Why do firms pay stock dividends: Is it just a stock split?

Abstract

This paper examines why firms choose to pay stock dividends. Using a sample of listed Chinese firms we find that older, more profitable firms with lower leverage, higher levels of retained earnings, private ownership prior to listing, that invest more in fixed assets and operate in regions with lower shareholder protection are more likely to pay stock dividends. Consistent with stock dividends substituting for stock splits, our evidence indicates that the initiation of a stock dividend is associated with a significant positive market reaction and increased analyst following. These results suggest that firms use stock dividends to attract analysts’ attention. In addition, the positive announcement effect for stock dividends increases with the size of the split factor, suggesting that management use stock dividends to keep the firm’s stock price within its acceptable trading range.

Keywords

Dividend policy stock dividend stock split

1. Introduction

What motivates a firm’s decision to pay a stock dividend? Firms have multiple methods to distribute back to shareholders, such as paying cash dividends, stock dividends, or repurchasing shares. In the past 50 plus years, a large body of literature has heavily focused on the analysis of cash dividends, although many puzzles still remain unsolved and there is no consensus conclusion. The main explanations for paying cash dividends include: (1) signaling future profitability (Bhattacharya, 1979; Miller and Rock, 1985); (2) reducing free cash flow and mitigating agency problems (Easterbrook, 1984; Jensen, 1986); (3) minority shareholders pressuring corporate insiders to disgorge cash (La Porta et al., 2000); and (4) catering to investors when they put a stock price premium on dividend payers (Baker and Wurgler, 2004).¹

Given that paying stock dividends, in essence, does not increase shareholders’ wealth directly, nor does it change the percentage ownership of current shareholders (as it simply increases the total number of shares outstanding), it is unclear why many firms still choose to pay stock dividends and how investors react to such events. In the current study, we investigate the motivation of paying stock dividends and related issues by exploring two competing hypotheses: stock split substitution versus cash dividend substitution.

Firms may use stock dividends as a substitute for stock splits for three possible reasons: (1) signaling managers’ revised expectations of the firm’s future prospects (Grinblatt et al., 1984; McNichols and Dravid, 1990); (2) communicating between firms and investors and attracting the attention of market participants (Almazan et al., 2008; Grinblatt et al., 1984); and (3) enabling the firm to adjust the share price to a more appropriate range and thus increasing the share’s liquidity (Copeland, 1979; Lin et al., 2009; McNichols and Dravid, 1990). Alternatively, stock dividends can also be used as a substitute for cash dividends for firms wishing to maintain a dividend history when they are unable or unwilling to pay a cash dividend (Lakonishok and Lev, 1987).

However, under normal market conditions, it is hard or impossible to identify whether a firm pays stock dividends substituting for stock splits or for cash dividends. The current study contributes to the literature by employing a unique Chinese sample that helps us unambiguously identify the motivation of paying stock dividends. Specifically, splits are prohibited under the Company Law of the People’s Republic of China. Thus, if a firm desires stock splits for any of the abovementioned considerations, issuing stock dividends is the only avenue for it.

This study also builds on the existing literature by investigating the influence of managements’ response to investor demand in setting the firm’s dividend policy. In the USA and many other countries, firms may use cash dividends as a strategy to mitigate potential agency problems. In contrast, paying cash dividends in China may be an outcome of agency problems rather than a means of mitigating agency problems, since Chinese firms are characterized by a large portion of non-tradable shares and the dividend is the primary source of investment return for the non-tradable shareholders.² The extent to which cash dividends are set to cater to the demands of non-tradable shareholders weakens the effectiveness as a signal.

Therefore, under the unique Chinese circumstances (i.e. prohibition of stock splits and preference of cash dividends by non-tradable shareholders), investors would react differently to stock dividends depending on whether the stock split substitution hypothesis or the cash dividend substitution hypothesis holds. If a firm pays stock dividends just for the purpose of establishing dividend history or substituting for cash dividends, investors are expected to react weakly or negatively to stock dividends. In contrast, if a firm pays stock dividends substituting for stock splits for the abovementioned reasons in the context of the Chinese environment, the stock split substitution hypothesis predicts a positive market reaction.

In addition to relying on the market reaction to distinguish between the two competing hypotheses, we also investigate how firm characteristics influence the choice between stock and cash dividends. In particular, stock dividends are likely to be paid by firms facing high levels of information asymmetry, which seek the attention of financial analysts (Almazan et al., 2008; Brennan and Hughes, 1991; Grinblatt et al., 1984). Unlike cash dividends, stock dividends do not affect the firm’s future cash flow or internal financing. However, in line with cash dividends, stock dividends reduce the current level of retained earnings. If the “attention” argument holds, we expect to see a positive announcement effect and an increase in analyst following when firms initiate stock dividends.

To understand our conjecture, we use a sample of 5153 dividend announcements by Chinese firms between 2003 and 2007, and plot the market adjusted cumulative abnormal return from t = −5 to t = 30 days for dividend payers and non-payers, where t = 0 is the announcement day of the proposed dividend distribution scheme. We further divide dividend 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 and stock payers). As is shown in Figure 1, the market responds in a similar way to both non-payers and cash dividend payers. However, the response to the announcement of stock dividends (either separately or combined with a cash dividend) is both positive and significant, suggesting that the market views stock dividends as good news, relative to both cash dividends and the absence of a dividend payment.³ The question remains, why do firms choose to pay stock dividends?

Figure 1.

Market adjusted cumulative abnormal returns around the announcement date of the proposed dividend distribution scheme.

Our evidence suggests that more profitable, lower levered, higher cash holding firms, and those that face stronger shareholder protection and undertake subsequent equity offerings, are more likely to pay dividends. In addition, we find that older firms with high levels of retained earnings, private ownership prior to listing, that invest more in fixed assets and operate in regions with lower shareholder protection, are more likely to pay stock dividends.

Overall, the results with respect to the market reaction to the announcement of a stock dividend are consistent with the stock split substitution hypothesis. We find that the positive and significant announcement effect for stock dividends is mainly related to the initiation of stock dividends. For those firms initiating stock dividends, analyst following increases significantly after the dividend announcements. In contrast, firms continuing to pay a stock dividend do not show 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 communicate with investors and attract the attention of analysts (Almazan et al., 2008; Grinblatt et al., 1984).

Finally, consistent with management making use of stock dividends to keep the firm’s stock price within its acceptable trading range, we find that the positive announcement effect for stock dividends increases with the size of the split factor (Lin et al., 2009; McNichols and Dravid, 1990).

2. Institutional background: discussion and implications

2.1. Dividend payment

There are two types of dividend payments in China: stock dividends and cash dividends. Stock dividends can be further classified into capitalization of reserve funds (CRFs) 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, in the same manner as cash dividends.⁴

The CRFs generally comes from either capital reserve funds or surplus reserve funds (i.e. the Surplus Reserve Fund (SRF) and the Other Surplus Reserve Funds (OSRF).⁵ It is viewed as a distribution of equity capital and, therefore, non-taxable.⁶ Since the CRF represents a distribution of equity capital rather than a dividend payment, we exclude CRF dividends from our sample.

2.1.1. Implications

The ownership structure of Chinese firms is characterized as state-controlled and closely held (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 et al., 2010; Chen et al., 2009; Gao and Kling, 2008). 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.⁷ Hence, non-tradable shareholders may pressure firms’ management to pay cash dividends. We proxy for investor demand using the number of non-tradable shares over total number of shares outstanding (NT/TS).

2.2. 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 the China Securities Depository and Clearing Corporate Limited (CSDCC) for assessment and approval. After approval from the CSDCC, the firm will announce a share registration date, ex-dividend date, and comprehensive details on the final PDS.

2.2.1. Implications

The prescribed announcement procedure provides a clear distinction between cash dividends and stock dividends. This allows us to examine the different forms of dividend announcements (i.e. stock or cash dividend) to confirm the influence of firm characteristics and institutional structures on dividend policy.

With respect to information asymmetry, it is well known that Chinese markets are characterized by opaque information disclosure (see Bai et al., 2004; Lin et al., 1998). Within this context, the existing literature suggests that stock dividends are an effective way of signaling a reassessment of firm value, whereby stock dividends are paid by firms facing high levels of information asymmetry and firms believing themselves to be undervalued (Brennan and Hughes, 1991; Grinblatt et al., 1984). In addition, within the context of the Chinese market, the payment of a stock dividend results in a change to the firm’s distributable funds. Specifically, retained earnings are transferred to equity capital, suggesting that stock dividends may be costly to the firm and can restrict the ability of the firm to pay cash dividends in the future. 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 et al., 2008).

We conjecture that undervalued firms are more likely to be high-growth firms, since these firms are more prone to information asymmetry problems and the growth option tends to be idiosyncratic to these firms. Thus, there is little to be learnt about the firm from its competitors or the market in general (Aboody and Lev, 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 use the dividend announcement to attract analysts, having 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.

This further suggests 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 signaling 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. In this study, we use the number of the Institutional Brokers’ Estimate System (I/B/E/S) analysts making one-year earnings forecasts as a proxy for financial analyst coverage.

Unlike stock dividends, cash dividends may not be growth related; hence, investors might react differently from stock dividend announcements. Furthermore, as we argued previously, paying cash dividends 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.

It may also be the case that stock dividends do not serve a signaling role. Rather, management makes use of stock dividends to achieve a stock split, keeping the firm’s stock price within its acceptable trading range. We proxy for trading range using the split factor attributable to stock dividends (Split), measured as the number of shares outstanding after the stock dividend divided by the number of shares outstanding before the stock dividend.⁸ We conjecture that the split factor will be positively related to dividend announcement returns.

2.3. Regulation of dividend policy and equity financing in China

In the past decade, most Chinese listed companies were in a rapid growth phase with a focus on capital accumulation and expansion. Hence, fewer initiations and lower dividend payments were generally observed (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 “Measures for the Administration of the Listed Company Issuing New Shares” 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 the CSRC) by the companies who failed to pay dividends 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 dividends was not mandatory as long as the companies’ Boards of Directors provided satisfactory explanations for failing to pay dividends.

In December 2004, the CSRC went one step further by stating that company refinancing will not be approved if dividends were not paid in the past three years. Moreover, in line with the “Administrative Measures for the Issuance of Securities by Listed Companies” 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.

In October 2008, the CSRC introduced a significant regulatory change designed to relate cash dividends directly and exclusively to refinancing. “The Decisions on Amending Some Provisions on Cash Dividends by Listed Companies” reinforced that companies seeking external finance 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 to previous rules in the “Administrative Measures for the Issuance of Securities by Listed Companies” in 2006, which allowed listed companies to pay dividends in either cash or stock. This change has the potential to significantly impact on the firm’s decision to pay stock dividends.⁹

2.3.1. 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 et al. (2009), this suggests a positive relationship between the level of tradable shares, subsequent equity financing, and dividend payments.

Accordingly, we argue that the initiation and continuation of dividends serve different signaling roles and have different market reactions. Initiation of dividends signals the market with respect to the revaluation of the firm’s growth potential, while continuation of dividends in part meets the requirement of establishing a history of dividend payments prior to external equity financing. 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).¹⁰

In addition, we anticipate differences in the forms of dividend, that is, cash or stock dividend. The choice of cash or stock dividends can be directly affected by the firm’s growth potential and investment. For high-growth firms that require external financing, paying a stock dividend is the optimal choice as it meets the dividend requirement for external financing while allowing for the retention of internal funds for ongoing investment.¹¹ In contrast, while paying a 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 dividends meet the government’s external financing requirement, we expect that growth firms are more (less) likely to pay a 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 analyzing the influence of firm characteristics on the propensity for paying various types of dividends is motivated by the existing empirical literature. In particular, prior studies suggest that the firm’s investment, capital structure, and dividend policies are interrelated. Following Aivazian et al. (2003), we include a set of firm variables that captures 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 that a direct link exists between growth, profitability, and financing needs (Aivazian et al., 2003). More specifically, high-growth firms are generally cash poor and pay lower cash dividends, while more profitable firms with slow growth are cash rich and pay higher dividends. 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 firm’s 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, and short-term investments 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 posit that young firms are less likely to pay cash dividends. As a result, we include firm age (Age), which is measured as 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 most 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 initial public offering (IPO), expressed in the number of 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 it takes a value of zero. We conjecture that previously state-owned firms paying cash dividends signal their “separation” from their previous state government 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 the total number of shares outstanding. In addition, we interact this measure with the Private dummy. To reflect the substitution between dividend payments and debt financing, we include firm’s financial leverage, measured as the book value of debt to equity ratio (DE).

The existing literature (Aivazian et al., 2003; La Porta et al., 2000; Mitton, 2004) suggests that stronger shareholder protection and corporate governance are associated with higher cash dividends rather than stock dividends. Using a sample of Chinese firms, Sen et al. (2009) find that 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, as the government largely controls directorships of Chinese listed firms (Fan et al., 2007).

Differing from previous studies, we focus on two measures of shareholder protection: the level of corruption 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 et al. (2011). This index measures the extent of corruption 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 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: the property rights protection index. Djankov et al. (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 et al., 2011) 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 privately controlled firm’s dividend policy, the property rights protection measure (PP) proxies for the effective pressure by individual shareholders to force corporations to distribute cash dividends. Each firm in the sample is assigned a corresponding provincial property rights protection value based on the location of the firm’s main operations.

The existing literature suggests that geographic location influences a firm’s information environment and, as a consequence, its dividend policy (John et al., 2011). 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 building. To mitigate this inefficient investment, investors are expected to demand higher cash dividends.

As noted by John et al. (2011), 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 China Stock Market and Accounting Research database (CSMAR). We first identify the proportion of the top 10 institutional holdings in each firm, then we sum this measure by province according to the location of the firm’s main operations, and each firm in the sample is assigned the corresponding provincial value.¹²

4. Data and sample selection

The data was collected from GTA Data Services (CSMAR), including top 10 institutional and retail shareholding information, firm characteristics, and company’s financial statement data. To ensure data accuracy and recovery of 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.¹³

The initial sample consists of 6584 firm-year annual dividend announcements during the period from 2003 to 2007.¹⁴ Among the 6584 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 cross-checking. The final sample consists of 5153 firm-year observations.

We identify firms as dividend payers or non-dividend payers. Non-dividend payers are firms that do not pay any type of dividend in a given year, while dividend 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 first investigate whether the 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.

5.1. Firm characteristics

The composition of the dividend announcement sample is reported in Table 1. The sample is categorized by the form of the dividend (cash, stock, or stock/cash combination) and whether the subsequent dividend payment increases, decreases, or remains unchanged relative to the current dividend payment.

Table 1.

Sample characteristics.

Panel A categorizes the sample by the form of the dividend: Non-payers (firms that do not pay any type of dividend); Cash Payers (firms that pay only cash dividends); and Stock Payers (firms that pay stock or stock plus cash dividends). For the sub-samples of Cash Payers and Stock Payers, Panel B reports the relative change in the dividend between the current and subsequent dividend payment (increase, decrease, no change). Panel C reports the mean announcement return (MCAR). MCAR is the market-adjusted cumulative abnormal return from t = −5 to t = 10, where t = 0 is the announcement date of proposed dividend distribution. Initiate denotes the sub-sample of firms initiating a dividend in the subsequent period. Changeto denotes the sub-sample of firms that switched from cash (stock) dividends to stock (cash) dividends in the subsequent period; Continue denotes the sub-sample of firms that paid the same type of dividends in both periods.

Panel A: Sample break-down based on dividend type
Non-payers	Cash Payers	Stock Payers		Total
2513	2365	275		5153
Panel B: Dividend changes (Div_t₊₁ – Div_t) by dividend type
	Increase	Decrease		No Change
Cash dividend payers	26.28%	52.67%		21.05%
Stock dividend payers
- Stock component	4.59%	87.24%		8.16%
- Cash component	23.97%	58.16%		17.86%
Panel C: Mean market-adjusted MCARs (%)
Stock Dividends		Dividend changes (Div_t₊₁ – Div_t)
		Increase	Decrease	No Change
- Initiate	7.56
- Changeto	3.43
- Continue	3.65	7.56	2.03	–
Cash Dividends
- Initiate	2.13
- Changeto	−0.06
- Continue	0.31	1.09	−0.43	0.93

Of the 5153 dividend announcements, 49% represent non-dividend payments, while 46% are cash dividends (Panel A). For both cash and stock dividends, a large proportion of dividend payments are followed by a smaller dividend in the next period. Among the sub-sample of cash payers, 53% of firms decrease the subsequent dividend payment while only 26% increase the subsequent dividend (Panel B). For stock dividend payers, 87% of the firms decrease the stock component, while 58% of firms decrease the cash component. With respect to announcement return (Panel C), we observe that the mean market-adjusted cumulative abnormal return around the announcement of a stock dividend is positive for both dividend increases and decreases, and is consistently greater than the corresponding cash dividend announcement.

Summary statistics (mean values) of firm characteristic variables are reported in Table 2. 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 (mean dividend per shares over mean share price) of 1.76%. In contrast, the mean stock dividend per share is equivalent to 3.97 yuan.¹⁵

Table 2.

Summary statistics.

This table presents mean values for each firm variable. DPS is dividend (measured in yuan) per share. For stock dividend payers, DPS is calculated as the sum of cash dividend per share and cash equivalent stock dividend per share. 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; Dividend payout is dividend per share over earnings per share; Split is a ratio of the number of outstanding shares after the stock dividend over the number of shares before the stock dividend; TA is the total assets; ROA is return on total asset and is expressed in percentage; DE is debt to equity ratio; Cash/TA = (cash + cash equivalent + short-term investment)/TA; FA growth rate is the annual growth rate of gross fixed assets; NT/TS is number of non-tradable shares (NT) relative to total number of shares outstanding (TS); GV/TS is number of government shares (GV) relative to total number of shares outstanding (TS); CSC/TA is total amount of cash spent on construction relative to total assets; Age is the number of days since the first day of listing, expressed in years; REPS is the retained earnings per share, used to measure a firm’s ability to pay stock dividends; ETC measures the total entertainment and travel costs of the firms in each province and is standardized by total sales of the firms (Cai et al., 2011), which measures management corruption for each province in China. Each firm in the sample is assigned a corresponding provincial ETC value based on the location of the firm’s main operations; PP denotes property rights protection at provincial level (Cai et al., 2011), which measures provincial government protection of property rights in China. Each firm in the sample is assigned a corresponding provincial PP value based on the location of the firm’s main operations; InstHolding is provincial institutional share holdings. Each firm in the sample is assigned a corresponding provincial value according to the location of the firm’s main operations; SEOAGE is number of years between a firm’s first seasoned equity offering and its IPO, expressed in years; % of Private is the percentage of firms that are privately owned before they went public; % of Cross-list is the percentage of firms that have their shares cross-listed in more than one stock exchange. With respect to dividends, non-payers are firms that do not pay any dividends in the current period, while payers are firms that pay dividends in the current year. Payers are further classified into cash dividend payers (that pay cash dividends only) and stock dividend payers (that pay stock or both stock and cash dividends). The sample period is from 2003 to 2007. ***, **, and * denote statistical significance at 1%, 5%, and 10% levels, respectively.

Variable	Whole sample (1)	Non-payers (2)	Payers (3)	Stock dividend payers (4)	Cash dividend payers (5)	t-statistics on the difference between (2) – (3)	t-statistics on the difference between (2) – (4)	t-statistics on the difference between (2) – (5)	t-statistics on the difference between (4) – (5)
DPS	0.07	0.00	0.13	3.97	0.14	−62.87***	−61.95***	−17.49***	16.91***
Dividend payout	40.13	0.00	74.14	1466.75	80.16	−6.23***	−6.01***	−14.87***	13.95***
Split				1.42
TA	2503.68	1821.86	3152.70	3089.74	3160.03	−15.93***	−15.28***	−5.86***	−0.31
ROA	0.02	0.00	0.05	0.06	0.05	−38.95***	−37.84***	−23.53***	4.84***
DE	0.79	1.05	0.55	0.65	0.53	7.53***	7.71***	5.30***	3.31***
Cash/TA	0.23	0.21	0.25	0.24	0.25	−10.76***	−10.63***	−3.92***	−0.94
FA growth	0.16	0.12	0.20	0.26	0.19	−9.37***	−8.33***	−6.43***	2.95***
NT/TS	0.58	0.56	0.59	0.55	0.59	−6.59***	−7.80***	1.85*	−4.91***
GV/TS	0.35	0.32	0.38	0.31	0.39	−9.16***	−10.06***	0.26	−4.52***
CSC/TA	0.06	0.05	0.08	0.09	0.08	−19.54***	−18.36***	−9.09***	2.53**
Age	6.39	7.15	5.68	6.34	5.60	16.19***	16.55***	3.60***	3.24***
REPS	0.23	−0.06	0.51	0.68	0.49	−35.12***	−33.64***	−23.92***	6.46***
ETC	0.01	0.01	0.01	0.01	0.01	8.42***	8.35***	3.51***	0.83
PP	0.60	0.59	0.61	0.59	0.61	−3.85***	−4.28***	0.54	−2.58**
SEOAGE	3.05	2.28	3.30	3.26	3.31	−7.12***	−7.17***	−2.25***	−0.81
InstHolding	0.06	0.05	0.07	0.06	0.07	−6.17***	−5.98***	−2.67***	−0.52
% of Private	30.29	38.00	22.95	34.18	21.65
% of Cross-list	8.67	8.60	8.75	5.82	9.09
# of Obs	5153	2513	2640	275	2365

It is clear from Table 2 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 in general consistent with the existing literature (Aivazian et al., 2003) and suggest that firm characteristics strongly influence dividend policy in China as in other emerging economies. Dividend paying firms have stronger shareholder protection (measured by lower management corruption and higher property rights protection), higher government shareholding, and operate in provinces characterized by higher institutional investor presence.

Relative to cash dividends, firms paying stock dividends are older, growing, more profitable, have more retained earnings and higher private ownership prior to listing, and invest more in fixed assets. The number of years between a firm’s first seasoned equity offering and its IPO (SEOAGE) is 3.26 years for stock dividend payers and 3.31 years for cash dividend payers, and the difference is not significant either economically or statistically.

To investigate whether our selected firm characteristic variables are likely to be subject to multicollinearity problems in our later regression analysis, we examine the correlations between the independent variables that are used in our analysis and report the results in Table 3. We find that these variables are generally not highly correlated with each other. The exception is the correlation between profitability (ROA) and retained earnings per share (REPS), non-tradable (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 number of years between a firm’s first seasoned equity offering and its IPO (SEOAGE).

Table 3.

Correlation matrix of variables.This table presents Pearson correlation of coefficients among the explanatory variables. Private is a dummy variable and equals one if a firm is not a state-owned firm prior to listing and zero otherwise; Cross-list is a dummy variable and equals one if a firm has cross-listed shares in either B-share markets in China or other overseas markets and zero otherwise; # of new offers is the total number of successful new issues conducted within the next three years after dividend announcement. MCAR is the market-adjusted cumulative abnormal return from t = −5 to t = 10, where t = 0 is the announcement date of the proposed dividend distribution. All other variables are as defined in Tables 1 and 2.

	Log (TA)	ROA	DE	Cash/TA	FA growth rate	NT/TS	GV/TS	CSC/TA	Log (Age)	REPS	Private	ETC	PP	Cross-list	# of new offers	InstHolding	Log (SEOAGE)	Split	MCAR
Log (TA)	1.00
ROA	0.19	1.00
DE	0.00	−0.18	1.00
Cash/TA	−0.08	0.18	−0.09	1.00
FA growth rate	0.13	0.08	−0.01	−0.08	1.00
NT/TS	−0.11	0.06	−0.04	0.02	0.01	1.00
GV/TS	0.20	0.08	−0.05	0.00	0.01	0.38	1.00
CSC/TA	0.13	0.25	−0.03	−0.15	0.18	0.04	0.03	1.00
Log(Age)	0.16	−0.14	0.09	−0.13	−0.12	−0.31	−0.13	−0.26	1.00
REPS	0.34	0.52	−0.21	0.10	0.17	−0.01	0.08	0.25	−0.19	1.00
Private	−0.27	−0.09	0.08	0.00	−0.01	−0.07	−0.55	−0.06	−0.04	−0.12	1.00
ETC	−0.06	−0.11	0.01	−0.07	0.00	−0.02	0.00	−0.06	0.01	−0.08	0.04	1.00
PP	−0.05	0.02	0.00	−0.05	0.01	0.03	−0.08	0.04	−0.10	0.05	0.10	−0.40	1.00
Cross-list	0.23	0.02	0.00	0.02	−0.04	−0.17	0.01	−0.04	0.19	−0.07	−0.08	−0.07	−0.12	1.00
# of new offers	0.15	0.20	0.01	−0.09	0.10	0.01	−0.02	0.16	−0.08	0.19	0.00	−0.02	−0.03	0.00	1.00
InstHolding	0.09	0.05	−0.04	0.18	0.00	0.05	0.06	−0.02	0.07	0.02	−0.11	−0.15	−0.39	0.12	0.01	1.00
Log(SEOAGE)	−0.02	0.11	0.00	−0.06	0.08	0.10	0.06	0.11	−0.34	0.11	0.00	0.03	−0.03	0.15	0.30	−0.05	1.00
Split	0.05	0.14	−0.01	0.03	0.06	−0.01	−0.02	0.08	−0.03	0.14	0.01	−0.01	−0.01	−0.01	0.06	−0.02	0.02	1.00
MCAR	0.08	0.08	−0.02	−0.01	−0.01	−0.11	−0.02	0.00	0.09	0.06	−0.04	0.00	0.00	−0.01	0.05	−0.02	0.00	0.09	1.00

5.2. Factors affecting dividend payment and types of dividend

We use a logistic regression 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 bull-market dummy variable that takes the value of one if the announcement date of a proposed dividend plan falls into a bull market, where the classification of bull and bear markets follows Yan et al. (2007), and industry dummy variables for 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 the number of years passed since the IPO and before the first SEO (SEOAGE).

Consistent with our expectations, the results in Columns 1 and 2 of Table 4 show that younger, larger, and more profitable firms are more likely to pay dividends. Further, 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. Contrary to the findings in the USA and other markets that mature firms are more likely to pay dividends, the negative coefficient on age combined with the positive coefficient on FA growth rate and on SEOAGE indicates that young and growing firms establish a dividend payment history to conform to the CSRC’s requirement for external equity financing.

Table 4.

Logistic regression on the likelihood of firms paying dividends.The table reports the results of logistic regression on the likelihood of firms paying dividends. Columns 1 and 2 examine the likelihood of firms paying dividends (i.e. any form of dividend payments) for the whole sample. The dependent variable equals one for dividend paying firms and zero otherwise. Columns 3 and 4 examine the likelihood of firms paying stock dividends using the sub-sample of stock-dividend-paying firms and non-payers. The dependent variable equals one for stock-dividend-paying firms and zero otherwise. Columns 5 and 6 estimate the likelihood of firms paying cash dividends using the sub-sample of cash-dividend-paying firms and firms paying no dividend (non-payers). The dependent variable equals one for cash-dividend-paying firms and zero otherwise. Columns 7 and 8 examine the likelihood of firms paying stock dividends using the sample of dividend payers. The dependent variable equals one for stock-dividend-paying firms and zero for cash-dividend-paying firms. All other variables are as defined in Tables 1 –3. All regressions include dummy variables for bull markets (the classification of bull and bear markets follows Yan et al. (2007) and industries (finance, utilities, properties, conglomerate, and industrial). Figures in parenthesis are Wald chi-squares. ***, ** and * denote significance at the 1%, 5% and 10% levels, respectively.

	Payers vs. Non-Payers		Stock vs. Non-Payers		Cash vs. Non-Payers		Stock vs. Cash
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Intercept	−13.15	−11.95	−9.70	−10.12	−13.98	−12.25	4.39	3.43
	(108.44)***	(55.06)***	(16.47)***	(15.46)***	(115.64)***	(54.87)***	(3.94)**	(2.11)
Log (TA)	0.62	0.50	0.38	0.38	0.65	0.52	−0.31	−0.34
	(115.40)***	(43.83)***	(12.05)***	(10.20)***	(120.33)***	(43.94)***	(9.45)***	(9.23)***

ROA	37.27	35.58	31.24	32.12	36.03	34.10	12.76	13.63
	(390.01)***	(213.21)***	(136.07)***	(116.57)***	(354.69)***	(192.66)***	(30.12)***	(30.07)***

DE	−0.38	−0.36	0.02	0.02	−0.45	−0.44	0.74	0.72
	(25.50)***	(14.84)***	(0.17)	(0.17)	(33.20)***	(19.63)***	(28.39)***	(22.42)***
Cash/TA	1.26	1.00	0.59	0.48	1.24	0.95	0.31	0.31
	(17.49)***	(5.95)**	(0.91)	(0.48)	(16.34)***	(5.22)**	(0.31)	(0.26)
FA growth rate	0.23	0.34	0.63	0.58	0.16	0.30	0.50	0.42
	(3.02)*	(3.50)*	(6.95)***	(4.13)**	(1.36)	(2.57)	(6.53)**	(3.28)*
NT/TS	0.88		−1.57	−2.13	1.26		−2.60	−2.60
	(7.94)***		(7.10)***	(10.63)***	(14.85)***		(22.13)***	(18.90)***
GV/TS		0.71				0.86
		(10.46)***				(14.79)***
CSC/TA	2.60	3.42	4.35	4.42	2.40	3.33	1.88	2.36
	(15.00)***	(14.73)***	(14.94)***	(11.74)***	(11.96)***	(13.51)***	(3.62)*	(4.72)**
Log(Age)	−0.76		−0.23	−0.27	−0.83		0.36	0.45
	(90.93)***		(1.94)	(2.13)	(103.05)***		(7.58)***	(9.72)***
REPS	1.04	1.02	1.15	1.12	0.95	0.93	0.81	0.66
	(100.01)***	(62.63)***	(43.32)***	(34.74)***	(80.98)***	(51.19)***	(23.67)***	(13.56)***
Private	−0.61		−0.01	0.03	−0.68		0.59	0.59
	(47.87)***		(0.00)	(0.03)	(55.03)***		(14.56)***	(11.93)***
ETC	−17.52	−29.86	−24.71	−15.90	−16.10	−30.67	−8.15	−12.86
	(5.02)**	(9.35)***	(2.24)	(0.86)	(4.05)**	(9.46)***	(0.30)	(0.77)
PP	0.83		−0.79		1.05		−1.70
	(8.94)***		(1.80)		(13.43)***		(10.64)***

Cross-list	0.01	0.02	−0.43	−0.31	0.12	0.07	−0.64	−0.46
	(0.01)	(0.01)	(1.36)	(0.64)	(0.58)	(0.08)	(4.56)**	(2.02)
# of new offers	0.14		0.01	0.04	0.15		−0.21	−0.19
	(3.15)*		(0.01)	(0.05)	(3.18)*		(3.06)*	(2.15)
InstHolding		1.24		2.42		1.11		−0.76
		(3.93)**		(6.77)***		(3.19)*		(0.98)
PrivateGV/TS*		−0.96				−0.90
		(3.35)*				(2.89)*
Log(SEOAGE)		0.29				0.32
		(9.25)***				(10.33)***

We find that corporate governance, shareholder protection, and institutional monitoring influence the likelihood of paying dividends. Consistent with our expectations, firms with lower management corruption and higher property rights protection are more likely to pay dividends, in particular cash dividends. Higher provincial institutional shareholdings are associated with a higher likelihood of paying a dividend, both stock and cash. The cross-listing dummy variable, however, does not appear to influence the likelihood of paying dividends.

Consistent with the earlier results of Wei et al. (2003), we find that although firms that were privately owned prior to listing are less likely to pay a dividend, they are more likely to pay a stock dividend if they choose to pay dividends. Firms with higher levels of non-tradable shares are more likely to pay cash dividends. This result is consistent with our argument that firms cater to the demands of non-tradable shareholders in setting their dividend policy.

With particular reference to the choice to pay stock dividends, the results in Columns 3 and 4 show that larger firms that are more profitable have higher levels of retained earnings and higher growth, and invest more in fixed assets, are more likely to choose to pay a stock dividend rather than forgoing a dividend. In addition, compared with cash payers, stock payers are smaller and make more use of debt financing (see Columns 7 and 8 of Table 4). These results are consistent with the argument that higher growth firms, with higher retained earnings and greater investment in fixed assets, tend to substitute stock dividends for cash dividends to meet the dividend history requirement for external financing while allowing for the retention of internal funds for ongoing investment. However, contrary to expectations, we find that the likelihood of paying a stock dividend is only weakly influenced by subsequent equity capital raisings. Further, to the extent that these high-growth firms are characterized as high information asymmetry, these results are consistent with the argument that these firms use stock dividends to attract the attention of analysts (Almazan et al., 2008; Grinblatt et al., 1984).

Finally, the choice of a stock dividend rather than a cash dividend is less likely for firms with a high proportion of non-tradable shares and state ownership prior to listing and for firms operating in regions with high protection of property rights, that is, high shareholder protection.

5.3. Announcement effect

With respect to the influence of firm characteristics on the market reaction to the announcement of the form and level of the dividend payment, Table 5 reports the cross-sectional analysis of dividend announcement returns, measured as the market-adjusted cumulative abnormal return from t = −5 to t = 10 days, where t = 0 is the announcement day of the proposed PDS. Columns 1 and 2 of Table 5 report the difference in market reactions between dividend payers and non-payers. A dummy variable, Payer, is included, and it takes a value of one if a firm pays any form of 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 and zero otherwise. As shown in Columns 1 and 2, the overall market reaction to the dividend scheme 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). In addition, we find no association between debt/equity ratio, growth, cash holding, fixed assets investment, and the market reaction to the dividend scheme announcement. However, we find that the larger the firm, the higher the ROA, and the stronger the market reaction around the announcement.

Table 5.

Dividend Announcement Returns.This table reports cross-sectional regression on market reaction to the dividend announcement. The dependent variable is market-adjusted cumulative abnormal return (MCAR) from t=−5 to t=10, where t=0 is the announcement date of the proposed profit distribution scheme. Columns 1 and 2 report the MCAR regressions for the whole sample, Payer is a dummy variable and equals one for dividend-paying firms and zero otherwise; ∆DProfit is a dummy variable and equals one if the change in a firm’s distributable profit is positive and zero otherwise. Columns 3 and 4 report the MCAR regressions for the sub-sample of dividend paying firms. Stock is a dummy variable and equals one for stock dividend payers and zero for cash dividend payers. Columns 5, 6 and 7 report market reaction to cash dividend initiation, continuation and change, respectively. Columns 8, 9 and 10 report market reaction to stock dividend initiation, continuation and change. Continue is a dummy variable that equals one for continuing stock dividend payers from last period and zero otherwise; Initiate equals one for firms that initiated stock dividend (i.e. no dividend was paid last period) and zero otherwise; Changetocash equals one for payers that switch from stock dividends in the last period to cash dividends in this period and zero otherwise. Changetostock equals one for payers that switch from cash dividends in the last period to stock dividends in this period and zero otherwise. All other variables are as defined in Tables 1 to 3. All regressions include dummy variables for bull markets and industries (finance, utilities, properties, conglomerate, and industrial). T-statistics are based on White’s Heteroskedasticity-robust standard errors. ***, ** and * denote statistical significance at the 1%, 5% and 10% level, respectively.

	Payers vs. Non-payers				Stock vs. Cash Dividend Payers
	(1)		(2)		(3)		(4)
	Coeff	t-stat	Coeff	t-stat	Coeff	t-stat	Coeff	t-stat
Intercept	−20.38	−3.98***	−20.43	−2.89***	−30.01	−4.84***	−14.91	−2.14**
Split	7.94	5.83***	7.28	3.85***	8.60	7.03***
Log (TA)	0.58	2.62***	0.59	1.84*	0.96	3.46***	0.92	2.75***
ROA	14.71	3.66***	13.30	2.18**	6.13	0.79	−2.32	−0.25
DE	−0.11	−1.52	−0.07	−0.86	−0.88	−1.91*	−1.07	−1.97**
Cash/TA	0.56	0.45	2.10	1.13	0.60	0.38	1.99	1.05
FA growth rate	−0.63	−1.19	−0.28	−0.33	−0.54	−0.92	−0.97	−1.33
NT/TS	−6.96	−5.18***			−4.73	−2.76***	−4.90	−2.44**
GV/TS			−1.31	−1.28
CSC/TA	−1.62	−0.59	−3.16	−0.79	−0.95	−0.31	0.87	0.23
Log(Age)	0.79	2.39**			1.00	2.83***	0.75	1.74*
REPS	−0.46	−1.42	−0.21	−0.46	−2.02	−3.67***	−1.78	−2.77***
Private	−0.59	−1.59			0.19	0.39	0.30	0.50
ETC	25.20	0.77	32.14	0.71	41.65	0.97	27.18	0.58
PP	0.76	0.63			1.22	0.84
Cross-list	−1.67	−2.73***	−0.46	−0.49	−1.77	−2.29**	−1.98	−2.12**
# of new offers	0.42	1.28			0.91	2.56**	0.81	2.00**
InstHolding			−4.70	−1.66*			−6.10	−2.45**
PrivateGV/TS*			0.59	0.26
Log(SEOAGE)			−0.24	−0.54
∆DProfit	0.45	1.21	1.15	2.23**	1.35	3.07***	1.64	3.15***
Payer	0.06	0.15	−0.34	−0.59
Stock							3.25	4.22***
Adj. R²	6.55%		3.98%		6.80%		4.94%

	Cash Dividend Payers Only
	(5)		(6)		(7)
	Coeff.	t-stat	Coeff.	t-stat	Coeff.	t-stat
Intercept	−25.85	−3.85***	−24.58	−3.67***	−24.74	−3.69***
Log (TA)	1.14	3.71***	1.09	3.56***	1.09	3.55***
ROA	14.34	1.68*	12.52	1.47	14.01	1.63
DE	−0.88	−1.70*	−0.89	−1.72*	−0.84	−1.63
Cash/TA	0.67	0.39	0.55	0.32	0.57	0.34
FA growth rate	−1.10	−1.52	−1.17	−1.61	−1.08	−1.49
NT/TS	−4.63	−2.43**	−4.54	−2.38**	−4.70	−2.47**
CSC/TA	−0.80	−0.23	−1.01	−0.29	−0.94	−0.27
Log(Age)	0.90	2.20**	1.02	2.50**	1.05	2.60***
REPS	−1.86	−3.01***	−1.83	−2.96***	−1.87	−3.02***
Private	0.05	0.09	0.03	0.05	0.07	0.12
ETC	58.36	1.24	60.27	1.28	57.72	1.23
PP	2.22	1.41	2.31	1.47	2.28	1.45
Cross-list	−1.52	−1.84*	−1.54	−1.86*	−1.62	−1.95*
# of new offers	1.13	2.91***	1.16	2.97***	1.14	2.94***
Initiate	1.33	2.00**
Continue			−0.41	−0.84
Changetocash					−0.49	−0.80
Adj. R²	4.58%		4.44%		4.43%

	Stock Dividend Payers Only
	(8)		(9)		(10)
	Coeff.	t-stat	Coeff.	t-stat	Coeff.	t-stat
Intercept	−6.79	−0.36	−2.81	−0.15	−2.05	−0.11
Split	10.84	4.78***	10.83	4.69***	11.24	4.91***
Log (TA)	−0.06	−0.06	−0.24	−0.27	−0.27	−0.31
ROA	5.76	0.27	4.02	0.19	−0.45	−0.02
DE	−0.96	−0.73	−0.66	−0.49	−0.96	−0.73
Cash/TA	0.79	0.15	0.34	0.06	0.94	0.18
FA growth rate	0.77	0.38	0.69	0.33	0.40	0.19
NT/TS	−4.98	−1.00	−5.48	−1.08	−4.26	−0.85
CSC/TA	−1.28	−0.14	−1.53	−0.17	−2.57	−0.29
Log(Age)	1.56	1.12	1.93	1.40	1.88	1.36
REPS	−1.04	−0.65	−1.18	−0.73	−1.04	−0.64
Private	1.36	0.89	1.68	1.01	1.48	0.97
ETC	−200.29	−1.30	−156.36	−1.02	−173.94	−1.13
PP	−7.34	−1.37	−7.79	−1.45	−7.59	−1.41
Cross-list	−5.56	−1.86*	−5.94	−1.96*	−5.31	−1.75*
# of new offers	−0.37	−0.33	−0.34	−0.31	−0.40	−0.35
Initiate	3.60	2.05**			−1.67	−1.18
Continue			−0.99	−0.56
Changetostock					1.84	1.34
Adj. R²	8.46%		6.94%		7.36%

Columns 3 and 4 investigate the market reactions to the payment of a dividend for the sub-sample of payers, which pay either cash or stock dividend. We find that the larger the firm, the lower the debt/equity ratio, the stronger the market reaction to the announcement of a dividend payment. In addition, the higher the proportion of non-tradable shares, the smaller the market reaction to the announcement of a dividend payment. Consistent with the argument that the payment of a dividend signals the firm’s financing choice, we find that the market reaction to the dividend announcement is positively related to subsequent equity capital raisings. In addition, we find no association between profitability (ROA) and the market reaction to the announcement of a dividend payment.

We compare the difference in market reactions between stock and cash dividends by including a dummy variable, Stock, which takes a value of one if a firm pays only a stock dividend and zero for cash dividends (Column 4). We find that the coefficient on Stock is consistently positive and significant at the 1% level, suggesting the announcement effect for stock dividends is significantly greater than that for cash dividends.

Columns 5–7 of Table 5 show the market reaction to the announcement of a cash dividend. We find that the higher the proportion of non-tradable shares, the smaller the market reaction to the announcement of a dividend payment. This is consistent with firms paying cash dividends as a way to compensate and cater to the demands of non-tradable shareholders, where this action by management is not seen as in the interest of public investors.

We compare the difference in market reactions to cash dividend payers that initiate a cash dividend, continue to pay a cash dividend, or change to a cash dividend, respectively. Three dummy variables are included in separate regressions: Continue takes a value of one for firms continuing to pay a cash dividend from last period and zero otherwise; Initiate equals one for newly initiated cash dividend payers (i.e. no dividend was paid in the last period) and zero otherwise; and Changetocash equals one for payers that switch from stock dividends in the last period to cash dividends in the current period and zero otherwise. Consistent with the existing literature (Michaely et al., 1995), the market response to the initiation of a cash dividend is strongly positive and significant. In contrast, the market response is insignificantly different from zero for firms continuing to pay cash dividends or choosing to change from stock to cash dividends. This result is consistent with our conjecture that the initiation of a cash dividend signals the firm’s financing choice to the market, whereas the continuation of a dividend does not serve the same signaling role.

Similarly, Columns 8–10 of Table 5 investigate the differences in market reactions to stock dividend payers that initiate a stock dividend, continue to pay a stock dividend, or change to a stock dividend. 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 the firm using the initiation of a stock dividend to differentiate itself in a market characterized by opaque information disclosure (see Bai et al., 2004; Lin et al., 1998), whereas the continuation of a dividend does not serve the same signaling role. Likewise, this result is also consistent with the initiation of a stock dividend to signal the firm’s financing choice to the market. In addition, the market response to dividend initiation is stronger for stock dividends than for cash dividends.

In addition, consistent with management making use of stock dividends to keep the firm’s stock price within its acceptable trading range, we find that the coefficient on the split factor is positive and significantly different from zero in all regressions.

Finally, taken with the observation from Column 7 that the market response is insignificantly different from zero for firms changing from stock to cash dividends, the result that the market response is insignificantly different from zero for firms changing from cash to stock dividends is consistent with the argument that following the initiation of a stock dividend firms consider stock dividends as a substitute for cash dividends, and that switching between them has no signaling role.

Table 6 investigates the differences in analyst coverage prior to and post the dividend announcement.¹⁶ Analyst coverage is defined as the number of IBES analysts making one-year forecasts on the firm. The data is obtained from I/B/E/S International for the period 2003–2007. The results show that the mean analyst coverage increases after the announcement of a stock dividend from 1.49 to 2.15, and the increase is statistically significant at the 5% level. Further, this increase is statistically significant only after the initiation of a stock dividend. These results are consistent with our argument that the initiation of a stock dividend allows management to disclose information allowing the market to verify the quality of the firm’s investments by analysts.

Table 6.

Descriptive statistics on analysts’ coverage.This table reports the mean number of analysts following prior to and post the dividend announcement. Analyst coverage is defined as the number of I/B/E/S analysts making one-year forecasts on the firm. The data is obtained from IBES International for the period 2003–2007. All other variables are as defined in Table 5. ***, ** and * denote statistical significance at the 1%, 5% and 10% levels, respectively.

	Number of observations	Mean analyst coverage		t-statistic (Post – Prior)
	Number of observations	Prior to dividend announcement	Post dividend announcement	t-statistic (Post – Prior)
Non-Payers	776	2.73	2.52	−0.66
Cash Dividend Payers	1430	1.95	2.16	1.56
Stock Dividend Payers	175	1.49	2.15	2.22**
- Initiate	26	0.40	1.35	2.55**
- Continue	38	2.00	2.87	1.15

Overall, these results are consistent with stock dividends substituting for stock splits. The results suggest that management choose to disclose information via stock dividends as a form of communication between firms and investors (Almazan et al., 2008) and use stock dividends to manage liquidity.

5.4. Fixed effects and cross-sectional estimates

This section examines the extent to which the cross-sectional and time-series variations in our explanatory variables drive 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 the 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 7 reports the result on year-fixed-effect regressions on both the likelihood of paying a dividend (Columns 1 and 2) and the choice between stock and cash dividends (Columns 3 and 4).

Table 7.

Logistic regression on the likelihood of firms paying dividends with year-fixed effect.This table reports the logistic regression on the likelihood of firms paying dividends with year-fixed effect. Columns 1 and 2 report the likelihood of firms paying dividends using the whole sample. The dependent variable equals one for dividend-paying firms and zero for firms paying no dividend. Columns 3 and 4 investigate the likelihood of firms paying stock dividends using the sub-sample of dividend payers. The dependent variable equals one for stock-dividend-paying firms and zero for cash-dividend-paying firms. All other variables are as defined in Tables 1 –4. All regressions include dummy variables for years, bull markets, and industries (finance, utilities, properties, conglomerate, and industrial). ***, ** and * denote statistical significance at the 1%, 5% and 10% levels, respectively.

	Payers vs. Non-Payers				Stock vs. Cash Dividend Payers
	(1)		(2)		(3)		(4)
	Coeff.	Wald ChiSq	Coeff.	Wald ChiSq	Coeff.	Wald ChiSq	Coeff.	Wald ChiSq
Intercept	−13.33	110.04***	−12.97	60.35***	4.72	4.38**	3.85	2.49
Log (TA)	0.63	117.35***	0.52	45.30***	−0.32	9.68***	−0.36	9.59***
ROA	36.91	377.34***	35.05	204.98***	13.74	33.56***	14.69	33.50***
DE	−0.38	25.79***	−0.36	14.87***	0.80	32.17***	0.79	25.77***
Cash/TA	1.33	19.24***	1.06	6.62**	0.13	0.06	0.08	0.02
FA growth rate	0.23	2.98*	0.36	3.74*	0.48	5.66**	0.40	2.86*
NT/TS	0.96	8.80***			−2.90	25.15***	−2.95	21.94***
GV/TS			0.67	9.21***
CSC/TA	2.61	14.94***	3.24	12.97***	1.58	2.48	2.03	3.36*
Log(Age)	−0.79	94.11***			0.40	9.11***	0.50	11.37***
REPS	1.03	96.18***	1.01	61.06***	0.93	30.19***	0.78	18.34***
Private	−0.60	46.37***			0.59	14.29***	0.61	12.70***
ETC	−17.01	4.70**	−29.58	9.09***	−7.65	0.26	−13.11	0.78
PP	0.84	8.93***			−1.63	9.47***
Cross-list	0.05	0.08	0.04	0.03	−0.73	5.81**	−0.57	2.93*
# of new offers	0.13	2.65			−0.18	2.10	−0.17	1.64
InstHolding			1.19	3.59*			0.81	1.07
PrivateGV/TS*			−0.97	3.40*
Log(SEOAGE)			0.30	9.45***

The regression estimates reported in Table 7 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 4.¹⁷

Likewise, to estimate the extent to which our announcement effect results reported in Table 5 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 dividend payers sub-sample in Table 8. The regression estimates reported in Table 8 indicate that the relationships between dividend announcement effects and firm characteristics are consistent with our earlier estimates (reported in Table 5) in both the cross-section and time-series regression.

Table 8.

Fama–MacBeth and year-fixed effect regression on the dividend announcement returns.This table reports the results of Fama–MacBeth and year-fixed effect regression on the dividend announcement returns. The dependent variable is market-adjusted cumulative abnormal return (MCAR) from t = −5 to t = 10, where t = 0 is the announcement date of the proposed profit distribution scheme. Panel A reports the MCAR regressions for the whole sample using the Fama–MacBeth (1973) two-stage approach (Columns 1 and 2) and the year-fixed effect model (Columns 3 and 4). Panel B reports the MCAR regressions for the payers sub-sample. The results based on the Fama–MacBeth two-stage approach are reported in Columns 5 and 6, whereas the year-fixed effect results are reported in Columns 7 and 8. Payer is a dummy variable and equals one for dividend-paying firms and zero otherwise; Stock is a dummy variable and equals one for stock dividend payers and zero for cash dividend payers. All other variables are as defined in Tables 1 –3 and 5. All regressions include dummy variables for bull markets and industries (finance, utilities, properties, conglomerate, and industrial). ***, ** and * denote statistical significance at the 1%, 5% and 10% levels, respectively.

Panel A: Payers vs. Non-Payers, whole sample
	Fama–MacBeth				Year-fixed effect
	(1)		(2)		(3)		(4)
	Coeff.	t-stat	Coeff.	t-stat	Coeff.	t-stat	Coeff.	t-stat
Intercept	−22.36	−3.28**	−11.18	−1.70	−21.96	−4.36***	−5.74	−0.80
Split	8.73	14.85***	7.87	22.93***	7.85	5.86***	7.09	3.82***
Log (TA)	0.62	2.35*	0.25	1.10	0.50	2.31 **	0.23	0.72
ROA	5.96	0.68	0.89	0.07	12.79	3.22***	11.88	1.98**
DE	−0.24	−3.24**	−0.34	−5.04**	−0.13	−1.76*	−0.09	−1.10
Cash/TA	1.08	1.01	1.97	1.60	1.08	0.88	2.14	1.18
FA growth rate	−0.45	−0.93	0.46	1.58	−0.34	−0.65	0.42	0.51
NT/TS	−2.24	−3.94**			−2.51	−1.85*
GV/TS			−0.41	−0.91			−0.14	−0.14
CSC/TA	−0.73	−0.54	−2.27	−0.84	0.35	0.13	−0.15	−0.04
Log(Age)	0.68	2.64**			0.81	2.49**
REPS	−0.23	−1.09	−0.04	−0.50	−0.45	−1.41	−0.18	−0.41
Private	−0.47	−1.89			−0.62	−1.69*
ETC	27.45	1.41	40.72	1.44	30.47	0.95	41.02	0.93
PP	0.67	1.37			0.72	0.61
Cross-list	−1.17	−2.32*	−0.10	−0.26	−1.25	−2.08**	−0.04	−0.05
# of new offers	0.91	3.80**			0.81	2.48**
InstHolding			−2.81	2.20			−3.81	−1.38
PrivateGV/TS*			1.00	0.88			1.83	0.81
Log(SEOAGE)			−0.29	−1.47			−0.39	−0.91
∆DProfit	0.32	2.35**	0.68	7.56***	−0.21	−0.53	0.51	1.01
Payer	0.05	0.12	−0.02	−0.08	0.13	0.34	−0.12	−0.22
Adj. R²	5.90%		3.00%		10.02%		8.27%

Panel B: Stock Dividend Payers vs. Cash Dividend Payers
	Fama–MacBeth				Year-fixed effect
	(5)		(6)		(7)		(8)
	Coeff.	t-stat	Coeff.	t-stat	Coeff.	t-stat	Coeff.	t-stat
Intercept	−28.49	−6.80***	−11.27	−2.08	−30.08	−4.87***	−8.99	−1.28
Split	9.12	16.62***			8.71	7.15***
Log (TA)	0.81	4.82***	0.71	2.86*	0.85	3.10***	0.79	2.39**
ROA	6.04	0.46	−0.68	−0.04	5.50	0.71	−2.78	−0.30
DE	−0.66	−2.55*	−0.74	−2.22	−0.91	−1.99*	−1.10	−2.03**
Cash/TA	1.24	1.09	2.40	2.37*	0.90	0.58	2.31	1.23
FA growth rate	−0.68	−1.10	−0.85	−2.11	−0.43	−0.73	−0.85	−1.18
NT/TS	−0.67	−1.96	−0.35	−0.52	−1.33	−0.76	−0.91	−0.44
CSC/TA	0.21	0.13	1.82	0.97	0.47	0.15	2.53	0.69
Log(Age)	0.86	2.74*	0.59	1.38	1.02	2.90***	0.75	1.76*
REPS	−1.89	−3.17**	−1.81	−2.34	−2.32	−4.18***	−2.11	−3.30***
Private	0.12	0.72	0.21	1.18	0.19	0.39	0.28	0.48
ETC	55.19	2.07*	35.59	1.08	47.82	1.12	35.48	0.76
PP	1.39	1.53			1.12	0.78
Cross-list	−1.21	−2.06	−1.43	−1.98	−1.25	−1.61	−1.37	−1.48
# of new offers	1.16	2.76*	1.04	1.59	1.20	3.38***	1.16	2.86***
InstHolding			−4.95	−2.13			−6.10	−2.48**
∆DProfit	0.93	3.84**	1.28	4.90**	0.92	1.94*	1.28	2.46**
Stock			−3.17	−8.93***			3.28	4.28***
Adj. R²	8.20%		5.71%		9.84%		7.27%

6. Conclusion

This paper examines why firms choose to pay 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 paying dividends. In particular, we find that older firms that are more profitable, have lower leverage, higher levels of retained earnings, and private ownership prior to listing, invest more in fixed assets, and operate in regions with weaker corporate governance and shareholder protection are more likely to pay stock dividends.

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. Further, for those firms initiating a stock dividend, the number of analysts 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 differentiate itself in a market characterized by opaque information disclosure, attracting the attention of analysts. In addition, we find evidence consistent with management making use of stock dividends to manage the liquidity of the firm’s stock price.

Overall, these results are consistent with stock dividends substituting for stock splits where management choose to disclose information via the initiation of stock dividends as a form of communication between firms and investors (Almazan et al., 2008), while allowing for the retention of internal funds for ongoing investments.

Footnotes

Acknowledgements

This paper has benefited from the useful comments and suggestions provided by Joseph Fan, Doug Foster, Jay Hartzell, Phong Ngo, Rik Sen, Tom Smith, Vivian Tai, Dogan Tirtiroglu, Sheridan Titman, Takeshi Yamada, participants at the 2009 Journal of Corporate Finance Special Conference on Emerging Markets, the 2011 Asian Finance Association Conference, the 2011 China International Finance Conference, and seminar participants at the University of Adelaide, Australian National University, and University of Melbourne. We are also appreciative of the research assistance provided by Elizabeth Zhu, Meifen Qian, and Bin Yu. An earlier version of this paper was titled “Stock versus cash: signaling or catering”.

Final transcript accepted 24 August 2014 by Tom Smith (AE Finance).

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

1.

A growing literature in both developed and emerging markets has sought to determine the link between firm characteristics, institutional structures, and dividend policy (Aivazian et al., 2003; Glen et al., 1995; La Porta et al., 2000; Mitton, 2004). Examples of empirical studies within specific countries include Kao and Chen (2013) using Taiwanese data and He (2012) utilizing a Japanese sample to investigate the relationship between product market competition and dividend policies. They both find that firms in more competitive industries pay more dividends and are more likely to increase dividends. By considering the institutional differences between tradable and non-tradable shares in China, Cheng et al. (2009) highlight that Chinese retail investors favor the stock dividend and that the cash dividend is preferred by non-tradable shareholders. In the context of the Australian market, Coulton and Ruddock (2011) find that dividend paying firms are larger, more profitable, and have less growth options than non-dividend paying firms. study long-run IPO performance and observe that Australian firms that initiated a dividend perform better than non-dividend payers.

2.

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 based on net asset value, not the market price (see Calomiris et al., 2010; Xiao, 2005). After China’s stock market reform was initiated in 2005 (where most firms completed their reforms by the end of 2007), non-tradable shares in firms were gradually converted to tradable shares over the next three years.

3.

Given the similarity in the market responses to the announcement of stock dividends and the combined cash and stock dividends, they are categorized into one group as stock dividend payers for the remainder of the paper.

4.

At present, shareholders are levied a 20% tax for both cash and stock dividends received; a policy that has remained unchanged since 1997. Please also note that capital gains in China are tax exempt.

5.

Capital reserve funds consist of accumulated donations, the difference in capital conversion raised from foreign currency transactions and, more importantly, premium on capital (i.e. the excess of the amount contributed by investors over the amount of the registered capital), etc.

6.

When the CRF comes from surplus reserve funds, the OSRF has to be used before 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.

7.

Given the fact that Chinese investors are levied a 20% 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.

8.

We recognize the use of this measure is confounded by the observation that the size of the split factor may itself be a signal. argue that managers use their private information about the firm to set the split factor, and this allows investors to make inferences about the managers’ private information from the observed split factor.

9.

We use 2003–2007 as our sample period to avoid the confounding effects of this regulatory change.

10.

We recognize the look-ahead bias introduced by our choice of the total number of successful new issues conducted within three years after the dividend announcement to proxy for the firm’s external equity financing and mitigate our conclusions accordingly.

11.

This relationship holds for our sample period.

12.

As a robustness check, we also examine our sample using individual firm’s institutional holdings and our conclusions remain unchanged.

13.

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 report of 2007, for example, are actually the dividend payment for the financial year of 2006.

14.

The sample starts in 2003 since detailed information on dividend announcements is not available prior to 2003. As previously discussed, we end the sample in 2007 to avoid the confounding effects of the regulatory changes associated with the form of the dividend history requirement that occurred in 2007.

15.

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.

16.

As a result of limitation on the availability, analyst coverage is only available for 775 firms and equivalent to 2381 firm-year observations.

17.

With the inclusion of year-fixed effects, the continued positive and significant coefficient for non-tradable shares suggests that this result is not mechanistically driven by the split share structure reform introduced in 2005.

References

Aboody

Lev

(2000) Information asymmetry, R&D, and insider gains. Journal of Finance 55: 2747–2766.

Aivazian

Booth

Cleary

(2003) Do emerging market firms follow different dividend policies from U.S. firms? Journal of Financial Research 26: 371–387.

Almazan

Banerjee

Motta

(2008) Attracting attention: Cheap managerial talk and costly market monitoring. Journal of Finance 63: 1399–1436.

Bai

Liu

. (2004) Corporate governance and market valuation in China. Journal of Comparative Economics 32: 599–616.

Baker

Wurgler

(2004) A catering theory of dividends. Journal of Finance 59: 1125–1165.

Bhattacharya

(1979) Imperfect information, dividend policy, and the “bird-in-hand” fallacy. Bell Journal of Economics 10: 259–270.

Brennan

Hughes

(1991) Stock prices and the supply of information. Journal of Finance 46: 1665–1691.

Cai

Fang

(2011) Eat, drink, firms and government: An investigation of corruption from entertainment expenditures of Chinese firms. Journal of Law and Economics 54: 55–78.

Calomiris

Fisman

Wang

(2010) Profiting from government stakes in a command economy: Evidence from Chinese asset sales. Journal of Financial Economics 96: 399–412.

10.

Chen

Jian

(2009) Dividends for tunneling in a regulated economy: The case of China. Pacific-Basin Finance Journal 17: 209–223.

11.

Cheng

Fung

Leung

(2009) Dividend preference of tradable-share and non-tradable-share holders in Mainland China. Accounting and Finance 49: 291–316.

12.

Copeland

(1979) Liquidity changes following stock splits. Journal of Finance 34: 115–141.

13.

Coulton

Ruddock

(2011) Corporate payout policy in Australia and a test of life-cycle theory. Accounting and Finance 51: 381–407.

14.

Djankov

La Porta

Lopez-de-Silanes

. (2003) Courts. Quarterly Journal of Economics 118: 453–517.

15.

Easterbrook

(1984) Two agency cost explanations of dividends. American Economic Review 74: 650–659.

16.

Fama

MacBeth

(1973) Risk, return, and equilibrium: Empirical tests. Journal of Political Economy 81: 607–636.

17.

Fan

JPH

Wong

Zhang

(2007) Politically connected CEOs, corporate governance and post-IPO performance of China’s newly partially privatized firms. Journal of Financial Economics 84: 330–357.

18.

Gao

Kling

(2008) Corporate governance and tunneling: Empirical evidence from China. Pacific-Basin Finance Journal 16: 591–605.

19.

Glen

Karmokolias

Miller

. (1995) Dividend policy and behavior in emerging markets: To pay or not to pay. IFC discussion paper 26. Washington, USA: International Finance Corporation (IFC).

20.

Grinblatt

Masulis

Titman

(1984) The valuation effects of stock splits and stock dividends. Journal of Financial Economics 13: 461–490.

21.

(2012) Agency problems, product market competition and dividend policies in Japan. Accounting and Finance 52: 873–901.

22.

How

Ngo

Verhoeven

(2011) Dividend initiations and long-run IPO performance. Australian Journal of Management 36: 267–286.

23.

Jensen

(1986) Agency cost of free cash flow, corporate finance, and takeover. American Economic Review 76: 323–329.

24.

John

Knyazeva

(2011) Does geography matter? Firm location and corporate payout policy. Journal of Financial Economics 101: 533–551.

25.

Kao

Chen

(2013) How product market competition affects dividend payments in a weak investor protection economy: Evidence from Taiwan. Pacific-Basin Finance Journal 25: 21–39.

26.

La Porta

Lopez-de-Silanes

Shleifer

. (2000) Agency problems and dividend policies around the world. Journal of Finance 55: 1–33.

27.

Lakonishok

Lev

(1987) Stock splits and stock dividends: Why, who and when. Journal of Finance 42: 913–932.

28.

Lin

Cai

(1998) Competition, policy burdens, and state-owned enterprise reform. American Economics Review 88: 422–427.

29.

Lin

Singh

(2009) Stock splits, trading continuity, and the cost of equity capital. Journal of Financial Economics 93: 474–489.

30.

McNichols

Dravid

(1990) Stock dividends, stock splits, and signaling. Journal of Finance 45: 857–879.

31.

Michaely

Thaler

Womack

(1995) Price reactions to dividend initiations and omissions: Overreaction or drift? Journal of Finance 50: 573–608.

32.

Miller

Rock

(1985) Dividend policy under asymmetric information. Journal of Finance 40: 1031–1051.

33.

Mitton

(2004) Corporate governance and dividend policy in emerging markets. Emerging Markets Review 5: 409–426.

34.

Sen

Ferris

(2009) Payout Policy, Ownership Concentration and Corporate Governance: Evidence from China (working paper). Singapore: Nanyang Technological University.

35.

Shao

Lin

(2004) A study on dividend policy of listed firms. Market Weekly 3: 45–47.

36.

Wei

Jiang

(2001) A survey on dividend policy in China. Economics Science 4: 79–87.

37.

Wei

Zhang

Xiao

(2003) Dividend Policy and Ownership Structure in China (working paper). Cardiff: Cardiff University.

38.

Xiao

(2005) Free cash flows, tunneling and cash dividend. Economic Science 2: 67–76.

39.

Yan

Powell

Shi

. (2007) Chinese stock market cyclical regimes: 1991–2006. Economics Letters 97: 235–239.