Document not found! Please try again

Culture, agency costs, and dividends

Culture, agency costs, and dividends

Journal of Comparative Economics 38 (2010) 321–339 Contents lists available at ScienceDirect Journal of Comparative Economics journal homepage: www...

272KB Sizes 0 Downloads 105 Views

Journal of Comparative Economics 38 (2010) 321–339

Contents lists available at ScienceDirect

Journal of Comparative Economics journal homepage: www.elsevier.com/locate/jce

Culture, agency costs, and dividends Jana P. Fidrmuc a, Marcus Jacob b,c,* a b c

Warwick Business School, University of Warwick, Coventry CV4 7AL, United Kingdom Harvard University, Department of Government, Cambridge, MA 02138, United States European Business School (EBS), Oestrich-Winkel, Germany

a r t i c l e

i n f o

Article history: Received 4 October 2009 Revised 22 April 2010 Available online 27 April 2010 JEL classification: G28 G35 K00 Z10 Keywords: Dividends Agency problems Culture Law and finance

a b s t r a c t Fidrmuc, Jana P., and Jacob, Marcus—Culture, agency costs, and dividends This paper presents a culturally rooted agency explanation for differences in dividend payout policies around the world. We conjecture that the social normative nature of culture influences the character of agency relations and determines the acceptance and legitimacy of different dividend payout strategies across different countries. By linking dividends to cultural differences across 5797 firms in 41 countries, our analysis shows that high individualism, low power distance, and low uncertainty avoidance are significantly associated with higher dividend payouts. A comprehensive set of robustness tests in which we control for legal institutions, share repurchases, corporate debt ratios, and ownership structures confirms that culture is a relevant factor when analyzing dividend distributions. Our results further show that legal institutions and culture as a social institution have complementary effects on dividend payouts. Overall, our finding that culture matters suggests important implications for a wide range of agency-based economic and capital market phenomena. Journal of Comparative Economics 38 (3) (2010) 321–339. Warwick Business School, University of Warwick, Coventry CV4 7AL, United Kingdom; Harvard University, Department of Government, Cambridge, MA 02138, United States; European Business School (EBS), Oestrich-Winkel, Germany. Ó 2010 Association for Comparative Economic Studies All rights reserved.

‘‘If we learn anything from the history of economic development, it is that culture makes almost all of the difference.” (Weber, 1930)

1. Introduction Miller and Modigliani (1961) declared them irrelevant, yet dividends are ubiquitous. This paradox is why dividend payouts have long inspired a search for explanations. The agency model of dividends maintains that dividend payouts help address the agency problems between corporate insiders and outside shareholders by distributing cash at the discretion of managers and forcing firms to interact more frequently with the capital market (Easterbrook, 1984; Zwiebel, 1996; Gomes, 2000). In this paper, we present a broader agency explanation of dividends that considers the preferences and behaviors of economic agents inherent in the cultural values. That is, we consider characteristics of agency relationships distinct from ‘‘deliberate managerial malfeasance that drives the literature’s usual framing of agency problems” (DeAngelo et al., 2008, p. 122). We posit that the social normative nature of culture significantly alters the character of agency relations (Ekanayake, 2004; Johnson and Droege, 2004) and determines the acceptance and legitimacy of different dividend payout strategies * Corresponding author. E-mail addresses: [email protected] (J.P. Fidrmuc), [email protected] (M. Jacob). 0147-5967/$ - see front matter Ó 2010 Association for Comparative Economic Studies All rights reserved. doi:10.1016/j.jce.2010.04.002

322

J.P. Fidrmuc, M. Jacob / Journal of Comparative Economics 38 (2010) 321–339

across countries. This way, culture as a social institution plays an important role in determining dividend payout policies around the world. This new perspective recognizes that in order to address agency problems, societies may adopt different institutional responses. These responses may be formal, through legal institutions as indicated in La Porta et al. (2000), but also may be informal or cultural. After all, as dividend distributions are largely at the discretion of companies; legal rights can only tangentially regulate dividend payouts. This leaves ample room for social regulation through informal institutions as highlighted in this paper. Culture is commonly defined as a set of shared values, norms, beliefs, and expected behaviors that are deeply embedded, unconscious, and often irrational (Herbig, 1994; Hofstede, 1980). Cultural values are conceptions of the desirable that guide the way social actors select actions, evaluate people and events, and explain their actions and evaluations (Schwartz, 1999). In this sense, cultural values ‘‘serve as guiding principles in people’s lives” (Schwartz, 1994, p. 88) and reflect what a group considers to be legitimate or illegitimate, acceptable or unacceptable, good or bad, normal or abnormal, or ethical or unethical (Breuer and Quinten, 2009; Hofstede, 2001). Hence, society-specific cultural values breed diverging preferences across countries for certain behaviors of social actors such as organizational leaders, controlling shareholders, and outside investors (Hofstede, 1980; Schneider and De Meyer, 1991). As a result, culture as ‘‘societal common knowledge” constrains economic interaction (Greif and Laitin, 2004). The idea that culture matters for economic interaction has recently been gaining more ground in economics. Guiso et al. (2009) summarize the contributions of culturally based explanations to our understanding of economic phenomena and show that cultural hypotheses are important for the explanation of fundamental economic issues such as rates of savings and international trade. Tabellini (2008) argues that culture is a core determinant of economic development. A growing body of empirical research on culture’s consequences has recently begun to appear also in the finance literature, showing that culture matters for a range of important capital market phenomena such as corporate capital structures (Chui et al., 2002), levels of cash holdings (Ramirez and Tadesse, 2009; Chang and Noorbakhsh, 2009), returns on momentum strategies (Chui et al., 2010), and the growth of informationally opaque industries (Huang, 2008).1 The central theme of agency theory is that those who control a corporation, be they managers or controlling shareholders, can use their power to divert corporate wealth to themselves to the detriment of outside investors. This diversion can take a number of forms ranging from excessive perquisite consumption, to transfer pricing, tunneling or the outright theft of corporate assets (Shleifer and Vishny, 1997). The agency model of dividends maintains that dividends help address agency problems by aligning managers’ interests with those of investors. This is because dividends distribute cash at the discretion of managers and force managers to enter the market for new capital more frequently (Easterbrook, 1984). Dividends are, however, not a unique instrument that could potentially fill the role of forcing managers to turn to capital markets for new securities. Share repurchases and the issuance of debt in series (so that a continuous stream of payments and a need for refinancing is created) could perform the same role and are in effect a substitute for dividends. Similarly, the quality of the legal protection of investors can provide a substitute for monitoring by the capital market (La Porta et al., 2000): firms in high legal protection countries may pay lower dividends as laws and courts potentially inflict monitoring. In contrast, firms in low protection countries may pay dividends to establish a reputation for the decent treatment of minority shareholders and through more frequent refinancing needs expose themselves to monitoring by outside capital markets. La Porta et al. (2000), however, show empirical support for a positive relationship between dividends and shareholder protection, that is, for an outcome model of dividends. According to the outcome model, dividends are paid as a result of investors’ pressure to release cash, supported by strong minority shareholder protection.2 Our approach subscribes to the substitute model of dividends. It acknowledges that dividend payout decisions are not only determined by an objective assessment of the severity of agency problems, but also by individuals’ values, behaviors, and preferences that inhere in the social fabric in which individuals’ daily transactions are embedded (Putnam, 2000). These values, behaviors, and preferences alter the nature and perceived severity of agency conflicts. Societies’ stances on distinct cultural value dimensions constitute the basis for specific social norms that serve as guiding principles for individuals and prescribe and proscribe behavior in specific circumstances (Hechter and Opp, 2001; Licht, 2008). Because individuals gain intrinsic value from behavior that is consistent with social norms, norms transmit cultural institutional pressure to promote accepted and expected behavior (Myers, 1999; Miller, 1999). For example, behavior consistent with social norms engenders positive feelings whereas inconsistent behavior evokes negative feelings such as shame, guilt, or disapproval (Licht, 2008). This suggests that compliance with social norms is triggered by normative attitudes that belong to individuals’ preference sets. This way culture as a social institution provides incentives and disincentives. It molds individuals’ behaviors, aspirations, objectives, and obligations (Hodgson, 2006, 2007; Dequech, 2009). Social preferences for individual behaviors are not enforced by the legal system, but rather by the approval or disapproval of other members of a group or community (Dequech, 2009). In this paper we argue that payout policies are a special case of social norms reflecting the acceptability or legitimacy of certain dividend payout strategies in a society. Across countries, social norms governing dividend payouts may vary because differing cultural value emphases breed different aspirations, behaviors, beliefs, and preferences, and therefore alter the nat1

For reviews of this literature see Breuer and Quinten (2009) and Reuter (2009). Companies may pay higher dividends where investor protection is high because minority shareholders can rely on such means as (a) voting for directors who offer better dividend policies, (b) selling shares to potentially hostile raiders who then gain control over non-dividend paying companies, or (c) suing companies that spend too lavishly on activities that primarily benefit corporate insiders. 2

J.P. Fidrmuc, M. Jacob / Journal of Comparative Economics 38 (2010) 321–339

323

ure and severity of agency conflicts.3 Importantly, our approach recognizes the weight of social norms not only on shareholders but also on those who control firms. Both shareholders and corporate insiders share the culturally acceptable levels of distribution. Those who control a corporation have good reason to comply with culturally acceptable dividend payouts because, first, doing so elicits a directly perceived benefit via the value of the conforming action as an end in itself (Weise, 2004). This may be due to greater approval, popularity, and/or recognition among investors and other stakeholders in the firm. Second, social norms are usually associated with the possibility of sanctions (Dequech, 2009). Competitive capital markets provide this possibility because firms are motivated to pay the culturally acceptable level of dividends as they compete for capital and need to ensure liquidity for their shares (Faccio et al., 2001; Farinha and López-de-Foronda, 2009). Thus, compliance with norms ultimately pays off because it increases future income through reputation. Perceiving dividends as a special case of social norms highlights the substitute model mechanism and allows us to derive key hypotheses about individuals’ preferences over corporate payouts, and thus about the relationship between culture and dividends. We analyze culture’s consequences for corporate dividend payout decisions using the national scores on Hofstede’s (1980) cultural dimensions of individualism, power distance and uncertainty avoidance. Based on these three cultural dimensions, we derive three hypotheses. First, we posit that in individualist societies, agency conflicts are inherently more severe because their members are more prone to pursue their own personal interests rather than adhere to others’ decisions and preferences. Therefore, outside shareholders may consider low payouts unacceptable and instead expect higher levels of distribution to discipline the self-interested corporate insiders. Second, we hypothesize that in high power distance countries, which tend to show greater acceptance of wealth and power differentials, the severity of agency conflicts is inherently lower. As a result, investors have a lower preference for dividends as a disciplining mechanism and find lower dividend payouts culturally more acceptable. In contrast, individuals in low power distance countries expect higher dividend payouts. This is because dividends, as means to distribute corporate profits to all shareholders, are compatible with a strong cultural value emphasis on equality and the reduction of power and wealth differentials. Finally, we conjecture that outside shareholders and corporate insiders in high uncertainty avoidance societies put heavy emphasis on the certainty that dividend payment expectations are met each period. Also, investors and insiders may prefer higher retained earnings because they are cash resources to hedge against unforeseen financial distress. High uncertainty avoidance societies, therefore, may accept and expect lower dividend payouts. Our analysis of a cross section of 5797 firms from 41 countries provides robust evidence that differences in dividend payout strategies have strong cultural origins. Specifically, we find that firms in countries that score high on individualism and low on power distance and uncertainty avoidance exhibit higher dividend payout ratios. Our results further suggest that legal institutions and culture as a social institution are important complementary determinants of payout policies across countries. Better legal protection in cultures that are naturally afflicted with more severe agency conflicts is associated with even higher dividend payouts. Overall, our paper contributes to the growing evidence in the literature that culture matters for a wide range of fundamental economic and capital market phenomena. The remainder of the paper is organized as follows. Section 2 introduces Hofstede’s (1980) cultural dimensions and derives our hypotheses. Section 3 describes the data. Section 4 discusses the results and confirms their validity through a series of robustness checks. We conclude in Section 5. 2. Hypotheses To measure cultural differences across countries we borrow measures of cultural norms and values developed in crosscultural psychology. These cultural measures build on a common postulate that all societies confront similar basic issues or problems when they come to regulate human activity (Licht et al., 2007). A society’s preferred ways of dealing with these basic issues lie at the essence of each culture and naturally point to its key cultural dimensions (Kluckhohn, 1962). It is thus possible to characterize the cultures of different societies by identifying the key issues (dimensions) and measuring the prevailing value emphases on these key dimensions (Licht et al., 2007). We opt for Hofstede’s (1980, 2001) cultural value dimensions. On the basis of an extensive research project on differences in national cultures among matched samples of IBM business employees across more than 50 countries, Hofstede (1980) develops a pioneering dimensional framework for characterizing culture. He identifies five independent dimensions of national cultural differences that are empirically validated and widely acknowledged: individualism versus collectivism, which is related to the integration of individuals into primary groups; power distance, which refers to solutions to basic problems of human inequality; uncertainty avoidance, which is related to the level of stress in a society in the face of an unknown future; masculinity versus femininity, refers to the distribution of roles between the genders; and finally, long-term versus short-term orientation, which is related to the importance attached to the future versus the past and present. Our empirical pre-analysis confirms that three of Hofstede’s five dimensions, namely individualism, power distance and uncertainty avoidance, are particularly relevant for dividend payout strategies across countries. We therefore focus on these three dimensions. Individualism versus collectivism captures the essential differences in a society’s resolution of the universal tension between the competing claims of individual self-fulfillment versus social conscience and action (Arrow, 1974; Shane and Venkataraman, 1996). High individualism scores (for example in the US) indicate a society in which the ties between individuals are loose and decisions are taken based on individual needs. Conversely, in collectivist societies (for example 3 Even though we emphasize these informal social norms, some countries may also rely on formal norms represented, for example, by mandatory dividend requirements.

324

J.P. Fidrmuc, M. Jacob / Journal of Comparative Economics 38 (2010) 321–339

China) individuals attach high emotional value to belonging to a group (Tajfel, 1981). As a result, individuals are more likely to adjust their behaviors and beliefs to match those of the wider group to which they belong (Mann, 1980). People in individualist societies are more likely than those in collectivist societies to vigorously pursue their personal objectives rather than adhere to others’ decisions and interests (Kohn, 1969; Hofstede, 1980). In such societies, individuals have a substantially greater need for personal achievement and the maximization of private benefits. This leads to more opportunistic behavior of corporate insiders even if it violates social harmony (Hofstede, 1980). Investors in these countries are accustomed to deal with insiders who want to do things their own way (Baum et al., 1993). As a result, investors may not accept low dividend payouts and invest only in firms with high dividend payouts. Firms, in turn, have a motivation to pay the expected level of dividends to allay greater expropriation concerns because they compete for capital in competitive markets and need to ensure liquidity for their shares (Faccio et al., 2001; Farinha and López-de-Foronda, 2009).4 On the other side of the spectrum, people in collectivist societies have less need for autonomy and they stress conformity and adherence to societal norms and regulations (Hofstede, 1980). Increased identification with a group makes corporate insiders less likely to differentiate between their own and others’ welfare, which triggers a sense of shared responsibility and obligation. As a result, individuals in collectivist countries tend to act in accordance with the interests of the wider group even when decisions or actions might be contradictory to their own interests. We thus posit that where collective values prevail, agency conflicts are inherently less severe and shareholders accept lower dividend payouts. We thus test the following hypothesis: Hypothesis 1. Companies in individualist countries tend to have higher dividend payouts. Power distance measures the extent to which the less powerful members of a society expect and accept that power is distributed unequally (Hofstede, 1980). Low power distance societies like Austria or Sweden stress shared power, equality and opportunity for everyone, and perceive inequality in power and wealth as a necessary evil that should be minimized. In contrast, high power distance countries (for example Mexico, Arab countries, and India) comprehend inequality as the basis of societal order where those in power emphasize their position and accentuate authority. We propose that in high power distance societies, where power and wealth differentials are the basis of societal order, lower dividend payouts are culturally more acceptable. This is because a greater tolerance for privilege differentials lowers investors’ perceived severity of agency conflicts and thus also reduces investors’ preference for dividends as a disciplining mechanism. In contrast, in low power distance countries with a strong cultural value emphasis on equality in wealth and power, investors may expect higher dividends because their need for equality intensifies agency tensions. This is particularly because dividend payments are a means through which corporate profits are distributed proportionally to all shareholders. High retained earnings, in contrast, may be less acceptable because they are associated with more corporate funds being left at the discretion of a few insiders, thus potentially perpetuating power and wealth differentials. For example, institutional investors and managers in the Netherlands, a low power distance country, have been found to force higher payouts (Renneboog and Szilagyi, 2006). Therefore, we test the following hypothesis: Hypothesis 2. Companies in low power distance countries tend to have higher dividend payouts. Uncertainty avoidance relates to the levels of anxiety in a society when it is confronted with an unknown future (Hofstede, 1980). High scores for uncertainty avoidance indicate societies with a low tolerance for unstructured situations, uncertainty, and ambiguity (for example Greece, Germany, and Japan). Individuals in high uncertainty avoidance societies are more skeptical about the potential rewards from risky ventures and apply higher discount rates to perceived risks. Low uncertainty avoidance scores (for example Singapore and Denmark), in contrast, indicate that a country’s residents have a higher tolerance for uncertainty and ambiguity, more readily accept change, and take more and greater risks (Hofstede, 1980). The most immediate association between high uncertainty avoidance and dividend payout strategies put forward in the literature reflects the fact that low dividends and higher retained earnings are investments in uncertain future profits and therefore associated with higher risk and ambiguity (Shao et al., forthcoming). Conversely, dividends take the form of cash and therefore are a ‘‘bird in the hand” (Gordon, 1959; Lintner, 1962). Uncertainty-avoiding cultures are thus often linked to high dividend payouts. This argument, however, fails to take into account that lower dividend payouts are more predictable. Shareholders and corporate insiders in uncertainty-avoiding cultures may prefer lower dividend payouts that facilitate a perception of certainty because it is easier to meet this level every period. After all, uncertainty avoidance is often expressed through a need for predictability (Hofstede, 1991). Moreover, in high uncertainty-avoiding societies investors and corporate insiders share a culturally rooted need for predictability in the form of financial stability within firms. This may be reflected in a greater concern about potential financial distress and the associated cash shortages that could force firms to raise funds on unfavorable terms or liquidate assets at large discounts (Chang and Noorbakhsh, 2009). As substantial cash dividend distributions reduce firms’ financial slack and flexibility, firms in uncertainty-avoiding societies may regard high dividend payouts as less accept4 Preferences and behaviors specific to firm leaders in individualist societies may further explain why they might accept higher dividends. Societies accentuating individual success may perceive the incapability to guarantee high dividend payouts as a personal failure of corporate insiders (Chui et al., 2002). As in individualist societies success stories are often closely associated with individual firm leaders or dominant owners, those who control corporations may share investors’ preferences for high dividend payouts to demonstrate their ability to generate shareholder value.

J.P. Fidrmuc, M. Jacob / Journal of Comparative Economics 38 (2010) 321–339

325

Table 1 Construction of the sample. Firms in the sample 58,670 45,965 37,109 2026 393 31 8 230 263 108 5797

Description Standard & Poor’s Capital IQ Public Companies Sample (June 2009 version) Thereof active and operating Missing dividend data, negative dividends, or dividends exceed sales in 2004 Financial firms (primary and/or secondary SIC between 6000 and 6999) Utility firms (primary and/or secondary SIC between 4900 and 4949) Missing sales or negative sales for 2001–2006 Firms listed in Luxembourg’s stock exchange Negative or missing normalized net income in 2004 Firms from countries that do not meet data requirements Firms with other missing data items Basic sample

able. This conjecture is consistent with the empirical evidence that firms in more uncertainty-avoiding countries maintain higher levels of cash to hedge against unpredictable adverse future events (Ramirez and Tadesse, 2009; Chang and Noorbakhsh, 2009). Individuals in low uncertainty avoidance countries, in contrast, ‘‘share an enhanced cultural inclination toward [. . .] confronting the unknown” (Licht et al., 2005, p. 246) and are more accustomed to dealing with ambiguity. We thus propose that in these societies insiders and investors are less concerned about the uncertainties associated with lower cash holdings, and therefore expect and accept higher dividend payouts. Hence, our third hypothesis is as follows: Hypothesis 3. Companies in low uncertainty avoidance countries tend to have higher dividend payouts. 3. Data Our main source of data is the Standard & Poor’s Capital IQ database, which provides data covering company information for 58,670 public companies, and of these, 46,186 companies in 122 countries are active and operating.5 Since accounting data are often reported and collected with a delay, our analysis uses accounting numbers from 2004. Table 1 summarizes the construction of our sample. First, we restrict the sample to firms with non-missing values for dividends to common and preferred shareholders and net income for 2004, as well as to firms with available sales and exchange rate data for the period from 2001 to 2006. Further, we restrict the sample to non-financial and non-utility firms. Finally, we eliminate firms trading in Luxembourg, firms with negative net income, firms with negative dividends or whose dividends exceed sales, and firms from countries for which we do not have scores on Hofstede’s culture dimensions. To ensure a meaningful basis for our cross-country regressions and the calculation of our country-level statistics, we also exclude countries for which we have less than five observations after the above sample adjustments. This brings about our basic sample of 5797 companies from 41 countries. Compared to the sample of 33 countries in La Porta et al. (2000), our empirical analysis covers a set of additional data, including for example China and India. These countries are not only of substantial relevance to the global economy but also add important variety in terms of cultural and financial market characteristics. Our sample also features five mandatory dividend countries (countries with a legal requirement that a certain fraction of earnings be paid out as dividends): Brazil, Chile, Greece, Peru, and the Philippines. We believe that excluding these countries may obscure important variability.6 Table 2 provides definitions and summarizes the construction of all variables. Central to our analysis is the dividend payout ratio. Since our analysis involves cross-country comparisons with accounting data adhering to different accounting standards, we opt to use three alternative measures of the dividend payout ratio: dividend to earnings, dividend to sales, and dividendto-cash flow. The numerator of all measures is the total cash dividend paid to common and preferred shareholders. The denominators are net earnings, sales, and cash flow from operations. The dividend-to-earnings ratio is more commonly used: it captures the essence of the payout policy in that it expresses the relative distribution of net income between dividends and retained earnings (La Porta et al., 2000). However, it has several problems: First, net earnings may depend on a country’s accounting conventions. Second, net earnings can easily be manipulated, which significantly impedes the measure’s comparability across countries. Third, a diversion of resources may occur before earnings are reported and thus the dividend-to-earnings ratio may overestimate the share of true earnings that is paid out in the form of dividends. As sales are far harder to manipulate or smooth through accounting practices and less likely to be subject to theft, we also employ the dividend-to-sales ratio. The tradeoff, however, is that sales are a very rough measure of funds available for distribution to equity holders, and the difference between sales and net earnings may vary across countries and correlate with cultural dimensions. For example, there is evidence in the literature that low uncertainty-avoiding cultures show greater achievement motivation and competitive aggressiveness (Hofstede, 1980; Lee and Peterson, 2000), which may result in lower profit margins relative to countries scoring high on uncertainty avoidance. Thus, we view sales as just a size deflator with no transparent economic interpretation attached 5

An ‘‘active company” is defined as a company that is not reorganizing, not out of business, not acquired, not no longer investing, and not in liquidation. We are grateful to an anonymous referee for this suggestion. Interestingly, four of the mandatory dividend countries in our sample have a Spanish tradition. The scores for these four countries on individualism, uncertainty avoidance and power distance are uniformly lower, higher and higher than the global average, respectively. All countries with compulsory dividend payouts rank among the highest for uncertainty avoidance. 6

326

J.P. Fidrmuc, M. Jacob / Journal of Comparative Economics 38 (2010) 321–339

Table 2 Variable definitions. Variable

Description/source

Dividend-to-earnings ratio

Total cash dividends paid to common and preferred shareholders as a fraction of earnings in fiscal year 2004. Earnings are measured after taxes and interest but before extraordinary items. Source – Standard & Poor’s Capital IQ database Total cash dividends paid to common and preferred shareholders as a fraction of sales in fiscal year 2004. Sales are net sales. Source – Standard & Poor’s Capital IQ database Total cash dividends paid to common and preferred shareholders as a fraction of cash flow from operations in fiscal year 2004. Source – Standard & Poor’s Capital IQ database Country individualism vs. collectivism (low individualism) score. Source – Hofstede (1980, 2001) Equals one if a firm’s country of origin scores above 51 (the sample median) on the individualism/ collectivism dimension, and zero otherwise. Source – Hofstede (1980, 2001) Country power distance score. Source – Hofstede (1980, 2001) Equals one if a firm’s country of origin scores above 58 (the sample median) on the power distance dimension, and zero otherwise. Source – Hofstede (1980, 2001) Country uncertainty avoidance score. Source – Hofstede (1980, 2001) Equals one if a firm’s country of origin scores above 64 (the sample median) on the uncertainty avoidance dimension, and zero otherwise. Source – Hofstede (1980, 2001) Natural logarithm of total assets in US dollars. Source – Standard & Poor’s Capital IQ database Net income in US dollars over total assets in US dollars. Source – Standard & Poor’s Capital IQ database Total debt outstanding in US dollars over total assets in US dollars. Source – Standard & Poor’s Capital IQ database Fraction of net income that a company is mandated to pay out in the form of dividends if the firm is listed in a mandatory dividend country. Source – PWC Worldwide Tax Summaries Equals one for mandatory dividend countries and zero otherwise. Used in unreported regressions with dividend to sales and dividend-to-cash flow as the dependent variable. Source – PWC Worldwide Tax Summaries Average annual percentage growth in real (net) sales over the period 2001–2006. Before computing sales growth, we translate net sales in local currency into US dollars using the average annual exchange rates for individual years and currencies. Net sales in US dollars are translated into real terms using the US GNP deflator. Sources – Standard & Poor’s Capital IQ and US Department of Commerce, Bureau of Economic Analysis The ratio of the value, to an outside investor, of US$1 distributed as dividend income to the value of US$1 received in the form of capital gains when kept inside the firm as retained earnings. The computation of this ratio is detailed in Appendix A. Sources – Organization for Economic Co-operation and Development. 2000– 2007. OECD Tax Database; PriceWaterhouseCoopers. 2006. Worldwide Tax Summaries; KPMG. 2004. Corporate Tax Rates Survey; Deloitte. 2004–2007. Tax Guides; BMF. 2005. Internationale Steuern im Vergleich. Monatsbericht des BMF, January Index created by examining and rating companies’ 1990 annual reports on their inclusion or omission of 90 items. These items fall into seven categories (general information, income statements, balance sheets, funds flow statement, accounting standards, stock data, and special items). A minimum of three companies in each country were studied. The companies represent a cross section of various industry groups; industrial companies represented 70%, and financial companies represented the remaining 30%. Source – International Accounting and Auditing Trends, Center for International Financial Analysis and Research Measures the extent to which public enforcement by the main government agency or official authority in charge of supervising securities is sufficient to elicit honesty from issuers. The index is a composite measure of proxies for: (1) characteristics of the supervisor of securities markets; (2) power of the supervisor to issue rules; (3) investigative powers of the supervisor of securities markets; and (4) (criminal) sanctions that may be directed to the issuer, distributor, and accountant in case of a defective prospectus. Source – La Porta et al. (2006) Measures the extent to which governmental authority is legitimately exercised only in accordance with written, publicly disclosed laws adopted and enforced in accordance with established procedure. The principle is intended to be a safeguard against arbitrary governance. Rule of law includes agents’ perceptions of the incidence of both violent and non-violent crime, the effectiveness and predictability of the judiciary, and the enforceability of contracts. The data averages rule of law estimates for the year 2004. Source – Kaufmann et al. (2007) Index formed by adding one when: (1) the law explicitly mandates or sets as a default rule that: (a) proxy solicitations paid by the company include a proxy form allowing shareholders to vote on the items on the agenda; or (b) a proxy form to vote on the items on the agenda accompanies notice to the meeting; or (c) shareholders vote by mail on the items on the agenda; (2) shareholders can not be required to deposit with the company or another firm any of their shares prior to a general shareholders meeting; (3) if the law explicitly mandates or sets as a default rule cumulative voting for candidates to the board of directors or supervisory boards and a mechanism of proportional representation; (4) if minority shareholders may challenge a resolution of both the shareholders and the board if it is unfair, prejudicial, oppressive, or abusive; (5) when the law or listing rules explicitly mandate or set as a default rule that shareholders hold the first opportunity to buy new issues of stock; and (6) when minimum percentage of share capital (or voting power) that the law mandates or sets as a default rule as entitling a single shareholder to call a shareholders’ meeting is less than or equal to 10%. Source – Djankov et al. (2008) Numerical measure of legal protection of minority shareholders against self-dealing by corporate insiders. Average of ex-ante and ex-post private control of self-dealing. Considers a fixed self-dealing transaction, and then measure the hurdles that the controlling shareholder must jump in order to get away with this transaction. Measures the intensity of regulation of self-dealing along a variety of dimensions, covering both public and

Dividend-to-sales ratio Dividend-to-cash flow ratio Individualism High individualism Power distance High power distance Uncertainty avoidance High uncertainty avoidance Size Return on assets Leverage Mandatory dividends Mandatory dividend dummy Sales growth

Tax advantage

Accounting standards

Public enforcement

Rule of law

Anti-director rights (AD rights)

Anti-self-dealing index (ASD index)

J.P. Fidrmuc, M. Jacob / Journal of Comparative Economics 38 (2010) 321–339

327

Table 2 (continued) Variable Share repurchases Private credit Free float Systematic risk Bankruptcy risk

Protestant Catholic Muslim Geography Contraception

Description/source private enforcement mechanisms, such as disclosure, approval, and litigation. Source – Djankov et al. (2008) US dollar value of common stock repurchases in 2004. Source – Standard & Poor’s Capital IQ database Ratio of credit from deposit-taking financial institutions to the private sector (IFS lines 22d and 42d) relative to GDP (IFS line 99b). Average of 1999 through 2003. Source – World Development Indicators Percentage of shares outstanding owned by investors other than strategic 5% investors and insiders. Source – Standard & Poor’s Capital IQ database Systematic risk is the slope of the 60-month regression line of the percentage price change of the stock relative to the percentage price change of the S&P 500. Source – Standard & Poor’s Capital IQ database Altman z-score in 2004. Calculated as 1.2 working capital/total assets + 1.4  retained earnings/total assets + 3.3 EBIT/total assets + 0.6 MV of equity/BV of liabilities + 1.0 sales/total assets. Source – Standard & Poor’s Capital IQ database Percentage of the population that is Protestant in 1980. Source – La Porta et al. (1999a,b) Percentage of the population that is Roman Catholic in 1980. Source – La Porta et al. (1999a,b) Percentage of the population that is Muslim in 1980. Source – La Porta et al. (1999a,b) Continent of the country Women who use modern methods of contraception in 2002. Calculated by dividing the total number of women of reproductive age (15–49) who are married or in union and who are currently using any modern methods of contraception by the total number of women of reproductive age (15–49) who are married or in union, multiplied by 100. Source – United Nations Population Division’s ‘World Contraceptive Use’

(La Porta et al., 2000). We include dividend-to-cash flow because it is a suitable measure of the actual amount of cash a company has left from its operations for cash distributions, capital investments and debt service. In all our regressions, we control for several firm- and country-level characteristics. First, we include firm size and sales growth to account for the fact that smaller, high growth firms may have better investment opportunities and thus lower dividend payouts relative to bigger and more mature companies (see, for example, Fama and French, 2001; Denis and Osobov, 2008). Also, agency problems may be more severe in larger, more mature firms with limited growth prospects and abundant free cash flow (Jensen, 1986). We measure firm size as the natural logarithm of the firm’s total assets. Our principal measure of investment opportunities is the real growth in annual sales over the 5-year period from 2001 to 2006, which has the advantage of being relatively independent of differences in accounting practices around the world. Second, dividend payouts may be influenced by differences in capital structure and profitability across firms and countries. Financial leverage is closely associated with the probability of bankruptcy and may substitute for the control function of dividend payouts (Jensen et al., 1992).7 Profitability is an important indicator of the firm’s availability of funds for potential distribution. In our regressions, we therefore include control variables for company leverage, measured by the ratio of total debt to total assets, and profitability, measured by the fraction of net income relative to total assets. At the country level, we control for the effect of mandatory dividend requirements by including a variable that measures the fraction of net income that public companies are mandated to pay out in the form of dividends. We also include a measure of the tax advantage of dividends based on Poterba and Summers (1984, 1985). Appendix A summarizes in detail our computations of the tax advantage control variable. Panel A in Table 3 summarizes our sample by presenting the number of observations for each country as well as country medians for the three payout ratios, values for the three cultural dimensions and several legal variables. Panel B provides correlation coefficients for our main variables. It shows that individualism and power distance are negatively correlated, while uncertainty avoidance is orthogonal to both individualism and power distance.

4. Results 4.1. Main results Table 4 shows our results with the dividend-to-earnings ratio as the dependent variable. We employ a random effects specification to explicitly account for the cross-correlation between error terms for firms in the same country (La Porta et al., 2000). Across all regression specifications we control for mandatory dividend requirements, country-specific dividend tax advantages and industry effects. At the firm level, we additionally control for firm size, return on assets, leverage, and investment opportunities. The results strongly support our hypotheses. When included as separate regressors, all three cultural dimensions have the predicted sign and are significant at the 1% level. First, in line with Hypothesis 1, firms in countries that score higher on individualism tend to have significantly higher dividend payout ratios (Model 1). This result supports our conjecture that in individualist countries, outside shareholders expect higher dividend distributions to discipline more self-interested corporate insiders. A regression specification with a dummy variable for individualism scores above the cross-country median (not reported) shows that high individualism results in dividend-to-earnings ratios that are on average 18% higher compared to collectivist countries. This is a sizable economic effect given the cross-country average of 48%. 7

We provide a detailed discussion of the effect of leverage and more thorough robustness checks in Section 4.3.

328

Number of firms Panel A: country level data Australia 222 Austria 28 Belgium 42 Brazil 24 Canada 176 Chile 62 China 140 Denmark 46 Finland 71 France 227 Germany 174 Greece 56 Hong Kong 336 India 425 Indonesia 46 Ireland 17 Israel 17 Italy 68 Japan 646 Malaysia 362 Mexico 32 Morocco 8 Netherlands 61 New Zealand 47 Norway 36 Pakistan 69 Peru 8 Philippines 23 Poland 40 Portugal 15

Individualism

Power distance

Uncertainty avoidance

Accounting standards

Public enforcement

90 55 75 38 80 23 20 74 63 71 67 35 25 48 14 70 54 76 46 26 30 46 80 79 69 14 11 32 60 27

36 11 65 68 39 63 80 18 33 68 35 60 68 77 78 28 13 50 54 104 81 70 38 22 31 55 64 94 68 63

51 70 94 76 48 86 30 23 59 86 65 112 29 40 48 35 81 75 92 36 82 68 53 49 50 70 87 44 93 104

75 54 61 54 74 52 n.a. 62 77 69 62 55 69 57 n.a. n.a. 64 62 65 76 60 n.a. 64 70 74 n.a. 38 65 n.a. 36

0.90 0.17 0.15 0.58 0.80 0.60 0.87 0.37 0.32 0.77 0.22 0.32 0.87 0.67 0.62 0.37 0.63 0.48 0.00 0.77 0.35 n.a. 0.47 0.33 0.32 0.58 0.78 0.83 n.a. 0.58

Rule of law

AD rights

1.82 1.81 1.51 0.33 1.80 1.18 0.39 1.97 1.93 1.41 1.73 0.81 1.37 0.00 0.84 1.58 0.75 0.65 1.34 0.55 0.40 0.04 1.77 1.92 1.97 0.86 0.62 0.65 0.42 1.19

4 2.5 3 5 4 4 1 4 3.5 3.5 3.5 2 5 5 4 5 4 2 4.5 5 3 2 2.5 4 3.5 4 3.5 4 2 2.5

ASD index 0.76 0.21 0.54 0.27 0.64 0.63 0.76 0.46 0.46 0.38 0.28 0.22 0.96 0.58 0.65 0.79 0.73 0.42 0.50 0.95 0.17 0.56 0.20 0.95 0.42 0.41 0.45 0.22 0.29 0.44

Median div. to earnings 0.68 0.43 0.32 0.48 0.42 0.63 0.42 0.34 0.61 0.31 0.39 0.44 0.50 0.29 0.38 0.40 0.56 0.37 0.21 0.39 0.28 0.75 0.54 0.77 0.63 0.51 0.62 0.54 0.46 0.59

Median div. to sales 0.04 0.01 0.01 0.04 0.03 0.04 0.04 0.02 0.03 0.01 0.01 0.03 0.03 0.02 0.03 0.02 0.05 0.02 0.01 0.02 0.02 0.05 0.02 0.05 0.03 0.03 0.08 0.04 0.02 0.02

Median div. to cash flow 0.38 0.14 0.15 0.24 0.22 0.33 0.24 0.18 0.32 0.16 0.18 0.27 0.25 0.14 0.23 0.21 0.33 0.17 0.10 0.22 0.16 0.25 0.25 0.44 0.24 0.25 0.23 0.28 0.15 0.16

J.P. Fidrmuc, M. Jacob / Journal of Comparative Economics 38 (2010) 321–339

Table 3 Descriptive statistics.

196 108 188 38 99 87 115 237 15 431 759

20 65 18 51 71 68 17 20 37 89 91

74 49 60 57 31 34 58 64 66 35 40

8 49 85 86 29 58 69 64 85 35 46

78 70 62 64 83 68 65 64 51 78 71

0.87 0.25 0.25 0.33 0.5 0.33 0.52 0.72 0.63 0.68 0.90

1.82 0.15 0.70 1.20 1.87 1.98 0.81 0.05 0.09 1.73 1.48

5 5 4.5 5 3.5 3 3 4 3 5 3

1.00 0.81 0.47 0.37 0.33 0.27 0.56 0.81 0.43 0.95 0.65

0.49 0.42 0.21 0.38 0.53 0.38 0.72 0.64 0.56 0.51 0.29

0.02 0.02 0.01 0.02 0.02 0.01 0.04 0.04 0.04 0.02 0.02

0.26 0.24 0.11 0.20 0.30 0.16 0.38 0.36 0.30 0.25 0.14

Average Median

141 68

50 51

54 58

62 64

64.26 64.00

0.53 0.58

0.91 1.19

3.65 4.00

0.54 0.47

0.47 0.46

0.03 0.02

0.23 0.24

0.486a 0.221 0.521a

0.171 0.373b 0.378b 0.178

0.659a 0.665a 0.201 0.515a 0.238

0.049 0.013 0.348b 0.444a 0.114 0.108

0.036 0.040 0.512a 0.449a 0.434a 0.139 0.510a

0.059 0.165 0.041 0.133 0.291c 0.066 0.039 0.260c

0.256c 0.089 0.114 0.079 0.384b 0.237 0.155b 0.308a

0.191 0.045 0.060 0.211 0.428a 0.343b 0.157 0.136

Panel B: correlation coefficients Individualism Power distance Uncertainty avoidance Accounting standards Public enforcement Rule of law AD rights ASD index Note. For variable definitions please refer to Table 2. a Statistical significance at the 1% level. b Statistical significance at the 5% level. c Statistical significance at the 10% level.

0.650a

0.147 0.091

J.P. Fidrmuc, M. Jacob / Journal of Comparative Economics 38 (2010) 321–339

Singapore South Africa South Korea Spain Sweden Switzerland Taiwan Thailand Turkey United Kingdom United States

329

330

J.P. Fidrmuc, M. Jacob / Journal of Comparative Economics 38 (2010) 321–339

Table 4 Cultural determinants of dividends. Variable Individualism

Model 1

Model 2

Uncertainty avoidance

Return on assets Leverage Mandatory dividends Sales growth Tax advantage Constant N

v2

Model 4 0.0006 (0.0017) 0.0064a (0.0011) 0.0026a (0.0005) 0.0808a (0.0186) 0.7242b (0.3492) 0.0820 (0.1051) 1.5939 (1.1182) 0.0428 (0.0498) 0.2727 (0.2017) 1.7970a (0.2250)

0.0067a (0.0016)

Power distance

Size

Model 3

0.0043a (0.0016)

0.0825a (0.0178) 0.7107b (0.3420) 0.0881 (0.1056) 1.4154 (1.1043) 0.0345 (0.0497) 0.1063 (0.1944) 0.9813a (0.1660) 5797 159.96a

0.0881a (0.0176) 0.7148b (0.3405) 0.0918 (0.1013) 1.3536 (1.0906) 0.0395 (0.0505) 0.3121c (0.1657) 1.7322a (0.2523) 5797 177.43a

0.0023a (0.0005) 0.0644a (0.0144) 0.6389b (0.3203) 0.0182 (0.1003) 1.3414 (1.1081) 0.0516 (0.0530) 0.3389b (0.1650) 1.3355a (0.1846) 5797 179.48a

5797 257.23a

Note. This table presents firm level regression results for the cross section of 41 countries. The dependent variable is the dividend-to-earnings ratio. We include country random effects and use industry dummies. All variables are described in Table 2. Robust standard errors are shown in parentheses. a Statistical significance at the 1% level. b Statistical significance at the 5% level. c Statistical significance at the 10% level.

Second, power distance is significant and negatively related to dividend payouts (Model 2). This is in line with our conjecture (Hypothesis 2) that in low power distance societies, where individuals have a greater distaste for privilege differentials and strongly emphasize equality, shareholders and insiders expect and accept higher dividend payouts. A dummy variable regression capturing the effect of scores above the cross-country median power distance (not reported) indicates that high power distance is associated with dividend-to-earnings ratios that are on average 21% higher relative to low power distance countries. Finally, Model 3 shows that higher uncertainty avoidance is significantly associated with lower dividend payments. As Hypothesis 3 suggests, shareholders and corporate insiders in societies with high uncertainty avoidance put greater emphasis on lower but predictable streams of dividends. In line with the previous evidence (Khambata and Liu, 2005; Ramirez and Tadesse, 2009; Chang and Noorbakhsh, 2009), these results also indicate that investors in high uncertainty avoidance countries accept lower cash distributions that accommodate companies with higher levels of cash and thus provide a cushion against potential unforeseeable financial hardship. The difference in average dividend-to-earnings ratios between high and low uncertainty-avoiding cultures is 12%.8 Model 4 includes all three cultural dimensions together. In interpreting these results, we have to acknowledge the correlations among the cultural variables. Panel B of Table 3 shows that individualism and power distance are highly negatively correlated, whereas uncertainty avoidance is orthogonal to both individualism and power distance. This is also reflected in Model 4. Individualism becomes statistically insignificant while power distance and uncertainty avoidance remain significant at the 1% level. However, when we drop power distance the statistical significance of individualism increases to above the 1% level. In sum, our results indicate that individualism, power distance and uncertainty avoidance are all significantly correlated with dividend payouts across countries. The results also show that among our cultural variables power distance has the highest explanatory power while individualism has the lowest. We repeat our analysis for dividend-to-cash flow and dividend-to-sales ratios and the results are slightly weaker than the ones presented above.9 First, the inherent crudeness of the dividend-to-sales measure takes its toll and reduces the statistical significance of the cultural dimensions. When we estimate Model 4 with dividend to sales as the dependent variable, power distance remains significant while uncertainty avoidance becomes insignificant on a conventional level. We believe this is due to the fact that also the ratio of earnings to sales is correlated with uncertainty avoidance. In fact, across our sample, firm earnings as a fraction of sales grows as uncertainty avoidance increases and drops back to low levels for very high uncertainty avoidance values. This inverse J relation probably reflects low profit margins and higher competition in low uncertainty avoidance countries and lower reported earnings in very high uncertainty avoidance countries (Hofstede, 1980; Lee and Peterson, 2000).10 As a consequence, we see that the relation between dividend to sales and uncertainty avoidance is not as strongly neg-

8 9 10

The regression specifications with dummy variables are not reported but are available upon request. These regressions are not reported but are available upon request. Note in Table 3 that higher uncertainty avoidance is associated with low investor protection, accounting standards and public enforcement.

J.P. Fidrmuc, M. Jacob / Journal of Comparative Economics 38 (2010) 321–339

331

ative as the relation for dividend to earnings reported in Table 4. As dividends are generally paid out of what is available for distribution at the bottom line (net earnings) and not based on firm top line performance (sales), our results for dividend to earnings are probably more reliable. Second, for the dividend-to-cash flow specification, both power distance and uncertainty avoidance are, as expected, significantly negative. Contrary to our hypothesis, however, the coefficient for individualism turns insignificant and negative. A closer investigation reveals that this negative relationship is driven by observations of relatively high dividend-to-cash flow ratios for the lowest quartile of individualism scores in our sample. For the second to the fourth quartile of individualism scores, the relationship between the dividend-to-cash flow ratio and individualism is statistically significantly positive (at the 1% level), as hypothesized. We cannot find any evidence in the literature according to which the observation that cash flow is considerably smaller relative to earnings in highly collectivist countries compared to moderately collectivist or highly individualist countries is driven by agency considerations. Overall, the results for dividend to sales and dividend-to-cash flow confirm our central conjecture that cultural dimensions are important determinants of payout policies around the world. 4.2. Robustness checks concerning legal variables As the first robustness test, we explore the possibility that our cultural variables may simply proxy for legal and other formal institutional determinants across countries. La Porta et al. (2000) show that minority shareholder rights are important determinants of payout policies and it is therefore possible that the inclusion of legal variables would decrease the significance of our cultural variables. Even though La Porta et al. (2000) emphasize the role of shareholder protection, we want to explore a wider set of legal institutions than are discussed in La Porta et al. (2006). We document the robustness of our results to the inclusion of countries’ accounting standards, public enforcement, rule of law and shareholder protection (anti-director rights and anti-self-dealing index) as additional legal controls. We complement the revised anti-director rights index with a newer anti-self-dealing index due to Djankov et al. (2008) that should be better suited for our analysis because it reflects investor protection against the ability of corporate insiders to divert corporate wealth to themselves. Detailed definitions of these variables are provided in Table 2. The results in Table 5 confirm that legal variables are significantly related to dividend payout policies. The coefficients for disclosure standards, public enforcement and anti-self-dealing rights are positive and significant, suggesting that, in line with La Porta et al.’s (2000) outcome model of dividends, better legal institutions result in higher dividend payouts. However, the rule of law and anti-director rights are not statistically significant. Importantly, the inclusion of the legal variables has mostly negligent effects on our cultural dimensions. As suggested by the correlation matrix (Table 3, Panel B), uncertainty avoidance is affected the most, especially through public enforcement in Model 6 where the coefficient for uncertainty avoidance becomes insignificant. Individualism is not significant in any of the models, but this is, similarly to the results presented in Table 4, due to its correlation with power distance. When included as the only cultural variable in the regressions (not reported), individualism remains significant at the 1% level in all models and is not affected by the inclusion of the legal indices. Power distance remains highly significant and the statistical significance even increases when we control for accounting standards and public enforcement in Models 5 and 6, respectively. The impact of public enforcement on our results deserves closer investigation. As Model 6 indicates, public enforcement in the form of a powerful authority in charge of the supervision of securities markets mitigates the effect of uncertainty avoidance on dividend payouts. To analyze this further, we include an interaction term of uncertainty avoidance with public enforcement (results not reported). The interaction term shows that dividend payouts are highest in countries with strong public enforcement and low uncertainty avoidance and lowest in countries with weak public enforcement and high uncertainty avoidance. Law and culture seem to perform complementary roles. A decomposition of the overall index of public enforcement into individual building blocks as defined in La Porta et al. (2006) shows that it is the regulator’s ability to impose sanctions (both criminal and non-criminal) that causes the strong overall effect of public enforcement: a higher ability to punish violations of securities laws is associated with higher dividends. This evidence shows that dividend payouts are highest in those countries where the social normative nature of culture prescribes high dividends, and at the same time securities laws are better enforced. It is also in line with Renneboog and Szilagyi (2006), who show that dividends often complement rather than substitute shareholders’ efforts to alleviate agency concerns. In sum, our results indicate that legal institutions and culture as a social institution are important complementary determinants of payout policies across countries. 4.3. Corporate debt ratios and payout policies Our second set of robustness checks addresses the issue that dividend and debt ratios are closely linked. After all, dividend payments reduce the value of equity and therefore increase financial leverage, and empirical evidence shows that dividend and debt policies may be determined simultaneously (Jensen et al., 1992), thus indicating potential endogeneity concerns. However, similar to Shao et al. (forthcoming), we argue that it is reasonable to consider financial leverage as exogenous in our setting because firms have less freedom to choose their capital structure relative to payout policies. This is for a number of reasons. First, debt ratios are remarkably stable over time and firms make their capital structure decisions for very complex reasons (Myers, 1984; Demirguc-Kunt and Maksimovic, 1999). Therefore, firms are likely to face given debt ratios when making their dividend payout decisions. Second, firms must often enter into complex contracts with bondholders and other creditors that govern the permissible dividend payout rate. Bondholders have a strong incentive to limit dividend payouts to avoid an adversarial move in the debt–equity ratio that raises their risk. Third, agency theory proposes that financial

332

J.P. Fidrmuc, M. Jacob / Journal of Comparative Economics 38 (2010) 321–339

Table 5 Robustness tests: legal variables. Variable

Model 5

Model 6

Model 7

Model 8

Model 9

Legal variable

Accounting standards

Public enforcement

Rule of law

Anti-director rights

Anti-self-dealing index

Individualism

0.0003 (0.0019) 0.0070a (0.0012) 0.0016a (0.0006) 0.0067a (0.0026) 0.0867a (0.0197) 0.7259b (0.3521) 0.0273 (0.1139) 1.7055 (1.1213) 0.0425 (0.0594) 0.2959 (0.2144) 1.3351a (0.3272)

Power distance Uncertainty avoidance Legal variable Size Return on assets Leverage Mandatory dividends Sales growth Tax advantage Constant N

v2

5477 261.90a

0.0008 (0.0015) 0.0081a (0.0011) 0.0003 (0.0012) 0.2953a (0.1103) 0.0789a (0.0186) 0.7464b (0.3579) 0.0643 (0.1038) 1.4177 (1.1447) 0.0498 (0.0513) 0.2747 (0.2015) 1.6024a (0.2719) 5749 290.17a

0.0004 (0.0017) 0.0058a (0.0010) 0.0025a (0.0006) 0.0338 (0.0341) 0.0824a (0.0184) 0.7176b (0.3476) 0.0736 (0.1070) 1.6158 (1.1010) 0.0408 (0.0492) 0.2651 (0.2031) 1.7265a (0.2435) 5797 258.55a

0.0005 (0.0016) 0.0063a (0.0011) 0.0027a (0.0004) 0.0138 (0.0306) 0.0815a (0.0193) 0.7295b (0.3525) 0.0789 (0.1034) 1.5787 (1.1392) 0.0438 (0.0512) 0.2597 (0.2244) 1.8469a (0.1962) 5797 278.63a

0.0005 (0.0017) 0.0068a (0.0010) 0.0014b (0.0007) 0.2021c (0.1077) 0.0807a (0.0186) 0.7213b (0.3493) 0.0631 (0.1056) 1.6705 (1.1124) 0.0436 (0.0498) 0.3599c (0.2126) 1.6887a (0.2399) 5797 269.17a

Note. This table presents firm level regression results for the cross section of 41 countries. The dependent variable is the dividend-to-earnings ratio. We include country random effects and use industry dummies. All variables are described in Table 2. Robust standard errors are shown in parentheses. a Significance at the 1% level. b Significance at the 5% level. c Significance at the 10% level.

leverage may serve as a substitute mechanism for dividends because it helps to control agency costs by reducing the cash at the disposal of insiders through regular interest payments (Easterbrook, 1984; Jensen, 1986). Overall, our arguments suggest causality to be primarily from capital structure decisions to dividend payout policies. Treating financial leverage as an exogenous variable in our analysis thus seems to be a reasonable approach. Nevertheless, as an additional check for potential endogeneity between dividend and debt ratios we perform several robustness tests (not reported). First, we rerun our regressions using 1- and 2-year lags of our leverage control variable. Second, following Dittmar and Mahrt-Smith (2007) we estimate reduced-form models that do not include leverage as an explanatory variable to check for a potential simultaneity effect when including leverage and the cultural dimensions in regressions at the same time. Third, we run separate regressions for subsamples of high and low leverage countries, defined as countries with leverage above and below the sample median of country medians. The results for all of these alternative specifications are very similar to those reported in Table 4 and confirm our hypotheses. Finally, firms in more shareholder-centered systems as opposed to more bank-centered systems may be more attentive to shareholder concerns and preferences and thus pay higher dividends. Therefore, we also control for country-level relative reliance on debt financing by including the average 1999–2003 ratio of credit from deposit-taking financial institutions to the private sector relative to GDP (private credit). Private credit enters insignificantly in all regression specifications and does not affect the sign or statistical significance of the cultural dimensions. Furthermore, we obtain information on countries’ financial architecture from Kwok and Tadesse (2006) and rerun our regressions separately for bank-based and market-based financial systems. The results again hold. However, as uncertainty avoidance is significantly related to the financial system dummy (Kwok and Tadesse, 2006), uncertainty avoidance is less significant for the bank- and market-based subsamples. 4.4. Robustness to different ownership structures Our hypotheses emphasize the acceptability of certain dividend payouts that inhere in the cultural values of societies and are commonly shared by both investors and corporate insiders. However, ownership structures vary across the world (La Porta et al., 1999a,b; Morck et al., 2005) and this variation has important implications for the nature of agency problems and therefore also for our analysis. In the US and the UK, stock ownership is often dispersed and the main agency problem is between managers and shareholders. Elsewhere, however, concentrated share holdings prevail and corporations are often controlled not by professional managers but by dominant owners.11 The predominant agency conflict thus lies between con11

Shleifer and Vishny (1997) and Morck et al. (2005) provide more detailed overviews of the diverse ownership structures prevailing across countries.

J.P. Fidrmuc, M. Jacob / Journal of Comparative Economics 38 (2010) 321–339

333

trolling or highly influential owners and minority shareholders. These differences in ownership structures and the associated differences in the nature of agency tensions across countries may potentially affect our results. Therefore, we opt for additional robustness checks accounting for ownership structures around the world. These robustness tests also control for the possibility that large shareholders may substitute for the disciplining role of dividends because they are relatively effective monitors themselves (Easterbrook, 1984).12 We perform three sets of robustness tests. First, we control for company-specific heterogeneity in ownership concentration by including free float as a control variable. Free float measures the percentage of shares outstanding owned by outside investors. Second, we run separate regressions for countries where dispersed ownership structures prevail and for countries with a dominance of concentrated shareholdings. Our cutoff point is the sample median of country medians for free float. Finally, we also run regressions excluding the US and the UK, which are the primary examples of dispersed ownership countries. The results from these additional robustness checks (not reported) confirm the results displayed in Table 4. The only exception is the coefficient for uncertainty avoidance, which is less significant for the low ownership concentration subsample when included together with individualism and power distance. However, uncertainty avoidance is statistically highly significant when included as the only cultural explanatory variable. 4.5. Robustness to alternative cultural measures Hofstede’s set of cultural dimensions is one of two sets of cultural measures widely accepted and frequently used in the literature. We therefore also test the robustness of our results to the alternative set of cultural dimensions developed by Schwartz (1994, 1999). Schwartz specifies three bipolar dimensions of culture that represent alternative resolutions to each of three problems that confront all societies. These are autonomy (intellectual and affective) versus embeddedness, egalitarianism versus hierarchy, and harmony versus mastery. The results using these alternative cultural dimensions are reported in Table 6 and strongly support our hypotheses. In particular, intellectual and affective autonomy are both highly correlated with Hofstede’s individualism dimension, and their corresponding coefficients in Models 10 and 11 are positive and highly significant. This indicates that high autonomy societies expect and accept higher payouts that help attenuate the more severe agency problems associated with insiders’ tendency to pursue their own interests more vigorously. By contrast, the coefficient for embeddedness is significantly negative, suggesting that in societies where people are embedded in the collectivity and striving toward the collectivity’s shared goals, even when decisions collide with their own interests, agency problems are less severe and the benefits of higher payouts are smaller. Furthermore, Hofstede’s power distance dimension and Schwartz’s egalitarianism versus hierarchy measure capture similar societal characteristics and are highly negatively correlated. Expectedly therefore, we find egalitarian cultures to have significantly higher dividend payout ratios, suggesting that where societies have a greater distaste for privilege differentials and strongly emphasize equality in wealth and power, both shareholders and insiders expect and accept higher levels of distribution. Analogously, we find that in hierarchical cultures dividend payouts are generally lower. Finally, we find mastery to be significant and negatively related to dividend payout ratios. However, when we include all the cultural dimensions together, the effect of mastery on dividend payouts becomes insignificant. Our results using Schwartz’s cultural measures contradict some of the evidence presented by Shao et al. (forthcoming). We believe that this is due to the more extensive country coverage of our data set, which contains 41 countries versus the 21 countries in Shao et al. (forthcoming), and thus allows for more compelling variation in country cultural characteristics. 4.6. Other robustness checks We also test the robustness of our findings to other potentially important determinants of dividend payouts at the firm level. First, it is possible that differences in dividend payout ratios may simply reflect cross-country variation in the prevalence of share repurchases. From an agency perspective, however, we believe that investors cannot consider share repurchases to be a credible substitute disciplining tool to dividend payouts ex-ante. This is because share repurchases are irregular in time and generally unanticipated by outside investors. Also, Jagannathan et al. (2000) find that repurchases do not appear to replace dividends but rather seem to serve the complementary role of paying out short-term cash flows, and are used by different firms. Similarly, Brav et al. (2005) maintain that firms repurchase when their stock has a good value. Finally, in contrast to dividends, share repurchases may be ‘‘unequal” monitoring devices, as some investors sell their shares to the company and others do not (Vermaelen, 2005). Nevertheless, to ensure that cross-country variation in the prevalence of share repurchases does not bias our results, we control for firms’ US dollar value of common stock repurchases in 2004. The share repurchase variable enters insignificantly in all regression specifications and does not diminish the statistical or economic significance of our cultural dimensions. Next, we control for systematic risk and bankruptcy risk by including the firm’s 5-year beta and the Altmann z-score as of 2004. Both control variables enter insignificantly in all specifications and do not affect the high statistical significance of the 12 Empirical evidence supports the substitute effect as it shows that concentrated ownership is associated with lower dividends (Rozeff, 1982; Moh’d et al., 1995; Renneboog and Trojanowski, 2006).

334

J.P. Fidrmuc, M. Jacob / Journal of Comparative Economics 38 (2010) 321–339

Table 6 Schwartz cultural value scores and dividends. Variable Intellectual autonomy

Model 10

Model 11

Model 12

Model 13

Model 14

Model 15

0.1785a (0.0040)

0.2867 (0.0001)

0.4297a (0.0000)

Egalitarian

0.3773a (0.0000) 0.2419a (0.0000)

Hierarchy Harmony

0.1944b (0.0240) 0.0167 (0.8663)

Mandatory dividends Sales growth Tax advantage Constant N

v2

0.1811 (0.1581) b

Mastery

Leverage

Model 18

0.1477 (0.1136) a

Embeddedness

Return on assets

Model 17

0.2421a (0.0003)

Affective autonomy

Size

Model 16

0.1136b (0.0113)

0.0791a (0.0000) 0.6339b (0.0437) 0.0318 (0.7511) 1.1796 (0.2787) 0.0438 (0.4026) 0.2267 (0.1610) 0.6953a (0.0050) 5720 167.43a

0.0821a (0.0000) 0.6748b (0.0391) 0.0965 (0.3540) 1.335 (0.2247) 0.0385 (0.4474) 0.1241 (0.4692) 0.3032 (0.1716) 5720 161.18a

0.0865a (0.0000) 0.6559b (0.0398) 0.0788 (0.4327) 1.124 (0.3020) 0.0414 (0.4164) 0.1738 (0.3047) 2.2791a (0.0000) 5720 170.96a

0.0777a (0.0000) 0.6688b (0.0406) 0.1029 (0.3103) 0.8917 (0.4163) 0.0423 (0.4103) 0.0653 (0.7300) 0.9574b (0.0123) 5720 214.80a

0.0824a (0.0000) 0.6582b (0.0403) 0.0435 (0.6684) 0.9879 (0.3699) 0.0358 (0.4761) 0.0743 (0.7209) 1.6025a (0.0000) 5720 188.76a

0.0737a (0.0000) 0.6443b (0.0455) 0.0095 (0.9220) 1.1568 (0.2995) 0.0484 (0.3567) 0.3338b (0.0448) 1.3011a (0.0054) 5720 171.41a

0.2187 (0.0246) 0.0740a (0.0000) 0.6390b (0.0447) 0.0123 (0.9046) 1.1157 (0.3055) 0.0462 (0.3827) 0.2773c (0.0938) 2.0813a (0.0000) 5720 209.42a

0.0804a (0.0000) 0.7102b (0.0394) 0.0996 (0.3094) 1.1897 (0.3079) 0.0359 (0.4743) 0.0876 (0.6511) 0.543 (0.2234) 5720 219.99a

0.0941 (0.6190) 0.0885a (0.0000) 0.6650b (0.0395) 0.0787 (0.4369) 1.0183 (0.3566) 0.0349 (0.4827) 0.0695 (0.7557) 1.8166a (0.0034) 5720 215.24a

Note. This table presents firm level regression results for the cross section of 39 countries. Schwartz value scores are not available for our sample countries Morocco and Pakistan. The dependent variable is the dividend-to-earnings ratio. We include country random effects and use industry dummies. Robust standard errors are shown in parentheses. a Statistical significance at the 1% level. b Statistical significance at the 5% level. c Statistical significance at the 10% level.

cultural dimensions. Furthermore, we make sure that we have not selected a particular year during national or international business cycles that is in some sense special, and re-estimate all regressions using 2005 values. Additionally, we re-estimate the 2004 and 2005 results using growth rates in total assets and earnings to check for the robustness of our growth measure based on sales. The results using these alternative specification strategies are again very similar in both sign and significance to the results reported in Table 4. A related concern is that our results are driven by firms from the US and Japan that account for almost one-fourth of the firms in our sample. However, the results from re-estimating all regressions without firms from Japan and the US yield even stronger evidence in favor of our hypotheses. Finally, we cross-check our findings with a dataset of 6982 firms in 41 countries sourced from the COMPUSTAT Global Industrial/Commercial database.13 Again, the results are very similar to the results presented in Table 4. 4.7. Causality checks Our final concern is that of reverse causality between dividend payouts and cultural dimensions. Given our research setup, reverse causality is highly unlikely for a number of reasons. First, in order for dividend payouts to cause differences in culture across countries, they would have to affect widespread beliefs about what is right and desirable. For this to be the case, dividends would have to either be central and salient, or, alternatively, strongly influence people’s day-to-day lives even if people are unaware of them (Licht et al., 2005). Hofstede’s data on cultural value dimensions, however, originate from respondents whose daily lives are unlikely to be more than tangentially affected by dividend payout strategies.14 Second, Hofstede collected his data on cultural values mostly during 1968–1973, while our dividend payout information is from 2004. Therefore, causality from dividend payout to cultural variable seems highly implausible. Finally, several scholars (Hofstede, 2001; Licht et al., 2005; Guiso et al., 2008) have emphasized the long-term persistence of culture even amid substantial changes in economic conditions or political institutions. 13 Even though the COMPUSTAT database provides more firm observations, we prefer Standard & Poor’s Capital IQ as it covers a greater range of company variables for a greater number of firms, which enables us to perform a more rigorous set of robustness checks. 14 Hofstede collects the work-related values of IBM employees as opposed to investment professionals.

335

J.P. Fidrmuc, M. Jacob / Journal of Comparative Economics 38 (2010) 321–339 Table 7 Robustness tests: instrumental variable estimation. Variable

Model 19

Model 20

Model 21

Individualism

Power distance

Uncertainty avoidance

1st stage

Return on assets Leverage Mandatory dividends Sales growth Tax advantage Catholic Muslim Protestant Continent

1st stage a

Cultural dimension Size

2nd stage

0.39 (0.64) 3.40 (2.21) 11.89b (5.36) 115.02b (42.70) 2.48 c (1.34) 15.90 (28.67) 0.08 (0.13) 0.25c (0.13) 0.34b (0.13) 12.83a (3.64)

0.0072 (0.0013) 0.0960a (0.0198) 0.6601c (0.3675) 0.0928 (0.1484) 1.5827 (1.0640) 0.0348 (0.0860) 0.0392 (0.2956)

2nd stage

1st stage a

0.66 (0.57) 1.16 (1.74) 7.88 (4.96) 44.85a (14.42) 0.65 (0.84) 3.29 (25.83) 0.05 (0.09) 0.39 (0.28) 0.36a (0.08) 6.42a (1.74)

0.0089 (0.0019) 0.0899a (0.0209) 0.6150 (0.4462) 0.1091 (0.1397) 0.9882 (1.0427) 0.0792 (0.0883) 0.1853 (0.3182)

Constant N F-test Hansen J-statistic

10.11 (27.59) 5797 32.27a

0.8207 (0.2522) 5797 0.33

a

87.01 (27.11)

5797 15.86a

0.0089c (0.0046) 0.0390 (0.0363) 0.8761c (0.4481) 0.0596 (0.1361) 2.2346 (1.9882) 0.0561 (0.0995) 0.4270 (0.4924)

0.62 (0.37) 0.23 (0.15)

Contraception a

2.89c (1.35) 3.51 (2.06) 3.06 (7.76) 8.91 (35.12) 2.01 (1.36) 24.08 (40.51)

2nd stage

a

1.8302 (0.3741) 5797 3.10

0.60 (0.35) 106.20a (43.89) 5547 4.40

1.7320a (0.4982) 5547 3.13

Note. This table presents firm level instrumental variable regression results for the cross section of 41 and 37 countries in Models 19–20 and Model 21, respectively. The corresponding cultural dimension is the instrumented variable. The dependent variable in the second stage is the dividend-to-earnings ratio. F-test refers to a test of joint significance of the excluded instruments in the first stage regression. Hansen J-statistic refers to the test of overidentifying restrictions under the null that the error term is uncorrelated with the instruments. We include country clusters and use industry dummies. All variables are described in Table 2. Robust standard errors are shown in parentheses. a Significance at the 1% level. b Significance at the 5% level. c Significance at the 10% level.

Nevertheless, as an additional robustness check we opt for an instrumental variable analysis that establishes an exogenous source of variation in culture to deal with potential causality concerns (Tabellini, 2008). A considerable challenge is to find valid instruments that are correlated with our cultural dimensions of individualism, power distance and uncertainty avoidance and at the same time are uncorrelated with the error term in our basic specification as shown in Table 4. Kwok and Tadesse (2006) propose three sets of instruments to isolate the exogenous component of uncertainty avoidance: religion, geography, and demography. Religion, measured as the fraction of the population that is Catholic, Muslim or Protestant, and geography turn out to be suitable instruments also for individualism and power distance. For uncertainty avoidance, we have to replace geography with contraception usage as it improves the exogeneity condition of the instruments. Table 7 shows that our results hold controlling for the endogeneity of the cultural dimensions. The coefficient estimates for individualism and power distance in Models 19 and 20 are both statistically significant at the 1% level and are higher in absolute value relative to the main results in Table 4. The coefficient for uncertainty avoidance exhibits lower statistical significance, but this may be due to a low correlation between uncertainty avoidance and the instruments as indicated by the F-test in Model 21, as well as insufficient independence between the instruments and the error process as indicated by the Hansen J-statistic. 5. Conclusion In this paper we extend the existing agency explanations of dividend payouts to encompass the preferences and behaviors of economic agents that inhere in the cultural value emphases of societies. We posit that the social normative nature of culture significantly alters the character of agency relations and governs the acceptability and legitimacy of different payout strategies across countries. Our results, based on Hofstede’s cultural dimensions and a data set of 5797 firms in 41 countries, provide strong evidence that culture substantially affects the dividend policies of firms around the world. Firms in countries that score high on individualism and low on power distance and uncertainty avoidance pay relatively higher dividends. These effects are economically significant and robust to a wide set of additional measurement and specification tests. Importantly, our results

336

J.P. Fidrmuc, M. Jacob / Journal of Comparative Economics 38 (2010) 321–339

from a combined analysis of cultural and legal variables do not alter our main conjecture but indicate that legal institutions and culture as a social institution have complementary effects on the dividend payouts of firms across countries. Our results contribute to the growing evidence in the literature that culture matters for a wide range of fundamental economic and capital market phenomena. With the empirical evidence of culture’s relevance for finance and economics mounting, the development of a solid theoretical foundation that reconciles existing empirical findings seems to be the next logical step. Such a theoretical foundation would also help to advance cultural finance’s recognition among financial economists and potentially lead to even more insightful new studies of culture’s consequences for a wide spectrum of corporate governance and capital market phenomena. Acknowledgments The authors would like to thank Daniel Berkowitz (the editor), Eva Liljeblom, Marieke van den Pool, Peter Roosenboom, Greg Trojanowski, Yupana Wiwattanakantang, and two anonymous referees for thoughtful comments on earlier drafts of this paper. We are also indebted to Martin Jacob for his invaluable input on taxation around the world. Earlier versions of this paper were presented at the European Financial Management Symposium 2009 Corporate Governance and Control in Cambridge, UK and at the 2nd Conference on Corporate Governance in Emerging Markets in São Paulo, Brasil.

Appendix A Table A1 Table A1 Construction of the tax advantage of dividends. Sources. Organization for Economic Co-operation and Development. 2000–2007. OECD Tax Database; PriceWaterhouseCoopers. 2006. Worldwide Tax Summaries; KPMG. 2004. Corporate Tax Rates Survey; Deloitte. 2004–2007. Tax Guides; BMF. 2005. Internationale Steuern im Vergleich. Monatsbericht des BMF, January. Country

(A)

(B)

Corporate tax

Australiaa Austria Belgiumb Brazil Canadac Chile Chinad Denmark Egypt Finland Francee Germanyf Greece Hong Kong Hungaryg Indiah Indonesiai Ireland Israel Italyj Japank Koreal Malaysia Mexico Morocco Netherlands New Zealand Norwaym Pakistann Panama Peru Philippines Poland Portugalo Singapore

(C)

(D)

(E)

(G)

(H)

Personal tax

Undistributed profits

Distributed profits

Capital gains

Dividends

Imputation rate

Value of $1 in dividends (1–B + E)  (1–D)

Value of $1 in capital gains (1–A)  (1–C/4)

Dividend tax preference (G/H)

.30 .34 .34 .34 .36 .17 .33 .30 .40 .29 .35 .40 .35 .18 .16 .36 .30 .13 .36 .33 .41 .30 .28 .33 .35 .35 .33 .28 .35 .30 .30 .32 .19 .28 .22

.30 .34 .34 .34 .36 .15 .33 .30 .40 .29 .35 .40 .35 .18 .16 .44 .30 .13 .52 .33 .41 .30 .28 .33 .42 .35 .33 .28 .42 .37 .34 .32 .19 .28 .22

.24 .00 .00 .15 .24 .45 .20 .43 .00 .29 .27 .00 .00 .00 .00 .00 .35 .20 .20 .13 .20 .00 .00 .00 .44 .00 .00 .28 .26 .27 .00 .10 .19 .00 .00

.49 .25 .15 .28 .46 .45 .20 .43 .00 .29 .56 .24 .00 .00 .35 .00 .35 .42 .25 .13 .10 .40 .28 .33 .44 .25 .39 .28 .00 .37 .30 .10 .19 .20 .00

.30 .00 .00 .00 .21 .15 .00 .00 .00 .29 .33 .00 .00 .00 .00 .00 .00 .00 .00 .00 .00 .19 .00 .33 .00 .00 .33 .28 .00 .00 .00 .00 .00 .00 .00

.52 .50 .56 .48 .46 .55 .54 .40 .60 .71 .43 .46 .65 .83 .55 .56 .46 .51 .36 .59 .53 .54 .52 .67 .33 .49 .61 .72 .59 .40 .46 .61 .66 .58 .78

.66 .66 .66 .64 .60 .74 .64 .62 .60 .66 .60 .60 .65 .83 .84 .64 .64 .83 .61 .65 .56 .70 .72 .67 .58 .66 .67 .72 .61 .65 .70 .66 .77 .73 .78

.78 .75 .85 .75 .76 .75 .84 .64 1.00 1.08 .72 .76 1.00 1.00 .65 .87 .71 .61 .59 .90 .95 .77 .72 1.00 .57 .75 .91 1.00 .96 .61 .66 .92 .85 .80 1.00

337

J.P. Fidrmuc, M. Jacob / Journal of Comparative Economics 38 (2010) 321–339 Table A1 (continued) Country

(A)

(B)

Corporate tax

South Africap Spainq Sweden Switzerlandr Taiwan Thailands Turkey United Kingdomt United Statesu

(C)

(D)

(E)

(G)

(H)

Personal tax

Undistributed profits

Distributed profits

Capital gains

Dividends

Imputation rate

Value of $1 in dividends (1–B + E)  (1–D)

Value of $1 in capital gains (1–A)  (1–C/4)

Dividend tax preference (G/H)

.30 .35 .28 .25 .25 .30 .33 .30 .40

.38 .35 .28 .25 .25 .30 .33 .30 .40

.10 .15 .30 .00 .40 .00 .00 .40 .15

.00 .45 .30 .40 .40 .10 .23 .35 .15

.00 .29 .00 .00 .10 .10 .00 .10 .00

.62 .51 .50 .45 .51 .72 .52 .52 .51

.68 .63 .67 .75 .68 .70 .67 .63 .58

.91 .82 .76 .60 .76 1.03 .78 .83 .88

Note. This table gives a fiscal year 2004 overview of the raw data and the calculations employed to derive the dividend tax advantage variable. Dividend tax advantage is the ratio of the value, to an outside investor, of US$1 distributed as dividend income to the value of US$1 received in the form of capital gains when kept inside the firm as retained earnings. For countries with no explicit 2004 tax data, we use most recent tax information. Consistent with La Porta et al. (2000) we use the tax rates faced by local residents who acquire minority stakes in publicly traded securities and hold their investments long enough to qualify for long-term capital gains tax. We combine federal and local taxes whenever possible. Furthermore, we follow Poterba’s (1987) assumption that the effective rate on capital gains is equivalent to one-fourth of the nominal rate. In order to compute the tax parameter, we follow La Porta et al. (2000) and use the criteria proposed by King (1977) to group the tax systems of the countries in the sample in three broad categories: the Classical System, the TwoRate System, and the Imputation system (see La Porta et al. (2000) and the OECD Tax Database for a more detailed description). a If the asset was acquired on or after 11:45 a.m. AEST on September 21, 1999 and has been held for at least 12 months, 50% of the nominal gain (with no indexing of costs for inflation) is included in the individual’s taxable income. b The corporate tax is levied at a rate of 33%, increased by a 3% crisis tax, which leads to a 33.99% rate. c The 36.1% corporate tax rate in Canada is computed as follows: 38% basic rate less 10% provincial abatement results in a 28% federal rate before surtax. A federal surtax of 4% is levied on this rate resulting in a 29% tax rate. Depending on the firm’s industry, a 7% general rate reduction or profits reduction applies. The resulting net federal tax rate of 22.1% is added to a ‘‘typical provincial rate” of 12–14%. Capital gains tax is the highest federal/provincial tax rate as applies in Newfoundland and Labrador at 24.3%. Gross-up provisions for dividends apply (gross-up dividend rate 125%). d The standard corporate income tax and local corporate income-tax rates are 30% and 3% respectively. e These rates apply to income earned in 2004, to be paid in 2005. For companies not paying the CSB (Contribution Sociale sur les Bénéficies), the corporate income-tax rates are 1.1 percentage points lower. The rate in column 2 shows the rate as from 1 July 2004 when the total prélèlement sociaux was increased from 10.0% to 10.3%. Capital gains arising from the sale of quoted or unquoted securities, as well as of shares in SICAV and FCP, are subject to a 16% rate (plus 11% social surcharges) where the proceeds of such sales exceed € 15,000 (2005 income). f German business profits are subject to two taxes, corporation tax and trade tax. Corporation tax is levied at a uniform rate of 25% and is then subject to a surcharge of 5.5% (the ‘‘solidarity levy”). The effective trade tax rate varies by location from – generally – just under 12% to just under 20% (around 18% for most larger cities). This tax is deductible as an expense for corporation tax. From January 1, 2002, only 50% of dividend income is taxable under German income tax law. The total tax burden is calculated as follows: .18 + (.25–.25 * .18) * 1.055 = .396275. g Distributed dividends that exceed a threshold equal to 30% of the value of the share are taxed at the shareholder level at a personal income tax rate of 35%. For dividends below this threshold, the rate is 20%. h Dividend distribution tax is levied at the rate of 12.8125% (12.5% plus surcharge of 2.5% of the tax) on the dividends distributed by the domestic company. i Capital gains as well as investment income are taxable as income at the individual income-tax rates (max. 35%). Dividends received by individuals from Indonesian companies are subject to a 15% withholding tax. In calculating tax liability, gross dividends are combined with other income received and the tax on the total income received is calculated using the progressive tax rates to 35%. The 15% withholding tax is credited against the total income tax due. j The rate of tax payable on capital gains from shareholding is 12.5% for non-qualifying shareholding in a company. An individual having qualifying shareholding or receiving dividends pays usual progressive tax rates on 40% of the capital gain. k Dividends distributed by listed corporations are withheld at a rate of 20% (10% for dividends distributed during the period between April 2003 and March 2009), and the taxpayer can choose not to include the dividend income in the tax return. On the other hand, if dividends are subject to an aggregate tax, the Credit for Dividends (to deduct 6.4–12.8% of dividend income from income tax and local inhabitants tax) is applicable. l Gross-up provisions apply (Gross-up dividend rate 119%). m Note the different calculation of the Value of $1 in Capital Gains in (H): (1–A), in accordance with the Norwegian ‘‘RISK-Method”. Source: Christiansen, V. (2004) Norwegian Income Tax Reforms, CESifo Dice Report, vol. 3, pp. 9–14. n Capital gains, realized within 1 year of acquisition are fully taxable; after 1 year, 75% of such gains are taxable and 25% are exempt. Dividend payment to a public company or an insurance company is subject to withholding tax rate of 5%. In all other cases withholding tax rate on dividend is 10%. The withholding of tax on dividend payment is considered as full and final discharge of tax liability. o The corporate tax rate is 25%, increased to 27.5% in most cases by a municipal surcharge (derrama) of 10%. Capital gains derived from the sale of shares held for more than 12 months are exempt from personal income tax. Only 50% of dividend income is taxable. Lisbon rates apply. p The corporate tax rate applicable to companies in 2004 is 30%. However, South Africa imposes an additional ‘Secondary Tax on Companies’ at the rate of 12.5% on any net dividends declared. The effect of this additional tax is that if a company distributes 100% of its retained earnings as a dividend, then an effective tax rate of 37.78% will apply. This does not apply to gold mining companies, which are taxed on a formula basis. Capital gains tax is charged at individual tax rates (40%) on 25% of the gain realized by an individual. q Capital gains from assets held for more than 1 year are included in the ‘‘special part” of the taxable base and are taxed at 15%. r The corporate income tax rate includes the church tax, while the personal income tax rate excludes it. The tax burden of income (and capital) varies from canton to canton. As a general rule, the approximate range of the maximum effective income tax rate on profit for federal, cantonal, and communal taxes is between 16.4% and 29.2%, depending on the company’s place of residence. The average tax rate is approx. 25%. s The withholding tax rate on dividends is 10% for individuals. However, individuals resident in Thailand are better off paying the normal progressive income-tax rates on dividends and claiming the credit, since the credit for those who earn more than Bt4m reduces the effective rate of taxation on dividends to below 10%. t Gross-up provisions apply (gross-up dividend rate 111.1%). u The US corporate tax rate includes a 6.6% (average) local tax rate on top of the adjusted central government corporate income tax rate of 32.7 tax rate. New York rates apply.

338

J.P. Fidrmuc, M. Jacob / Journal of Comparative Economics 38 (2010) 321–339

References Arrow, Kenneth, 1974. Limits of organization. W.W. Norton, New York. Baum, J. Robert, Olian, Judy D., Erez, Miriam, Schnell, Eugene R., Smith, Ken G., Sims, Henry P., Scully, Judith S., Smith, Ken A., 1993. Nationality and work role interactions: a cultural contrast of Israeli and US entrepreneurs’ versus managers’ needs. Journal of Business Venturing 8, 449–512. Brav, Alon, Graham, John R., Harvey, Campbell R., Michaely, Roni, 2005. Payout policy in the 21st century. Journal of Financial Economics 77, 483–527. Breuer, Wolfgang, Quinten, Benjamin, 2009. Cultural Finance. Working Paper. . Chang, Kiyoung, Noorbakhsh, Abbas, 2009. Does national culture affect international corporate cash holdings? Journal of Multinational Financial Management 19, 323–342. Christiansen, Vidar, 2004. Norwegian income tax reforms. CESifo Dice Report 3, 9–14. Chui, Andy C.W., Lloyd, Alison E., Kwok, Chuck C.Y., 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., Titman, Sheridan, Wei, John K.C., 2010. Individualism and momentum around the world. Journal of Finance 65, 361–392. DeAngelo, Harry, DeAngelo, Linda, Skinner, Douglas J., 2008. Corporate payout policy. Foundations and Trends in Finance 3, 95–287. Demirguc-Kunt, Asli, Maksimovic, Vojislav, 1999. Institutions, financial markets, and firm debt maturity. Journal of Financial Economics 54, 295–336. Denis, David J., Osobov, Igor, 2008. Why do firms pay dividends? International evidence on the determinants of dividend policy. Journal of Financial Economics 75, 123–138. Dequech, David, 2009. Institutions, social norms, and decision-theoretic norms. Journal of Economic Behavior & Organization 72, 70–78. Dittmar, Amy K., Mahrt-Smith, Jan, 2007. Corporate governance and the value of cash holdings. Journal of Financial Economics 83, 599–634. Djankov, Simeon, La Porta, Rafael, Lopez-de-Silanes, Florencio, Shleifer, Andrei, 2008. The law and economics of self-dealing. Journal of Financial Economics 88, 430–465. Easterbrook, Frank, 1984. Two agency cost explanations of dividends. American Economic Review 74, 650–659. Ekanayake, Samson., 2004. Agency theory, national culture and management control systems. The Journal of American Academy of Business Cambridge 4, 49–54. Faccio, Mara, Lang, Larry H.P., Young, Leslie, 2001. Dividends and expropriation. American Economic Review 91, 54–78. Fama, Eugene F., French, Kenneth R., 2001. Disappearing dividends: changing firm characteristics or lower propensity to pay? Journal of Financial Economics 60, 3–44. Farinha, Jorge, López-de-Foronda, Óscar, 2009. The relation between dividends and insider ownership in different legal systems: international evidence. The European Journal of Finance 15, 169–189. Gomes, Armando, 2000. Going public with asymmetric information, agency costs, and dynamic trading. Journal of Finance 55, 615–646. Gordon, Myron J., 1959. Dividends, earnings and stock prices. Review of Economics and Statistics 41, 99–105. Greif, Avner, Laitin, David, 2004. A theory of endogenous institutional change. American Political Science Review 98, 633–652. Guiso, Luigi, Sapienza, Paola, Zingales, Luigi, 2008. Social capital as good culture. Journal of the European Economic Association 6, 295–320. Guiso, Luigi, Sapienza, Paola, Zingales, Luigi, 2009. Cultural biases in economic exchange? Quarterly Journal of Economics 124, 1095–1131. Hechter, Michael, Opp, Karl-Dieter, 2001. Introduction. In: Hechter, M., Opp, K.D. (Eds.), Social Norms, xi–xx. Russell Sage Foundation. Herbig, Paul A., 1994. The innovation matrix: culture and structure prerequisites to innovation. Quorum, Westport, CT. Hodgson, Geoffrey M., 2006. What are institutions? Journal of Economic Issues 40, 1–25. Hodgson, Geoffrey M., 2007. Institutions and individuals: interaction and evolution. Organization Studies 28, 95–116. Hofstede, Geert, 1980. Culture’s Consequences: International Differences in Work-related Values. Sage Publications, Beverly Hills, CA. Hofstede, Geert, 1991. Culture and Organizations: Software of the Mind. McGraw-Hill, New York. Hofstede, Geert, 2001. Culture’s Consequences, second ed. Sage Publications, Beverly Hills, CA. Huang, Rocco, 2008. Tolerance of uncertainty and the growth of informationally opaque industries. Journal of Development Economics 87, 333–353. Jagannathan, M., Stephens, C.P., Weisbach, M.S., 2000. Financial flexibility and the choice between dividends and stock repurchases. Journal of Financial Economics 57, 355–384. Jensen, Michael C., 1986. Agency cost of free cash flow, corporate finance, and takeovers. American Economic Review Papers and Proceedings 76, 323–329. Jensen, Gerald R., Solberg, Donald P., Zorn, Thomas S., 1992. Simultaneous determination of insider ownership, debt, and dividend policies. Journal of Financial and Quantitative Analysis 27, 247–263. Johnson, Nancy B., Droege, Scott, 2004. Reflections on the generalization of agency theory: cross-cultural considerations. Human Resource Management Review 14, 325–335. Kaufmann, Daniel, Kraay, Aart, Mastruzzi, Massimo, 2007. Governance matters VI: governance indicators for 1996–2006. World Bank Policy Research Working Paper No. 4280. Khambata, Dara, Liu, Wei, 2005. Cultural dimensions, risk aversion and corporate dividends policy. Journal of Asia–Pacific Business 64, 31–43. King, Mervyn, 1977. Public Policy and the Corporation. Chapman and Hall, London. Kluckhohn, Clyde, 1962. Universal categories of culture. In: Tax, Sol (Ed.), Anthropology Today. University of Chicago Press, Chicago, pp. 304–320. Kohn, Melvin L., 1969. Class and Conformity: A Study in Values. Dorsey Press, Homewood, IL. Kwok, Chuck C.Y., Tadesse, Solomon, 2006. National culture and financial systems. Journal of International Business Studies 37, 227–247. La Porta, Rafael, Lopez-de-Silanes, Florencio, Shleifer, Andrei, 1999a. Corporate ownership around the world. Journal of Finance 54, 471–517. La Porta, Rafael, Lopez-de-Silanes, Florencio, Shleifer, Andrei, Vishny, Robert, 1999b. The quality of government. Journal of Law, Economics, and Organization 15, 222–279. LaPorta, Rafael, Lopez-de-Silanes, Florencio, Shleifer, Andrei, Vishny, Robert, 2000. Agency problems and dividend policies around the world. Journal of Finance 55, 1–33. La Porta, Rafael, Lopez-de-Silanes, Florencio, Shleifer, Andrei, 2006. What works in securities laws. Journal of Finance 61, 1–32. Lee, Sang M., Peterson, Suzanne J., 2000. Culture, entrepreneurial orientation, and global competitiveness. Journal of World Business 35, 401–416. Licht, Amir N., 2008. Social norms and the law: why peoples obey the law. Review of Law and Economics 4, 715–750. Licht, Amir N., Goldschmidt, Charan, Schwartz, Shalom H., 2005. Culture, law, and corporate governance. International Review of Law and Economics 25, 229–255. Licht, Amir N., Goldschmidt, Charan, Schwartz, Shalom H., 2007. Culture rules: the foundations of the rule of law and other norms of governance. Journal of Comparative Economics 35, 659–688. Lintner, John, 1962. Dividends, earnings, leverage, stock prices and supply of capital to corporation. Review of Economics and Statistics 44, 243–269. Mann, Leon, 1980. Cross-cultural studies of small groups. In: Triandis, Harry C., Brislin, Robert W. (Eds.), Handbook of Cross Cultural Psychology, vol. 5. Allyn and Bacon, Boston, pp. 155–209. Miller, Dale T., 1999. The norm of self-interest. American Psychologist 54, 1–8. Miller, Merton H., Modigliani, Franco, 1961. Dividend policy, growth and the valuation of shares. Journal of Business 34, 411–433. Moh’d, Mahmoud, Perry, Larry G., Rimbey, James N., 1995. An investigation of the dynamic relationship between agency theory and dividend policy. Financial Review 30, 367–385. Morck, Randall, Wolfenzon, Daniel, Yeung, Bernard, 2005. Corporate governance, economic entrenchment, and growth. Journal of Economic Literature 43, 655–720. Myers, Stewart C., 1984. The capital structure puzzle. The Journal of Finance 39, 575–592. Myers, David G., 1999. Social Psychology. McGraw-Hill, New York.

J.P. Fidrmuc, M. Jacob / Journal of Comparative Economics 38 (2010) 321–339

339

Poterba, James M., 1987. Tax Evasion and Capital Gains Taxation. Working Paper. . Poterba, James M., Summers, Lawrence, 1984. New evidence that taxes affect the valuation of dividends. Journal of Finance 39, 1397–1415. Poterba, James M., Summers, Lawrence, 1985. The economic effects of dividend taxation. In: Altman, Edward, Subramanyam, Martin (Eds.), Recent Advances in Corporate Finance. Richard D. Irwin Publishers, Homewood, Ill., pp. 227–284. Putnam, Robert D., 2000. Bowling Alone: The Collapse and Revival of American Community. Simon & Schuster, New York. Ramirez, Andres, Tadesse, Solomon A., 2009. Corporate cash holdings, uncertainty avoidance, and the multinationality of firms. International Business Review 18, 387–403. Renneboog, Luc D.R., Szilagyi, Peter G., 2006. How relevant is dividend policy under low shareholder protection? Discussion Paper 73, Tilburg University, Center for Economic Research. Renneboog, Luc D.R., Trojanowski, Grzegorz, 2006. Control structures and payout policy. Managerial Finance 33, 43–64. Reuter, Charles-Henri., 2009. A Survey of Culture and Finance. Working Paper. . Rozeff, Michael S., 1982. Growth, beta, and agency costs as determinants of dividend payout ratios. Journal of Financial Research 5, 249–259. Schneider, Susan C., De Meyer, Arnoud, 1991. Interpreting and responding to strategic issues: the impact of national culture. Strategic Management Journal 12, 307–320. Schwartz, Shalom H., 1994. Beyond individualism/collectivism: new cultural dimensions of values. In: Kim, Uichol, Triandis, Harry C., Kagitcibasi, Cigdem, Choi, Sang-Chin, Yoon, Gene (Eds.), Individualism and Collectivism: Theory, Method and Applications. Sage Publications, Thousands Oaks, CA, pp. 85– 119. Schwartz, Shalom H., 1999. A theory of cultural values and some implications for work. Applied Psychology: An International Review 48, 23–47. Shane, Scott A., Venkataraman, Shane, 1996. Renegade and rational championing strategies. Organization Studies 17, 751–777. Shao, Liang, Kwock, Chuck, Guedhami, Omrane, forthcoming. National culture and dividend policy. Journal of International Business Studies. Shleifer, Andrei, Vishny, Robert W., 1997. A survey of corporate governance. Journal of Finance 52, 737–783. Tabellini, Guido, 2008. Institutions and culture. Journal of the European Economic Association 6, 255–294. Tajfel, Henri, 1981. Human Groups and Social Categories: Studies in Social Psychology. Cambridge University Press, New York. Vermaelen, Theo, 2005. Share repurchases. Foundations and Trends in Finance 1, 173–268. Weber, Max, 1930. The Protestant Ethic and the Spirit of Capitalism. Harper Collins, New York, NY. Weise, Peter, 2004. Kultur und die Vereinheitlichung der Sozialwissenschaften. In: Blümle, Gerold, Goldschmidt, Nils, Klump, Rainer, Schauenberg, Bernd, Senger, Harro von (Eds.), Perspektiven einer kulturellen Ökonomik. LIT, Muenster, pp. 427–440. Zwiebel, Jeffrey, 1996. Dynamic capital structure under managerial entrenchment. American Economic Review 86, 1197–1215.