Corporate Governance and Cash Holdings: Empirical Evidence. from an Emerging Market

Transcription

1 Corporate Governance and Cash Holdings: Empirical Evidence from an Emerging Market I-Ju Chen Division of Finance, College of Management Yuan Ze University, Taoyuan, Taiwan Bei-Yi Wang Division of Finance, College of Management Yuan Ze University, Taoyuan, Taiwan JEL classifications: G30; G34 Keywords: Cash holdings, Ownership structure, Board structure, Information transparency Address correspondence to I-Ju Chen, Division of Finance, College of Management, Yuan Ze University, 135, Far-East Rd., Chung-Li, Taoyuan, Taiwan. Tel: ext.3664; Fax: ; I-Ju Chen acknowledges the funding from the National Science Council in Taiwan (NSC H and H MY2).

2 Corporate Governance and Cash Holdings: Empirical Evidence from an Emerging Market Abstract The average cash holding for Taiwanese listed firms is more than doubles from 1990 to This study examines the relationship between firm-level corporate governance metrics and cash holding. We construct our hypotheses based on the agency motive of cash holding. We employ ownership structure, board attributes, and information transparency as proxies for the nature of corporate governance for Taiwanese firms. Firms with higher divergence between control and cash flow rights, unaffiliated, and not controlled by families, are more inclined to accumulate cash. In addition, firms with higher board independence and smaller board size tend to hold more cash. Firms with higher information transparency will actually have larger cash reserves. Our evidence suggests that corporate governance structures are important in determining cash holdings for Taiwanese publicly-traded firms. Key words: Cash holdings, Ownership structure, Board structure, Information transparency ii

3 Corporate Governance and Cash Holdings: Empirical Evidence from an Emerging Market 1. Introduction In the wake of financial tsunami, investors lose confidence of holding financial assets, and companies realize the importance of cash flow management. In this situation, people will choose to hoard more cash reserves to protect themselves from future cash flow uncertainty. This also raises the question that what factors determine the corporate cash holdings. Corporate cash holding is an important factor in a firm s assets. Many studies explore the determinants of corporate cash holdings (Bates et al., 2009; Harford et al., 2008; Kim et al., 1998; Opler et al., 1999; Ozkan and Ozkan, 2004). From transaction and precautionary motives, firms hold more cash to protect themselves against future cash flow uncertainty if the costs of external financing are high (Almeida et al., 2004; Han and Qiu, 2007; Opler et al., 1999). From the tax motive viewpoint, Foley et al. (2007) find that U.S. multinational corporations that incur larger tax expenses with repatriating foreign earnings are more likely to accumulate cash. Agency theory suggests that corporations in countries with weak shareholder protection have higher cash holdings (Dittmar et al., 2003; Pinkowitz et al., 2006). Their results indicate that entrenched managers are more likely to hoard cash rather than make dividend payments to shareholders. Moreover, the value of cash holdings is worth less when agency problems are severe (Dittmar and Mahrt-Smith, 2007; Pinkowitz et al., 2006). A recent study states that there is a secular increase in the cash holdings of U.S. firms from 1980 to 2006 (Bates et al., 2009). Bates et al. (2009) investigate why average cash ratios for U.S. firms are more than doubles during the sample period. Their results show that the main reasons for this dramatic increase are that net working capital has fallen, cash flow volatility has increased, capital expenditures have decreased, and R&D expenditures have improved. Moreover, following Jensen (1986), we would expect that firms with agency problems will hoard cash if they do not have good investment opportunities and the entrenched managers do not want to return cash to shareholders. However, Bates et al. (2009) find no evidence that agency problems contribute to the increase in cash ratios for U.S. firms. Therefore, this study provides empirical tests of the influence of agency problems on cash holdings in Taiwanese firms. Several corporate governance issues are prominent in Asia according to Claessens and Fan (2002). First, they report that corporate ownership is highly concentrated in Asian 1

4 corporations, unlike U.S. firms, whose shares are widely held. This indicates that typical Asian corporations are tightly held by one or several members. More importantly, the company is often affiliated with a business group controlled by the same family. This implies that the family effectively controls the firms via pyramid structures and cross-holdings, making the identification of the controlling owners are complicated task (Claessens et al., 2000). Furthermore, voting rights of controlling owners frequently exceed cash flow rights in Asian corporations. Second, compared with U.S. firms, internal governance mechanisms such as board structures are weak in Asian markets. Yeh (2002) document that boards of directors are often dominated by insiders, and hardly have outsider existence in Taiwan. A single leadership structure (the chairman of the board is also the CEO) is generally present in Taiwan corporations. Third, Asian firms typically have lower levels of transparency and disclosure quality, which may be the outcome of poor corporate governance (Fan and Wong, 2002). They argue that this low level of transparency is related to agency problems and relationship-based transactions. Although high ownership concentration is common among Asian corporations, the extensiveness of the cross-shareholding or pyramid structures is quite popular in Taiwan (Yeh, Lee and Woidtke, 2001). Further, the degree of shareholder protection in Taiwan is much lower than countries in U.S. and U.K., which suggests that managers are less likely to concern with shareholder wealth and make decisions that destroy firm value (Dittmar et al., 2003)., Many countries in Asia concerned over corporate governance since the Asian crisis in 1997, and Taiwan authorities imposed the corporate governance best-practice principles in 2002 and launched information transparency and disclosure rankings system in 2003 to strengthen corporate governance practices. This motivates us to explore whether managerial incentives affect firms cash holdings and how governance influences the level of cash holdings for Taiwanese publicly traded firms. While previous studies focus on the determinants of cash holding for U.S. firms, we examine ownership structure, board attributes, and information transparency as proxies for the nature of corporate governance for Taiwanese firms. The specific proxies for ownership structure are divergence between ownership and control rights, group affiliation, and family control. We hypothesize that firms with divergence between ownership and control rights will have higher cash holding. Following substantial literatures, we hypothesize that the influence of group affiliation and family control on cash holding is ambiguous. Based on the board structure, we contend that firms accumulate more cash when the proportion of independent directors on the board is higher, and the chairman of the board is also the CEO 2

5 (single leadership structure). In addition, we hypothesize that the relationship between board size and cash holding is ambiguous. The final hypothesis is that firms with a higher level of information transparency will have lower cash holding. We construct our sample using the Taiwan Economic Journal (TEJ) database for the period 1990 to In the empirical analysis, we employ several alternative definitions of the cash ratio, including (1) cash to total assets; (2) cash to net assets (where net assets equals book assets minus cash); (3) current assets to total assets; and (4) current assets to net assets. The explanatory variables that we use follow our hypotheses and are motivated by specific corporate governance issues in Asia. The control variables follow Opler et al. (1999). Using various proxies for the nature of corporate governance, we run regressions to estimate the determinants of cash holding in Taiwan. We first find that there is a secular increase from 1990 to 2011 in the average cash ratio for Taiwanese publicly traded firms. The average cash to total assets is more than doubles over the sample period. In univariate comparisons, we find that firms with higher cash holding are less likely to be group-affiliated, and not controlled by families. For the board structure of high cash balances firms, the board size is smaller and more independent directors are involved in the boards. In addition, these high cash balances firms tend to have higher degree of information transparency compared to firms with low cash balances. Furthermore, we find that firms with high divergence between control and cash flow rights are more likely to accumulate cash in univariate analysis. However, we find no evidence that firms with single leadership structure are associated with more cash holdings. The remainder of this paper is organized as follows. In Section 2, we briefly review the theoretical determinants of cash holding and the literature on the relationship between cash holding and agency problems. In Section 3 we develop the hypotheses. We describe our sample and research design in Section 4. Section 5 discusses the empirical results of the univariate investigation and regression estimations. Section 6 concludes. 2. Literature review 2.1 Studies on cash holding A substantial literature discusses the determinants of cash holding. Opler et al. (1999) provide comprehensive reviews of the determinants of corporate cash holding for publicly traded firms in the U.S. They find evidence supportive of a tradeoff model of cash holding, 3

6 indicating that firms trade off the costs and benefits of holding cash and determine the optimal cash balances. In addition, several studies have examined the different factors affecting corporate cash policy for the U.S. firms. Foley et al. (2007) show that firms with higher tax costs when repatriating earnings appear to increase cash holding. Han and Qiu (2007) find that cash flow volatility has a positive impact on cash holding for financially constrained firms, while this relationship does not exist for unconstrained firms. Faulkender and Wang (2006) investigate the marginal value of cash holdings that arises from differences in corporate financial policy. Their results show that the marginal value of cash declines with more cash holding, higher leverage, and better access to capital markets. A more recent study, Bates et al. (2009) document that there is a secular increase in the cash holdings of U.S. firms, and they find evidence supportive of precautionary motive for this increase. Opler, Pinkowitz, Stulz, and Williamson (1999) examine the determinants and implications of corporate cash holding. They consider two explanations for cash holding: the tradeoff theory and the financing hierarchy theory. The tradeoff theory suggests that firms trade off the costs and benefits of holding cash to determine optimal cash levels. In this context, they consider the transaction costs motive, the effect of asymmetric information, and the agency costs of external financing inducing a demand for cash holding. The costs of holding cash includes the cost-of-carry and possible tax disadvantages. However, there are two benefits from holding cash. First, firms can avoid the transaction costs involved in raising external funds and do not have to liquidate assets to make payments. Second, firms can reserve cash to hedge against the risk of future cash shortfalls. Keynes (1936) describes the first benefit as the transaction motive for cash holdings, and the second one as the precautionary motive. However, the financing hierarchy theory suggests that there is no optimal amount of cash. This theory suggests that when firms have sufficient cash flow to finance new investment, they repay debts and accumulate cash. Conversely, when firms lack the internal funds to finance investment, they reduce their accumulated cash and issue debt. Therefore, managers are indifferent to holding cash and taking on debt, and cash reserves are the residuals of the investment and financing decisions (Dittmar et al., 2003). In order to test these theories, they construct a sample of publicly traded U.S. firms from 1971 through The regression results provide evidence for the tradeoff theory. Firms hold more cash when they are smaller, have higher investment and R&D expenditures, better investment opportunities, more volatile cash flows, and lower net working capital. These are 4

7 characteristics that either increase the cost of cash shortfalls or increase the cost of outside funds. These findings imply that both transaction costs and the precautionary motive are important factors in the tradeoff theory. Foley et al. (2007) examine the impact of taxes on cash holding, a topic that has received less attention in previous literatures. Specifically, they test the hypothesis that the magnitude of cash holding is partly associated with the tax burden of repatriations. The sample covers the period 1982 to 2004 for US firms. According to Opler et al. (1999), they use the natural logarithm of the ratio of cash to net assets as their main dependent variable. Their empirical results show that firms with higher tax costs when repatriating earnings appear to increase cash holdings. This result is robust across different measures of the repatriation tax burden, as well as across different levels of cash holding and changes in cash holding. Moreover, the tax burden of repatriations makes firms hold more cash abroad. However, their results lack the evidence to conclude that firms facing a higher repatriation tax burden hold less cash domestically. They further test whether affiliates in countries with high repatriation tax costs hold more cash than other affiliates of the same firm. Their empirical results confirm this contention. Financially constrained firms hold less cash, and their cash holdings are less sensitive to repatriation tax costs. Finally, the sensitivity of cash holdings to repatriation taxes is particularly pronounced for technology-intensive firms. Han and Qiu (2007) provide a direct analysis of the interaction between a firm s precautionary cash holding, cash flow uncertainty, and financial constraints. They propose that there is a positive relationship between cash holding and future cash flow volatility for financially constrained firms because these firms are sensitive to cash flow volatility and tend to accumulate cash reserves for precautionary savings. The sample consists of U.S. publicly traded companies from 1997 to 2002 using quarterly information. Their theoretical model is based on that of Almeida et al. (2004) which investigate whether changes in the expected value of future cash flows contribute to a firm s cash holding behavior with a perfect external hedging market. In order to test this hypothesis, they separate firms into financially constrained and unconstrained groups, based on four indices (firm size, dividend payout, debt rating, and commercial rating), and use cash flow volatility as the main explanatory variable. They then run regressions for the financially constrained and unconstrained firms respectively. The results indicate that cash flow volatility has a significantly positive impact on the cash holdings of financially constrained firms, consistent with the predictions of the model. 5

8 However, this relationship does not exist for unconstrained firms. In other words, the intertemporal trade-off gives constrained firms the incentive to accumulate precautionary savings when cash flow risks cannot be fully diversified. Faulkender and Wang (2006) investigate whether corporate financial policy has an effect on the cross-sectional variation in the marginal value of cash holdings. Their data consists of 87,127 firm-years over the period 1971 to They construct three hypotheses. First, as a firm increases the level of its cash position, the value of an additional dollar of cash decreases. Second, in firms with high leverage, an extra dollar of cash holdings is less valuable for shareholders than in firms with low leverage. Third, in financially constrained firms, an extra dollar of cash holdings is more valuable to shareholders. In order to test these hypotheses, they run regressions to estimate the excess equity return on the changes in cash holdings, a firm s profitability, financing policy, and investment policy over the fiscal year. The findings for cash level and leverage are consistent with the first two hypotheses, suggesting that firms with little cash on hand are likely to raise external funds and receive the highest benefits from additional internal funds. The marginal value of cash decreases with the level of leverage. The authors also find that shareholders value the marginal dollar of cash higher in constrained firms than in unconstrained firms, supporting the third hypothesis. Bates, Kahle, and Stulz (2009) investigate the reasons for the secular increase in the cash holdings of U.S. firms from 1980 to They first sort firms based on different criteria to see whether this increase concentrates among certain firms. The results show that the secular increase in cash ratios concentrates among firms that do not pay dividends and among firms in industries that experience a large increase in cash flow volatility, supporting the precautionary motive. Moreover, they also find that firms in more recent IPO listing cohorts are more likely to experience an increase in cash ratios, which is consistent with the evidence in Brown and Kapadia (2007). After showing that there is a dramatic increase in cash ratios, they test the reasons for that increase. Based on the model developed by Opler, Pinkowitz, Stulz, and Williamson (1999), they find that the main reasons for the increase in cash holdings are that net working capital has fallen, cash flow volatility has increased, capital expenditures have decreased, and R&D expenditures have improved. Furthermore, they also examine whether agency problems could explain this secular increase for U.S. firms. Using the GIM index measured as managerial entrenchment and a valuation 6

9 regression developed by Fama and French (1998), the results show that there is no evidence that agency problems contribute to the increase in cash ratios. 2.2 Relationship between cash holdings and the agency problem Several studies have tested an agency cost explanation for cash holdings. The agency costs hypothesis suggests that entrenched managers in firms with high free cash flow are not willing to pay out cash to shareholders when firms have poor investment opportunities, and thus increase cash holdings. Dittmar et al. (2003) and Kalcheva and Lins (2007) examine the importance of corporate governance in determining cash holding with an international sample. They find that firms in countries with poor shareholder protection significantly have higher cash holdings, which supports the agency costs hypothesis. Similar results are found in Pinkowitz et al. (2006), who use institution quality and financial development as proxies for corporate governance. Dittmar and Mahrt-Smith (2007) and Harford, Mansi, and Maxwell (2008) provide evidence indicating that firms with weaker corporate governance hold smaller cash reserves because the entrenched managers hoard excess cash and spend excess cash quickly. Kusnadi (2011), investigating firms in Singapore and Malaysia, suggests that managers of firms with a pyramidal ownership structure and controlled by families are more likely to accumulate cash. Dittmar, Mahrt-Smith, and Servaes (2003) examine the relationship between corporate governance and cash holdings. Following Opler et al. (1999), they discuss two views of cash holdings: the tradeoff model and the financing hierarchy model. In addition, they examine whether the cash holding of a firm depends on the level of shareholder protection. Their sample includes more than 11,000 firms from 45 countries in They first compare average cash holdings across countries using a variety of shareholder protection and capital market development measures. They then run pooled regressions to investigate the determinants of cash holdings across countries. Dittmar, Mahrt-Smith, and Servaes (2003) also find that corporations in countries where shareholders enjoy less protection have significantly higher cash holdings. Moreover, other determinants of cash holding, such as investment opportunities and asymmetric information, appear to be less important in these countries. The empirical results show that firms hold more cash when it is easier to raise funds in countries with poor shareholder protection, which is consistent with the importance of agency costs. There is no evidence that managers hold more cash simply because it is more difficult to access capital markets in countries with poor shareholder protection. In sum, firms hold less cash in countries where 7

10 shareholders enjoy greater rights and when external capital markets are highly developed, which supports the agency costs hypothesis. Kalcheva and Lins (2007) revisit the theoretical arguments on agency problems. They construct three hypotheses based on the previous literatures. First, firms with entrenched managers will hold more cash, particularly when country-level shareholder protection is poor. Second, a firm s value will be lower when entrenched managers hold high levels of cash, especially when country-level shareholder protection is weak. Third, when firms with entrenched managers pay dividends in countries with poor investor protection, firm value will be higher. The sample consists of 5,102 firms from 31 countries in Moreover, they construct three measures of managerial entrenchment. The first is the percentage of control rights held by the management group and its family. The second variable is a dummy variable that equals one when the management group and its family is the largest blockholder of a firm s control rights. The final variable is also a dummy variable equaling one when the management control rights exceed other blockholders control rights by 20%. For country-level shareholder protection, they use the measure for anti-director rights of La Porta et al. (1998). In order to test the hypotheses, they run regressions using cash holdings and Tobin s Q as the main dependent variables. The results of the analysis of Kalcheva and Lins (2007) show that outside investors discount the value of cash holding when firms have entrenched managers and investor protection against expropriation is weak. They also find that dividend payments enhance firm value when managerial agency problems exist. In sum, these findings are consistent with their hypotheses. Pinkowitz, Stulz, and Williamson (2006) examine two theories about the determinants of cash holdings across countries. The first is the tradeoff theory, which suggests that shareholder-wealth maximizing managers face different tradeoffs, while the second is the agency costs theory, which argues that large shareholders in countries with poor investor protection have incentives to make decisions that expropriate private benefits from control. These two theories predict that firms hold more cash in countries with poor institutions and financial development. The sample spans 35 countries from 1988 to Using the method of Fama and MacBeth (1973), they run cross-sectional regressions for each year and the time series coefficients are used to make inferences. The primary dependent variable in the regression is excess cash holdings and the independent variables are country characteristics. 8

11 The results of the analysis of Pinkowitz, Stulz, and Williamson (2006) show that the determinant of cash holdings is consistent with both theories, indicating that cash holdings are negatively related to the quality of institutions and financial development. Furthermore, firms in countries with higher GDP per capital hold more cash. In other words, cash holdings of firms are positively related to economic development. They further examine the relation between the cash holdings of a firm and the value of that firm. The results show that the cash holding of firms in countries with poor institutions are worth less to minority shareholders than firms in countries with better institutions, which indicates that agency costs play an important role in how minority investors value cash held by corporations. Dittmar and Mahrt-Smith (2007) investigate how corporate governance has an impact on the value and eventual use of cash reserves. They examine the potential value destruction because of negligence and profusion in the management and how good corporate governance helps to prevent it. Next, they investigate whether poorly governed firms dissipate excess cash more quickly, while well-governed firms control their cash better. Their sample includes all US publicly traded firms from 1990 to For corporate governance data, they focus on internal and external measures, including the degree of managerial entrenchment due to takeover protection and large shareholder monitoring. The methodology follows that of Faulkender and Wang (2005) and estimates the impact of changes in cash holdings on changes in firm value. The results show that the value of a dollar of cash is greater for a well-governed firm than a poorly-governed firm, indicating that governance has a substantial impact on firm value through cash holdings. Furthermore, Dittmar and Mahrt-Smith (2007) investigate how governance affects the level and use of excess cash. Based on Jensen (1986), they hypothesize that poorly governed firms dissipate cash more quickly than well-governed firms, spending cash in ways that reduce the firm s accounting returns and operating performance. The empirical results are consistent with this hypothesis. They further find that governance affects cash policy via the decision to use excess cash but not by the decision to accumulate it. In short, poorly governed firms waste cash and thus destroy firm value. Harford, Mansi, and Maxwell (2008) examine how firm governance affects the use of cash. They start with a sample of 11,645 firm-year observations for 1872 firms in the US. They use a number of corporate governance measures to gauge the firm s agency costs, including a GIndex of antitakeover provisions, ownership concentration (insider and institutional), compensation to top management (pay sensitivity), and board structure (size 9

12 and independence). Using these governance metrics, they find that firms with weaker corporate governance actually have smaller cash holdings. They further test the relation between investment decisions and corporate governance. They find that firms with poor corporate governance increase investments through acquisition and capital expenditures but reduce R&D investment when they allocate excess cash. Furthermore, how firms choose to distribute their excess cash to shareholders also depends on their governance. The payout results show that firms with weaker governance choose to repurchase instead of increasing dividends, indicating that such firms want to avoid making future payouts. Harford, Mansi, and Maxwell (2008) also examine whether the difference in investment and payout decisions is reflected in future profitability. Their results show that there is a positive relationship between shareholder rights and profitability. This relation is more pronounced when combined with excess cash holdings. This indicates that firms with low shareholder rights and excess cash have lower profitability and valuations because these firms managers choose to spend cash quickly on acquisitions and capital expenditures, rather than hoard it. Kusnadi (2011) examines the relationship between firm-level corporate governance and cash holdings, along with their combined effects on firm value, for firms in Singapore and Malaysia. Different from previous studies, he uses board attributes and measures of ownership concentration as proxies for a firm s internal governance. According to Claessens et al. (2000), they suggest that most companies in East Asia are organized into family-controlled groups and that the families maintain ultimate control of the firms via pyramid structures or cross-holding. Specifically, the separation of cash flow and control rights in a pyramidal ownership structure results in both an incentive effect and an entrenchment effect. The author thus predicts that an increase in the incentive effect is associated with smaller cash holdings, while an increase in the entrenchment effect is related to larger cash holdings. For the valuation implications of cash holdings, he hypothesizes that firms with poor internal governance and hold large amounts of excess cash are likely to have a negative firm value. The sample consists of 276 listed firms in Singapore and Malaysia covering the period from 2000 to The board attributes include board size, single leadership structure, and the proportion of independent non-executive directors on each firm s board. A single leadership structure means that the chairman of the board and the CEO are the same person. The measures of ownership concentration consist of insider ownership, and two dummy variables are created to mark firms with insider ownership exceeding the 20% threshold and those with a pyramidal 10

13 ownership structure. In order to test the hypotheses, he runs regressions using the natural logarithm of the ratio of cash and cash equivalents to net assets and the ratio of market value of equity plus total liabilities to net assets as the main dependent variables. The results show that managers of firms with a pyramidal ownership structure and family-controlled firms are more likely to accumulate cash reserves, supporting the entrenchment effect. In addition, the incremental value of holding excess cash is found to be negative for firms with a single leadership structure, firms with a pyramidal ownership structure, and family-controlled firms. In sum, the board characteristics and ownership concentration are important determinants of cash holding for firms in Singapore and Malaysia. Pinkowitz, Stulz, and Williamson (2006) investigate whether the value of corporate cash holdings is lower in countries with weaker investor protection due to the greater ability of controlling shareholders to extract private benefits from cash holdings in these countries, a contention of agency theories. They further examine that investors value dividends in countries with poor investor protection at a premium because in such countries investors expect the cash to be partly consumed as private benefits. The sample consists of 35 countries for the period 1988 to In order to test the hypotheses, they follow a valuation regression which is developed by Fama and French (1998). Instead of having a continuous measure of investor protection, they split the whole sample in half each year and use a dummy variable that equals one in countries with above-median investor protection. They measure investor protection using a corruption index obtained from an international country risk guide and an anti-director index based on LLSV (1998). The corruption index assesses the risk of corruption for high government officials. It indicates that countries with rampant corruption are those with poor investor protection. The anti-director index values from zero to six, where countries with a value of six are those with the best protection of minority shareholder rights. Moreover, they take GDP per capita as a measure of economic development since measures of investor protection are highly correlated with economic development. Based on the regression results, they find that cash contributes significantly greater to firm value in countries with better investor protection and the dividend payout is worth more in weak investor protection countries, strongly supporting these hypotheses. In sum, agency problems truly play an important role in firm value across countries. 3. Hypotheses 11

14 3.1 Ownership structure Divergence between ownership and control rights The definition of ownership in this study relies on cash-flow rights, and control relies on voting rights based on a chain of literatures (Claessens et al., 2000; Claessens et al., 2002). La Porta et al. (1999) and Claessens et al. (2000) show that the controlling shareholders of publicly traded firms are often able to control a firm s operations with a relatively small stake in its cash-flow rights in most countries. This situation will exacerbate the entrenchment effect (Claessens et al., 2002; La Porta et al., 1999). The entrenchment effect indicates that higher managerial ownership may entrench managers, as they are increasingly less subject to control by board of directors and to discipline by the market (Morck et al., 1988). Thus, when the divergence between ownership and control increases, controlling shareholders have the incentive to expropriate wealth by seeking personal benefits at the expense of minority shareholders (Bebchuk et al., 2000; Harris and Raviv, 1988; La Porta et al., 1999; Shleifer and Vishny, 1997). Kuan et al. (2011) further find that there is a positive relationship between the divergence between ownership and control rights and cash holding. Therefore, we predict that an increase in the entrenchment effect is associated with larger cash holdings: H1a. Firms with greater divergence between ownership and control rights will have higher cash holdings Group affiliation The evidence to date on the benefits and costs of group affiliation is mixed. Compared to independent firms, group affiliation is associated with greater use of internal factor markets, including financial markets. Khanna & Palepu (2000) and Shin & Park (1999) argue that the internal financial markets within a business group enable member firms with the best projects to obtain resources when a country s external capital markets are not well developed. The internal markets can provide funds to firms that have better growth opportunities, but which are financially constrained. In addition, the internal financial markets can reduce transaction costs. Based on these benefits of group affiliation, group-affiliated firms may choose to hoard less cash than unaffiliated firms. However, Claessens et al. (2006) empirically examine the benefits and agency costs of group affiliation for a large sample in East Asian countries. They find that there are gains from group affiliation but that agency problems are important in shaping the benefits and costs of group affiliation. Indeed, the tunneling hypothesis emphasizes the agency problem between controlling shareholders and minority shareholders within group affiliated firms. It contends that a business group has the potential to make controlling shareholders pay more 12

15 attention to their own wealth but pay less attention to minority shareholders when corporate governance is weak (Bae et al., 2002; Bertrand et al., 2002). Bae et al. (2002) find that acquisitions by Korean business groups (chaebols) are used as a way for controlling shareholders to increase their own wealth at the expense of minority shareholders, consistent with the tunneling hypothesis. Bertrand et al. (2002) also find that groups in India are used by controlling shareholders to tunnel resources away from minority investors, suggesting that firms with large controlling shareholder may channel corporate resources to benefit themselves but provide little benefits to minority owners. Following the tunneling hypothesis, we infer that group-affiliated firms tend to have serious agency problems and consequently hold more cash reserves. H1b. The relationship between group affiliation and cash holding is ambiguous Family control Family-controlled firms are prevalent in many East Asian countries (Claessens et al., 2000). Yeh et al. (2001) find that the interest conflict between controlling and minority shareholders within family-controlled firms is severe among Taiwanese listed firms. They argued that the average control by the family shareholders is high enough to influence a firm s decision making; therefore, they force the firm to adopt policies that satisfy their personal interests rather than those of the minority shareholders. Kuan et al. (2011) and Ozkan and Ozkan (2004) further find that family-controlled firms hold higher levels of cash than nonfamily-controlled firms. However, following Chrisman et al. (2007), Chrisman et al. (2004), and Fama and Jensen (1983), family-controlled firms typically monitor and provide incentives to family managers, thus mitigating agency problems between managers and shareholders. Compared to nonfamily-controlled firms, family members hold substantial ownership in their companies, thus have strong incentives to monitor management in order to protect their interests (Anderson and Reeb, 2003). Furthermore, family members intend to have a strong commitment to their firms. This will lead to success in enhancing their wealth and the reputation of their companies. As a result, families are more likely to curb wasteful expenditures, which in turn mitigate the agency cost of cash holdings. H1c. The relationship between family control and cash holdings is ambiguous. 3.2 Board structure 3.2.1Board independence A generally accepted perspective is that outside (non-executive) directors are appointed 13

16 to act in the shareholders interests (Rosenstein and Wyatt, 1997; Mayers et al., 1997) and outside directors have incentives to effectively monitor management decision-making (Fama and Jensen, 1983). Therefore, independent directors could mitigate managerial entrenchment and expropriation of firm resources, which implies that firms with more outside directors can reduce agency problems between managers and shareholders. Thus, we argue that firms with a higher proportion of independent directors on the board have lower cash holdings: H2a. A higher level of board independence is associated with smaller cash holdings Board size The literature for the impact of board size on agency problems is mixed. Previous studies suggest that increased board size has two different effects: greater monitoring and less efficient decision-making. Jensen (1993) and Yermack (1996) suggest that smaller boards are more efficient, as decision making becomes slower when more people are involved. It implies that smaller board is associated with stronger board structure, and thus reduces the agency problems. On the contrary, Harris and Raviv (2006) find that larger boards provide better monitoring services when managers opportunities to expropriate private benefits are high. Moreover, Kusnadi (2011) finds that the relationship between board size and cash holding is negative for the samples in Singapore and Malaysia. H2b. The relationship between board size and cash holdings is ambiguous Leadership structure Studies by Baliga et al. (1996) and Tsui et al. (2001) show that a dual leadership structure (where the CEO is not also the chairman) offers more effective control than a single leadership structure. Jensen (1993) argued that the combination of CEO and chairman positions leads to less monitoring of top management and thus more severe agency problems. Kusnadi (2011) further suggest that a single leadership structure is more likely to hold large cash reserves in Singapore and Malaysia. Thus, we hypothesize that separating the CEO and chairman of the board positions will result in lower cash holdings: H2c. Single leadership structure is associated with higher cash holdings. 3.3 Information transparency Information transparency and disclosure are known to be integral to corporate governance. A higher level of transparency and disclosure can reduce the information asymmetry between a firm s management and shareholders and thus mitigate agency 14

17 problems in corporate governance (Patel et al., 2002; Bailey et al., 2006). Patel et al. (2002) show that Asian emerging markets present greater transparency and disclosure after the financial crisis. Moreover, according to the trade-off theory of cash holdings, higher information transparency enables firms to more easily raise external funds for capital investments, which implies that firms with higher information transparency do not need to retain much cash (Kusnadi, 2004). Kusnadi (2004) finds that corporate transparency is negatively associated with corporate cash holdings in Singapore. Thus, we hypothesize: H3. A higher level of information transparency is related to lower corporate cash holdings. 4. Data and Methodology 4.1 Sample selection The sample firms employed in this study include firms listed on the Taiwan Stock Exchange (TSE) and in the Over-The-Counter market (OTC) for the period 1990 to Data are obtained from the Taiwan Economic Journal (TEJ) database. We require firms that have no missing values for cash and cash equivalents, and remove firms whose net assets equal zero in order to obtain adequate cash ratios. We exclude financial firms because they may carry cash to meet capital requirements rather than for the economic reasons studied here. The final sample consists of 22,587 firm-year observations over the sample period. Appendix A gives a detailed description of each variable. 4.2 Variable descriptions Cash holding We employ several alternative definitions of cash holding, including (1) cash to total assets, (2) cash to net assets (where net assets equals total assets minus cash) (3) current assets to total assets, (4) current assets to net assets. We define the first two ratios as the cash ratio, and the others as the current assets ratio. In this study, we compare different definitions of cash holdings to check the secular trend of the typical firms during the sample period. All the ratios of cash holdings are winsorized at the 1% level on either tail to minimize the influence of outliers on the results Governance variables We employ multiple measures of corporate governance as discussed in section 3, which include ownership structure, board structure, and information transparency. The specific variables for ownership structure are the divergence between control rights 15

18 and cash flow rights, group affiliation, and family control. This study defines control rights following La Porta et al. (1999), adding the direct and indirect voting rights of controlling shareholders in a firm s control chain. The direct voting right is represented by the shares directly held by the ultimate controlling shareholders, while the indirect voting right is represented by the shares held by the ultimate controlling shareholder through a pyramid structure or cross-holdings. The cash flow right is defined as the ownership of the ultimate controlling shareholder. The divergence between control rights and cash flow rights equals control rights minus cash flow rights, which captures the degree of agency problems. The criteria to identify a firm s membership in a business group in the TEJ include: (1) the primary shareholders are the same as those in another affiliated firm or belong to the same family (the primary shareholders refer to the largest shareholder, or those who own at least 5% of the firm s shares), (2) one-third of the firm s board of directors are identical to those of other affiliated firms, (3) the CEO or the general manager is the same as that of another, affiliated firm, (4) the firm is controlled by or is subordinated to an affiliated firm. We set a dummy that equals one if the firm is group-affiliated, and zero, otherwise. The fundamental criterion that TEJ uses to define a family-controlled firm is at least two family members serving as directors, supervisors, or managers. TEJ has four requirements for identifying whether firms are controlled by single family: (1) the chairman of the board and the general manager are from the same family; (2) the number of seats controlled by internal directors exceeds 50% while outside director seats are less than 33%; (3) the number of seats controlled by directors is greater than 33% and at least three family ultimate controlling members serve as directors, supervisors, and general manager; and (4) controlling shareholdings are greater than necessary controlling shareholdings. The family control dummy equals one if a firm is controlled by a family, and zero, otherwise. Due to the limitations of the TEJ database, group affiliation data and family control data begins from 1999, and the divergence between control rights and cash flow rights starts in This study uses board independence, board size, and chair duality as measures of the firm s board structure. The directors on each board are classified as either inside directors or independent outside directors. Directors who are not related to any of the firm s top managers are classified as independent. Board independence is defined as the ratio of independent directors to total directors. Board size is the number of directors on each firm s board. Chair duality is defined to mark firms in which the chairman of the board and the general manager are the same person. A value of one labels a single leadership structure. In 2003, the Securities and Futures Institute (SFI), entrusted by the TSEC and the GTSM, 16

19 launched the Information Transparency and Disclosure Rankings System (ITDRS) to evaluate the degree of corporate transparency and information disclosure of all TSEC/GTSM-listed companies in Taiwan. The ITDRS evaluates the degree of corporate transparency by identifying 88 attributes grouped into five categories: (1) compliance with mandatory disclosures (11 attributes); (2) timeliness of reporting (16 attributes); (3) quality of information disclosed in annual reports (4 attributes); (4) quality of financial forecasts (46 attributes); and (5) quality of information disclosed on corporate websites (11 attributes). This study employs a ranking system to measure the degree of information transparency for each firm. The IDTRS evaluation system ranks the listed companies according to five grades, A+, A, B, C, and C-, since the third evaluation year This paper uses values from 5 (A+) to 1 (C-) to represent this ranking. In 2003 and 2004, the ITDRS ranked all the listed and OTC companies, and reported the top one-third firms that have the highest information transparency. Thus, we define the top one-third firms as taking a value of four, while the others take a value of two. In 2011, the IDTRS evaluation system was changed to seven grades, A++, A+, A, A-, B, C, and C-. In order to maintain consistency, a firm is given a 5 if its grade is A++ or A+, 4 if its grade is A or A-, and 3, 2, or 1 for B, C, and C-, respectively. Due to the limitations of the ranking system, the information transparency data begins from Financial variables Given that cash holdings are firm specific, we use financial variables to control for firm-specific effects. The financial variables in this study follow those of Opler, Pinkowitz, Stulz, and Williamson (1999). The financial variables include firm size, leverage, market to book ratio, cash flow ratio, net working capital to total assets, R&D to sales, capital expenditure ratio, cash flow volatility, and dividend dummy. Firm size is measured as the natural log of total assets. Leverage is measured as long-term debt plus debt in current liabilities divided by total assets. The market to book ratio, proxy for growth opportunities, is measured as (book value of assets minus book value of equity plus the market value of equity) divided by book value of assets. We define cash flow ratio as earnings after interest, dividends, and taxes, but before depreciation, divided by total assets. Net working capital to total assets, a proxy for liquidity, is the ratio of current assets net of cash minus current liabilities divided by total assets. R&D to sales is the ratio of research and development expenditures to sales. The capital expenditure ratio is defined as the ratio of capital expenditures to non-cash assets, where capital expenditures are the sum of changes in fixed assets and depreciation. In order to control for the precautionary motive for 17

20 cash holdings, we add cash flow volatility to our regression models. We measure cash flow volatility as the standard deviation of cash flow to total assets for the previous 5 years, and we require at least three observations for each firm-year. We define a dividend dummy that equals one if a company pays dividends in a given year and zero otherwise. All of the control variables are winsorized at the 1% level on either tail to reduce the influence of outliers. 4.3 Research Design We first start with univariate analysis in which we divide all the sample firms into subsamples based on the corporate governance metrics in each year. The purpose of univariate analysis is to examine whether there is a significant difference in cash holdings between subsamples and test the hypotheses presented in section 3. The research design is as follows (using the divergence between control and cash flow rights as example): (1) calculate the median divergence between the control and cash flow rights in each year; (2) divide the sample firms into two groups based on the median divergence between control and cash flow rights; (3) calculate the average cash holdings in each subsample; and (4) perform a t-test to examine the difference between the two groups. For the group affiliation, family control, and chair duality metrics, we can directly separate our sample into two groups without calculating the median. We use various definitions of cash holdings and provide the results in Table For the board independence metric, since there are no independent directors on the boards of Taiwanese publicly traded firms before 2002, the univariate analysis of board independence begins from Furthermore, this study presents univariate comparisons of key descriptive variables by cash to total assets quartile. We are interested in whether the characteristics of companies which hold high cash balances, such as those companies in the fourth quartile, differ from those with low cash balances, such as those in the first quartile. The quartiles are constructed each year, and we calculate the means and medians of measures of corporate governance according to the cash to total assets quartile. Then, we test the hypothesis that the fourth quartile firms differ significantly from the first quartile firms using a t-test. The results are presented in Table 11.We then examine the relation between corporate governance and cash holdings and control for firm-specific variables in a multivariate setting using a time-series model. The main dependent variables used in the regressions are cash to total assets and cash to net assets. The independent variables are governance-related variables and firm-specific factors affecting cash holdings. The variables of interest in this study are the corporate governance proxies that discussed in section 3. We estimate the following baseline model: 18