Cash versus Stock Dividends: Signalling or Catering. Abstract. What motivates the firm s choice of cash or stock dividends? Using a sample of listed

Transcription

1 Cash versus Stock Dividends: Signalling or Catering Abstract What motivates the firm s choice of cash or stock dividends? Using a sample of listed Chinese firms, we investigate the firm s choice of cash or stock dividends and market reactions to the announcement of these dividend choices. Chinese listed companies represent a unique sample characterized by state-controlled and closely held firms as well as significant differences in economic development and legal structures at the provincial level together with corporate regulations that require firms seeking external financing to show a history of dividend payment. We find that firms that are more profitable, have lower leverage, higher cash holding, stronger shareholder protection and state ownership prior to listing are more likely to pay dividends and cash dividends, in particular. In addition, we find that growing firms with high levels of retained earnings and investing more in fixed assets are more likely to pay stock dividends. Firms appear to cater to investor demands in setting dividend policy; hence firms with a large proportion of nontradable shares are more likely to pay cash dividends. Consistent with the use of stock dividends to attract the attention of analysts, we find that the announcement of a stock dividend initiation is associated with a significant positive market reaction and increased analyst following. In contrast, these effects are not evidenced for firms continuing to pay a stock dividend. Key words: Dividend policy: Cash dividend; Stock dividend: China; Emerging markets JEL Classifications: G32, G35

2 1. Introduction What motivates the firm s choice of cash or stock dividends? There is a growing literature in both developed and emerging markets that has sought to determine the link between firm characteristics, institutional structures and corporate financial decisionsboth capital structure and dividend policy (Aivazian, Booth and Cleary, 2003; Booth, Aivazian, Demirguc-Kunt, and Maksimovic, 2001; Fan, Titman and Twite, 2009; Glen, Karmokolias, Shah, 1995; La Porta, Lopez-de-Silanes, Shleifer, Vishny, 2000; Mitton, 2004). With respect to dividend policy, the existing literature suggests that like their developed economy counterparts, the dividend policy of firms in emerging economies is influenced by profitability, debt, and the market-to-book ratio (Aivazian et al., 2003). However, emerging economy firms pay higher dividends and seem to be more affected by asset mix than their developed economy counterparts (Aivazian et al., 2003). In addition, research suggests that in emerging economies, stronger corporate governance and shareholder protection is associated with higher dividends (Mitton, 2004; La Porta et al., 2000). Using a sample of Chinese firms dividend announcements between 2003 and 2007, this study builds on this literature in two important ways. First, we focus on the influence of management response to investor demand in setting the firm s dividend policy. Chinese listed companies represent a unique sample of firms. The financial system in China is characterized by state-controlled, closely held firms (Lin et al., 1998) with a large proportion of non-tradable shares, which account for almost 60% of the total number of shares outstanding (Calomiris, Fisman and Wang, 2008; Gao and Kling, 2008; Chen et al., 2009). While the non-tradable shares reduce stock market liquidity, they also 1

3 put pressure on firms management to pay cash dividends to cater to the demands of this investor group. For these non-tradable shareholders, the cash dividend payment is generally the primary source of return since their shares cannot be traded in the open markets. Second, we examine differences in the choice to pay a cash or stock dividend. In particular, we are able to discriminate between alternative explanations for the payment of dividends focusing on stock dividends as an effective way of attracting the attention of financial analysts. We conjecture that stock dividends are paid by firms facing high levels of information asymmetry and firms believing themselves to be undervalued (Grinblatt, Masulis, and Titman, 1984; Brennan and Hughes, 1991), a form of informal communication cheap talk (Almazan, Banerjee and Motta, 2008). Finally, we use the market s response to dividend announcements to confirm the role of firm characteristics and institutional factors in firm dividend policy. To understand our motivation, we plot the market adjusted cumulative abnormal return from t=-5 to t=30 for dividend payers and non-payers, where t=0 is the announcement date of the proposed dividend distribution scheme. We further divide the payers into three groups: firms paying only stock dividends (stock-payers), firms paying only cash dividends (cash-payers), and those paying both cash and stock dividends (cash & stock payers). As can be seen from Figure 1, the market responds in a similar way to both non-payers and cash dividend paying firms. However, the response to the announcement of stock dividends is both positive and extreme, suggesting that the market views stock dividends as good news, relative to both cash dividends and the absence of a dividend payment. 2

4 [Figure 1 about here] Our evidence suggests that more profitable, lower levered, higher cash holding firms and firms undertaking subsequent equity offerings are more likely to pay dividends. In addition, we find that growing firms with high levels of retained earnings and investing more in fixed assets pay stock dividends and these high growth firms are most likely to be undervalued. In contrast, cash dividend paying firms tend to have lower growth, be less profitable, have lower retained earnings and invest less in fixed assets. Consistent with management catering to non-tradable investor demands in setting dividend policy, we find that firms with a higher proportion of non-tradable shares are more likely to pay cash dividends. Finally, the positive and significant announcement effect for stock dividends is focused on the initiation of stock dividends. For those firms initiating a stock dividend, analyst following increases significantly after the dividend announcement. In contrast, firms continuing to pay a stock dividend do not evidence an increase in analyst coverage post the dividend announcement. These results are consistent with the argument that firms use the initiation of stock dividends to attract the attention of analysts (Grinblatt, Masulis, and Titman, 1984; Almazan, Banerjee and Motta, 2008). The remainder of the paper is organized as follows. Section 2 discusses institutional background and its implications. Section 3 introduces the set of firm level variables that influence dividend policy. Section 4 describes the data and sample selection. Section 5 presents the empirical results, and Section 6 concludes. 3

5 2. Institutional Background: Discussion and Implications We begin by discussing the institutional background within the Chinese stock markets and, together with their empirical implications, highlight how these institutional structures can potentially affect a firm s dividend policy. Specifically, we consider institutional structures that relate to shareholder protection, institutional investors, dividend payment (the repatriation of profits, type of dividend and dividend declarations) and equity financing. A. Shareholder Protection The existing literature (Aivazian and Booth, 2003; La Porta, Lopez-de-Silanes, Shleifer, Vishny, 2000; Mitton, 2004) suggests that stronger shareholder protection and corporate governance should be associated with higher cash dividends. Using a sample of Chinese firms, Sen, Le and Ferris (2009) find the propensity to pay dividends is associated with weaker governance, where governance is proxied by the percentage of outside directors on the board. This proxy is problematic however, in the case of Chinese listed firms, since directorships of Chinese listed firms are largely controlled by the government (Fan, Wong and Zhang, 2007). Differing from previous studies, we focus on two measures of shareholder protection: the level of corruption within the firm and the extent of property rights protection. We measure the level of corporate corruption using the Entertainment and Travel Costs (ETC) index composed by Cai, Fang and Xu (2009). This index measures the extent of corrupt payments by the firm to government officials, aggregated by provincial regions. Higher index values are associated with higher levels of entertainment and travel payments (i.e., higher corruption). In the context of the firm s 4

6 dividend policy, the ETC index proxies for the threat of all or part of shareholder rights being expropriated by management or public officials. Thus, a large ETC implies weak corporate governance. Each firm in the sample is assigned a corresponding provincial ETC value based on the location of the firm s main operations. We also include a second measure, namely, property rights protection index. Djankov, La Porta, Lopez-de-Silanes and Shleifer (2003) identify the protection of property rights as a key factor shaping the structure of contracts, where legal systems evolved to protect private property rights facilitating the ability of individual shareholders to protect their cash flow entitlements. The index of property rights protection (also developed by Cai, Fang and Xu (2009)) measures the extent to which the legal system of each provincial region protects the property rights of individuals; a high index value implies a high level of protection of individual property rights. In the context of the firm s dividend policy, the property rights protection measure (PP) proxies for the effective pressure by shareholders to force corporations to distribute cash. Each firm in the sample is assigned a corresponding provincial property rights protection value based on the location of the firm s main operations. B. Institutional Investors The existing literature (John et al., 2007) suggests that geographic location influences a firm s information environment and, as a consequence, its dividend policy. The existence of asymmetric information, together with the weaker monitoring of management in the more remote provinces increase the likelihood of management using free cash flow to undertake inefficient investment decisions, for example, empire 5

7 building. To mitigate this inefficient investment, investors are expected to demand higher cash dividends. As noted by John et al. (2007), the presence of institutional investors facilitates the revelation and dissemination of information and improves the monitoring of management. However, these roles may be impeded by the firm s location, in particular, the remoteness of the firm relative to the location of the institutional investors. To test this relationship we measure the level of institutional investment as the provincial institutional shareholdings (InstHolding), obtained from GTA Data Services (CSMAR). The measure first identifies the proportion of the top ten institutional holdings in each firm, summing by province according the location of the firms main operations. Each firm in the sample then is assigned the corresponding provincial value. C. Dividend Payment Profits Repatriation in China Based on the Company Law of the People s Republic of China (Company Law), prior to the annual audit and the subsequent settlement of taxes to determine the company s after-tax profits, there are a number of accounting items that need to be estimated and presented in the company s accounting statements as mandatory fund dispersion. These include amounts for the Surplus Reserve Fund (SRF), the Employee Welfare and Bonus Fund (EWBF), and the Other Surplus Reserve Fund (OSRF). The proportion allocated to these funds shall be determined in accordance with the Company Law or by the Board of Directors. 6

8 Surplus Reserve Fund (SRF) The Company Law specifies that the amount of the SRF to be more than 10% of after-tax profits. The SRF is normally used to cover any losses suffered by the company. However, when the accumulated amount of the SRF exceeds 50% of the company s total registered capital, no further allocation is required. In addition to being used as a temporary advance payment to cover the losses of the company, the SRF can also be used to increase the company s capital and to expand its production. Employee Welfare and Bonus Fund (EWBF) According to the Company Law, the amount of the EWBF must be between 5-10% of after-tax profits. EWBF can be used as an irregular reward to employees and other collective staff welfare. The fund shall be used in part as an extraordinary bonus for employees (i.e. an invention bonus and an end-of-year bonus) and in part given to the trade union of the company for various welfare activities (e.g., childcare, construction or renovation of employee housing, etc.). Other Surplus Reserve Fund (OSRF) The Company Law does not specify the minimum amount to be contributed to the OSRF. Hence, the proportion allocated is determined in accordance with contract provisions or by the Board of Directors. The OSRF may be used to purchase fixed assets, increase the amount of working capital available and expand the company s production. Only after the allocations to these three reserve funds, can the companies declare dividend distributions from the remaining distributable profits. Profits not distributed 7

9 from previous years may be added to the current year s distributable profits. It is worthwhile noting that when surplus reserve funds (i.e. SRF and OSRF) are used to cover losses of the companies, the OSRF must be utilized before the SRF. However, dividend cannot be paid if the losses of the companies cannot be covered by current year s profits as well as the SRF and OSRF funds. Types of Dividend Payments in China There are two types of dividend payment in China: cash dividend and stock dividend. Stock dividend can be further classified into capitalization of reserve funds (CRF) and pure stock dividends. The pure stock and CRF dividends differ in both accounting and tax treatments, depending on whether the source of the dividends is reserve funds or retained earnings. Pure stock dividends are distributed from retained earnings and are viewed as a distribution of earnings. Hence, they are taxable, just like cash dividends. 1 The capitalization of reserve funds (CRF) generally comes from either capital reserve funds or surplus reserve funds (i.e. the SRF and the OSRF). 2 It is viewed as a distribution of equity capital, therefore, nontaxable. 3 Since the CRF represents a 1 At present, shareholders are levied a 20 percent tax for both cash and stock dividends received; a policy that has remained unchanged since Please also note that capital gains in China are tax exempt. 2 Capital reserve funds consist of accumulated donations, the difference on capital conversion raised from foreign currency transaction and more importantly, premium on capital (i.e. the excess of the amount contributed by investors over the amount of the registered capital), etc. 3 When the CRF comes from surplus reserve funds, the OSRF has to be used first and then the SRF. However, the balance of the SRF cannot fall below 25% of registered capital. In practice, however, capitalization of surplus reserve funds is rare in China. During our sample period, only one case of capitalization of surplus reserve funds was observed 8

10 distribution of equity capital rather than a pure stock dividend payment, we exclude CRF dividends from our sample. 4 Implications With respect to information asymmetry, it is well known that Chinese markets are characterised by opaque information disclosure (see Bai et al., 2004; Lin, Cai and Li, 1998). The existing literature suggests that stock dividends are an effective way of attracting the attention of financial analysts and stock dividends are paid by firms facing high levels of information asymmetry and firms believing themselves to be undervalued (Grinblatt, Masulis, and Titman, 1984; Brennan and Hughes, 1991). Within the context of the Chinese market, the payment of a stock dividend results in a change in the firm s distributable funds, that is retained earnings are transferred to equity capital, suggesting that stock dividends may be costly to the firm. An alternative view that does not rely on costly signals argues that management may choose to disclose information via stock dividends as a form of cheap talk communication between firms and investors (Almazan, Banerjee and Motta, 2008). In this study, we use the number of IBES analysts making one-year forecasts on the firm as a proxy for financial analyst coverage. Further, we conjecture that undervalued firms are more likely to be high growth firms since these firms are more prone to an information asymmetry problem and the growth option tends to be idiosyncratic to these firms. It follows that there is little to be learnt about the firm from its competitors or the market in general (Aboody and Lev, 4 The CRF is always announced with Profit Distribution Scheme in the shareholders meetings, thus, Chinese investors tend to not distinguish between the CRF and the pure stock dividend. Hence, the CRF may always be regarded as an alternative type of stock dividend (see Yu, 2004). However, for the purposes of our study, we exclude CRF from our sample. 9

11 2000). Accordingly, we argue that both stock dividend and non-dividend paying firms may face growth opportunities and exhibit high levels of information asymmetry, but stock dividend paying firms have achieved operating milestones that allow analysts to verify the quality of the investments. Therefore, we conjecture that stock dividend paying firms have higher growth, are more profitable, have higher retained earnings and invest more in fixed assets. Likewise, we conjecture that cash dividend paying firms are more stable and hence have relatively lower growth, are less profitable, have lower retained earnings and invest less in fixed assets. Dividend Declaration Procedures Based on the firm s financial performance at the end of the previous year, at a declaration date, the Board of Directors proposes a Profit Distribution Scheme (PDS), which is announced at the same time as the annual report. The proposed PDS is then submitted to the Annual General Meeting (AGM) for approval. The announcement includes both the form (cash or stock or both) and level of the proposed dividend (zero and above). Following the AGM, the final PDS is submitted to China Securities Depository and Clearing Corporate Limited (CSDCC) for assessment and approval. After the approval of CSDCC, the firm will announce a share registration date, ex-dividend date and comprehensive details on the final PDS. 10

12 Implications Given the prescribed announcement procedure, we examine differences in investor reactions to different forms of dividend announcements (i.e., cash or stock dividend) to confirm the influence of firm characteristics and institutional structures on dividend policy. Investor s reaction to cash and stock dividend announcements are expected to be quite different. 5 For high growth firms that require external financing, paying stock dividend is the optimal choice since it meets the dividend requirement for external financing while allowing for the retention of internal funds for ongoing investment. Thus, paying stock dividend may signal that the firm has growth potential. Furthermore, as we argued previously, stock dividends are paid by firms facing high levels of information asymmetry and firms believing themselves to be undervalued. In this context, stock dividends are an effective way of attracting the attention of financial analysts, suggesting that the market response to the initiation of a stock dividend will differ from the decision to continue paying a stock dividend. In particular, we argue that the initiation of a stock dividend signals the market allowing for verification of the quality of the investments by analysts. However, the continuation of a stock dividend does not serve the same signalling role. Thus, the growth signal embedded in paying stock dividends combined with the role played by analysts suggests that the initiation of stock dividends are associated with positive market reactions. However, unlike paying stock dividend, paying cash dividend may not be growth related, thus, investors may react differently from stock dividend announcements. 5 Given the fact that Chinese investors are levied a 20 percent tax for both cash and stock dividends received while capital gains are tax exempt, this suggests that investor choice between cash and stock dividends is tax neutral. 11

13 Furthermore, as we argued previously, paying cash dividend may be a way that firms compensate and cater to non-tradable shareholders. This action by management is not seen as in the interest of public investors; hence, we expect that cash dividend announcements will be associated with negative market reactions. D. Regulation on Dividend Policy and Equity Financing in China It is generally known that firms in developed economies tend to have stable dividend polices. In contrast, most Chinese listed companies are in a rapid growth phase with a focus on capital accumulation and expansion. Hence, fewer initiations and lower dividend payments are generally observed for Chinese firms (Shao and Lin, 2004). As a result, the China Securities Regulatory Commission (CSRC) has sought to encourage listed companies to establish long term dividend payment policies with an aim to promote cash dividend payments. In March 2001, the CSRC released the Administration Measures for New Shares Issuance by Listed Companies and stated that the underwriters of common equity must pay close attention to the refinancing applications (i.e. right issues, seasoned offerings and convertible bond issues, as defined by CSRC) by the companies who failed to pay dividend in the past three years and whose Boards of Directors provided no justifiable explanation. This was the first time in China that the CSRC formally attempted to regulate dividend payments. However, the CSRC did not specify the types of dividend (i.e. cash dividend or stock dividend) to be paid. More importantly, the payment of dividend was not mandatory as long as the companies Boards of Directors provided satisfactory explanations for failing to pay dividends. 12

14 In December 2004, the CSRC went one step further by stating that company refinancing will not be approved if cash dividends were not paid in the past three years. Moreover, in line with the Administration Measures for New Shares promulgated in May 2006, the CSRC further stated that the issuance of new shares by listed companies shall be consistent with the provision that the companies accumulatively distributed cash or stock dividends over the past three years must be higher than 20% of the average realized annual distributable profits. 6 Implications Given the strong incentive of the Chinese government to regulate dividend payments of listed companies since 2001, we expect that firms will establish a history of dividend payments prior to seeking external equity financing. As identified by Chen, Jian and Xu (2009), this suggests that a positive relationship is expected to exist between the level of tradable shares, subsequent equity financing and dividend payments. We argue that the initiation of a dividend signals the firms financing choice to the market, however, the continuation of a dividend does not serve the same signaling role. We proxy for the firm s external equity financing using the total number of successful new issues conducted within three years after the dividend announcement (# of new offers). 6 While outside our sample period, the CSRC in October 2008, released the Noticed on Amendment in Regulations for Listed Companies Cash Dividend, aiming to urge listed companies to distribute cash dividends. It reinforced that the companies that are looking for refinancing must comply with a provision where the company s accumulatively distributed cash dividends in the past three years must be higher than 30% of the average realized annual distributable profits. In contrast, a previous rule in 2006 allowed listed companies to pay dividend in either cash or stock, while under the new regulation, the dividends can only be distributed in cash. 13

15 In addition, we anticipate differences in the forms of dividend, i.e., cash or stock dividend. The choice of cash or stock dividend can be directly affected by a firm s growth potential and investment. For high growth firms that require external financing, paying stock dividend is the optimal choice since it meets the dividend requirement for external financing while allowing for the retention of internal funds for ongoing investment. 7 In contrast, while paying cash dividend also meets the government s requirement for external financing, it consumes cash and reduces internal funds available for investment opportunities, ceteris paribus. Thus, to the extent that raising external financing is costly and stock dividend meets the government s external financing requirement, we expect that growth firms are more (less) likely to pay stock (cash) dividend. Further, the incentive to seek analyst attention is likely to be strongest at the time of external equity financing, when the incentive to reduce information asymmetry is the greatest. Given the above, we would expect a positive relationship between subsequent equity capital raisings and stock dividend payments. 3. Firm Characteristics and Dividend Policy Our choice of analysing firm characteristics on the propensity for paying various types of dividends is motivated by the existing empirical literature. In particular, the studies that suggest the firm s investment, capital structure, and dividend policies are interrelated. For example, a large empirical literature shows that a firm s dividend decision, which directly affects free cash flow, can affect investment choices (see Higgins, 1981; Lang and Litzenberger, 1989). In addition, Aivazian et al. (2003) report 7 This relationship holds for most of our sample period. 14

16 that similar to their developed economy counterparts, the dividend policy of firms in emerging economies is influenced by profitability, debt, and the growth opportunity. Following Aivazian et al. (2003) we include a set of firm variables that capture factors known to influence dividend policy. Firm size is a proxy for financial market access; larger firms tend to have better market access and should be able to pay higher cash dividends. In our study, firm size is measured by total assets (TA). The literature also suggests a direct link exists between growth, profitability and financing needs. More specifically, high growing firms are generally cash poor and pay lower cash dividends, while more profitable firms with slow growth are cash rich and have higher dividend payouts. We utilize two proxies for growth opportunities: fixed asset growth rate (FA growth rate), defined as the annual growth rate of gross fixed assets, and the total amount of cash spent on construction over total asset (CSC/TA). Return on assets (ROA) is used to measure profitability. The firms ability to pay dividends is influenced by both distributable profits and cash holdings. We proxy for the level of distributable profits using retained earnings per share (REPS), and a dummy variable, DProfit. DProfit takes a value of one if there is an increase in a firm s distributable profits, and zero otherwise. To measure a firm s cash position, cash holding (Cash/TA) is defined as the sum of cash, cash equivalent, shortterm investment, scaled by total assets. In contrast to developed markets, Chinese listed companies tend to be younger and in a rapid growth phase with a focus on capital accumulation and expansion (Shao and Lin, 2004; Wei and Jiang, 2001). Hence, we hypothesis that young firms are less likely to pay cash dividends. As a result, we include firm age (Age) which is measured as 15

17 number of days since the first day of listing, expressed in years. The requirement that firms establish a history of dividend payments prior to seeking external equity financing suggests that the more relevant timeframe for the firm may be the occurrence of the first successful new equity issue. Hence, we include an alternative measure of firm age, SEOAGE, which is measured as the number of years between a firm s first seasoned equity offering and its IPO, expressed in years. To reflect the level of state control of corporations in the Chinese markets, we include a dummy variable (private) that takes a value of one if a firm is a private firm (i.e., not a state-owned enterprise) before listing, otherwise is zero. We conjecture that firms paying cash dividends signal their separation from their previous state governments, and in particular, that the level of cash flows attributable to the state government has been reduced from their pre-listing levels. To the extent that payment of a cash dividend is a substitute for the reduced cash flows, we posit that state owned firms have a strong incentive to pay cash dividends. To test this relationship we include the level of government ownership (GV/TS), measured as the number of government shares relative to total number of shares outstanding. In addition, we interact this measure with the private dummy. In addition, reflecting the substitution between dividend payments and debt financing we including the firm s financial leverage, measured as the book-value debt to equity ratio (DE). Finally, we examine the influence of management response to investor demand in setting the firm s dividend policy and proxy for investor demand using the number of non-tradable shares over total number of shares outstanding (NT/TS). The ownership 16

18 structure of Chinese firms is characterized by state-controlled, closely held firms (Lin et al., 1998) with a large proportion of non-tradable shares, where non-tradable shares account for almost 60% of the total number of shares outstanding (Calomiris, Fisman and Wang, 2008; Gao and Kling, 2008; Chen et al., 2009). While the non-tradable shares reduce stock market liquidity, they also put pressure on firms management to pay cash dividends to cater to the demands of this investor group. For these non-tradable shareholders, the cash dividend payment is generally the primary source of return since their shares cannot be traded in the open markets Data and Sample Selection The data was collected from GTA Data Services (CSMAR). Along with other data items, GTA maintains records on dividend announcement information, top ten institutional and retail shareholding information, firm characteristic and company s financial statement data. Companies also file this information with their listing markets, i.e., Shanghai and Shenzhen Stock Exchanges. To ensure data accuracy and recover missing information, we cross-checked GTA data with various sources, such as databases compiled by SinoFin Financial Information, China Merchants Securities, China Finance Information, and Eastmoney.com. Note that Chinese companies generally report their proposed annual dividend payment (or non-payment) information in their financial report from January to April in the following year. Thus, proposed dividend distribution plans reported in the financial 8 Trading in non-tradable shares is limited to the purchases by state owned (or controlled) institutions through negotiation or auction and requires approval from the government. More specifically, the transfer price is determined based on net asset value, not the market price (see Calomiris et al., 2008; Xiao, 2005). 17

19 report of 2007, for example, are actually the dividend payment for the financial year of The initial sample consists of 6,584 firm-year annual dividend announcements during the period from 2003 to The sample starts in 2003 since detailed information on dividend announcements is not available prior to Among the 6,584 firm-year dividend announcements, 315 observations have missing information on financial report dates and these observations are removed from the sample. Finally, we apply various filters to mitigate possible data errors that could not be recovered through crosschecking. The final sample consists of 5,153 firm-year observations. We identify firms as dividend payers or non-payers. Non-payers are firms that do not pay any type of dividend in the current period while payers are firms that pay any type of dividend (cash or stock or both) in the current year. Within dividend payers, we identify stock dividend payers as those firms that pay stock or both stock and cash dividends and cash dividend payers as those firms that pay cash dividends only. 5. Empirical Results In this section, we investigate whether institutional variables and firm characteristics discussed in sections 2 and 3 influence the likelihood of paying dividends, different forms of dividends and market reactions to various dividend announcements. 9 Note that dividend announcement in a given year is for prior year s earnings distribution. 18

20 A. Firm Characteristics Summary statistics (mean values) of firm characteristic variables are reported in Table 1. We observe a large difference between cash and stock dividend payments. The mean dividend per share is only 0.14 yuan for cash dividend payers, which is equivalent to a dividend yield of 1.76%. In contrast, the mean stock dividend per share is equivalent to 3.97 yuan. 10 [Table 1 about here] It is clear from Table 1 that dividend paying firms are younger, larger, more profitable, have lower leverage, higher cash holdings, and spend more cash on construction than non-dividend paying firms. In addition, dividend paying firms have higher retained earnings, reflecting a greater ability to pay stock dividends. These results are consistent with the existing literature (Aivazian and Booth, 2003) and suggest that as with other emerging economies, firm characteristics strongly influence dividend policy in China. Dividend paying firms have stronger shareholder protection (measured by lower management corruption and higher property rights protection), higher government shareholding and operate in provinces characterised by higher institutional investor presence. To investigate whether our selected firm characteristics variables are likely to be subject to collinearity problems in our later regression analysis, we examine the correlations between the independent variables that are used in our analysis. We find that these variables are generally not highly correlated with each other. The exception is the 10 For stock dividends, dividend per share is the sum of any cash dividend and the cash equivalent stock dividend expressed on a per share basis. The cash equivalent stock dividend per share is computed as P*( n (1 + n)), where P is share price and n is number of stock dividend per share 19

21 correlation between profitability (ROA) and retained earnings per share (REPS), nontradable (NT/TS) and government shareholdings (GV/TS), private ownership prior to listing (private) and government shareholdings (GV/TS), provincial institutional share holdings (InstHoldings) and provincial property rights protection (PP); and number of years since IPO (Age) and the number of years between a firm s first seasoned equity offering and its IPO (SEOAGE). [Table 2 about here] A logistic regression is used to model the dividend choice as a function of firm characteristics. In addition to the institutional and firm variables described in Sections 2 and 3, we include a cross listing dummy variable that takes a value of one if the firm has cross-listed shares in either B-share markets in China or other overseas markets. To control for stock market conditions and industry effects, all regressions include a bullmarket dummy variable that takes a value of one if the announcement date of a proposed dividend plan is defined as a bull market, where the classification of bull and bear markets follows Yan et al. (2007), and industry dummy variables for the finance, utilities, properties, conglomerate and industrial industries. As a consequence of multicollinearity we cannot include all the variables in a single regression. Rather, we estimate a separate regression including government ownership (GV/TS), institutional shareholdings (InstHolding) and period to first SEO (SEOAGE). [Table 3 about here] Consistent with our expectations, the results for Columns 1 and 2 of Table 3 show that younger, larger, and more profitable firms are more likely to pay dividends. Further, 20

22 these firms tend to have lower leverage, higher cash holding and retained earnings, and spend more cash on construction. However, asset growth is only weakly positively related to the likelihood of paying dividends. Consistent with firms conforming to the requirement that firms establish a history of dividend payments prior to seeking external equity financing, firms undertaking subsequent equity offerings are more likely to pay dividends and the longer the period between the firm s IPO and its initial secondary equity offering (SEOAGE), the more likely the firm is to pay dividends. In addition, we find that corporate governance and shareholder protection influence the likelihood of firms paying dividends, and cash dividends in particular (see Column 3 and 4). Consistent with our expectations, firms with lower management corruption, higher provincial institutional shareholdings and higher property rights protection are more likely to pay dividends. The cross-listing dummy variable, however, does not appear to influence the likelihood of paying dividends. In general, the choice of paying cash dividends over not paying a dividend exhibits similar results (Columns 3 and 4 of Table 3). Consistent with the earlier results of Wei, Zhang, and Xiao (2003) we find that firms that were both state owned prior to listing and have higher government ownership at the time of the dividend announcement are more likely to pay cash dividends. This is consistent with our argument that state owned firms use the payment of cash dividends to signal lower cash flows to their respective state governments. Finally, the finding that firms with higher levels of non-tradable shares are more likely to pay cash dividends is consistent with our catering argument where firms cater to the demands of non-tradable shareholders in setting their dividend policy. 21

23 In terms of the choice to pay stock dividends, the results in Columns 5 and 6 show that larger and high growth firms are more likely to choose to pay a stock dividend rather than forgoing a dividend. In addition, these firms appear to be more profitable, are highly leveraged, exhibit higher growth, investing more in fixed assets and hold higher levels of retained earnings than firms choosing to pay a cash dividend (See Columns 7 and 8 of Table 3). These results are consistent with the argument that higher growth firms with higher retained earnings and greater investment in fixed assets use stock dividends to attract the attention of analysts (Grinblatt, Masulis, and Titman, 1984). Contrary to expectations, we find that the likelihood of paying a stock dividend is only weakly influenced by subsequent equity capital raisings. Finally, the choice of a stock dividend rather than a cash dividend is less likely for firms with high proportion of non-tradable shares and for firms operating in regions with high protection of property rights, that is, high shareholder protection. B. Announcement Effect With respect to the influence of firm characteristics on the market reaction, Table 4 reports the cross-sectional analysis of dividend announcement returns. Columns 1 and 2 of Table 4 investigate the difference in market reactions between dividend payers and non-payers. A dummy variable, Payer, is included. It takes a value of one if a firm pays any dividend and zero otherwise. We also include DProfit as a control for the confounding effect of earnings announcements. DProfit is a dummy variable that takes a value of one if the change in a firm s distributable profit is positive, zero otherwise. As shown in Columns 1 and 2, the overall market reaction to the announcement of a 22

24 dividend payment is indifferent to whether the firm is a payer or non-payer, given the insignificance of the dummy variable, Payer. This result is not surprising given the relatively small sample of stock dividend payers and the similarity observed in market reactions to the announcement of cash dividends and non-payers (see Figure 1). [Table 4 about here] We now compare the difference in market reactions between cash and stock dividends, including a dummy variable, Cash, which takes a value of one if a firm pays only a cash dividend and is zero for stock dividends. We find that the coefficients of Cash are consistently negative and significant at 1% level, suggesting the announcement effect for stock dividends is significantly greater than that for cash dividends (Columns 3 and 4). Further, the higher the proportion of non-tradable shares, the lower is the market reaction to the dividend announcement. This is consistent with firms paying cash dividends as a way to compensate and cater to non-tradable shareholders, where this action by management is not seen as in the interest of public investors. Consistent with the argument that the initiation of a dividend signals the firm s financing choice to the market, we find that the market reaction to the dividend announcement is positively related to subsequent equity capital raisings. Columns 5, 6 and 7 of Table 4 investigate the differences in market reactions to stock dividend payers that initiate a dividend, continue to pay a dividend or change between cash and stock dividends. Three dummy variables are included individually: Continue, takes a value of one for firms continuing to pay a stock dividend from last period and zero otherwise; Initiate equals one for newly initiated stock dividend payers 23

25 (i.e. no dividend was paid last period) and zero otherwise; and Changeto equals one for payers that switch from cash dividends in the last period to stock dividends in current period and zero otherwise. The market response to the initiation of a stock dividend is strongly positive and significant. In contrast, the market response is insignificantly different from zero for firms continuing to pay stock dividends or choosing to change from cash to stock dividends. This result is consistent with our conjecture that the initiation of a stock dividend signals the market the firm s financing choice to the market, whereas, the continuation of a dividend does not serve the same signaling role. Table 5 investigates the differences in analyst coverage prior to and post the dividend announcement. 11 Analyst coverage is defined as the number of IBES analysts making one-year forecasts on the firm. The data is obtained from IBES International for the period The results show that the mean analyst coverage increases after the announcement of a stock dividend, from 1.49 to 2.15, and the difference is statistically significant at the 5% level. Further, this increase occurs only after the initiation of a stock dividend. While the level of analyst coverage does not change around the announcement for cash dividends and non-payers. These results are consistent with our argument that the initiation of a stock dividend signals the market allowing for verification of the quality of the investments by analysts., [Table 5 about here] 11 Given the availability, the data on analyst coverage is only available for 775 firms equivalent to 2,381 firm-year observations. 24

26 C. Fixed-Effects and Cross-Sectional Estimates This section examines the extent to which the cross-sectional and time-series variation in our explanatory variables drives our results. Up to this point our emphasis has been on the cross-sectional variation in the likelihood of paying dividends and announcement effects. However, firms dividend payments can vary over time, and some of that variation may be explained by the year to year changes in our explanatory variables. To estimate the extent to which our dividend payment results are generated from the cross section versus the time series, Table 6 reports the result on year-fixed effect regressions on both the likelihood of paying a dividend (Columns 1 and 2) and the choice between cash and stock dividends (Columns 3 and 4). [Table 6 about here] The regression estimates reported in Table 6 indicate that the relationships between dividend policy and firm characteristics are significant in the time-series, and are consistent with our earlier estimates reported in Table Likewise, to estimate the extent to which our announcement effect results reported in Table 4 are generated from the cross section versus the time series we estimate both Fama-MacBeth (1973) and year-fixed effect regressions. Specifically, we report Fama-MacBeth (1973) and year-fixed effect regressions for both the full sample and the sub-sample of dividend payers in Table 7. The regression estimates reported in Table 7 indicate that the relationships between dividend announcement effects and firm 12 With the inclusion of year-fixed effects, the continued positive and significance coefficient for nontradeable shares suggests that this result is not mechanistically driven by the split share structure reform introduced in

27 characteristics are consistent with our earlier estimates (reported in Table 4) in both the cross-section and time-series. [Table 7 about here] 6. Conclusion The current study examines firm dividend policy in China; the world s largest emerging economy. At the outset we posed the question of what motivates the firm s choice of cash or stock dividends. Using a sample of Chinese firms dividend announcements between 2003 and 2007, we find that both firm characteristics and institutional structures influence the propensity of Chinese firms to pay dividends. In particular, we find that more profitable, lower leveraged, higher cash holding firms and firms undertaking subsequent equity offerings are more likely to pay dividends. In addition, we find that stronger corporate governance and shareholder protection is associated with dividends payments. That is, firms with stronger shareholder protection, state ownership prior to listing and operating in provinces characterised by higher institutional investor presence are more likely to pay cash dividends. While growing firms with high levels of retained earnings and higher investment in fixed assets pay stock dividends and seek new external equity financing. Further, consistent with management responding to investor demands in the setting of dividend policy, we find that firms with higher proportions of non-tradable shares are more likely to pay cash dividends. Finally, the announcement effect for stock dividends is significantly greater than that for cash dividends and this effect is driven by the initiation of stock dividends. For 26

28 those firms initiating a stock dividend, analyst following increases significantly after the dividend announcement. In contrast, firms continuing to pay a stock dividend do not evidence an increase in analyst coverage post the dividend announcement. These results are consistent with the argument that firms use the initiation of stock dividends to attract the attention of analysts. 27

29 References Aboody, D., and Le, B., Information asymmetry, R&D, and insider gains. Journal of Finance 55, Aivazian, V., Booth, L., and Cleary, S., Do Emerging Market Firms Follow Different Dividend Policies From U.S. Firms? Journal of Financial Research, Almazan, A., Banerjee, S., and de Motta, A., Attracting Attention: Cheap Managerial Talk and Costly Market Monitoring. Journal of Finance 63, Bai. C., Liu Q., Lu, J., Song, F., and Zhang, J., Corporate governance and market valuation in China. Journal of Comparative Economics 32, Bechmann, K.L., and Raaballeb, J., The Differences Between Stock Splits and Stock Dividends: Evidence from Denmark. Working paper, Copenhagen Business School. Booth, L., Aivazian, V., Demirguc-Kunt, A., and Maksimovic, V., Capital Structures in Developing Countries. Journal of Finance 56, Brennan, M.J. and Hughes, P.J., Stock prices and the supply of information. Journal of Finance 46, Cai, H., Fang, H., and Xu, L., Eat, drink, firms and government: an investigation of corruption from entertainment expenditures of Chinese firms, working paper at Peking University. Calomiris, C., Fisman, R., and Wang, Y., Profiting from government stakes in a command economy: evidence from Chinese asset sales. Forthcoming at Journal of Financial Economics. Chen, D., Jian, M., and Xu, M., Dividends for Tunnelling in a Regulated Economy: The Case of China. Pacific Basin Finance Journal 17, Chen, G., Firth, M., and Gao, N., The Information Content of Concurrently Announced Earnings, Cash Dividends, and Stock Dividends: An Investigation of the Chinese Stock Markets. Journal of International Financial Management and Accounting 123, Djankov, S., La Porta, R., Lopez-de-Silanes, F., and Shleifer, A., Courts, Quarterly Journal of Economics 118, Easterbrook, F., Two agency-cost explanations of dividends. American Economic Review 74,

30 Fama, Eugene F., and James D. MacBeth, 1973, Risk, return, and equilibrium: Empirical tests, Journal of Political Economy 81, Fama, Eugene F., and French, Kenneth R., Testing Trade-off and Pekcing Order Predictions about Dividends and Debt, Review of Financial Studies 15, Fan, J.P.H., Titman, S., and Twite, G., An International Comparison of Capital Structure and Debt Maturity Choices. Working paper, Australian National University. Fan, J.P.H., Wong, T.J., and Zhang, T., Politically Connected CEOs, Corporate Governance and Post-IPO Performance of China s Newly Partially Privatized Firms. Journal of Financial Economics 84, Gao, L., Kling, G., Corporate governance and tunneling: empirical evidence from China. Pacific Basin Finance Journal 16, Glen, J.D., Karmokolias, Y., Miller, R.R., and Shah, S., Dividend policy and behavior in emerging markets: To pay or not to pay. IFC Discussion paper 26, International Finance Corporation. Grinblatt, M. S., Masulis, R.W., and Titman, S., The Valuation Effects of Stock Splits and Stock Dividends. Journal of Financial Economics, 13, Higgins, R., Sustainable growth under inflation. Financial Management 10, Jensen, M., Agency costs of free cash flow, corporate finance and takeovers. American Economic Review, 76, John, K., Knyazeva, A., Knyazeva, D., Geography and Dividend Policy, Working paper, New York University. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., and Vishny, R., Agency Problems - and Dividend Policies around the World, Journal of Finance 55, Lang, L. and Litzenberger, R., Dividend Announcements: Cash Flow Signalling vs. free Cash Flow Hypothesis. Journal of Financial Economics, Lin, J., Cai, F. and Li, Z., Competition, policy burdens, and state-owned enterprise reform. American Economics review 88, Mitton, T., Corporate Governance and Dividend Policy in Emerging Markets. Emerging Markets Review, Sen, N., Le, Z., and Ferris, S.P., Payout Policy, Ownership Concentration and Corporate Governance: Evidence from China. Working paper, Nanyang Technological University. 29