Political institutions, stock market liquidity and firm dividend policy: Some international evidence

Political institutions, stock market liquidity and firm dividend policy: Some international evidence

Journal Pre-proofs Political Institutions, Stock Market Liquidity and Firm Dividend Policy: Some International Evidence Karen M.Y. Lai, Walid Saffar, ...

656KB Sizes 0 Downloads 0 Views

Journal Pre-proofs Political Institutions, Stock Market Liquidity and Firm Dividend Policy: Some International Evidence Karen M.Y. Lai, Walid Saffar, Xindong Kevin Zhu, Yiye Liu PII: DOI: Reference:

S1815-5669(19)30120-1 https://doi.org/10.1016/j.jcae.2019.100180 JCAE 100180

To appear in:

Journal of Contemporary Accounting & Economics

Received Date: Revised Date: Accepted Date:

26 March 2018 14 November 2019 17 December 2019

Please cite this article as: Lai, K.M.Y., Saffar, W., Zhu, X.K., Liu, Y., Political Institutions, Stock Market Liquidity and Firm Dividend Policy: Some International Evidence, Journal of Contemporary Accounting & Economics (2019), doi: https://doi.org/10.1016/j.jcae.2019.100180

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2019 Published by Elsevier Ltd.

POLITICAL INSTITUTIONS, STOCK MARKET LIQUIDITY AND FIRM DIVIDEND POLICY: SOME INTERNATIONAL EVIDENCE

Karen M.Y. Lai* Department of Accounting, Deakin University, Melbourne Walid Saffar School of Accounting and Finance, The Hong Kong Polytechnic University Xindong (Kevin) Zhu Department of Accountancy, City University of Hong Kong Yiye Liu Department of Accountancy, City University of Hong Kong

JEL Classification : G18; G38; P16; G35 ; G12 Keywords: Political institutions, legal institutions, dividend policy, stock market liquidity Acknowledgements: We thank Helen Lu, Wilson Tong and seminar/conference participants at Monash University, Sunway Campus and 2014 AFAANZ for their helpful comments. We also acknowledge the support given by the Internal Competitive Research Grant at the Hong Kong Polytechnic University (GUB27). *Corresponding Author: Karen M.Y. Lai, Deakin University, 221 Burwood Highway. Burwood. Australia. 3125, +61 3 9246 8650, [email protected]

POLITICAL INSTITUTIONS, STOCK MARKET LIQUIDITY AND FIRM DIVIDEND POLICY: SOME INTERNATIONAL EVIDENCE

Abstract In this cross-country study, we draw on the dividend liquidity hypothesis and the political economy literature to examine whether political institutions affect the relationship between stock market liquidity and a firm’s dividend policy. In countries with weak political institutions, we expect that investors are less able to demand higher dividends for stocks with low liquidity. Using a sample of 52 countries, we show that the negative association between stock market liquidity and dividends is more pronounced in countries with sound political institutions, consistent with the “outcome” model of dividends. These results are stronger in countries with better legal institutions and weaker for firms with financial constraints.

JEL Classifications: G18; G38; P16; G35; G12 Keywords: Political institutions, legal institutions, dividend policy, stock market liquidity

1

I. INTRODUCTION Banerjee et al. (2007) draw on the MM dividend irrelevance proposition (Miller and Modigliani, 1961) to examine whether there is a link between firm dividend policy and stock market liquidity. They find that in the US, firms with less liquid stocks are more likely to pay dividends, relative to firms with more liquid shares. The reason for this is that managers of firms with low liquidity (high market friction) pay dividends to improve the firm’s market valuation. Banerjee et al. (2007) refer to this relationship as the liquidity dividend hypothesis. In this paper, we extend this literature by examining the liquidity/dividend relationship in an international setting where countries have very different institutional settings compared with the US (Gugler, 2003). For example, while the US has relatively sound political and legal institutions, other countries, especially developing countries, have weak institutions.1 Since financial constraints have been identified as an important dimension in corporate finance (Farre-Mensa and Ljungqvist, 2016), we also consider whether financial constraints moderate the relationship between the dividend liquidity hypothesis and political institutions. This study is motivated by the following factors. First, Banerjee et al. (2007) do not consider the specific mechanisms through which investors’ demands affect dividends. However, several other studies suggest that investors in countries with strong institutional environments (e.g., sound political institutions) are in a better position to demand higher dividends for stocks with low liquidity. For instance, prior studies (e.g., Brav et al., 2005; Choy et al., 2011; Huang et Anecdotal evidence and reports suggest that over-regulation, solicitation of bribes, confiscatory taxation, and outright expropriation of assets are likely to affect operations in countries with weak political institutions (Stulz, 2005). A good case in point is Argentina, which some describe a country with about one hundred years of ineptitude due to the lack of institutions to create a successful market environment for investor trading activity and stock market liquidity (The Economist, February 15th, 2014). The Economist article draws attention to the signs of political institutional weaknesses, which include constant interruptions to democracy and the habit of tinkering from presidents to allow them to serve more terms. 1

2

al., 2015) show that firms in countries with strong shareholder rights are associated with higher dividends. Therefore, we contend that unlike the US, where there is evidence of substitution between liquidity and dividends in a relatively strong institutional environment (Banerjee et al., 2007), there are many other jurisdictions with poor institutional arrangements where this substitution relationship is less likely to exist or is weaker since investors are less able to demand higher dividends for stocks with low liquidity. While recent studies recognize the importance of political institutions in financial markets in an international setting (Boubakri et al., 2014; Fukuyama, 2014; Qi et al., 2010), there is relatively little systematic empirical evidence to back up these propositions. Second, if indeed firms with low liquidity have less ability to pay dividends in countries with

weak

political

institutions,

this

may

have

implications

for

market

capitalization/development in the economy. The absence of the firm’s ability to substitute dividends for poor liquidity could adversely affect market valuation and the ability of the firm to raise funds. In low liquidity markets, investors are also uncertain about executing large transactions, thus adversely affecting the development of the stock market, which in turn could negatively affect overall economic development. These ideas are consistent with Levine and Zervos (1998), who argue that stock markets become larger and more liquid and volatile following liberalization of capital controls and dividend flows. Therefore, an understanding of whether political institutions can increase the likelihood that dividends can substitute for low liquidity is important from the perspective of implications for stock market development and economic growth. Third, recent studies highlight the importance of political institutions as a regulator of the financial market environment. Surprisingly, there is relatively little integration of this

3

importance into the finance literature, despite the fact that it is known to influence a firm’s success or even survival (Roe and Siegel, 2011). While prior studies attempt to explain differences in finance and corporate policies across countries in terms of corporate and securities law institutions, legal origin, and trade openness, more recent studies call for a consideration of other more important factors, such as political institutions. For example, Roe and Siegel (2011: 3) posit that “primary institutions of investor protection such as courts, legal rules and regulators cannot function well in unstable political environments.” Moreover, there is growing evidence of the critical role of political institutions in financial markets (e.g., Haber et al., 2008; Henisz, 2000; Keefer, 2008; Knyazeva et al., 2013; Lederman et al., 2005; Roe and Siegel, 2011). More recently, Qi et al. (2010: 204) indicate that “some scholars argue that our understanding of how legal institutions impact financial development and economic growth may be incomplete, and provide evidence supporting the primary importance of political institutions.” Based on prior studies (e.g., Roe, 2006; Roe and Siegel, 2011), we conjecture that a country’s political institutions can affect the stock liquidity/dividend relationship through the channel of bargaining power. These prior studies suggest that the structure of political institutions affects the way a country protects its investors and enforces property rights. In other words, political institutions affect investors’ bargaining power in terms of the demand for dividends, which is likely to affect firms’ dividend policies. When countries have strong political rights, managers consider investors’ demand for dividends in determining dividend policies (Brav et al., 2005). However, in countries with weak political institutions, diffuse investors have little or no bargaining power to demand higher dividends even when there is lower liquidity in the stock market as a result of lower levels of investor rights and weak

4

property rights. Thus, in countries with weak political institutions, the relationship between dividend and liquidity may be weaker or not exist. The above arguments are consistent with the “outcome models” of how dividends are paid, as suggested by LLSV (La Porta et al., 2000a). Since the extant literature suggests a link between political institutions and financial outcomes, researchers are left to identify the most relevant political institutions and why they matter. Following Henisz (2012), we select the Political Constraints Index as our measure of political institutions, where higher scores indicate greater political constraints and the government’s ability to give assurances, and thus reflecting the soundness of political institutions.2 Based on a suggestion from prior studies (e.g., Borner, et al., 1995; North, 1981; Weingast, 1993) that a government’s ability to credibly commit not to interfere with private property rights is the essential component of political institutions that drives long term economic impact, Henisz derives the political constraints index and demonstrates that the feasibility of a change in policy can predict the cross-national variation in economic growth.3 Moreover, the index captures investors’ ex-ante views of restrictions on government behavior rather than ex-post government performance and performs well when the sample includes autocratic regimes (e.g., Qi et al., 2010). This measure is highly correlated with the degree of government commitment to private property rights and provides a good proxy for the extent to which polities can provide an environment for investors to demand dividend payments and/or to freely trade shares (Boubakri et al., 2014).

Political rights are also tied to ownership structures of firms (Roe 2003). Therefore, we also consider anti-director rights as an alternative measure of political institutions in our robustness test. 3 This is derived from a structural model of political interaction that incorporates the institutional constraints on the number of independent veto points in the political system and the distribution of political preferences across and within the executive, legislative, judicial, and sub-federal branches of government. 2

5

Using a sample of 311,849 firm-year observations in 52 countries between 1992 and 2016, we find that the negative relationship between dividend and liquidity is stronger for firms in countries with sound political institutions. Our results are robust to several tests, including different measures of stock liquidity, dividend policies and political constraints, alternative controls, and alternative samples. To address the potential endogeneity issue, we conduct an event study analysis around major changes in political constraints and provide evidence that the liquidity/dividend relationship is stronger (weaker) following a major improvement (deterioration) in the strength of political institutions (political constraints and political rights). Further, we find that the effect of political constraints on the relationship between liquidity and dividends is more pronounced in countries with better legal institutions. In summary, our study suggests that a combination of sound political institutions and strong legal protection does make a difference in the investors’ ability to extract dividends from low-liquidity firms. Moreover, our results also suggest that when firms are facing financial constraints, investors have less bargaining power for more dividends when stock liquidity is low even in countries with strong political institutions.4 This study contributes to several strands of the literature. First, we extend prior studies conducted in the U.S. (e.g., Banerjee et al., 2007) by providing evidence regarding the relationship between liquidity and dividend policies in an international setting. This is important given the globalization in trade and investments and the need to better understand how different political and institutional arrangements in a cross-section of countries can affect corporate finance practices. This additional insight should be useful to international regulators

Prior studies (e.g., DeAngelo and DeAngelo, 1990; Chen and Wang, 2012; Pathan et al., 2015) show that firms with higher external financing costs and constrained access to finance are likely to adopt a conservative dividend policy. 4

6

and country-level policymakers. Further, we extend prior studies which show that differences in political institutions, investor protection, laws and legal enforcement in different countries affect dividend policies (La Porta et al., 2000b; Haber, 2005; Haber et al., 2008; Keefer, 2008; Qi et al., 2010; and Roe and Siegel, 2011). Second, we extend recent work on the role of political institutions in determining financial outcomes. For example, this study complements the study by Qi et al. (2010), who show that a major determinant of debt in a cross country setting is political rights, and Boubakri et al. (2014), who show that sound political institutions are related to low cost of equity capital. In an earlier study, Eleswarapu and Venkataraman (2006) suggest that politically unstable environments generate high equity trading costs.

We extend this literature by presenting

evidence that political constraints are associated with international differences in the dividend/liquidity relationship. This evidence is not available in the extant literature. Third, we add to the literature that emphasizes the interdependence between the political and legal institutions (Fukuyama, 2014). For example, Qi et al. (2010) document that political and legal institutions are substitutes in explaining the cost of debt, and Boubakri et al. (2013) report that the effects of political institutions on corporate risk-taking is more pronounced in countries with high levels of corruption. The present study relates to these works on the relative importance of political institutions and legal institutions in financial markets by identifying the underlying channels through which political institutions may impact the dividend/liquidity substitutability in stock markets. More specifically, this paper shows that sound political institutions, by itself, without appropriate legal remedies may not be sufficient in providing a good investment environment.

7

Last but not least, we contribute to the finance literature by showing that the negative relationship between dividend and stock liquidity with strong political institutions is weaker for high financially constrained firms. Firms that are financially constrained are unable to pay dividends even if they are in politically stable environments. Our findings have enriched the existing knowledge of the effects of financial constraints firms and have improved our understanding regarding the relationship between dividend and stock liquidity in countries with strong political institutions. We believe that financial constraints constitute an important dimension that needs to be considered in evaluating the role of political rights in the liquidity/dividend relationship. The remainder of the paper is organized as follows. Section II reviews the literature and develops the hypotheses. Section III discusses the sample. Section IV reports the main results and robustness tests. Section V concludes.

II. BACKGROUND AND HYPOTHESES DEVELOPMENT 2.1. Political Institutions and Corporate Outcomes Stulz (2005), in his “twin agency model,” offers a theoretical framework to explain the link between government expropriation (e.g., autocratic regimes) and managerial diversion and decision-making. Stulz argues that those who control the state can use their power to improve their welfare. In addition, governments play an important role in managerial decision making by affecting the firms’ operating environment through, among other things, over-regulation, solicitation of bribes, and expropriation of the firms’ assets. When state expropriation (under weak political constraints) is high, those who control a firm can use their power for their own

8

benefit. He conjectures that these are twin problems rather than two separate problems. They flourish together because they feed on each other. Recent research in accounting and finance also suggests that political institutions affect real firm outcomes. Bushman et al. (2004), for example, show that higher transparency is observed in countries where political institutions are stronger, and Ben-Nasr et al. (2012) find that low political freedom is associated with high costs of external financing. Further, Boubakri et al. (2013) demonstrate that countries with poor political institutions are associated with overregulation, which discourages firms from taking more risks as investment risks increase. Qi et al. (2010) show that political institutions explain cross-country differences in the cost of debt, while Boubakri et al. (2014) more recently find evidence that sound political institutions are associated with international differences in the cost of equity capital. In the same way, countries with weak political institutions are more likely to have narrow capital markets since investors and managers are wary of government policies (Caprio et al., 2013). Accordingly, participation by outside investors will be low, and this leads to a smaller float of equity. Thus, in general, the extant literature suggests links between political institutions and firm outcomes. However, the links between political institutions and dividend policies are less clear. In theory, if weak political constraints are associated with high costs of external financing, then firms are less likely to pay out dividends in order to retain cash for financing future investment opportunities (Huang et al., 2015). In a similar vein, Julio and Yook (2012) and Huang et al. (2015) show that temporary political uncertainty can lead to delay in corporate investment and reduction in payouts. Conversely, Guedhami et al. (2017) find that firms in less politically free countries pay out more cash. They argue that low political freedom countries are associated

9

with the extraction of bribes and state expropriation. As a result, this would dampen a firm’s expected investment prospects and increase cash payouts. Given that recent studies on the effect of political institutions and corporate payouts provide mixed results and extant research seems to overlook the impact of political constraints on the liquidity-dividend relationship, this study aims to fill the gap in this area of research. 2.2. Hypotheses development As pointed out earlier, the primary channel by which political institutions can affect the dividend-liquidity relationship is through bargaining power. We next discuss the channel. Investor bargaining power: Starting with the notion that political institutions affect the level of investor protection and property rights enforcement (Roe, 2006; Roe and Siegel, 2011), we argue that in countries where political institutions are sound, investors have stronger bargaining power and can demand higher dividends when there is lower liquidity. Prior studies also suggest that managers consider investors’ dividend demand in determining dividend policies (Brav et al., 2005). This is consistent with Grossman, Hart, and Moore’s residual control rights framework (summarized in Hart, 1995), in which investors can demand cash only because they have power. Further, Roe and Siegel (2011) suggest that unstable polities are less likely to protect investors, implying that investors cannot depend on firms to pay dividends when liquidity is low or to freely and confidently trade shares when dividends are low. Other studies show that investors, especially minority shareholders, can enforce dividend policies in countries with strong investor protection (i.e., the outcome model; e.g., La Porta et al., 2000a). In other words, in countries with sound political institutions, investors can demand higher dividends in the event of low liquidity and perhaps accept lower dividends in the event of high liquidity

10

because of their strong bargaining power. However, the opposite may be true in countries with low investor protection, which are more likely to have weak political institutions. Thus, we predict that the negative association between liquidity and dividend to be stronger (weaker) in countries with sound (weak) political institutions where investors are able (unable) to demand higher dividends. This leads to our first hypothesis: H1: Strong political institutions lead to a stronger negative relationship between dividend and liquidity, all else being equal. Legal institutions: While the political economy literature (Roe and Siegel, 2011; Fukuyama, 2014) suggests that political institutions may have a bearing on legal institutions, in practice, countries with sound political institutions may have varying levels of legal institutions. Thus, we also argue that legal institutions should be considered as moderating the effect of political institutions on the negative relationship between dividend and liquidity. Recent studies (e.g., Qi et al., 2010) suggest that political institutions may complement or substitute for legal institutions. Political institutions may also affect the liquidity/dividend relationship through their impact on legal institutions and the consistency of the legal system (e.g., Roe and Siegel, 2011). Rajan and Zingales (2003: 18) argue that the essential ingredients of a developed financial system include the following: (1) respect for property rights, (2) an accounting and disclosure system that promotes transparency, (3) a legal system that enforces arm’s length contracts cheaply, and (4) a regulatory infrastructure that protects consumers, promotes competition, and controls egregious risk-taking.

11

The nature of the political system and the extent of corruption in the government affect its laws and law enforcement.5 Milhaupt and Pistor (2007) emphasize that the political economy determines the extent of a country’s stability and law enforcement. In countries with weak legal institutions, securities markets tend to expropriate outside investors. This legal system tends to be more formalized and less adaptable, resulting in a poorer environment for contracts (Levine, 2005). If the enforcement of laws is arbitrary, and managerial powers are concentrated in the hands of a few, then public trust in capital markets will be adversely affected (Eleswarapu and Venkataraman, 2006). An effective system of legal protection enables minority shareholders to use their legal powers to force companies to disgorge cash and preclude insiders from using too much of the company earnings to benefit themselves (La Porta et al., 2000a).6 Given the strand of literature documenting the significance of legal institutions in financial development and the growing importance of the political system, we also consider how these two types of institutions interact in the liquidity/dividend relationship. H2: The effect of political constraints on the relationship between liquidity and dividend is moderated by legal institutions. Financial Constraints: In this section, we consider the role of financial constraints in the relationship between political institutions and the dividend/liquidity relationship. A firm with financial constraints is likely to use internal funds for additional capital investment (both current and future) instead of external financing due to particularly high opportunity cost of investment (Almeida and Campello 2010). In other words, firms that are under credit constraints are likely to demand liquid assets such as cash and working capital and are likely to Prior studies also show that political institutions are related to the degree of corruption (Lederman et al., 2005; Qi et al., 2010). 6 For example, in countries with weak investor rights, firms pay lower dividends. See also Choy et al. (2011). 5

12

experience higher costs when they try to raise external funds (He et al., 2016). Thus, firms with credit constraints are likely to dampen corporate liquidity and reduce dividend payout. This argument is consistent with Campello et al.’s (2010) finding that constrained firms choose to cut their planned dividend distributions deeper during the financial crisis in 2008. Further, according to the residual dividend policy theory (Preinreich, 1932; Sage, 1937; Miller and Modigliani, 1961), firms only pay out dividends if they have excess cash after financing all profitable investments. Thus, we argue that financially constrained firms may not pay out dividends despite poor stock liquidity, even in countries with strong political institutions. Therefore, we also consider whether financial constraints affect the negative relationship between stock liquidity and dividends in countries with strong political institutions. H3: The effect of political constraints on the relationship between liquidity and dividend is moderated by different levels of financial constraints.

III. SAMPLE, VARIABLES, AND DESCRIPTIVE STATISTICS 3.1. Sample We collect our trading activity data from Compustat Global Security Daily files for the period between 1992 and 2016. We then match the firms with the required financial data in Compustat Global Fundamentals Annual files. We exclude (1) financial and utility firms due to the nature of their financial metrics, (2) observations with missing country-level data7, and (3) firms without consolidated financial statements. The sample includes both active and non-active firms to mitigate concerns about the survivorship bias. Our final sample includes 311,849

7 We also remove countries with less than 100 observations. In a sensitivity test, we add them back and find similar results.

13

observations, with 39,908 unique firms covering 52 countries. Table 1 describes the sample in detail. [Insert Table 1 about here] Table 2 provides the distribution of our sample across countries. Most of the sample is composed of firms from the US (27.38%), Japan (16.73%), China (8.07%), the UK (5.86%), and India (5.02%). The other countries in our sample each have less than 5% of the observations. This geographical spread is important because it suggests that our sample represents different development levels of legal, political, and institutional environments. [Insert Table 2 about here] 3.2. Variables The Appendix provides definitions and data sources for the variables in our study, which belong to four categories: dividend policy variables, trading activity measures, political variables, and firm- and country-level controls. 3.2.1. Dividend Policy Variables. Following Banerjee et al. (2007), we consider DIV as a dummy variable in the main analysis that equals one if the firm’s common stock has paid positive ordinary cash distributions for a given year. This is the proxy for dividend policy, which we derive from the merge of Compustat Global and Compustat North America. Following Choy et al. (2011), we employ DP as another proxy for dividend policy, representing the industry-adjusted dividend payout ratio computed as the firm’s dividend payout ratio minus the median of its country and industry. We define the dividend payout ratio as the total cash dividend paid to common and preferred shareholders deflated by earnings.

14

3.2.2. Trading Activity Measures. Following prior studies (Banerjee et al., 2007; Chordia et al., 2001; Datar et al., 1998), we use the percentage of the shares turnover, TURN, measured by the number of shares traded divided by the number of shares outstanding in a calendar year to measure trading activity. To assess the robustness of our results, we also measure a stock’s trading activities using DVOL, the natural logarithm of annual traded dollar volume in the security adjusted by the Consumer Price Index (Brennan et al., 1998; Chordia et al., 2001; Banerjee et al. 2007); NOTRD, the proportion of days with zero traded volume (Dumas and Luciano, 1991; Banerjee et al. 2007); and ILLIQ, the average ratio of the absolute daily return to daily dollar volume (Amihud et al., 1997; Amihud, 2002; Banerjee et al. 2007). 3.2.3. Political Variable To examine the relationship between political institutions and the dividend/liquidity relationship, we select Henisz’s (2012) political constraints index (POLCON) as a measure of the quality of political institutions. This index ranges from 0 to 1, with higher scores indicating greater political constraints and hence, stronger political institutions. This measure has been widely used in the literature and presents numerous advantages.8 First, Henisz (2000: p.1) concurs with prior studies (e.g., North, 1990) that a major determinant of political institutions is the government’s ability to credibly commit to not interfering with private property rights (Henisz, 2000; North, 1990), and thus a good measure of political institutions should have the ability to distinguish between the varying levels of constraints on policy change. Henisz’s index considers several characteristics of political rights, including the extent of constraints on veto

8 See for example, Stulz (2005); Qi et al. (2010) and Boubakri et al. (2013) for the use of the political constraints index.

15

players in the executive, legislative, judicial, and sub-federal branches of government and the distribution of preferences across and within those branches. Second, it captures investors’ exante views of restrictions on government behavior rather than ex-post government performance (e.g., Qi et al., 2010). Third, the index is available for more than 200 countries and over a long period, which covers all of our sample period. 3.2.4. Control Variables We control for firm and country characteristics that affect dividend policies (Banerjee et al., 2007; Choy et al., 2011). At the firm level, we include five control variables. First, we control for firm size, measured by the natural logarithm of total assets denominated in US dollars for a given year (SIZE). Second, we control for profitability measured by the earnings-to-assets ratio (ROA). Third, we control for growth opportunities proxied by the market-to-book ratio (MB). Finally, we control for leverage measured by the ratio of total debt to total assets (LEV) and stock return volatility (VOLATILITY) measured by the standard deviation of daily returns in the previous year. We include the natural logarithm of real GDP per capita (LNGDPC) as a countryspecific control variable to capture the country’s level of development. This ensures that our political variables do not only capture the effect of “rich” versus “poor” countries (Choy et al., 2011). Finally, we include country, year, and industry dummies to control for the different fixed effects of these variables. Country dummies should at least partially mitigate the endogeneity issue raised by the potential omitted country-level variables. 3.3. Descriptive Statistics Table 2 also summarizes the descriptive statistics of the key variables for each country. Consistent with previous studies, Belgium, France, Switzerland, Australia, Canada, and the US have strong political constraints and sound political institutions. Political constraints are weak

16

or even non-existent in many countries, such as China. Our dependent variable, DIV, varies between 0.22 in Jordan and 0.90 in Kenya. The turnover ratio (TURN) also seems to vary by country. Table 3 Panel A reports the descriptive statistics for the variables used in our main regression analysis. In terms of country-level variables, the mean (median) value for our main proxy for political institutions (POLCON) is 0.639 (0.748), with a standard deviation of 0.262. These statistics indicate that political institutions are not homogenous across our sample, and thus confirm that a cross-country analysis is appropriate for our investigation. The results further show that our sample includes countries with varying degrees of economic development as measured by the logarithm of GDP per capita. [Insert Table 3 about here] In terms of the firm-specific variables, the dependent variable DIV has a mean and standard deviation of 0.549 and 0.498, respectively. The liquidity variable TURN has a mean, median, and standard deviation of 0.717, 0.271, and 1.500, respectively. Our sample includes small and large firms, as well as high- and low-leverage firms. Firm size (SIZE) proxied by the natural logarithm of the mean (median) of total assets is 5.860 (5.678), equivalent to around $350.72 million USD ($292.36 million USD). The sample firms have an average leverage ratio of 0.569 and a mean ROA of 0.002, and a relatively high level of growth opportunities, with a mean (MB) of 2.335. Table 3 Panel B shows the Pearson correlations among the regression variables and pvalues estimated from two-tailed tests. Our dependent variable, the likelihood of a firm’s paying dividend (DIV) is significantly and negatively correlated with our liquidity measure (TURN), which is consistent with Banerjee et al. (2007). DIV is also significantly correlated with

17

the political constraints measure (POLCON). In addition, the magnitudes of the Pearson correlations do not suggest serious multicollinearity issues in our analysis.

IV. ANALYSIS AND RESULTS In this section, we report our results for the impact of political institutions (as well as firm and country characteristics) on the relationship between firm dividend policy and liquidity using a pooled multivariate regression framework. Following Banerjee et al. (2007), we perform annual logistic regressions using robust standard errors corrected for clustering at the firm level (Petersen, 2009) to explain the dividend policies of firms in different countries. Because the number of firms varies across countries, the individual observations are weighted with the inverse of the number of firms from the corresponding country. Specifically, we estimate the following model (subscripts are suppressed for notational convenience): DIV = β1+β2TURN + β3POLCON + β4 TURN *POLCON + β5 FIRMCONTROLS + β6 COUNTRY CONTROL +  Y 1 YEAR + Y 1



K 1 K 1

IND +  C 1 CNT + C 1

(1)

where DIV is an indicator variable that equals one if the common stock of the firm has paid positive ordinary cash distributions for a given year. POLCON is Henisz’ (2012) index of political constraints, and TURN is the number of shares traded divided by the number of shares outstanding in a calendar year. FIRM CONTROLS refers to the set of firm-level control variables (SIZE, LEV, MB, ROA, and VOLATILITY). COUNTRY CONTROL includes the level of development (LNGDPC). YEAR, IND, and CNT are dummies that control for year, industry, and country fixed effects, respectively, and  is an error term. The industry classification is based on the two-digit SIC code. We focus on the coefficient β4 in the analysis, which measures the sensitivity of the dividend/liquidity relationship to the quality of the political institutions

18

prevalent in the country. A negative value indicates that sound political institutions are likely to increase the distribution of dividends in firms with less liquid stocks and vice versa. 4.1. Main Regression Analysis Table 4 reports the results for the multivariate analysis. Model (1) presents the results for the relationship between firm dividend policy and stock market liquidity. Similar to Banerjee et al.’s (2007) findings, we observe a significantly negative coefficient on share turnover TURN (coefficient = -0.164 with z-statistic = -18.81), in line with the liquidity dividend hypothesis. This finding generalizes Banerjee et al.’s (2007) results from the US market to a more recent crosscountry analysis. Several control variables have the expected signs. Consistent with prior research, firm size (SIZE) and firm profitability (ROA) are positive and significantly correlated with DIV, suggesting that large and profitable firms are more likely to pay dividends. Further, high-growth and highly leveraged firms are less likely to pay dividends, consistent with contracting theory (Jensen, 1986; Smith and Watts, 1992). The logarithm of GDP per capita (LNGDPC) is negative and significant at the 1% level, suggesting that firms in developed markets are less likely to pay dividends compared to those in emerging markets. The association between DIV and VOLATILITY is also negative and significant, consistent with Hoberg and Prabhala’s (2009) view that risk decreases the propensity to pay dividends. [Insert Table 4 about here] Model (2) incorporates the political institutions proxy (POLCON), and the relationship between stock market liquidity and dividend payment remains unchanged. In Model (3), when we add an interaction term between POLCON and stock turnover (TURN), we find a more pronounced negative relationship between stock market liquidity and dividend payment in countries with tighter political constraints (TURN*POLCON = -0.069 with z-statistic= -4.45).

19

This result is economically significant. These findings confirm our expectation of a more distinct relationship between stock liquidity and dividend payment in countries with more political constraints. In other words, investors in firms with illiquid stocks are more likely to receive higher dividend payments in countries with sound political institutions. In Models (4) and (5), we split the sample into two groups, one for countries with weak political institutions (Model 4) and the other for countries with strong political institutions (Model 5) based on the median of the variable POLCON. We find that the coefficient of TURN is -0.096 for the first group and 0.380 for the second group, a difference that is statistically significant at the 1% level. This suggests that firms with illiquid (liquid) stocks are more (less) likely to pay dividends in countries with strong political institutions. Both the split sample design and the interaction terms support our first expectation that a country’s political institutions affect the liquidity dividend hypothesis. 9 4.2. Channels Tests To show that political institutions affect liquidity/dividend effect through the channel bargaining power, we re-run our regression models using anti-self-dealing index ANTISELF based on Djankov et al. (2008), as a proxy for investor bargaining power. The bargaining power proxy focuses on shareholder rights, which offers minority shareholders legal protection against controlling shareholders’ potential expropriation activities. Strong shareholder rights protection reduces large shareholders’ tunneling incentives and, consequently, increase the bargaining power

of

minority

shareholders.

Therefore,

the

effect

of

political

institutions

on

Since some countries may have more than one stock exchange, therefore we re-perform our main analysis by including a control variable EXCHANGES. EXCHANGES equals to 1 if a country has more than one stock exchange and 0 otherwise. The results (untabulated) show that TURN*POLCON remains negative and significant. 9

20

liquidity/dividend relationship should be stronger for firms in countries with better shareholder rights protection. ANTISELF captures regulations of corporate self-dealing transactions along three dimensions: disclosure, procedures to approve transactions, and facilitation of private litigation in cases of suspected self-dealing. The index ranges from 0 to 1, with higher values indicating higher levels of investor protection against self-dealing by controlling shareholders and corporate insiders. In columns (1) and (2) of Table 5, we present the results for low and high ANTISELF samples, respectively. We obtain a more negative and significant result for the interaction term TURN*POLCON for environments with high ANTISELF, suggesting that the relationship we document is mainly valid for countries in which shareholders have strong bargaining power. [Insert Table 5 about here] Overall, the results from these analyses provide insights on the channel through which the relationship between political institutions and liquidity/dividend operates. 4.3. Endogeneity Issue Following Boubakri et al.’s (2014) approach, we address the potential endogeneity issue by identifying events of major changes in political constraints and examine whether the association between liquidity and dividends changes around these events. We identify four major changes in political constraints: Indonesia in 1999, Pakistan in 1999, Argentina in 2002, and Thailand in 2006. Specifically, POLCON increased dramatically in Indonesia in 1999 from 0.246 to 0.420 and significantly decreased in Argentina in 2002 from 0.719 to 0.336, in Pakistan in

21

1999 from 0.280 to 0, and in Thailand in 2006 from 0.440 to 0.260.10 We consider these events as exogenous shocks to the political constraints and estimate Model 1 in Table 4 for each of the four countries for the two years before and two years after the event. As shown in Table 6, the coefficient of TURN decreased in Indonesia after political rights improved, from 0.126 (zstatistic of 0.19) to -0.879 (z-statistic of -2.08), suggesting a higher dividend for illiquid stocks after the change. The coefficients of TURN, however, increased in Argentina from -5.285 (zstatistic of -1.67) to 2.558 (z-statistic of 0.79), in Pakistan from -2.269 (z-statistic of -1.71) to -0.470 (z-statistic of -1.19), and in Thailand from -0.611 (z-statistic of -3.68) to -0.386 (z-statistic of -2.25) following the decrease in political constraints. These findings suggest that the relationship between liquidity and dividends weakened along with the political rights in these three countries. [Insert Table 6 about here] 4.4. Alternative Variables Table 7 provides additional support for our evidence using different proxies for the dependent variable and the main liquidity measure. In Models (1) and (2), we replicate Equation (1) using dividend payout ratio (DP) as the dependent variable, with and without the interaction term between liquidity and political institutions. Our results remain unchanged.11 In the remaining models of Table 7, we consider alternative variables for the liquidity measure. First, in Models (3) and (4), we test Equation (1) using trading volume (DVOL) instead of TURN These countries experienced political changes (see Boubakri et al., 2014: 335-336). For example, political rights in Indonesia improved in 1999 after the resignation of Suharto and the election of a new parliament, the first since 1955. Other countries experienced a sharp decrease in political constraints: in 1999, the Army General Pervez Musharraf conducted a coup d’état in Pakistan, which removed Prime Minister Nawaz Sharif; in Argentina, the political constraints fell in 2002 as a result of the economic crisis and political turmoil; and in 2006, the Royal Thai Army conducted a coup that overthrew Prime Minister Thaksin Shinawatra and his caretaker government. 11 There are fewer observations due to missing data for DP. 10

22

as the liquidity measure and obtain similar results. Firms with lower trading volume (DVOL) are more likely to pay dividends (Model 3). The interaction term, DVOL*POLCON in the regression of Model (4), loads negatively and is statistically significant at the 10% level, supporting our prediction that the relationship is more pronounced when political institutions are sound. Second, we re-examine Equation (1) using an inverse measure of liquidity, the proportion of non-trading days (NOTRD). The results shown in Models (5) and (6) show that firms with a higher proportion of non-trading days (NOTRD) are more likely to pay dividends, and the positive association is enhanced in countries with sound political institutions. Finally, we adopt another inverse measure of liquidity (ILLIQ) to test Equation (1), and the results are presented in Models (7) and (8). The positive coefficient of ILLIQ in Model (7) and the positive coefficient of ILLIQ*POLCON in Model (8) suggest that higher illiquidity is associated with higher dividend payment, and the positive association is stronger in countries with more sound political institutions, which supports our Hypothesis H1. [Insert Table 7 about here] Further, to test the robustness of our findings, we use the anti-director rights index (ANTIDIRECTOR) from LLSV (1998) as an alternative proxy to measure political institutions. Since the index places much emphasis on shareholder voting and minority investor protection, it can capture “the government’s ability to credibly commit to not interfering with private property rights,” which is a key component of political institutions that are likely to affect economic outcomes (Weingast, 1993 and Henisz 2000). The results in Table 8 show that the interaction term TURN*ANTIDIRECTOR is negative and significant, consistent with our previous results. In addition, when we partition the sample according to the median of ANTIDIRECTOR, the subgroup with high anti-director rights has a more negative coefficient

23

(TURN) than the subgroup with low anti-director rights. The difference of 0.211 is statistically significant at the 1% level. Overall, the results in Tables 7 and 8 reinforce our earlier evidence showing that the relationship between liquidity and dividends is stronger for firms headquartered in countries with sound political institutions. [Insert Table 8 about here] 4.5. Additional Control Variables Table 9 reports results for specifications that control for additional omitted variables to ensure that these variables are not driving our results. We include these variables separately in Models (1) through (5) and include them all together in Model (6). In Model (1), we control for investor legal protection (INVPROT)12 since prior studies (Beck et al., 2003; Choy et al., 2011) suggest that legal origins are important in explaining financial development and corporate finance policies. The coefficient TURN*POLCON remains negative and significant. In Model (2), following Banerjee et al. (2007), we use the proportional change in assets for year t (GROWTH) apart from market to book ratio (MB) as a proxy for growth opportunities. We repeat the analysis using both GROWTH and MB and still find a negative and significant coefficient for TURN*POLCON, as shown in Model (2) of Table 9.13 We also repeat the analysis by including a slope coefficient from the market model estimate (BETA) and the standard deviation of the residuals from the market model (STDRET). This is to proxy for systematic and idiosyncratic risk since the firm risk is a significant determinant of the propensity to pay dividends (Hoberg and Prabhala, 2009). As shown in Model (3) of Table 9, the coefficient of the interaction term We use a factor score based on five investor protection measures: an indicator variable equal to 1 if the country has a common law tradition, and 0 otherwise (COMMON); an index of anti-director rights (ANTIDIRECTOR); a measure of disclosure requirements (DISREQ); a liability standard index (LIAB); and an index of public enforcement of securities laws (PUBENF). 13 The correlation between GROWTH and MB is 0.11. In a sensitivity test, we remove MB from the model, and results are qualitatively similar. 12

24

(TURN) and (POLCON) remains negative and significant, corroborating our earlier findings. Following Choy et al. (2011), we include a country-level tax advantage variable (TAXADV) in Model 4 because tax advantages/disadvantages of dividend payments versus capital gains influence a firm’s dividend policy. We compute the dividend tax advantage variable as the ratio of the value of US $1 distributed as dividend income (to an outside investor) to the value of US $1 received in the form of capital gains when kept as retained earnings (La Porta et al., 2000a). Our results remain qualitatively similar. Choy et al. (2011) demonstrate that a country’s political economy, in particular, the type of electoral system, plays an important role in dividend policies. Thus, Table 9 also presents the results of an analysis that adds the degree of proportionality of a country’s voting system (PROP) (Pagano and Volpin, 2005) in the main regression analysis. The results of Model (5) in Table 9 show that the coefficient of the interaction term between political constraints measure (POLCON) and stock market liquidity (TURN) remains negative and significant after including PROP. Finally, in Model (6), we re-run the analysis and include INVPROT, GROWTH, BETA, STDRET, TAXADV, and PROP simultaneously in the regression, and the result still holds.14 Overall, the effect of political institutions on the association between stock market liquidity and dividends is consistent with our expectation: a higher degree of political rights has a strong effect on the negative relationship between liquidity and dividends. [Insert Table 9 about here] 4.6. Additional Tests

14 Since INVPROT, PROP, and TAXADV are time-invariant in our sample period for each country, we remove the country fixed effects from the regression models with INVPROT, PROP, and TAXADV.

25

To ensure that our results are not driven by US firms, since they dominate the sample (27.38%), in unreported regressions, we repeat the analyses and exclude US observations. The results are qualitatively similar to our main results. Additionally, to mitigate concerns that one particular country is driving our results, in unreported regressions, we rerun the analyses and exclude one country at a time from the basic regression (Model 3 of Table 4). Our results hold, even for the reduced subsample of firms. This evidence suggests that no one country, even if overrepresented, drives the results. 4.7. The Impact of Legal Institutions In this section, we test our second hypothesis H2 and consider the role of cross-country differences in legal institutions and whether they influence the association between political institutions and the dividend-liquidity relationship. Roe (2006) and Roe and Siegel (2011) argue that political institutions affect legal institutions and the constancy of the legal system. To examine whether corporate finance policy will depend on the political and legal institutions, we identify how these two interact to influence the relationship between stock market liquidity and dividend policy. Motivated by Hail and Leuz (2006), Boubakri et al. (2012), and Boubakri et al. (2014), among others, we consider several proxies of country-level legal institutions quality. First, we consider the law and order index measure (LAW&ORDER) from the International Country Risk Guide (ICRG) to control for the quality of the legal environment (i.e., the effectiveness of the country’s regulatory enforcement). The index LAW&ORDER ranges from 0 to 6, with higher values indicating higher quality of legal institutions. We also consider the likelihood of contract repudiation by the government (REPUD) (La Porta et al., 1998), where higher scores represent a lower risk of repudiation. Finally, we include the level of corruption control (CONTROLCORR), taken from Kaufmann et al. (2010) and the World Bank. This proxy

26

measures the extent to which public power is exercised for private gain, including both petty and grand forms of corruption. The higher the score, the stricter the control of corruption. In Table 10, we present the split sample results using the set of the quality of legal institutions variables. We find across all the legal variables that the interaction term TURN*POLCON is negative and statistically significant in the subsample of firms located in countries with sound legal institutions (Columns 2, 4, and 6). Reinforcing this evidence, the difference in the TURN*POLCON coefficients between the samples of weak and sound legal institutions is statistically significant at conventional levels for all three conditioning variables. These results are generally consistent with our prediction in H2 that political institutions have a stronger (weaker) influence on the relationship between stock market liquidity and dividends for firms in countries with strong (worse) legal institutions. [Insert Table 10 about here] 4.8. The role of financial constraints In this section, we investigate whether financial constraints moderate the association between political institutions and the dividend-liquidity relationship. The difficulties of obtaining funds for financially constrained firms are likely to hinder their ability to carry out their optimal investment and growth trajectories. Therefore they may manage liquidity by saving cash out of cash flows (Almeida et al. 2004). We thus expect a financially constrained firm with low stock liquidity to pay lower dividends even in countries with strong political institutions. We adopt a measure of financial constraint (FINCON) using the size-age index following Hadlock and Pierce (2010) and Cheng et al. (2014). We partition the sample into a lower degree of financial constraint (below-median FINCON) and a higher degree of financial constraint (above-median FINCON) and rerun the main regressions. As shown in Columns (1)

27

and (2) in Table 11, the interaction terms TURN*POLCON are negative and statistically significant in both higher and lower financial constraint subsamples, but higher in magnitude in the subsample with a lower degree of financial constraint. In addition, the difference in the TURN*POLCON coefficients between the subsamples of lower and higher financial constraints is statistically significant. Overall, these results are consistent with our expectations in the third hypothesis, that the impact of political constraint on the dividend-liquidity relationship is stronger in firms with a lower degree of financial constraints. 15 [Insert Table 11 about here] 4.9. Country Level Analysis In this section, we report the results of testing at the country level to buttress our main findings. If political institutions and corporate governance are weak, we expect that investors cannot demand high dividend payouts, even with low stock liquidity. We propose that the association between dividends and stock market liquidity in countries with poor political institutions and weak corporate governance will not hold. To test this assumption, we split our sample into two categories: six countries (Australia, Canada, France, Spain, Switzerland, and the US) with strong political institutions (POLCON) and strong corporate governance (CG);16 and six countries (Argentina, Columbia, Mexico, Nigeria, Philippines, and Sri Lanka) with weak POLCON and weak CG. To determine whether the country has a high or low level of corporate governance, we use a CG factor score based on a factor analysis of six distinct corporate

In addition, we also consider a firm with a higher degree of financial constraint when it reports negative net income in a given year (LOSS=1) and a lower degree of financial constraint when nonnegative net income is reported (LOSS=0). Our results (untabulated) show that only the subsample with low financial constraints LOSS=0 remains negative and significant. 16 To better differentiate countries with strong and weak political institutions, a country is classified as having strong political institutions if, on average POLCON, is greater than 0.80, and classified as having weak political institutions if, on average POLCON, is less than 0.50. 15

28

governance measures, ANTISELF, REPUD, CONTROLCORR, DISREQ, CIFAR, and SUE. The different definitions are provided in the Appendix. A higher CG factor score represents stronger corporate governance. The results are consistent with our expectations. As shown in Table 12, five out of six countries have strong CG and strong POLCON and are more likely to have a negative and significant association between liquidity and dividends, while countries with weak CG and weak POLCON are more likely to have insignificant results. These results are broadly consistent with our earlier findings. [Insert Table 12 about here] In addition, since country-level economic conditions may affect government policies that may, in turn, affect our main findings, we also consider the role of economic downturns. We perform another country-level analysis by partitioning our sample on the basis of economic cycles; that is, we split our samples into countries experiencing or not experiencing economic downturns. Following Qi et al. (2017), we define a firm-year observation as recession if any two consecutive seasonally adjusted quarterly real GDP growth rates during the fiscal year were negative (RECESSION = 1) and 0 otherwise. Our results in Table 13 Columns (1) and (2) show that the interaction terms TURN*POLCON are negative and statistically significant in both RECESSION=1 and RECESSION=0 subsamples, but the magnitude is higher in the subsample when RECESSION=0, indicating that in countries with strong political institutions, the negative relationship between stock liquidity and dividends is less pronounced during economic downturns. [Insert Table 13 about here] V. CONCLUSION

29

This paper examines whether there is an association between the liquidity dividend hypothesis and political rights in a cross-country setting. Using a sample of 311,849 firm-year observations in 52 countries between 1992 and 2016, we find that the negative relationship between dividend and liquidity is stronger for firms in countries with high political constraints (sound political institutions). The results support our conjecture that political institutions affect the liquidity dividend hypothesis through the bargaining power dimension. Following North and Weingast (1989) and Henisz (2004), we argue that weak political institutions and policy reversals undermine political credibility. Investor and firm managers’ wariness of unpredictable political environments are likely to affect stock market liquidity, preventing shareholders from extracting dividend payments from corporate insiders. Recent studies show that political and legal institutions are interdependent and might complement or substitute for one another (e.g., Qi et al., 2010). Thus, we also examine the interdependence of political and legal institutions and find that political constraints have a more pronounced effect on the relationship between liquidity and dividends in countries with better institutions governing investor protection. More interestingly, the results support previous findings that the legal system does not work independently of the political system (e.g., Milhaupt and Pistor, 2007). Our results are robust to a battery of tests, including alternative measures of liquidity, political institutions, and controlling for additional variables. Finally, we find that financial constraints moderates the association between political institutions and the dividend-liquidity relationship. Our paper extends the literature on the liquidity dividend hypothesis by examining the effect of a country’s political institutions. We also contribute by providing a better understanding of the channels by which political institutions impact stock market liquidity and dividend policy. We join recent studies in providing more evidence that political institutions are

30

vital to the development of liquid capital markets and the overall development of capital markets (e.g., Boubakri et al., 2013; Eleswarapu and Venkataraman, 2006; Roe and Siegel, 2011). Our findings have important policy implications for regulators and governments attempting to design policies to improve investment and business environments. Our study is subject to a number of limitations. First, some studies show that share repurchases consume cash that firms could distribute as dividends (Grullon and Michaely, 2002; Jagannathan et al., 2000). However, the increased popularity of open market repurchases is seen only in developed markets, and thus we are unable to test whether the results are driven by the increased repurchase activity of firms. Second, Banerjee et al. (2007) suggest that managerial stock options may incentivize managers to avoid paying cash dividends. However, the related data is unavailable, so we could not examine the effect of shares reserved for conversion on the association between stock market liquidity and firm dividend policy.

31

References Almeida, H., Campello, M., 2010. Financing frictions and the substitution between internal and external funds. Journal of Financial and Quantitative Analysis 45, 589–622. Almeida, H., Campello, M., Weisbach, M., 2004. The Cash Flow Sensitivity of Cash. Journal of Finance 59, 1777-804. Amihud, Y., 2002. Illiquidity and stock returns: cross-section and time-series effect. Journal of Financial Markets 5: 31-56. Amihud, Y., Mendelson, H., Lauterbach, B., 1997. Market microstructure and securities values: evidence from the Tel Aviv Stock Exchange. Journal of Financial Economics 45: 365-390. Banerjee, S., Gatchev, V.A., Spindt, P.A., 2007. Stock market liquidity and firm dividend policy. Journal of Financial and Quantitative Analysis 42, 369-398. Beck, T., Demirguc-Kunt, A., Levine, R., 2003. Law, endowments and finance. Journal of Financial Economics 70, 131-181. Ben-Nasr, H., Boubakri, N., Cosset, J., 2012. The political determinants of the cost of equity. Journal of Accounting Research 50, 605-646. Borner, S., Brunetti, A., Weder, B., 1995. Political credibility and economic development. St. Martin’s Press New York. Boubakri, N., El Ghoul, S., Saffar, W., 2014. Political rights and equity pricing. Journal of Corporate Finance 27, 326-344. Boubakri, N., Guedhami, O., Mishra, D., Saffar, W., 2012. Political connections and the cost of equity capital.” Journal of Corporate Finance 18: 541–59. Boubakri, N., Mani, S.A., Saffar, W., 2013. Political institutions, connectedness, and corporate risk-taking. Journal of International Business Studies 44, 195-215. Brav, A., Graham, J.R., Campbell, R.H., Michaely, R., 2005. Payout policy in the 21st century. Journal of Financial Economics 77, 483–527. Brennan, M.J., Chordia, T., Subrahmanyam, A., 1998. Alternative factor specifications, security characteristics, and the cross-section of expected stock returns. Journal of Financial Economics 49, 345–373. Bushman, R.M., Piotroski, J.D., Smith, A.J., 2004. What determines corporate transparency? Journal of Accounting Research 42, 207-252. Campello, M., Graham, J.R., Harvey, C.R., 2010. The real effects of financial constraints: Evidence from a financial crisis. Journal of Financial Economics 97, 470-487. Caprio, L., Faccio, M., McConnell, J.J., 2013. Sheltering corporate assets from political extraction. Journal of Law, Economics, and Organization 29, 332–354. Chen, S., Wang, Y., 2012. Financial constraints and share repurchases. Journal of Financial Economics 105, 311-331. Cheng, B., Ioannou, I., Serafeim, G., 2014. Corporate social responsibility and access to finance. Strategic Management Journal 35, 1-23.

32

Chordia, T., Subrahmanyam, A., Anshuman, V.R., 2001. Trading activity and expected stock returns. Journal of Financial Economics 59, 3–32. Choy, H., Gul, F.A., Yao, J., 2011. Does political economy reduce agency costs? Some evidence from dividend policies around the world. Journal of Empirical Finance 18, 16-35. Datar, V., Naik, N., Radcliffe, R., 1998. Liquidity and stock returns: an alternative test. Journal of Financial Markets 1, 203–219. DeAngelo, H., DeAngelo, L., 1990. Dividend policy and financial distress: An empirical investigation of troubled NYSE firms. Journal of Finance 45, 1415-1431. Djankov, S., LaPorta, R., Lopez-de-Silanes, F., Shleifer, A., 2008. The law and economics of selfdealing. Journal of Financial Economics 88, 430-465. Dumas, B., Luciano, E., 1991. An exact solution to a portfolio choice problem under transactions costs. Journal of Finance 46, 577-595. Eleswarapu, V. R., Venkataraman, K., 2006. The impact of legal and political institutions on equity trading costs. Review of Financial Studies 19, 1081-1111. Farre-Mensa, J., Ljungqvist, A., 2016. Do measures of financial constraints measure financial constraints? Review of Financial Studies 29, 271-308. Fukuyama, F., 2014. Political order and political decay: From the Industrial Revolution to the Globalisation of Democracy. Farrar, Straus and Giroux, New York. Grullon, G., Michaely, R., 2002. Dividends, share repurchases, and the substitution hypothesis. Journal of Finance 57, 1649-1684. Guedhami, O., Kwok, C.C.Y., Shao, L., 2017. Political freedom and corporate payouts. Journal of Corporate Finance 43, 514-529. Gugler, K., 2003. Corporate governance, dividend payout policy, and the interrelation between dividends, R&D, and capital investment. Journal of Banking and Finance 27, 1297-1321. Haber, S., 2005. Political institutions and financial development: Evidence from the economic histories of Mexico and the United States. Stanford Center for International Development Working Paper No 268. Haber, S., North, D., Weingast, B., 2008. Political institutions and financial development. Stanford University Press, Stanford, CA. Hadlock, C.J., Pierce, J.R., 2010. New evidence on measuring financial constraints: moving beyond the KZ index. Review of Financial Studies 23, 1909-1940. Hail, L., Leuz, C., 2006. International differences in the cost of equity capital: do legal institutions and securities regulation matter? Journal of Accounting Research 44, 485-531. Hart, O., 1995. Firms, contracts, and financial structure. Oxford University Press: London. He, Z., Chen, X., Huang, W., Pan, R., Shi, J., 2016. External finance and dividend policy: a twist by financial constraints. Accounting and Finance 56, 935-959. Henisz, W., 2000. The institutional environments for economic growth. Economics and Politics 12, 1-31.

33

Henisz, W., 2004. Political institutions and policy volatility. Economics and Politics 16, 1-27. Henisz, W., 2012. The political constraints index (POLCON) data set. Philadelphia: University of Pennsylvania. Hoberg, G., Prabhala, N.R., 2009. Disappearing dividends, catering, and risk. Review of Financial Studies 22, 79-116. Huang, T., Wu, F., Yu, J., Zhang, B., 2015. Political risk and dividend policy: evidence from international political crises. Journal of International Business Studies 46, 1–22. Jaganathan, M., Stephens, P.C., Weisbach, M.S., 2000. Financial flexibility and the choice between dividends and stock repurchases. Journal of Financial Economics 57, 355-384. Jensen, M., 1986. Agency costs of free cash flow, corporate finance, and takeovers. American Economics Review 76, 323-329. Julio, B., Yook, Y., 2012. Political uncertainty and corporate investment cycles. Journal of Finance 67, 45–84. Kaufmann, D., Kraay, A., Mastruzzi, M., 2010. The worldwide governance indicators: methodology and analytical issues. World Bank Policy Research. Working Paper 5430. Keefer, P., 2008. Beyond legal origin and checks and balances: political credibility, citizen information, and financial sector development. In: Haber, S., North, D., and Weingast, B. (2008). Political institutions and financial development. Stanford University Press, Stanford, CA, 125155. Knyazeva, A., Knyazeva, D., Stiglitz, J.E., 2013. Ownership change, institutional development and performance. Journal of Banking and Finance 37, 2605-2627. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R., 1998. Law and finance. Journal of Political Economy 106, 1113–1155. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R., 2000a. Agency problems and dividend policies around the world. Journal of Finance 55, 1–33. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R., 2000b. Investor protection and corporate governance. Journal of Financial Economics 58, 3–27. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R., 2006. What works in securities laws? Journal of Finance 61, 1-32. Lederman, D., Loayza, N., Soares, R., 2005. Accountability and corruption: political institutions matter. Economics and Politics 17, 1-35. Levine, R., 2005. Finance and growth: Theory and evidence. Handbook of Economic Growth. The Netherlands: Elsevier Science. Levine, R., Zervos, S., 1998. Capital control liberalization and stock market development. World Development 26, 1169-1183. Miller, M.H., Modigliani, F., 1961. Dividend policy, growth, and the valuation of shares. Journal of Business 34, 411–433. Milhaupt, C., Pistor, K., 2007. Law and capitalism: what corporate crises reveal about legal systems and economic development around the world. University of Chicago Press, Chicago, IL.

34

North, D.C., 1981. Structure and Change in Economic History. Norton & Co. New York, NY. North, D.C., 1990. Institutions, Institutional Change, and Economic Performance. New York: Cambridge University Press. North, D.C., Weingast, B.R., 1989. Constitutions and commitment: The evolution of institutions governing public choice in Seventeenth-Century England. Journal of Economic History 49, 803832. Pagano, M., Volpin, P., 2005. The political economy of corporate governance. American Economic Review 95, 1005–1030. Pathan, S., Faff, R., Méndez, C.F., Masters, N., 2015. Financial constraints and dividend policies. Australian Journal of Management 41, 484-507. Petersen, M.A., 2009. Estimating standard errors in finance panel data sets: comparing approaches. Review of Financial Studies 22, 435-480. Preinreich, G.A., 1932. Stock yields, stock dividends and inflation. The Accounting Review 7, 273–289. Qi, Y., Roth, L., Wald, J.K., 2010. Political rights and the cost of debt. Journal of Financial Economics 95, 202-226. Qi, Y., Roth, L., Wald, J.K., 2017. Creditor protection laws, debt financing, and corporate investment over the business cycle. Journal of International Business Studies 48, 477-497. Rajan, R., Zingales, L., 2003. The great reversals: the politics of financial development in the 20th century. Journal of Financial Economics 69, 5–50. Roe, M.J., 2003. Political determinants of corporate governance. Oxford University Press. Roe, M.J., 2006. Legal origins, politics and modern stock markets. Harvard Law Review 120, 460-527. Roe, M.J., Siegel, J.I., 2011. Political stability: effects on financial development, roots in the severity of economic inequality. Journal of Comparative Economics 39, 279-309. Sage, G.H., 1937. Dividend policy and business contingencies. Harvard Business Review 15, 245–252. Smith, C., Watts, R., 1992. The investment opportunity set and corporate financing, dividend and compensation policies. Journal of Financial Economics 32, 263-292. Stulz, R., 2005. The limits of financial globalization. Journal of Finance 52, 1595-1638. Weingast, B., 1993. Constitutions as governance structures: the political foundations of secure markets. Journal of Institutional and Theoretical Economics 149, 286-311. World Bank, 1997. The State in a Changing World. Oxford University Press, New York.

35

APPENDIX Variable Definitions Source

DVOL

Definition An indicator variable that equals 1 if the common stock of the firm has paid positive cash distributions for a given year, and 0 otherwise. Industry-adjusted dividend payout ratio, where dividend payout ratio is computed as total cash dividend paid to common and preferred shareholders deflated by earnings and industry-adjusted dividend payout ratio is computed as the firm’s dividend payout ratio minus the median of its country and industry. The percentage of the share turnover measured by the number of shares traded divided by the number of shares outstanding for a given year. The natural logarithm of annual traded dollar volume in the security adjusted by the Consumer Price Index.

NOTRD

The proportion of days with zero traded volume.

As above

ILLIQ

The average ratio of the absolute daily return to daily dollar volume.

As above

SIZE

The natural logarithm of total assets for a given year.

As above

ROA

Net income divided by total assets for a given year.

As above

MB

Market value of equity divided by the book value of common equity for a given year.

As above

LEV

Total liabilities divided by total assets for a given year.

As above

VOLATILITY

As above

LNGDPC

The measure of volatility for a given year, calculated as the standard deviation of daily returns in the one-year period. The measure of the degree of political constraints of a country. Derived from a model of political interaction that incorporates information on the number of independent branches of government with veto power, and the distribution of preferences across and within those branches. Government branches considered include chief executives, lower house of legislature, higher house of legislature, judiciary, and sub-federal branches. Higher scores indicate stronger political constraints and sound political institutions. The natural logarithm of GDP per capita for a given year.

ANTISELF

Average of ex-ante and ex-post private control of self-dealing.

Djankov et al. (2008)

ANTIDIRECTOR

An index aggregating shareholder rights concerning voting and minority protection. The index ranges from 0 to 5, with higher values indicating stronger protection of minority shareholders against insider expropriation. Measure of investor legal protection, which is a factor score based on factor analysis of five investor protection measures, including COMMON, ANTIDIRECTOR, DISREQ, LIAB, and PUBENF. Higher INVPROT represents stronger investor protection. COMMON is a dummy variable, coded as 1 if a country’s legal origin is based on English common law, and 0 otherwise. LIAB is a liability standard index measures the liability standard for investors to recover damages from issuers of securities, directors, and auditors when there have been misleading

La Porta et al. (1998)

Variable DIV DP

TURN

POLCON

INVPROT

36

Compustat Global & North America As above As above As above

Henisz (2012)

The World Bank

La Porta et al. (2006)

GROWTH

disclosures in financial statements. A higher value is considered to be related to stronger investor protection. PUBENF is a public enforcement of securities laws index, measuring to what extent regulatory agencies have investigative authority and the ability to punish firms and auditors that violate security laws. A higher value is considered to be related to stronger investor protection. One-year growth rate of total assets for a given year.

BETA

Slope coefficient from the market model estimated using daily return data over a given year.

Compustat Global & North America As above

STDRET

The standard deviation of the residuals from the market model estimated using daily return data over a given year.

As above

TAXADV

The relative tax advantage of dividend versus capital gain in a country, computed as the ratio of the value of US$1 distributed as dividend income (to an outside investor) to the value of US$1 received in the form of capital gains when kept by the firm as retained earnings. Degree of proportionality of a country’s voting system as constructed by Pagano and Volpin (2005). It is computed as Proportional Representation – Plurality – Housesys + 2. Assessment of the law and order tradition in the country. This variable ranges from 0 to 6, with higher scores indicating greater rule of law in the country. An assessment of the likelihood of contract repudiation by the government. A lower score represents a higher risk of repudiation. An assessment of one country’s control of corruption. Higher scores indicate stricter control of corruption.

La Porta et al. (2000a)

PROP LAW&ORDER REPUD CONTROLCORR

FINCON LOSS DISREQ

CIFAR

SUE

A measure of financial constraint, calculated as the SA index (see Cheng et al. 2014). In detail, SA index = (0.737*firm size) + (0.043*firm size^2) – (0.040*firm age). Higher SA index represents higher financial constraints. An indicator variable, coded as 1 if a firm reports negative net income in a given year, and 0 otherwise. The measure of disclosure requirements relating to the prospectus, compensation of directors and key officers, ownership structure, inside ownership, contracts outside the ordinary course of business, and transactions between the issuer and its directors, officers, and/or large shareholders. The index ranges from 0 to 1, with higher scores indicating more extensive disclosure requirements. An index created by examining and rating companies’ 1995 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 3 companies in each country was studied. A higher score represents a higher quality of accounting information. Index of the procedural difficulty in recovering losses from the auditor in a civil liability case for losses due to misleading statements in the audited financial information accompanying the prospectus. SUE equals one when investors are only required to prove that the audited financial information accompanying the prospectus contains a misleading statement. Equals two-thirds when investors must also prove that they relied on the prospectus and/or that their loss was caused by the misleading statement. Equals one-third when investors must also prove that the

37

The World Bank ICRG (2012) La Porta et al. (1998) Kaufmann et al. (2010), The World Bank Compustat Global & North America As above La Porta et al. (2006)

Bushman et al. (2004)

La Porta et al. (2006)

RECESSION

auditor acted with negligence. Equals zero if restitution from the auditor is unavailable or the liability standard is intent or gross negligence. A dummy variable that equals one if any two consecutive seasonal adjusted quarterly real GDP growth rates were negative during the firm’s fiscal year, and zero otherwise.

38

International Monetary Fund

Table 1 Sample Selection

Observations

Number of distinct firms

677,389

60,175

Observations with missing country-level data17

-330,100

-16,187

Observations with missing firm-level data

-35,440

-4,080

Primary sample

311,849

39,908

  Non-financial firms from Compustat Global Security Daily merged with Compustat Global Fundamentals Annual file for the period between 1992 and 2016 Less:

17

Countries with less than 100 observations are also excluded from our sample.

39

Country Argentina Australia Austria Belgium Brazil Canada Chile China Colombia Croatia Denmark Egypt Finland France Germany Greece Hungary India Indonesia Ireland Israel Italy Japan Jordan Kenya Korea Lithuania Luxembourg Malaysia Mexico Netherlands New Zealand Nigeria Norway Pakistan Peru Philippines Poland Portugal Russia Singapore South Africa Spain Sri Lanka Sweden Switzerland Taiwan Thailand Turkey UK USA Venezuela Total

N 799 15,458 625 1,011 2,805 6,373 1,066 25,153 174 151 1,115 173 1,452 6,068 5,989 1,673 235 15,643 2,404 694 1,711 1,796 52,160 561 281 7,098 242 292 9,517 1,152 1,855 840 299 1,958 2,109 504 1,318 2,253 395 463 5,700 2,892 879 806 3,390 2,129 11,287 4,031 1,091 18,272 85,388 119 311,849

% 0.26% 4.96% 0.20% 0.32% 0.90% 2.04% 0.34% 8.07% 0.06% 0.05% 0.36% 0.06% 0.47% 1.95% 1.92% 0.54% 0.08% 5.02% 0.77% 0.22% 0.55% 0.58% 16.73% 0.18% 0.09% 2.28% 0.08% 0.09% 3.05% 0.37% 0.59% 0.27% 0.10% 0.63% 0.68% 0.16% 0.42% 0.72% 0.13% 0.15% 1.83% 0.93% 0.28% 0.26% 1.09% 0.68% 3.62% 1.29% 0.35% 5.86% 27.38% 0.04% 100.0%

DIV 0.53 0.39 0.61 0.60 0.69 0.25 0.88 0.68 0.80 0.62 0.58 0.79 0.81 0.63 0.53 0.65 0.53 0.81 0.53 0.27 0.42 0.64 0.81 0.22 0.90 0.66 0.60 0.48 0.62 0.58 0.60 0.67 0.80 0.49 0.84 0.65 0.43 0.41 0.62 0.54 0.61 0.46 0.67 0.66 0.53 0.64 0.60 0.69 0.51 0.24 0.36 0.74 0.59

Table 2 Descriptive Statistics by Country TURN 0.09 0.29 0.15 0.15 0.28 0.29 0.09 2.30 0.21 0.08 0.26 0.29 0.21 0.19 0.25 0.41 0.30 0.45 0.26 0.19 0.40 0.34 0.46 0.41 0.07 1.23 0.10 0.32 0.25 0.14 0.36 0.10 0.07 0.37 0.34 0.14 0.16 0.25 0.18 0.11 0.30 0.16 0.36 0.18 0.30 0.24 0.88 0.47 1.12 0.31 0.98 0.26 0.35

SIZE 6.13 3.75 6.03 5.92 7.05 4.15 6.68 6.66 8.22 5.98 5.46 6.48 5.74 6.02 5.50 5.40 6.28 6.34 4.91 6.00 5.02 6.65 7.32 3.67 6.88 5.96 4.61 7.01 4.45 7.29 6.69 4.69 4.80 5.76 5.88 5.79 4.78 4.40 7.03 7.46 4.63 5.25 7.71 3.34 4.87 6.60 5.03 4.43 6.42 5.10 5.77 9.13 5.83

ROA 0.04 -0.07 0.01 0.02 0.05 -0.10 0.05 0.04 0.05 0.03 0.01 0.10 0.04 0.02 0.00 0.01 0.06 0.05 0.04 0.00 -0.01 0.01 0.02 0.04 0.08 0.01 0.04 0.02 0.02 0.04 0.03 0.01 0.08 0.00 0.08 0.07 0.02 0.03 0.02 0.06 0.03 0.06 0.04 0.05 -0.01 0.03 0.03 0.04 0.05 -0.01 -0.03 0.02 0.03

40

MB 2.29 2.71 1.80 1.95 4.09 2.28 4.83 3.47 2.57 1.38 2.82 2.49 2.38 2.33 2.18 1.41 1.49 2.02 2.35 2.53 3.73 1.95 1.43 1.66 2.27 1.25 1.59 3.76 1.37 2.96 2.67 2.67 4.36 3.36 1.80 1.78 2.08 2.14 2.19 1.63 1.57 2.48 3.04 2.06 2.94 2.73 1.84 1.55 2.08 2.75 2.68 2.86 2.40

LEV 0.45 0.30 0.54 0.49 0.48 1.40 0.43 0.32 0.34 0.46 0.50 0.46 0.49 0.59 0.49 0.54 0.37 0.39 0.52 0.49 0.44 0.56 0.42 0.32 0.32 0.54 0.44 0.44 0.42 0.47 0.57 0.44 0.62 0.49 0.44 0.43 0.40 0.43 0.72 0.42 0.43 0.48 0.63 0.46 0.50 0.50 0.43 0.45 0.44 0.49 0.87 0.35 0.49

VOLATILITY 0.03 0.04 0.03 0.03 0.04 0.05 0.02 0.04 0.02 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.04 0.04 0.03 0.02 0.03 0.03 0.03 0.04 0.03 0.04 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.04 0.04 0.03 0.02 0.03 0.04 0.03 0.02 0.04 0.03 0.03 0.03 0.03 0.03 0.03 0.04 0.04 0.03

POLCON 0.36 0.82 0.75 0.87 0.68 0.81 0.72 0.04 0.37 0.75 0.73 0.24 0.75 0.86 0.81 0.66 0.72 0.65 0.29 0.76 0.74 0.70 0.67 0.12 0.43 0.75 0.78 0.72 0.71 0.44 0.77 0.73 0.35 0.75 0.25 0.19 0.44 0.71 0.74 0.70 0.04 0.44 0.84 0.28 0.76 0.86 0.71 0.49 0.62 0.74 0.84 0.45 0.60

LNGDPC 9.04 10.54 10.55 10.54 8.77 10.42 9.04 8.34 8.52 9.41 10.75 7.61 10.54 10.41 10.49 10.09 9.09 7.14 7.30 10.58 10.19 10.42 10.55 8.02 6.74 9.79 9.15 11.33 8.67 8.92 10.47 10.17 7.06 11.07 6.84 8.09 7.26 9.22 9.84 9.17 10.40 8.43 10.19 7.70 10.60 10.89 9.71 8.05 9.04 10.37 10.70 8.61 9.36

Table 3 Descriptive Statistics and Correlation Matrix Panel A: Descriptive Statistics Variable

N

Mean

Std. Dev

25%

Median

75%

DIV

311,849

0.549

0.498

0.000

1.000

1.000

TURN

311,849

0.717

1.500

0.098

0.271

0.759

POLCON

311,849

0.639

0.262

0.445

0.748

0.851

SIZE

311,849

5.860

2.473

4.230

5.678

7.378

ROA

311,849

0.002

0.130

-0.006

0.026

0.064

MB

311,849

2.335

2.332

0.807

1.502

2.845

LEV

311,849

0.569

12.104

0.183

0.434

0.632

VOLATILITY

311,849

0.034

0.020

0.021

0.029

0.040

LNGDPC

311,849

9.920

1.152

9.497

10.453

10.699

Panel B:

Pearson Correlations DIV

TURN

POLCON

SIZE

ROA

MB

LEV

VOLATILITY

DIV

1

TURN

-0.045

1

POLCON

-0.127

-0.260

1

SIZE

0.387

0.122

-0.107

1

ROA

0.363

-0.015

-0.087

0.363

1

MB

-0.085

0.121

-0.052

-0.054

-0.010

1

LEV

-0.013

-0.005

-0.004

-0.052

-0.035

-0.011

1

VOLATILITY

-0.334

0.130

-0.073

-0.436

-0.428

0.030

0.054

1

LNGDPC

-0.163

-0.026

0.436

0.015

-0.188

-0.018

0.010

0.010

Bold text indicates two-tail significance at the .10 level or less. See Appendix for variable definitions.

41

Table 4 Stock Market Liquidity, Firm’s Dividend Policy and Political Constraints Model Dependent Variable= Intercept TURN POLCON

TURN*POLCON

?

-

(1)

(2)

(3)

(4)

(5)

DIV

DIV

DIV

15.067***

16.096***

16.297***

DIV Low POLCON 19.983***

DIV High POLCON 28.778***

(22.98)

(24.34)

(24.66)

(26.42)

(21.25)

-0.164***

-0.165***

-0.147***

-0.096***

-0.380***

(-18.81)

(-18.97)

(-17.62)

(-10.91)

(-15.01)

0.740***

0.796***

(10.83)

(11.35)

?

-

-0.069*** (-4.45)

SIZE

ROA

MB

LEV

VOLATILITY

LNGDPC

+ + ? ?

0.359***

0.358***

0.359***

0.365***

0.408***

(53.67)

(53.47)

(53.85)

(42.70)

(40.65)

6.096***

6.064***

6.073***

6.790***

5.039***

(59.22)

(58.83)

(58.78)

(48.97)

(35.58)

-0.031***

-0.030***

-0.029***

-0.014***

-0.043***

(-6.95)

(-6.60)

(-6.48)

(-2.57)

(-6.30)

-0.053

-0.108**

-0.105**

-0.765***

0.002***

(-1.19)

(-2.26)

(-2.20)

(-11.71)

(5.37)

-26.114****

-26.483***

-26.417***

-22.400***

-34.610***

(-46.89)

(-47.37)

(-47.41)

(-34.99)

(-33.78)

-2.058***

-2.228***

-2.260***

-2.632***

-2.876***

(-41.99)

(-43.22)

(-43.58)

(-41.53)

(-28.77)

Dif. In TURN*POLCON 0.284*** (p-value < 0.01) Industry Dummies Yes Yes Yes Yes Yes Year Dummies Yes Yes Yes Yes Yes Country Dummies Yes Yes Yes Yes Yes N 311,849 311,849 311,849 157,023 154,826 Pseudo R2 0.506 0.458 0.507 0.520 0.556 *, ** and *** represent two-tailed significance at level of 10%, 5% and 1% respectively. N denotes the number of observations. See Appendix for variable definitions. Z-statistics are shown in parentheses for country weighted PROBIT regressions, calculated based on robust standard errors clustered at the firm level.

Table 5 Stock Market Liquidity, Firm’s Dividend Policy and Political Constraints: Bargaining Power Model

(1)

Dependent Variable= Intercept

?

TURN

-

POLCON

?

TURN*POLCON

-

SIZE

+

ROA

+

MB

-

LEV

-

VOLATILITY

?

LNGDPC

?

Low DIV 22.150*** (18.13) -0.137*** (-7.25) -0.452*** (-2.40) -0.102*** (-4.16)

(2) ANTISELF High DIV 6.460*** (2.90) -0.207*** (-11.96) 0.221*** (2.38) -0.199*** (-7.91)

0.415*** (32.42) 8.194*** (40.49) -0.035*** (-4.19) -1.376*** (-16.27) -41.571*** (-30.83) -2.653*** (-27.71)

0.335*** (28.99) 3.928*** (26.70) -0.023*** (-2.97) 0.001*** (6.18) -27.913*** (-27.43) -1.146*** (-14.14)

Dif. In TURN*POLCON 0.097*** (p-value < 0.01) Industry Dummies Yes Yes Year Dummies Yes Yes Country Dummies Yes Yes N 112,193 169,656 Pseudo R2 0.431 0.431 *, ** and *** represent two-tailed significance at level of 10%, 5% and 1% respectively. N denotes the number of observations. See Appendix for variable definitions. Z-statistics are shown in parentheses for PROBIT regressions, calculated based on robust standard errors clustered at the firm level.

43

Table 6 Stock Market Liquidity, Firm’s Dividend Policy and Political Constraints: Event Study Analysis Model Country

(1)

(2) INDONESIA Low High POLCON POLCON

Dependent Variable= Intercept

?

TURN

-

SIZE

+

ROA

+

MB

-

LEV

-

VOLATILITY

?

(3) (4) ARGENTINA High Low POLCON POLCON

(5)

PAKISTAN

(6)

High POLCON

Low POLCON

(7)

(8) THAILAND High Low POLCON POLCON

DIV

DIV

DIV

DIV

DIV

DIV

DIV

DIV

2.308* (1.80) 0.126 (0.19)

2.653*** (3.09) -0.879** (-2.08)

-1.361 (-0.59) -5.285* (-1.67)

4.960** (1.88) 2.558 (0.79)

0.079 (0.12) -2.269* (-1.71)

1.036 (0.54) -0.470 (-1.19)

12.386 (0.07) -0.611*** (-3.68)

2.296*** (3.69) -0.386** (-2.25)

0.339** (2.12) 3.990** (2.29) -0.044 (-0.27) -1.658 (-1.25) -33.067** (-2.19)

0.046 (0.47) 2.646** (1.98) -0.043 (-0.69) -1.836** (-2.18) -27.793*** (-2.88)

1.135** (2.30) 5.844** (2.05) -0.954 (-1.13) -12.465*** (-2.31) 22.733 (0.61)

0.242 (0.99) 3.810 (0.84) 0.305* (1.78) -7.258*** (-2.76) -22.11*** (-3.05)

0.302 (0.64) 16.768 (1.55) -0.074 (-0.19) -2.359 (-0.58) 14.984 (0.41)

0.226 (0.73) 10.068** (2.12) 0.541* (1.69) -4.358** (-1.98) 14.106 (0.63)

0.321** (2.24) 7.349*** (3.65) 0.122 (1.37) -2.151*** (-2.88) -48.829*** (-4.16)

0.200** (2.20) 8.681*** (4.18) 0.012 (0.22) -1.181* (-1.75) -52.554*** (-5.31)

Industry Yes Yes Yes Yes Yes Yes Yes Dummies Year Dummies Yes Yes Yes Yes Yes Yes Yes Dummies N 183 229 56 98 41 99 582 Dummies Pseudo R2 0.222 0.156 0.548 0.317 0.302 0.240 0.343 *, ** and *** represent two-tailed significance at level of 10%, 5% and 1% respectively. N denotes the number of observations. See Appendix definitions. Z-statistics are shown in parentheses for PROBIT regressions, calculated based on robust standard errors clustered at the firm level.

44

Yes Yes 705 0.239 for variable

Table 7 Alternative Regression Models Model

(1) (2) Alternative dependent variable

Dependent Variable=

(3)

(4)

(5) (6) Alternative liquidity measures

(7)

(8)

DP

DP

DIV

DIV

DIV

DIV

DIV

DIV

15.980*** (24.54)

17.154*** (26.06)

15.860*** (24.70)

16.703*** (25.87)

15.385*** (23.23)

10.429*** (11.98)

Intercept

?

0.055 (0.35)

-0.302** (-1.99)

TURN

-

-0.006*** (-11.54)

-0.004*** (-6.78)

POLCON

?

-0.309*** (-39.10)

TURN*POLCON

-

-0.013*** (-13.19)

DVOL

-

DVOL*POLCON

-

NOTRD

-

NOTRD*POLCON

+

ILLIQ

-

ILLIQ*POLCON

+

SIZE

+

0.002*** (5.15)

0.001*** (3.14)

0.377*** (50.43)

0.380*** (50.23)

0.354*** (52.08)

0.354*** (51.57)

0.405*** (44.76)

0.429*** (36.77)

ROA

+

0.164*** (25.35)

0.097*** (12.91)

6.158*** (60.08)

6.128*** (58.86)

6.123*** (59.45)

6.098*** (58.76)

6.174*** (58.86)

5.398*** (40.12)

MB

-

-0.002*** (-6.52)

-0.004*** (-10.96)

-0.030*** (-6.64)

-0.027*** (-5.94)

-0.037*** (-8.31)

-0.036*** (-8.00)

-0.020*** (-4.23)

-0.033*** (-5.13)

0.981*** (11.35)

-0.039*** (-8.70)

0.631*** (9.26)

1.376*** (14.56)

-0.041*** (-9.06) -0.003* (-1.84) 0.766*** (12.62)

0.568*** (8.45) 0.333*** (3.62) 0.064*** (11.14)

0.066*** (8.65) 0.038*** (5.24)

45

LEV

-

0.001*** (2.82)

0.001*** (3.06)

-0.002 (-0.13)

-0.052 (-1.11)

-0.013 (-0.37)

-0.062*** (-1.32)

-0.118*** (-2.42)

-0.365*** (-5.64)

VOLATILITY

?

-2.217*** (-48.40)

-2.372*** (-48.85)

-28.620*** (-50.14)

-29.134*** (-50.69)

-28.369*** (-50.50)

-28.649*** (-50.74)

-29.339*** (-51.65)

-32.921*** (-43.72)

LNGDPC

?

0.001 (0.19)

-0.060*** (-13.44)

-2.126*** (-44.08)

-2.323*** (-45.12)

-1.753*** (-8.79)

-2.321*** (-45.33)

-2.121*** (-42.81)

-1.595*** (-22.02)

Industry Dummies Yes Yes Yes Yes Yes Yes Yes Yes Year Dummies Yes Yes Yes Yes Yes Yes Yes Yes Country Dummies Yes Yes Yes Yes Yes Yes Yes Yes N 233,196 233,196 312,407 312,407 312,704 312,704 286,016 286,016 Pseudo R2 / R2 0.094 0.102 0.504 0.505 0.505 0.501 0.506 0.493 *, ** and *** represent two-tailed significance at level of 10%, 5% and 1% respectively. N denotes the number of observations. See Appendix for variable definitions. Z-statistics / T-statistics are shown in parentheses for country weighted PROBIT / OLS regressions, calculated based on robust standard errors clustered at the firm level.

46

Table 8 Stock Market Liquidity, Firm’s Dividend Policy and Anti-director Rights Model Dependent Variable=

Intercept TURN ANTIDIRECTOR

?

?

(1)

(2)

(3)

DIV

10.719***

DIV Low Anti-director Rights 5.792***

DIV High Antidirector Rights -0.033

(10.04)

(4.43)

(-0.03)

-0.140***

-0.207***

-0.418***

(-10.22)

(-13.05)

(-15.51)

0.419***

0.431***

0.423***

(55.69)

(39.53)

(39.86)

6.029***

7.529***

4.291***

(53.62)

(49.90)

(27.81)

-0.057***

-0.053***

-0.032***

(-11.03)

(-7.51)

(-27.81)

-0.352***

-1.275***

0.002***

(-6.71)

(-18.01)

(6.14)

-28.830***

-41.698***

-20.633***

(-46.11)

(-39.60)

(-23.15)

-1.646***

-2.360***

-0.201*

(-27.06)

(-26.79)

(-1.82)

0.246 (1.58)

TURN*ANTIDIRECTOR

-

-0.356*** (-11.85)

SIZE

ROA

MB

LEV

VOLATILITY

LNGDPC

+ + ? ?

Dif. In TURN*ANTIDIRECTOR 0.211***(p-value < 0.01) 0.097*** Industry Dummies Yes Yes Yes (pYear Dummies Yes Yes Yes value < Country Dummies No Yes Yes 0.01) N 283,062 151,319 131,743 Pseudo R2 0.537 0.537 0.512 *, ** and *** represent two-tailed significance at level of 10%, 5% and 1% respectively. N denotes the number of observations. See Appendix for variable definitions. Z-statistics are shown in parentheses for country weighted PROBIT regressions, calculated based on robust standard errors clustered at the firm level.

47

Table 9 Stock Market Liquidity, Firm’s Dividend Policy and Political Constraints: Additional Control Variables Model Dependent Variable= Intercept

?

TURN

-

POLCON

?

TURN*POLCON

-

SIZE

+

ROA

+

MB

-

LEV

-

VOLATILITY

?

LNGDPC

?

INVPROT

?

GROWTH

-

BETA

?

STDRET

?

TAXADV

+

PROP

-

(1)

(2)

(3)

(4)

(5)

(6)

DIV

DIV

DIV

DIV

DIV

DIV

12.699*** (16.47) -0.260*** (-16.21) 1.009*** (11.37) -0.092** (-4.58)

12.423*** (17.90) -0.105*** (-5.79) 0.526*** (3.53) -0.505*** (-16.88)

20.780*** (29.98) -0.097*** (-11.54) -1.080*** (-11.77) -0.120*** (-7.29)

-3.443*** (-16.49) -0.230*** (-13.73) 1.049*** (10.83) -0.138*** (-7.00)

-5.273 (-0.14) -0.228*** (-13.55) 1.149*** (12.20) -0.113*** (-5.73)

-5.187*** (-6.09) -0.212*** (-11.15) 0.442*** (2.67) -0.304*** (-9.87)

0.410*** (55.46) 6.018*** (53.86) -0.056*** (-10.90) -0.402*** (-7.53) -28.706*** (-45.93) -1.717*** (-28.53) -0.352*** (-2.41)

0.364*** (43.95) 5.740*** (48.27) -0.042*** (-8.02) -0.164*** (-2.86) -39.917*** (-42.30) -1.757*** (-27.94)

0.406*** (52.11) 6.067*** (56.28) -0.008* (-1.67) -0.478*** (-8.98) -23.626*** (-38.03) -2.139*** (-40.39)

0.378*** (42.19) 5.206*** (38.13) -0.044*** (-6.80) -0.284*** (-4.39) -27.950*** (-37.09) -1.828*** (-24.42)

0.380*** (39.75) 5.422*** (4.23) -0.027*** (-4.23) -0.535*** (-8.27) -31.786*** (-39.93) -1.688*** (-23.50)

0.391*** (40.15) 5.701*** (44.07) -0.030*** (-5.05) -0.372*** (-5.87) -24.607*** (-15.38) -2.326*** (-33.79) -1.688* (-1.81) -0.206*** (-6.43) -0.335*** (-15.06) -35.058*** (-14.69) 91.104*** (7.62) 13.096*** (0.95)

-0.298*** (-10.33) -0.377*** (-19.79) -0.841*** (-28.38) 41.967*** (17.05 ) 19.118 (0.59)

Industry Dummies Yes Yes Yes Yes Yes Yes Year Dummies Yes Yes Yes Yes Yes Yes Country Dummies No Yes Yes No No No N 287,211 297,908 281,713 272,216 261,984 260,166 Pseudo R2 0.532 0.496 0.510 0.502 0.524 0.515 *, ** and *** represent two-tailed significance at level of 10%, 5% and 1% respectively. N denotes the number of observations. See Appendix for variable definitions. Z-statistics are shown in parentheses for country weighted PROBIT regressions, calculated based on robust standard errors clustered at the firm level.

48

Table 10 Stock Market Liquidity, Firm’s Dividend Policy and Political Constraints: Partition Analysis Model Legal Factors Dependent Intercept Variable=

?

TURN

-

POLCON

?

TURN*POLCON

-

SIZE

+

ROA

+

MB

-

LEV

-

VOLATILITY

?

LNGDPC

?

(1) (2) LAW&ORDER Low High DIV DIV -8.149*** -32.192*** (-6.12) (-13.70) -0.311*** -0.257*** (-11.54) (-14.63) -0.582*** 1.826*** (-3.84) (16.06) -0.175*** -0.399*** (-3.83) (-14.29)

(3)

Low DIV 19.172*** (8.53) -0.312*** (-16.19) 0.959*** (6.77) -0.242*** (-7.30)

High DIV 3.040*** (4.78) -0.142*** (-8.12) -0.343*** (-3.86) -0.415*** (-16.53)

(5) (6) CONTROLCORR Low High DIV DIV -2.680*** 8.179*** (-2.45) (3.53) -0.157*** -0.438*** (-8.66) (-21.42) 0.568*** 0.194* (3.30) (1.73) -0.154*** -0.330*** (-5.17) (-9.49)

0.342*** (24.38) 6.545*** (28.23) -0.029*** (-3.06) -1.218*** (-12.82) -33.900*** (-25.86) -0.566*** (-4.43)

0.464*** (43.28) 6.339*** (40.63) -0.040*** (-5.27) -0.765*** (-9.01) -27.256*** (-29.71) -3.175*** (-24.73)

0.402*** (77.99) 5.613*** (53.81) -0.031*** (-8.14) -0.264*** (-6.89) -31.711*** (-53.80) -0.449*** (-6.35)

0.418*** (36.66) 7.302*** (40.57) -0.011 (-1.47) -1.709*** (-21.13) -29.630*** (-31.17) -1.358*** (-14.23)

0.400*** (40.05) 4.891*** (32.62) -0.038*** (-5.28) -0.006 (-0.15) -25.729*** (-28.49) -3.311*** (-30.55)

REPUD

(4)

0.447*** (46.42) 4.992*** (36.88) -0.044*** (-6.70) 0.002*** (5.12) -25.478*** (-30.68) -1.604*** (-16.70)

Dif. In TURN*POLCON 0.224*** (p-value<0.01) 0.173*** (p-value < 0.01) 0.176*** (p-value<0.01) Industry Yes Yes Yes Yes Yes Yes Dummies Year Dummies Yes Yes Yes Yes Yes Yes Country Yes Yes Yes Yes Yes Yes Dummies N 97,230 209,836 139,999 149,298 141,162 165,904 Pseudo R2 0.434 0.530 0.610 0.530 0.509 0.576 *, ** and *** represent two-tailed significance at level of 10%, 5% and 1% respectively. N denotes the number of observations. See Appendix for variable definitions. Z-statistics are shown in parentheses for PROBIT regressions, calculated based on robust standard errors clustered at the firm level.

49

Table 11 Stock Market Liquidity, Firm’s Dividend Policy and Political Constraints: Financial Constraints

Model Dependent Variable= Intercept

?

TURN

-

POLCON

?

TURN*POLCON

-

(1) Low FINCON DIV 17.435*** (16.48) -0.325*** (-23.59) 0.991*** (8.84) -0.149*** (-6.78)

(2) High FINCON DIV 11.473*** (13.65) -0.191*** (-14.65) 0.203** (2.03) -0.056*** (-2.57)

0.370*** 0.368*** (30.60) (33.71) ROA + 5.735*** 5.941*** (30.52) (49.36) MB -0.036*** -0.005 (-5.39) (-0.90) LEV 0.048 -0.354*** (0.61) (-5.88) VOLATILITY ? -22.722*** -28.151*** (-25.28) (-39.56) LNGDPC ? -2.455*** -1.733*** (-29.04) (-27.36) Dif. In TURN*POLCON 0.093*** (p-value < 0.01) Industry Dummies Yes Yes Year Dummies Yes Yes Country Dummies Yes Yes N 155,709 156,140 Pseudo R2 0.414 0.529 *, ** and *** represent two-tailed significance at level of 10%, 5% and 1% respectively. N denotes the number of observations. See Appendix for variable definitions. Z-statistics are shown in parentheses for PROBIT regressions, calculated based on robust standard errors clustered at the firm level. SIZE

+

50

Table 12 Country-level analysis on the relationship between liquidity and dividend for countries with strong POLCON and strong CG versus countries with weak POLCON and weak CG

Dependent Variable= DIV Strong POLCON and Strong CG countries Australia Canada France Spain Switzerland US

N

Coefficient of TURN

Z-statistic clustered at the firm level

15,458 6,373 6,068 879 2,129 85,388

-0.751*** -0.922*** -3.163*** -0.369 -2.066*** -0.422***

-2.62 -4.02 -7.78 -1.13 -3.70 -13.96

Weak POLCON and Weak CG countries Argentina 799 -1.671*** -3.35 Columbia 174 -2.494 -0.62 Mexico 1,152 -0.989 -0.86 Nigeria 299 -2.125 -1.36 Philippines 1,318 -0.144 -0.39 Sri Lanka 806 -0.725 -1.58 Note: 1. Countries presented here only include those countries with available data for all legal protection and accounting quality measures; 2. Low and high CG groups are determined based on the ranking of all legal protection and accounting quality measures used in our paper, i.e., CG is a factor score based on factor analysis of the six corporate governance measures, including ANTISELF, REPUD, CONTROLCORR, DISREQ, CIFAR, and SUE.

51

Table 13 Stock Market Liquidity, Firm’s Dividend Policy and Political Constraints in the business cycle Model Dependent Variable= Intercept TURN

POLCON

TURN*POLCON SIZE

ROA

MB

LEV

VOLATILITY

LNGDPC

?

(1)

(2)

DIV RECESSION=1 22.780***

DIV RECESSION=0 12.714***

(13.21)

(16.21)

-0.196***

-0.210***

(-4.92)

(-12.43)

-0.167

1.353***

(-0.59)

(13.53)

-0.181***

-0.354***

(-3.21)

(-14.38)

0.496***

0.408***

(38.49)

(50.61)

5.378***

5.770***

(26.99)

(46.68)

-0.044***

-0.051***

(-4.33)

(-9.34)

-1.189***

-0.207***

(-13.62)

(-3.57)

-26.248***

-29.569***

(-21.01)

(-40.50)

-2.508***

-1.550***

(-15.34)

(-23.48)

?

+ + ? ?

Dif. in TURN*POLCON

0.173***

(p-value<0.01)

Industry Dummies Yes Yes Year Dummies Yes Yes Country Dummies Yes Yes N 48,747 221,727 Pseudo R2 0.543 0.551 *, ** and *** represent two-tailed significance at level of 10%, 5% and 1% respectively. N denotes the number of observations. See Appendix for variable definitions. Z-statistics are shown in parentheses for country weighted PROBIT regressions, calculated based on robust standard errors clustered at the firm level.

52