: the literature examining the role of culture in

business and management, and the literature examining cross-country corporate risk-taking.

There is a substantial literature examining whether differences in national culture can help

explain cross-country differences in various management practices. For example, researchers have shown

significant association between culture and corporate compensation practices (Schuler and Rogovsky

(1998)), capital structure decisions (Chui, Lloyd, and Kwok (2002), and Li, Griffin, Yue, and Zhao

(2011)), disclosure (Hope (2003)), earnings management (Han, Kang, Salter, and Yoo (2010)),

entrepreneurship (Kreiser, Marino, Dickson, and Weaver (2010)), dividend policy (Shao, Kwok, and

Guedhami (2010)), mergers and acquisitions and foreign direct investments (Siegel, Licht, and Schwartz

(2010, 2011)), bank lending (Giannetti and Yafeh (2010), and Zheng, Ghoul, Guedhami, and Kwok

(2011)), and venture capital activity (Li and Zahra (2011)).

Our paper is also related to the small strand of literature examining corporate risk-taking in an

international setting.1 Claessens, Djankov, and Nenova (2000) employ 12 indicators to capture corporate

risk-taking including measures of cash flow risk, financial leverage, and liquidity. They show that

companies in common law countries, countries with stronger protection of property rights, and those in

market-based financial systems take less risk. John, Litov, and Yeung (2008) focus on the relation

between investor protection and corporate risk-taking. Using a large panel of manufacturing companies

from 39 countries, they show significant positive associations between measures of investor protection

and risk-taking, and between risk-taking and growth. Laeven and Levine (2009) examine risk taking by

banks, their ownership structures, and national bank regulations. They find that banks with more powerful

1 There is a much larger literature examining risk-taking by US firms, for example, Rajgopal and Shevlin (2002),

Coles, Daniel, and Naveen (2006), Sanders and Hambrick (2007), Hilary and Hui (2009), Low (2009), and

Bargeron, Lehn, and Zutter (2010).


6/47

4

(controlling) owners tend to take greater risks, consistent with theories predicting that shareholders have

stronger incentives to increase risk than non-shareholding managers and creditors. Acharya, Amihud, and

Litov (2010) show that having strong creditor rights in a country leads firms to reduce risk, measured by

undertaking diversifying acquisitions, lower cash-flow risk, and lower leverage.

Our study contributes to the international corporate risk-taking literature in several important

ways. First, we identify three channels through which culture influences firm-level risk-taking, and

provide new evidence on the significant role of culture in risky corporate decisions. Second, we identify

important conditioning variables that limit or accentuate the influence of culture on firm-level risk-taking.

Finally, we employ a hierarchical linear modeling approach to isolate the effects of firm-level and

country-level variables in our empirical analysis.

Our Hypotheses

Our fundamental proposition is that national culture operates both directly on corporate risk-

taking and indirectly through firm-level and country-level characteristics. For our measures of cultural

values, we use the most well-studied indicators: Hofstedes (1980, 2001) individualism and uncertainty

avoidance (e.g., Kwok and Tadesse (2006), Beckmann, Menkhoff, and Suto (2008), Chui and Kwok

(2008), Chui, Titman, and Wei (2010), Han et al. (2010), Kreiser et al. (2010), and Li and Zahra (2011)).

Figure 1 outlines our thought experiment regarding the directional effects of our two cultural

values and their channels of influence on risky corporate decisions. We distinguish three levels of

influence: country level (national formal institutions including laws, regulations, and market

development), firm level (corporate organization structures including compensation practices and

ownership), and individual level (managerial attitude to risk including focus on payoffs versus risk and

subjective discount rates).

Country-level influences

Countries differ in the type and level of formal institutions they provide to regulate and facilitate

financial risk-taking (La Porta, Lopez-de-Silanes, Shleifer, and Vishny (1997, 1998)). Formal institutions

such as investor protection, rule of law, and market development have been shown to affect the level of

corporate risk-taking within a country (e.g., John et al. (2008), and Acharya et al. (2010)). Following


7/47

5

Licht, Goldschmidt, and Schwartz (2005, 2007), and Kwok and Tadesse (2006), we argue that these

formal institutions are shaped by national cultural values including individualism and uncertainty

avoidance.

Because individualistic societies emphasize individual freedom, autonomy, and self-interested

competition, they require formal institutions that protect the rights of competing parties (Licht et al.

(2005)). In contrast, collectivist societies emphasize strong informal ties among in-groups and rely on

informal networks and relationships rather than formal institutions to protect against opportunism (Li and

Zahra (2011), see path (1) in Figure 1 Panel A).

Because uncertainty avoidant societies emphasize social conformity and rule-following, they are

less comfortable with market-based financial systems that are characterized by uncertainty and ambiguity

(Kwok and Tadesse (2006)). In contrast, low uncertainty avoidant societies are more comfortable with

unpredictable outcomes and therefore accepting market-based financial systems (Kwok and Tadesse

(2006), Beckmann et al. (2008), and Li and Zahra (2011), see path (1) in Figure 1 Panel B).

Based on the above discussion, we expect that individualism is positively, and uncertainty

avoidance is negatively, related to formal institutional development, especially market-based financial

systems, which in turn encourages corporate risk-taking.

Countries also differ in the relative protection they offer to creditors and shareholders (La Porta et

al. (1997, 1998), and Djankov, La Porta, Lopez-de-Silanes, and Shleifer (2008)). This difference is

relevant to corporate risk-taking because in a modern corporation with limited liabilities, there are

intrinsic conflicts of interest between creditors and shareholders, with the former receiving fixed payoffs

and hence being risk-averse, and the latter enjoying the upside potential and hence being relatively risk-

seeking (Galai and Masulis (1976), and Myers (1977)). Given that high uncertainty avoidant societies are

uncomfortable with market-based financial systems, we expect less shareholder protection, which in turn

discourages corporate risk-taking (John et al. (2008), see path (2) in Panel B).

Firm-level influences

Standard agency theories suggest that managers are more risk-averse than shareholders because

of their career concerns and lack of diversification (Jensen and Meckling (1976), and Demsetz and Lehn

(1985)). In response, many corporate boards have instituted equity-based compensation schemes to align


8/47

6

the risk preferences of managers and shareholders (Jensen and Meckling (1976), and Haugen and Senbet

(1981)). A number of US-based studies have demonstrated that equity-based pay encourages managerial

risk-taking (e.g., Rajgopal and Shevlin (2002), Coles et al. (2006), Sanders and Hambrick (2007), and

Low (2009)).2 Individualism, with its emphasis on individual freedom and self-interest, is consistent with

the practice of equity-based managerial compensation (Schuler and Rogovsky (1998), see path (2) in

Panel A). Uncertainty avoidance, in contrast, with its emphasis on avoiding situations characterized by

uncertainty and ambiguity, is inconsistent with this practice (Schuler and Rogovsky (1998), see path (3)

in Panel B). Thus, we expect this firm-level compensation practice to be more common in high

individualism and low uncertainty avoidance countries, which in turn encourages corporate risk-taking.

Individual-level influences

Psychological research has identified a number of reasons for expecting differences in risk-taking

between national cultures. Individualistic societies emphasize primacy of the individual over the group,

which has two important implications for risk-taking. Gelfand et al. (2002, p. 835) identify these two

aspects of individualism as follows (italics added): The self is served in individualistic cultures by being

distinct from and better than others, in order to accomplish the culturally mandated task of being

independent and standing out.

First, individualistic managers are free to make risky decisions using their own judgment (Kreiser

et al. (2010)), and are motivated to stand out from other managers to demonstrate their autonomy. Thus,

these managers will seek out innovative risky projects that will demonstrate their unique self-image. This

leads to the decision rules that overweight risky payoffs, relative to less individualistic managers (see path

(3) in Panel A).

Second, individualistic managers, because of their self-enhancing belief that they are more skilled

and have a higher level of outcome control than other managers (Sedikides, Gaertner, and Toguchi (2003),

and Yamaguchi, Gelfand, Ohashi, and Zemba (2005)), underestimate the level of uncertainty in risky

decisions. Thus, these managers will accept innovative risky projects due to their belief that they can

2 There is almost no large-scale cross-country comparison of compensation practices across countries due to data

limitation. Fernandes, Ferreira, Matos, and Murphy (2010) is a notable exception with pay data for top executives in

14 countries for the year 2006.


9/47

7

control such projects. This leads to the decision rules that require a lower risk premium (i.e., a lower

discount rate) for risky projects, relative to less individualistic managers (see path (4) in Panel A).

Due to the influences of both managerial autonomy and self-enhancement, we expect that

individualism is directly and positively related to corporate risk-taking.

Hofstede (1991, p. 112) defines uncertainty avoidance as the extent to which the members of a

culture feel threatened by uncertain or unknown situations. In a later paper, Hofstede (2001, p. 148)

notes that uncertainty avoidance does not equal risk avoidance. Low uncertainty avoidant managers are

comfortable with the unpredictability and ambiguity inherent in innovative projects, for which there are

no rules to fall back on. In contrast, high uncertainty avoidant managers feel anxious in the presence of

uncertainty and ambiguity and hence either will avoid innovative projects or will require a higher risk

premium (i.e., a higher discount rate, Li and Zahra (2011), see path (4) in Panel B). Thus, we expect that

uncertainty avoidance is directly and negatively related to corporate risk-taking.

In summary, our first two hypotheses regarding the effects of cultural values on corporate risk-

taking are as follows:

H1: There is a positive association between national levels of individualism and firm-level corporate

risk-taking.

H2: There is a negative association between national levels of uncertainty avoidance and firm-level

corporate risk-taking.

Boundary conditions

Previous studies have shown that formal institutions and culture operate both independently and

in combination on management decisions. Hope (2003) shows that Hofstedes cultural values behave

differently across different legal regimes (common law versus civil law) when the dependent variable is

firm-level disclosure quality. Han et al. (2010) find that the interaction terms between investor protection

and individualism, and between investor protection and uncertainty avoidance, are positively associated

with the magnitude of earnings discretion. Li and Zahra (2011) show that formal institutions have a

positive effect on venture capital activity, but this effect is weakened in high uncertainty avoidant

societies and in more collectivist societies.


10/47

8

In settings where managers have greater discretion, we expect a stronger role of cultural values in

corporate risk-taking. By contrast, in settings where managers are more constrained by rules, we expect a

limited role of cultural values in corporate risk-taking. Specifically, we expect that because firms engaged

in extensive earnings smoothing have greater managerial discretion (Han et al. (2010)), such firms should

show a stronger effect of cultural values on risk-taking. Further, we expect that cultural values will have

less scope for influence in large firms because large firms rely more on highly controlled management

systems.

Thus, our third and fourth hypotheses regarding the moderating effects of certain firm

environments on the relation between cultural values and corporate risk-taking are as follows:

H3: The influences of culture are strengthened in firms actively engaged in earnings smoothing.

H4: The influences of culture are weakened in large firms.

3. Key Variable Construction and Our Sample

Measures of Cultural Values

The two cultural values we use in this paper are constructed from the world-wide sample obtained

by Hofstede (1980, 2001). Appendix I provides a detailed discussion of the construction of both indices.

It is noteworthy that the specific items defining each cultural value are highly distinct from the context of

corporate decision-making we study in this paper. For example, the most heavily weighted item in

constructing the uncertainty avoidance index is Competition between employees usually does more harm

than good. This, like other items in the index, represents an abstract guideline for appropriate behavior

and is not directly translatable into making risky corporate decisions. Nonetheless, each of our two

cultural values has a natural interpretation in terms of promoting or restraining risk-taking behavior in a

corporate context, as described in our hypothesis development section above.

The Hofstede value dimensions were derived from a sample of IBM employees in the 1970s, well

before the beginning of our sample periodreducing endogeneity concerns. Any changes in cultural

values that have occurred over the past forty years would weaken our conjectured linkage between

cultural values and corporate decisions. Similarly if IBM employees do not share the same cultural values

as corporate managers, this would also weaken the conjectured linkage between culture and corporate


11/47

9

risk-taking. Finding robust effects of cultural values on corporate risk-taking would support the general

thesis that cultural values are enduring norms or guidelines that are widely shared within nations.

Measures of Corporate Risk-Taking

Motivated by prior work, we employ three measures of corporate risk-taking. Std(ROA) is the

degree of risk-taking in firms operations based on the volatility of corporate earningsriskier corporate

operations lead to more volatile earnings (John et al. (2008), and Acharya et al. (2010)). For each firm

with available earnings and total assets for at least five years over the 1997-2006 period, we compute the

deviation of the firms EBITDA/total assets from the country average for the corresponding year. We then

calculate the standard deviation of this measure for each firm.

Our second measure is R&D, which is commonly employed as a measure of risky corporate

policies (Bhagat and Welch (1995), Coles et al. (2006), and Bargeron et al. (2010)). R&D investments are

risky because they have a low probability of success and their benefits are distant and uncertain (Palmer

and Wiseman (1999)). R&D is constructed as the average ratio of R&D expenditures over total assets for

the period 1997-2006.

Our third measure of risk-taking is the amount of long-term debt (LTD) a company takes on

(Claessens et al. (2000), and Li et al. (2011)). Ceteris paribus, companies with high levels of long-term

debt will find themselves pressured for interest payments, and will suffer from greater interest rate risk,

too little working capital, and ultimately, bankruptcy risk. LTD is constructed as the average ratio of long-

term debt over total assets for the period 1997-2006.

In brief, our first measure of risk-taking captures the overall risk taken by the firm; R&D captures

risk-taking in long-term corporate investment decisions; and LTD captures risk-taking in long-term

financing decisions; with each firm providing a single value over the ten-year sample period.

Measures of Investor Protection and Institutional Development

To characterize investor protection in each country, we use four measures. The anti-self-dealing

index is a measure of legal protection of minority shareholders against expropriation by corporate insiders.

This measure is constructed by Djankov et al. (2008) with a focus on private enforcement mechanisms


12/47

10

such as disclosure, approval, and litigationthat govern a specific self-dealing transaction.3 Rule of law

is an indicator of the effectiveness of regulatory enforcement. The data on rule of law are taken from La

Porta et al. (1998). High accounting disclosure standards lead to better investor protectionthey make

the diversion of corporate resources more difficult. The disclosure data are also retrieved from La Porta et

al. (1998), who tabulate the original data from the Center for International Financial Analysis and

Research. Creditor rights is the sum of four provisions: the absence of automatic stay in reorganization,

the requirement for creditors consent or the minimum dividend required for a debtor to file for

reorganization, the ranking of secured creditors first in reorganization, and the removal of incumbent

management upon filing for reorganization. The data on creditor rights are taken from La Porta et al.

(1998), who record creditor rights provisions across 49 countries.

Finally, we include three indicators of institutional development: country private credit, country

market capitalization, and country GDP growth volatility. Country private credit is constructed as the

average ratio of the value of credits by financial intermediaries to the private sector to GDP for the period

1997-2006. Country market capitalization is constructed as the average ratio of stock market

capitalization to GDP for the period 1997-2006. Country GDP growth volatility is the standard deviation

of annual GDP growth rates for the period 1997-2006. All these three variables are obtained from the

World Banks World Development Indicators. Appendix I provides a detailed description of all variables.

Sample Overview

Our main data source is the Compustat Global Vantage database for the period 1997-2006. The

sample is chosen based on the requirement that firm-level data are available to compute our risk-taking

measures and country-level data are available for cultural values and key country characteristics. To

remove the effect of outliers, we winsorize all firm-level variables at the one percent level in both tails of

the distribution. Our final sample consists of 7,250 firm-level observations from 35 countries. Appendix

II lists the two-digit SIC industries (SIC codes 2000-3999) covered by our sample and presents the

industry distribution across sample countries to provide insight into industry concentration. In all our

3 It is worth noting that our main result on shareholder protection is not affected if we use La Porta et al.s (1998)anti-director rights index. Further, all four investor protection variables are directly associated with a countrys legalorigin as shown by La Porta et al. (1998). To avoid multicollinearity, we opt not to include the legal origin variablesin our model specification.


13/47

11

model estimation we include two-digit industry fixed effects to control for cross-country differences in

industry concentration and their effects on corporate risk-taking.

Table 1 Panel A provides a summary of our sample. The number of firms included per country

varies from 10 firms (Peru) to 1,659 (Japan). We show that among our 35 countries, the three countries

with the highest score on individualism are: US (0.91), Australia (0.90), and UK (0.89); while the three

countries with the lowest score on individualism are: Peru (0.16), Taiwan (0.17), and South Korea (0.18).

The three countries with the highest score on uncertainty avoidance are: Greece (1.12), Portugal (1.04),

and Belgium (0.94); while the four countries with the lowest score on uncertainty avoidance are:

Singapore (0.08), Denmark (0.23), Hong Kong (0.29), and Sweden (0.29).

Table 1 Panel B provides the summary statistics for firm-level variables. The mean (median)

values for our three risk-taking measuresstd(ROA), R&D, and LTDare 6.20 percent (4.68 percent),

2.83 percent (0.58 percent), and 12.46 percent (9.72 percent), respectively.

Table 1 Panel C presents the Pearson correlations between the dependent and independent

variables using firm-level observations. Among our three measures of corporate risk-taking, we show that

there is a positive correlation between the standard deviation of operating income and R&D, a small

positive correlation between the standard deviation of operating income and long-term debt, and a small

negative correlation between long-term debt and R&D. Because each measure of risk-taking assesses

distinct and potentially independent corporate decisions, we do not require strong correlation between

these measures. Between our two cultural value variables, we show that there is a negative correlation

between individualism and uncertainty avoidance.

Among our seven formal institution variables, we find significant pair-wise correlations among

almost all pairs. In particular, there is a strong positive correlation between the anti-self-dealing index and

creditor rights, a negative correlation between rule of law and GDP growth volatility, and a positive

correlation between disclosure and market capitalization.4

Between the two cultural values and seven formal institution variables, we first show that there

are significant pair-wise correlations among all pairs. Most notably, there is a strong positive correlation

4 In unreported analysis, we include one variable out of the above three pairs and find that the significant effects ofculture remain.


14/47

12

between individualism and rule of law, and a negative correlation between individualism and GDP

growth volatility. There is a strong negative correlation between uncertainty avoidance and disclosure,

and between uncertainty avoidance and market capitalization. The significant correlations between

cultural values and measures of formal institutions are consistent with Licht et al. (2005, 2007), who

argue that a countrys corporate governance and legal standards reflect its national culture background.

4. Our Empirical Specifications

Multilevel Data and Hierarchical Linear Models

Our data structure is multilevel. At the country level, we have firms from 35 different countries.

At the firm level, we have over 7,000 firms. From a modeling perspective, it is important to distinguish

the effects that take place at the country level from those that take place at the individual firm level, both

to understand the role of country- versus firm-level determinants, and to appropriately model their

interactions.

We employ a hierarchical nested form of the general linear model to explore our multilevel data

(see Raudenbush and Bryk (2002) or Goldstein (2003) for an introduction to hierarchical linear models

(HLM)). In our data, the set of firms within countries form the base-level observations and countries serve

as the higher-level observations.

There are three distinct benefits from using an HLM in our setting. First, the HLM framework

using a country mean-centered approach to firm-level variables cleanly separates the variance in firm-

level risk-taking determined by the firm- versus country-level explanatory variables.

Second, the HLM framework corrects for the distortion introduced by varying sample sizes across

countries. Under the OLS specification, it is common for the coefficient on a country-level variable to be

spuriously significant just because of the large sample size at the firm level. This problem is accentuated

when countries differ markedly in the number of firms they contribute to the sample. Unlike the OLS

regression where each firm-level observation receives equal weight, the HLM regression simultaneously

models regressions at both the firm-level and the country-level; with the country-level regression


15/47

13

weighted by the precision of the firm-level data, which is inversely related to the sample size within a

country.

Third and finally, the HLM framework accurately incorporates cross-level interactions between

the firm- and country-level variables. The power of HLMs comes from their ability to correctly pool firm-

level effects across countries while also examining country-level relations.

In summary, compared to the typical OLS regressions used to examine cross-country firm-level

data, the HLM analyses employed in this paper will provide insight into whether the effects obtained in

the cross-sectional regressions take place at the firm- or the country-level, and will also provide a stricter

and more appropriate test of the relevant coefficients.

Mean-Centering the Data

We pre-process the data to help decompose the country- and firm-level variations in corporate

risk-taking. First, we center every independent variable by its grand mean (averaged across firms and

countries), so that every transformed variable has a mean of zero. Second, we create country-level mean

values (averaged within a country) on those grand-mean-centered variables that vary by firm 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 1 and subtracting the corresponding within-country means in step

2. We name these firm-level deviations separately from their corresponding country-level means by

adding the suffix _firmdev. By centering the firm-level variables within-country and adding the

country-level means to the set of predictors, we completely separate the covariances within- and between-

country. Furthermore, this decomposition allows us to explore the potentially differential effects of firm

characteristics such as firm size at the (individual) firm- and (average) country-level. Finally, using mean-

centered independent variables makes estimation of the interactions more efficient, and also makes

interpretation of the intercept values clear: the expected value of the dependent variable when all

independent variables are at their means. In the end, our model specification contains some variables that

have only country-level values (such as cultural values and investor protection variables), and others that

have country-level and firm-level values (such as earnings smoothing and firm size); where the country-

level values are all grand mean-centered and the firm-level values are all country mean-centered.


16/47

14

Model Specification

To explore the relation between cultural values and corporate risk-taking, we regress firm-level

observations of risk-taking in corporate operations on variables that capture firm characteristics,5 country-

level cultural values, and country-level investor protection and institutional development variables. Our

HLM specification is as follows:

._

__Pr

__

_

__

__

__

__

_

_-

,18

1615

1413

1211

109

,8,7

,6,5

,4,3

,2

,1,

jij

jj

jj

jj

jj

jiji

jiji

jiji

ji

jiji

ectryVolatilityGrowthGDPCountry

ctrytionCapitalizaMarketCountryctryCreditivateCountry

ctryRightsCreditorctryDisclosure

ctryLawofRuleIndex_ctrydealingAnti-self-

ctryAvoidanceyUncertaintctryismIndividual

ctrymeanGrowthSalesInitialctrymeanEarningsInitial

ctrymeanSizeFirmInitialctrymeanSmoothingEarnings

firmdevGrowthSalesInitialfirmdevEarningsInitial

firmdevSizeFirmInitial

firmdevSmoothingEarningsMeasuretakingRisk

(1)

For firm i from countryj,Risk-taking Measure can be std(ROA), R&D, or LTD. Departing from

the standard OLS regression, we decompose firm characteristics into firm-level deviations and country-

level means to help understand the differential firm- and country-level effects, and to more appropriately

estimate the significance of the country-level culture and formal institution variables (which vary only

across countries). This model is estimated using an iterative maximum likelihood fitting procedure

available in the MLWin program.

Finally, to capture the conditioning effects of earnings smoothing and firm size on the relation

between cultural values and corporate risk-taking, we add four interaction terms to Equation (1): Each of

the two cultural values interacts with earnings smoothing and firm size measured as firm-level deviations.

5. Main Results

Multivariate Tests

5 Initial values are used for firm size, earnings, and sales growth. This mitigates the endogeneity concern that firm-

level characteristics and corporate risk-taking are jointly determined (John et al. (2008)).


17/47

15

Table 2 presents the estimation results.6 For comparison, we present the OLS regression results in

addition to our main results under the HLM specification as in Equation (1). When std(ROA) is the

dependent variable, we show that three out of the four firm characteristics, with the exception of initial

sales growth, measured at both the firm-level deviation and the country-level are negatively and

significantly associated with risky corporate decisions, implying that these effects hold both when

comparing across firms and when comparing across countries. The effects are expected: Greater earnings

smoothing is associated with lower variability of accounting returns, and larger firms and firms with

higher earnings are associated with lower operating risks.7

The coefficients on the two cultural value variables are all significant and with the predicted sign:

Individualism is positively and significantly associated with firm-level risk-taking, and uncertainty

avoidance is negatively and significantly associated with firm-level risk-taking, consistent with our

hypotheses H1 and H2.8 Out of seven formal institution variables, there is a negative and significant

association between rule of law and firm-level riskiness, and a positive and significant association

between GDP growth volatility and firm-level riskiness.

The economic impact of our cultural values on corporate risk-taking is noteworthy: A one

standard deviation increase in individualism increases the risk-taking proxy by 21.6% of its mean; and a

one standard deviation increase in uncertainty avoidance decreases the risk-taking proxy by 10.0% of its

mean. In contrast, a one standard deviation increase in rule of law (GDP growth volatility) decreases

(increases) the risk-taking proxy by 5.7% (7.9%) of its mean.

When R&D is the dependent variable, we show that among the firm characteristics, both earnings

smoothing and initial earnings measured at both the firm-level deviation and the country-level mean are

negatively and significantly associated with R&D. Thus, a firm with greater earnings smoothing (higher

earnings) than another would be expected to have lower R&D expenditures, and a country with higher

6

All our main results are based on manufacturing firms (SIC codes 2000-3999). It is worth noting that in unreportedanalysis, we show that the significant association between cultural values and corporate risk-taking remains usingfirms from all industries (results available upon request).7 Han et al. (2010) show that cultural values directly affect the extent of earnings management. By includingearnings management in our basic model specification, we focus on the direct effects of cultural values on risk-taking, allowing for their indirect effects to be subsumed in earnings management. Further, it is worth noting thatour measure of risk-taking is based on the standard deviation of earnings, not the level of earnings which itself isaffected by culture.8 It is worth noting that including the cultural variables one at a time in the regression model does not change theirsignificant association with corporate risk-taking (results available upon request).


18/47

16

average earnings smoothing (higher average earnings) than another would also be expected to have lower

average R&D expenditures. One other firm characteristicinitial firm size measured at the firm-level

deviationis negatively and significantly associated with R&D, although this variable measured at the

country-level mean is not consistently and significantly related to R&D.

More importantly, the coefficient on individualism is significant and with the predicted sign,

while the coefficient on uncertainty avoidance has the predicted sign but is not statistically significant. A

one standard deviation increase in individualism increases the risk-taking proxy by 25.1% of its mean.

We also note that there is a positive and significant association between private credit and R&D. A one

standard deviation increase in private credit increases the risk-taking proxy by 21.8% of its mean.

When LTD is the dependent variable, we show that among the firm characteristics, only initial

firm size measured at both the firm-level deviation and the country-level mean is positively and

significantly associated with long-term debt. This positive relation is consistent with the standard capital

structure theory where large firms are perceived to be more credit worthy and hence are able to take on

more debt (including long-term debt). Two other firm characteristicsinitial earnings and initial sales

growth measured at the firm-level deviationare negatively and significantly associated with long-term

debt. The negative relation between profitability measured at the firm-level deviation and long-term debt

is consistent with the pecking order theory of capital structure (Myers (1984)): More profitable firms tend

to employ less debt. The negative relation between growth measured at the firm-level deviation and long-

term debt is consistent with the trade-off theory of capital structure: Fast growing firms are more likely to

have funding needs in the future and hence this leads to a low level of debt (including long-term debt) to

mitigate future underinvestment. We also note that the coefficient on initial sales growth measured at the

country-mean level is positive and significant, implying that countries that have higher average initial

sales growth tend to have higher average long-term debt.

Both cultural value variables have the correct sign, and the coefficient on uncertainty avoidance issignificant: A one standard deviation increase in uncertainty avoidance decreases the risk-taking proxy by

14.6% of its mean. We also note that there is a positive and significant association between rule of law

and long-term debt, a negative and significant association between creditor rights and long-term debt, and

a negative and significant association between stock market capitalization and long-term debt. A one


19/47

17

standard deviation increase in rule of law (creditor rights, and stock market capitalization) increases

(decreases) the risk-taking proxy by 7.3% (14.3%, and 5.2%, respectively) of its mean.

Comparing between the OLS regressions and our HLM regressions, we note that uncertainty

avoidance (in predicting R&D), individualism (in predicting LTD), rule of law (in predicting R&D), and

creditor rights (in predicting std(ROA) and R&D) lose significance under the HLM specification,

highlighting the tendency of standard regression models to over-reject the null hypothesis when testing

the effects of country-level variables in an individual firm-level regression (Goldstein (2003, p. 24)).

Overall, the two cultural values have consistent effects on corporate risk-taking, supporting our

hypotheses H1 and H2: There is a positive association between individualism and firm-level riskiness,

and a negative association between uncertainty avoidance and firm-level riskiness. In contrast, due to

moderate to high correlations among formal institution variables (see Table 1 Panel C), they have limited

explanatory power in predicting corporate risk-taking.

Distinguishing the Direct and Indirect Effects of Culture

The evidence thus far captures only the direct effects of cultural values on risky corporate

decisions. It is also of interest to investigate the specific channels through which cultural values might

indirectly affect corporate risk-taking. Tabellini (2008) argues that cultural values in the distant past give

rise to informal institutions that reflect the level of trust in society, and these in turn give rise to formal

institutions. Licht et al. (2005, 2007) show that both individualism and uncertainty avoidance predict rule

of law. We conjecture that firms appetite for risk may be influenced by both their cultural values and

these formal institutional outcomes.

Table 3 Panel A presents the pair-wise Pearson correlation coefficients between the cultural value

variables, the investor protection variables, and the institutional development variables. Panel B presents

the parallel regression results where the two cultural value variables are used as the explanatory variables.

We show that individualism is negatively and significantly related to the anti-self-dealing index

and GDP growth volatility, and is positively and significantly related to rule of law and disclosure.

Cultures low on individualism emphasize strong group cohesion and encourage people to take more


20/47

18

responsibility for each others well-being; hence they rely less on the judicial system. Cultures high on

individualism rely on formal institutions to protect individual rights. This would predict a positive

association between individualism and rule of law (Licht et al. (2005)). Our result in Panel B is consistent

with Licht et al. (2007).

We show that uncertainty avoidance is negatively and significantly related to five out of all seven

measures of investor protection and institutional development. The cultural concept of uncertainty

avoidance emphasizes conformity and granting power to authorities, leading to less reliance on formal

contracts and avoidance of judicial proceedings if possible (Licht et al. (2005)). Our findings above are

consistent with that interpretation.

In summary, the evidence from both the multilevel model (see Table 2) and the country-level

model (see Table 3) supports our conjecture on the indirect effects of culture that there is a chain of

influences from cultural values to formal institutions to corporate risk-taking.

Economic Significance of Direct, Indirect, and Total Effects of Culture

To assess the economic significance of the effects of cultural values on risky corporate decisions,

we examine the consequence of a change in each cultural value on measures of corporate risk-taking.

Specifically, we first compute the change in individualism (IND) from the 25th percentile to the 75th

percentile using our sample of 35 countries: IND = 75th percentile 25th percentile = 0.47. Similarly, we

compute the change in uncertainty avoidance (UA) from the 25th percentile to the 75th percentile: UA =

75th percentile 25th percentile = 0.38. We then examine the effects on different corporate risk-taking

measures as a result of the above specific change in each cultural value in Table 4, which decomposes the

total effects into direct and indirect effects.

Row (1) in Table 4 presents the coefficients from the indirect effect regression in Table 3 Panel

B. Row (2) reports the product of the Row (1) coefficients and the percentile change in individualism:

IND (uncertainty avoidance: UA). Row (3) presents the coefficients from the direct effect regression in

Table 2. Row (4) reports the product of the Row (2) and Row (3) coefficients, which is the indirect effect


21/47

19

as a result of specified changes in one cultural value. The sum of indirect effects is the sum of all

coefficients in Row (4). The direct effect is the product of the coefficient on individualism (uncertainty

avoidance) in Table 2 and the percentile change in individualism: IND (uncertainty avoidance: UA).

The total effect is the sum of indirect and direct effects.

Panel A presents the economic significance of the effects of culture on std(ROA). We show that

when individualism is increased from the 25th percentile to the 75th percentile, the direct effect is to

increase std(ROA) by 2.34%, and the indirect effect through firm and country characteristics is to

decrease std(ROA) by 1.42%. It is worth noting that the direct and indirect effects of individualism on

std(ROA) are offsetting. The total effect is to increase std(ROA) by 0.92%, consistent with our hypothesis

H1. Given that the sample mean (median) std(ROA) is 6.20% (4.68%), these effects are of clear

economic significance.

For the cultural value of uncertainty avoidance, when it is increased from the 25th percentile to the

75th percentile, the direct effect is to decrease std(ROA) by 1.01%, and the indirect effect through firm

and country characteristics is to decrease std(ROA) by 1.15%. The total effect is to decrease std(ROA) by

2.16%, consistent with our hypothesis H2.

Panel B presents the economic significance of the effects of culture on R&D. We show that when

individualism is increased from the 25th percentile to the 75th percentile, the direct effect is to increase

R&D by 1.24%, and the indirect effect through firm and country characteristics is to increase R&D by

0.57%. The total effect is to increase R&D by 1.82%. Given that the sample mean (median) ratio of R&D

expenses to total assets is 2.83% (0.58%), the direct and indirect effects are of economic significance. For

the cultural value of uncertainty avoidance, when it is increased from the 25th percentile to the 75th

percentile, the direct effect is to decrease R&D by 0.60%, and the indirect effect through firm and country

characteristics is to increase R&D by 0.46%. The direct and indirect effects of uncertainty avoidance on

R&D are offsetting. The total effect is to decrease R&D by 0.14%.

Panel C presents the economic significance of the effects of culture on long-term debt. We show

that when individualism is increased from the 25th percentile to the 75th percentile, the direct effect is to


22/47

20

increase long-term debt by 1.74%, and the indirect effect through firm and country characteristics is to

increase long-term debt by 1.56%. The total effect is to increase long-term debt by 3.30%. Given that the

sample mean (median) ratio of long-term debt to total assets is 12.46% (9.72%), these effects are of clear

economic significance. For the cultural value of uncertainty avoidance, when it is increased from the 25th

percentile to the 75th percentile, the direct effect is to decrease long-term debt by 2.96%, while the indirect

effect through firm and country characteristics is to increase long-term debt by 1.26%. The direct and

indirect effects of uncertainty avoidance on long-term debt are offsetting. The total effect is to decrease

long-term debt by 1.70%.

Boundary Conditions

To examine the moderating role of certain firm characteristics on the influence of cultural values

on risky corporate decisions, we add interaction terms between earnings smoothing and firm size, and our

two cultural value measures to Equation (1).

Table 4 presents results on the interaction terms.9 As expected, the positive influences of

individualism on risk-taking are strengthened in firms with greater earnings smoothing, consistent with

our hypothesis H3. This effect is significant for two of our risk-taking measures. Our result confirms that

extensive earnings smoothing is a manifestation of enhanced managerial discretion (Han et al. (2010)),

which in turn facilitates cultural tendencies of corporate managers towards firm-level risk-taking. Similar

effects are not found for uncertainty avoidance.

Also as expected, both the positive influence of individualism and the negative influence of

uncertainty avoidance on risk-taking are mitigated in larger firms, consistent with our hypothesis H4. This

effect is significant for two of our risk-taking measures. These findings support our conjecture that large

firms with highly disciplined financial management systems are less subject to the influences of their

cultural backgrounds in corporate risk-taking.

9 In unreported analysis, we find that the results on firm- and country-level variables when all these interaction termsare included remain the same as those reported in Table 2. For space considerations, we opt not to report them inTable 4 (results available upon request).


23/47

21

Overall, the evidence in Table 4 supports our hypotheses H3 and H4 that certain firm

characteristics moderate the role of cultural values in corporate risk-taking. Prior studies such as Hope

(2003) and Han et al. (2010) show that formal institutions can be important conditioning variables for the

influence of culture. In this paper, we provide new evidence suggesting that in addition to country-level

formal institutional characteristics, firm characteristics can also be important moderators of the cultures

role in corporate decisions.

6. Additional Investigation

We implement the following checks on our main results (all results are available upon request).

First, using Schwartzs (1994) culture value of harmony, we show that there are significantly

negative associations between harmony and all three measures of corporate risk-taking.

Second, we add a set of religion variables, defined as the share of population whose primary

religion is Protestantism, Catholicism, Islam, or Buddhism, to our main model specification in Equation

(1), and find that the relation between cultural values and corporate risk-taking is unaffected.

Third, we remove countries with less than 20 observations (Argentina, Peru, and Portugal) and as

a result end up with 7,208 firm observations from 32 countries. Using this sample, we re-estimate

Equation (1). We find that our main results on the significant role of cultural values in corporate risk-

taking remain unchanged.

Fourth, as our Table 1 shows, Japan and the US contribute 1,659 and 1,514 firms to our sample

respectively, representing 44% of sample firms. It is important to check if our main findings remain after

excluding firms from these two countries. We find that the effects of cultural values are somewhat

weaker than those in Table 2. In particular, the coefficient on uncertainty avoidance loses its significance

when the dependent variable is std(ROA), the coefficient on individualism loses its significance when the

dependent variable is R&D, and the significance of the coefficient on uncertainty avoidance is reduced

from the 1% level to the10% level when the dependent variable is long-term debt.

Finally, it is natural to expect that cultural values might play a weaker role in firms that have

international exposure, including both product markets and shareholder base. To identify multinational


24/47

22

firms in our sample, we rely on one data entryforeign exchange gain/lossto proxy for sample firms

overseas exposure. There are 2,024 firms with foreign exchange exposures among our sample firms. We

find that after removing this substantial sample of multinational firms, the significant correlation between

cultural values and corporate risk-taking is somewhat weakened. In particular, the significance of the

coefficient on uncertainty avoidance is reduced from the 5% level to the 10% level when the dependent

variable is std(ROA), and the coefficient on individualism loses its significance when the dependent

variable is R&D. To identify cross-listed firms in our sample, we start with Citigroups list of Depository

Receipt issuers around the world and supplement that information with the list of foreign firms on the

NYSE, the Nasdaq, and the London Stock Exchange. We are able to identify 453 cross-listed companies

in our sample. We find that after removing cross-listed firms, the effects of cultural values remains the

same as those in Table 2.

10

7. Conclusions

In this paper, we present strong evidence supporting the important role of cultural influences in

corporate risk-taking: Individualism has positive and significant direct effects, whereas uncertainty

avoidance has negative and significant direct effects on corporate risk-taking. Furthermore, we

demonstrate that culture influences formal institutional development, which in turn influences risky

corporate decision-making. These indirect effects sometimes are consistent in sign and reinforcing, and

sometimes are opposite in sign and offsetting the direct effects of culture. The total effects of each

cultural value on corporate risk-taking are in the hypothesized direction and of economic importance.

Finally, we find that greater earnings smoothing accentuates the effect of individualism on risk-taking,

and larger firm size mitigates the effects of individualism and uncertainty avoidance on risk-taking.

Our paper identifies specific economic and psychological channels through which culture exerts

its influence on risky corporate decision-making, and represents a novel demonstration of how informal

10 In this paper, we focus on the direct influence of culture on the decision making of corporate managers. Theremay also be an indirect influence of culture on managers through their shareholders cultural preferences.Empirically, it is difficult to disentangle these two channels of cultural influences on corporate risk-taking. Usingthe subsample without cross-listed firms and hence firms with few foreign shareholders, we obtain stronger resultson the influence of culture on corporate decisions, which might suggest that there is an indirect channel of culturalinfluences through the shareholders.


25/47

23

institutions such as culture can affect firms appetite for risk and consequently their risky corporate

decisions. It also captures both the direct and indirect influences of culture and establishes boundary

conditions on the influence of culture on risky corporate decisions.

The findings in our paper are relevant to the general management community. Standard economic

theories suggest that corporate decisions should be determined only by economic considerations such as

profit maximization. We show that in reality, cultural values do often guide the way companies from

around the world make risky investment and financing decisions, leading to decisions that deviate from

optimal practice. Our results support a growing awareness among management scholars that even in

market economies with sophisticated professional managers, intangible factors such as culture matter in

high-stakes corporate decisions.

Our paper also complements the existing literature on cultural differences in individual financial

risk-taking. Most psychological studies in the area compare the behavior of college students in a small

number of countries, most often contrasting the US (or similar Western nations) with either China or

Japan (or similar Asian nations). Research on culture and individual decision-making in particular

focuses on one of three topics: national differences in self-enhancement (with Asian countries showing

less overt self-enhancement, e.g., Sedikides et al. (2003), and Yamaguchi et al. (2005)), overconfidence in

judgment (with Asian countries showing more overconfidence, e.g., Yates, Lee, Shinotsuka, Patalano, and

Sieck (1998)), or risk-aversion. A series of articles by Weber and Hsee (1998, 1999), provide the most

direct evidence on national differences in individual risk-taking. They find that financial risk-aversion, as

measured by small-stakes gambles, is less pronounced in China than in the US.

Our findings from a large sample of countries and examining the outcomes of high-stakes

corporate investment decisions indicate that contrary to evidence from individual risk-taking in China and

the US, greater individualism is linked to greater corporate risk-taking. We hope that the findings from

our paper help raise awareness of the richer context in which corporate decisions are made, and motivatefurther investigation of the channels through which culture has its influences.


26/47

24

Appendix I: Variable Definitions

Country-level cultural value variables:

Individualism: The index is a weighted sum of the following four statements, with the second and thirditems given positive weights and the first and last items given negative weights:

1) Have sufficient time for your personal or family life2) Have good physical working conditions (good ventilation and lighting, adequate work space, etc.)3) Have security of employment4) Have an element of variety and adventure in the job

High individualism is indicated by ratings of of very little or no importance to items (2) and (3), andratings of of utmost importance to items (1) and (4). Individualism refers to the strength of the tiespeople have to others within the community. A high score on individualism indicates a loose connectionwith people. In countries with a high individualism score there is a lack of interpersonal connection andlittle sharing of responsibility, beyond family and perhaps a few close friends. A society with a lowindividualism score would have strong group cohesion, and there would be a large amount of loyalty andrespect for members of the group. The group itself is also larger and people take more responsibility for

each others well-being.

Uncertainty avoidance: The index is a weighted sum of the following one question and three statements,with the first two items given positive weights and the last two items given negative weights:

1) How often do you feel nervous or tense at work?2) One can be a good manager without having precise answers to most questions that subordinates

may raise about their work3) Competition between employees usually does more harm than good4) A companys or organizations rules should not be brokennot even when the employee thinks it

is in the companys best interestHigh uncertainty avoidance is indicated by answering always to the first question, and ratings ofstrongly disagree to item (2), and ratings of strongly agree to items (3) and (4). Uncertainty avoidance

captures the degree to which the members of a society feel uncomfortable with uncertainty and ambiguity.This feeling leads them to beliefs promising certainty and to maintaining institutions protectingconformity. Strong uncertainty avoidance societies maintain rigid codes of belief and behavior and areintolerant towards deviant persons and ideas. Weak uncertainty avoidance societies maintain a morerelaxed atmosphere in which practice counts more than principles and deviance is more easily tolerated.

Firm-level risk-taking variables:

Standard deviation of ROA (std(ROA)): Following John, Litov, and Yeung (2008), we computecompany earnings volatility

5|)1

(1

1)(

1 1

2,,,,,

TAdjROAT

AdjROAT

ROAstdT

t

T

tcitcici ,

where

tcN

k tck

tck

tctci

tcitci

A

EBITDA

NA

EBITDAAdjROA

,

1 ,,

,,

,,,

,,,,

1. tcN , indexes the firms within country c and

year t, and tciEBITDA ,, is defined as depreciation (item #11) plus operating income after depreciation


27/47

25

(item #14), and tciA ,, is the contemporaneous total assets (item #89). More specifically, for each firm

with available earnings and total assets for at least five years in 1997-2006 we compute the deviation ofthe firms EBITDA/Assets from the country average (for the corresponding year) and then calculate the

standard deviation of this measure for each firm. tciAdjROA ,, is winsorized at the one percent level in

both tails of the distribution.

Source: Compustat Global Vantage Database.

R&D: The average ratio of R&D expenses (item #52) to total assets (item #89) with available data for atleast five years in 1997-2006.Source: Compustat Global Vantage Database.

LTD: The average ratio of long-term debt (item #106) to total assets (item #89) with available data for atleast five years in 1997-2006.Source: Compustat Global Vantage Database.

Firm-level control variables:

Earnings smoothing: Following John, Litov, and Yeung (2008), our measure of earnings smoothing isthe ratio of firm-level standard deviations of operating income after depreciation (item #14) and operatingcash flow where both variables are scaled with lagged total assets (Earnings smoothing1). The higher thevalue of this measure, the lower earnings smoothing is. To facilitate interpretation, we thus consider the

modified measure Earnings smoothing = 1 Earnings smoothing1, for which higher values indicatehigher level of earnings smoothing. Operating cash flow is equal to operating income after depreciation(item #14) minus accruals, where accruals are calculated as

DEPTPSTDCLCashCAAccruals )()( , where CA is total current assets (item

#75), Cash is cash or cash equivalents (item #60), CL is total current liabilities (item #104), STD is short-term debt (item #94), TP is income taxes payable (item #100), and DEP is depreciation (item #11).Source: Compustat Global Vantage Database.

Initial firm size: The natural logarithm of total assets (item #89) measured in millions of US$ retrieved asof the first year of entry of the company in the sample.Source: Compustat Global Vantage Database.

Initial earnings: The ratio of EBITDA (item #11 + item #14) to total assets (item #89) retrieved as of thefirst year of entry of the company in the sample.Source: Compustat Global Vantage Database.

Initial sales growth: The sales growth rate as of the first year of entry of the company in the sample.Source: Compustat Global Vantage Database.

Country-level control variables:

Anti-self-dealing index: Aggregation of ex-ante and ex-post private control of self-dealing. Ex-anteprivate control of self-dealing is an index of approval by disinterested shareholders and ex-ante disclosure,and ranges from 0 to 4. Ex-post private control of self-dealing is an index of disclosure in periodic filingsand ease of proving wrongdoing. It ranges from 0 to 1. The data are from Djankov et al. (2008).


28/47

26

Rule of law: The assessment of the law and order tradition of the country (La Porta et al. (1998)).Calculated as average of the months of April and October of the monthly index between 1982 and 1995.Scale from zero to 10, with lower scores for less tradition for law and order.Source: International Country Risk Guide.

Rating of accounting disclosure standards (Disclosure): Index that is created by examining and rating

companies 1990 annual report on their inclusion or omission of 90 items. These items fall into sevencategories (general information, income statements, balance sheets, fund flow statement, accountingdisclosure standards, stock data, and special items). A minimum of three companies in each country werestudied (La Porta et al. (1998)).Source: Center for International Financial Analysis and Research.

Creditor rights: The index is formed by adding 1 when: (1) the country imposes restrictions, such ascreditors consent or minimum dividends to file for reorganization; (2) secured creditors are able to gainpossession of their security once the reorganization petition has been approved (no automatic stay); (3)secured creditors are ranked first in the distribution of the proceeds that result from the disposition of theassets of a bankrupt firm; and (4) the debtor does not retain the administration of its property pending theresolution of the reorganization. The index ranges from 0 to 4 (La Porta et al. (1998)).

Country private credit: Average of the ratio of the value of credits by financial intermediaries to theprivate sector to GDP for the period 1997-2006.Source: World Development Indicators at http://devdata.worldbank.org/dataonline/.

Country market capitalization: Average of the ratio of stock market capitalization to GDP for the period1997-2006.Source: World Development Indicators at http://devdata.worldbank.org/dataonline/.

Country GDP growth volatility: standard deviation of annual GDP growth rates for the period 1997-2006.Source: World Development Indicators at http://devdata.worldbank.org/dataonline/.


29/47

27

Appendix II. Industry Classification and Distribution

This table presents the industry classification and distribution of sample firms. Panel A lists the two-digit SICmanufacturing industry classification and description. Panel B presents the manufacturing industry distributionacross countries.

Panel A: Two-digit SIC Manufacturing Industry Classification and Description

SIC Industry Description

20 Food and kindred products

21 Tobacco products

22 Textile mill products

23 Apparel and other textile products

24 Lumber and wood products

25 Furniture and fixtures

26 Paper and allied products

27 Printing and publishing28 Chemicals and applied products

29 Petroleum and coal products

30 Rubber and misc. plastics products

31 Leather and leather products

32 Stone, clay, and glass products

33 Primary metal industries

34 Fabricated metal products

35 Industrial machinery and equipment

36 Electronic and other electric equipment

37 Transportation equipment

38 Instruments and related products39 Misc. manufacturing industries


30/47


31/47

29

References:

Acharya, Viral V., Yakov Amihud, and Lubomir Litov, 2010, Creditor rights and corporate risk-taking,Journal of Financial Economics forthcoming.

Bargeron, Leonce, Kenneth Lehn, and Chad Zutter, 2010, Sarbanes-Oxley and corporate risk-taking,

Journal of Accounting and Economics 49, 34-52.

Beckmann, Daniela, Lukas Menkhoff, and Megumi Suto, 2008, Does culture influence asset managersviews and behavior?Journal of Economic Behavior & Organization 67, 624-643.

Bhagat, Sanjai, and Ivo Welch, 1995, Corporate research & development investments internationalcomparisons,Journal of Accounting and Economics 19, 443-470.

Chui, Andy C.W., and Chuck C.Y. Kwok, 2008, National culture and life insurance consumption,Journal of International Business Studies 39, 88-101.

Chui, Andy C.W., Alison E. Lloyd, and Chuck C.Y. Kwok, 2002, The determination of capital structure:

Is national culture a missing piece to the puzzle?,Journal of International Business Studies 33,99-127.

Chui, Andy C.W., Sheridan Titman, and K.C. John Wei, 2010, Individualism and momentum around theworld,Journal of Finance 65, 361-392.

Claessens Stijn, Simeon Djankov, and Tatiana Nenova, 2000, Corporate risk around the world, WorldBank Policy Research Working Paper 2271.

Coles, Jeffrey L., Naveen D. Daniel, and Lalitha Naveen, 2006, Managerial incentives and risk-taking,Journal of Financial Economics 79, 431-468.

Demsetz, Harold, and Kenneth Lehn, 1985, The structure of corporate ownership: Causes andconsequences,Journal of Political Economy 93, 1155-1177.

Djankov, Simeon, Rafael La Porta, Florencio Lopez-de-Silanes, and Andrei Shleifer, 2008, The law andeconomics of self-dealing,Journal of Financial Economics 88, 430-465.

Fernandes, Nuno, Miguel A. Ferreira, Pedro Matos, and Kevin J. Murphy, 2010, The pay divide: (Why)are US top executives paid more? University of Southern California working paper.

Galai, Daniel, and Ronald Masulis, 1976, The option pricing model and the risk factor of stock,Journalof Financial Economics 3, 53-81.

Gelfand, Michele J., Marianne Higgins, Fumio Murakami, Susumu Yamaguchi, Lisa H. Nishii, Jana L.Raver, and Alexandria Dominguez, 2002, Culture and egocentric perceptions of fairness inconflict and negotiation,Journal of Applied Psychology 87, 833-845.

Giannetti, Mariassunta, and Yishay Yafeh, 2010, Do cultural differences between contracting partiesmatter? Evidence from syndicated bank loans,Management Science forthcoming.

Goldstein, Harvey I., 2003,Multilevel Statistical Models, 3rd edition, Edward Arnold: London.


32/47

30

Han, Sam, Tony Kang, Stephen Salter, and Yong Keun Yoo, 2010, A cross-country study on the effectsof national culture on earnings discretion,Journal of International Business Studies 41, 123-141.

Haugen, Robert A., and Lemma W. Senbet, 1981, Resolving the agency problems of external capitalthrough options, Journal of Finance 36, 629-647.

Hilary, Gilles, and Kai Wai Hui, 2009, Does religion matter in corporate decision making in America?Journal of Financial Economics 93, 455-473.

Hofstede, Geert H., 1980, Cultures Consequences: International Differences in Work-Related Values,Sage, Thousand Oaks, CA.

Hofstede, Geert H., 1991, Cultures and Organizations: Software of the Mind, McGraw-Hill, Berkshire,England.

Hofstede, Geert H., 2001, Cultures Consequences: Comparing Values, Behaviors, Institutions, andOrganizations Across Nations, 2nd edition, Sage, Thousand Oaks, CA.

Hope, Ole-Kristian, 2003, Firm-level disclosures and the relative roles of culture and legal origin,Journalof International Financial Management and Accounting 14, 218-248.

House, Robert J., Paul J. Hanges, Mansour Javidan, Peter W. Dorfman, and Vipin Gupta, 2004, Culture,Leadership and Organizations: The GLOBE Study of 62 Societies, Sage, Thousand Oaks, CA.

Jensen, Michael C., and William H. Meckling, 1976, Theory of the firm: Managerial behavior, agencycost and ownership structure,Journal of Financial Economics 3, 305-360.

John, Kose, Lubomir Litov, and Bernard Yeung, 2008, Corporate governance and risk-taking,Journal ofFinance 63, 1679-1728.

Kreiser, Patrick, Louis D. Marino, Pat Dickson, and K. Mark Weaver, 2010, Cultural influences onentrepreneurial orientation: The impact of national culture on risk taking and proactiveness inSMEs,Entrepreneurship Theory and Practice 34, 959-983.

Kwok, Chuck C. Y., and Solomon Tadesse, 2006, National culture and financial systems,Journal ofInternational Business Studies 37, 227-247.

Laeven, Luc, and Ross Levine, 2009, Bank governance, regulation and risk taking,Journal of FinancialEconomics 93, 259-275.

La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer and Robert W. Vishny, 1997, Legaldeterminants of external finance,Journal of Finance 52, 1131-1150.

La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer and Robert W. Vishny, 1998, Law andfinance,Journal of Political Economy 106, 1113-1155.

Li, Kai, Dale Griffin, Heng Yue, and Longkai Zhao, 2011, National culture and capital structuredecisions: Evidence from foreign joint ventures in China,Journal of International BusinessStudies 42, 477-503.


33/47

31

Li, Yong, and Shaker A. Zahra, 2011, Formal institutions, culture, and venture capital activity: A cross-country analysis,Journal of Business Venturing in press.

Licht, Amir N., Chanan Goldschmidt, and Shalom H. Schwartz, 2005, Culture, law, and corporategovernance,International Review of Law and Economics 25, 229-255.

Licht, Amir N., Chanan Goldschmidt, and Shalom H. Schwartz, 2007, Culture rules: The foundations ofthe rule of law and other norms of governance,Journal of Comparative Economics 35, 659-688.

Low, Angie, 2009, Managerial risk-taking behavior and equity-based compensation,Journal of FinancialEconomics 92, 470-490.

Myers, Stewart, C., 1977, The determinants of corporate borrowing,Journal of Financial Economics5,147-175.

Myers, Stewart C., 1984, The capital structure puzzle,Journal of Finance 39, 575-592.

Palmer, Timothy B., and Robert M. Wiseman, 1999, Decoupling risk taking from income stream

uncertainty: A holistic model of risk, Strategic Management Journal 20, 1037-1062.

Rajgopal, Shivaram, and Terry Shevlin, 2002, Empirical evidence on the relation between stock optioncompensation and risk taking,Journal of Accounting and Economics 33, 145-171.

Raudenbush, Stephen W., and Anthony S. Bryk, 2002,Hierarchical Linear Models: Applications andData Analysis Methods, Advanced Quantitative Techniques in the Social Science Series, Sage,Thousand Oaks, CA.

Sanders, W.M. Gerard, and Donald C., Hambrick, 2007, Swinging for the fences: The effects of CEOstock options on company risk taking and performance,Academy of Management Journal 50,1055-1078.

Schuler, Randall S., and Nikolai Rogovsky, 1998, Understanding compensation practice variations acrossfirms: The impact of national culture,Journal of International Business Studies 29, 159-177.

Schwartz, Shalom H., 1994, Beyond individualism-collectivism: New dimensions of values., in U. Kim,H.C. Triandis, C. Kagitcibasi, S.C. Choi, and G. Yoon, eds.:Individualism and Collectivism:Theory, Method and Application, Sage, Newbury Park, CA.

Sedikides, Constantine, Lowell Gaertner, and Yoshiyasu Toguchi, 2003, Pancultural self-enhancement,Journal of Personality and Social Psychology 84, 60-79.

Shao, Liang, Chuck C.Y. Kwok, and Omrane Guedhami, 2010, National culture and dividend policy,Journal of International Business Studies 41, 1391-1414.

Siegel, Jordan I., Amir N. Licht, and Shalom H. Schwartz, 2010, Egalitarianism, cultural distances, andFDI: A new approach, Harvard Business School working paper.

Siegel, Jordan I., Amir N. Licht, and Shalom H. Schwartz, 2011, Egalitarianism and internationalinvestment,Journal of Financial Economics forthcoming.


34/47

32

Tabellini, Guido, 2008, Presidential address institutions and culture,Journal of the European EconomicAssociation 6, 255-294.

Weber, Roberto, and Colin Camerer, 2003, Cultural conflict and merger failure: An experimentalapproach,Management Science 49, 400-415.

Weber, Elke U., and Christopher K. Hsee, 1999, Models and mosaics: Investigation of cross-culturaldifferences in risk perception and risk preference, Psychonomic Bulletin and Review 6, 611-617.

Weber, Elke U., and Christopher K. Hsee, 1998, Cross-cultural differences in risk perception, but cross-cultural similarities in attitudes towards perceived risk,Management Science 44, 1205-1217.

Weber, Yaakov, Oded Shenkar, and Adi Raveh, 1996, National and corporate cultural fit in merger-acquisitions: An exploratory study,Management Science 42, 1215-1227.

Yamaguchi, Susumu, Michele Gelfand, Megumi M. Ohashi, and Yuriko Zemba, 2005, The culturalpsychology of control: Illusion of personal versus collective control in the U.S. andJapan,Journal of Cross-Cultural Psychology 36, 750-761.

Yates, J. Frank, Ju-Whei Lee, Hiromi Shinotsuka, Andrea L. Patalano, and Winston R. Sieck, 1998,Cross-Cultural Variations in Probability Judgment Accuracy: Beyond General KnowledgeOverconfidence?, Organizational Behavior and Human Decision Processes 74, 89-117.

Zheng, Xiaolan, Sadok El Ghoul, Omrane Guedhami, Chuck C. Y., Kwok, 2011, Collectivism andcorruption in bank lending, University of South Carolina working paper.


35/47

Figure 1. Channels of Influence of Cultural Values on Corporate Risk-Taking

Panel A: Individualism and Corporate Risk-Taking

High Individualism

Institutional context

Due to acceptance ofcompetition and personal

ambition, high individualismcountries embrace a set

of laws and formal institutionsthat encourages market-based

financial systems

Market-based institutionsencourage corporate risk-

taking by reducingenvironmental risk

Channel: Operates throughrule of law and disclosure

The direction of influence is positive (high individualism leads to morecorporate risk-taking)

(1) (2) (3)

Agency theory of risk-taking

Assumes managers are risk-averse

and need to be compensated for risk-

taking

Due to preferences for individual

competition-based compensation,

equity-based pay is expected to be

more common in high individualism

countries

Managerial autonomy

High individualism managersprefer freedom to make

decisions and eager to take creditfor success

Manageri

High indbelieve t

making deof uncerta

Equity-based pay mitigatesmanagerial risk-aversion and

thus encourages corporaterisk-taking

The motivation to stand out from

the crowd encourages high

individualism managers to seek out

more innovative projects with high

but uncertain long-term payoff

The bsuperi

high indunder

unc

Channel: Operates throughequity-based compensation

practices

Channel: Overweights riskypayoffs associated with

innovative projects

Channdisco

Country level Firm level Individual level


36/47

Panel B: Uncertainty Avoidance and Corporate Risk-Taking

Low Uncertainty Avoidance

Institutional context

Due to comfort withunpredictable outcomes, low

uncertainty avoidance countriesembrace a set of laws and formal

institutionsthat encourages market-based

financial systems

Market-based institutionsencourage corporate risk-

taking by reducingenvironmental risk

Channel: Operates throughmarket development and

disclosure

The direction of influence is negative (low uncertainty avoidance leads tomore corporate risk-taking)

(

Culture vs Corp. Risk-taking

Documents

Transcript of Culture vs Corp. Risk-taking