DIVIDENDS ON COMMON AND PREFERRED SHARES: THE RELATIONSHIP WITH THE OWNERSHIP CONCENTRATION IN RUSSIAN PUBLIC COMPANIES

Abstract

This paper investigates the relationship between the dividend policy and ownership structure in Russian public companies. A study of the link between dividends and ownership concentration is based on the sample of public companies with dual class share structure. These shares were traded on the Russian Trading System (RTS) in the period of 2003-2009. The authors explore a broad range of factors related to the ownership concentration. This study allows making conclusions on the impact of the ownership concentration on the dividend policy. Moreover, there is evidence that this impact differs for the dividends on ordinary and preferred shares.

Headnote

Abstract

This paper investigates the relationship between the dividend policy and ownership structure in Russian public companies. A study of the link between dividends and ownership concentration is based on the sample of public companies with dual class share structure. These shares were traded on the Russian Trading System (RTS) in the period of 2003-2009. The authors explore a broad range of factors related to the ownership concentration. This study allows making conclusions on the impact of the ownership concentration on the dividend policy. Moreover, there is evidence that this impact differs for the dividends on ordinary and preferred shares.

Keywords

dividends, corporate governance, public companies, dual class shares, ownership concentration

(ProQuest: ... denotes formula omitted.)

1. Introduction

The growing attention to the agency problem and to the role of dividends in mitigating an agency problem is explained, in particular, by the growing number of corporate scandals related to violation of shareholders' rights and to managers using insider information for personal enrichment. The corporate scandals, having resounded around the world during the recent decade, revealed serious drawbacks in corporate governance systems. Thus, the study of dividend policy, as a way to mitigate the agency problem and the factors affecting it, is of relevance in general and in the context of the Russian market, for example, in view of existence of such factors aggravating the agency problem as high concentration of ownership and weak legal protection of shareholders' rights.

There is a limited number of studies based on Russian companies and dealing with the problems range in question. This research, in a way filling up the gap in the studies in place, is based on data on companies having two classes of shares simultaneously traded on RTS stock exchange during 2003-2009.

2. Dividend policy, corporate governance and dual class shares structures

According to Jensen and Meckling (Jensen, Meckling, 1976), agency costs arise when the decisions taken by the managers entail destruction of the company's value and the investors have to incur monitoring costs. One of the mechanisms capable to reduce agency costs is decrease of the cash flow at the managers' disposal. Availability of an ample free cash flow results in managers adopting more investment programs even if with low net present value (NPV). Thus, payment of extra funds to shareholders in the form of dividends is more appropriate to avoid such funds spending in a way ineffective from the point of view of the company value maximization. Easterbrook (Easterbrook, 1984) notes that dividends are the main means of reduction of the cash flow available to managers and play an important role in the agency problem mitigation.

Another argument in favour of dividend payment is the assertion that firm pay dividends in order to create a favourable image and avoid underestimate of the company in situations when the firm has to raise external capital. Increasing the dividends paid, the company has to resort to the capital market to raise funds to make investments. This entails thorough evaluation of the management's decisions by potential investors which also mitigates the agency problem.

According to studies (such as, for example: (Shleifer, Vishny, 1997; La Porta et al., 2000)), pyramidal structures of property, cross-holdings, issue of two or more types of shares and other mechanisms used for changing the proportionality and distribution of risk and control within the company represent a source of agency costs. These instruments are capable to increase private benefits from control and the conflict of interests between major and minority shareholders and, accordingly, may affect the company's dividend policy. R.La Porta (La Porta et al., 2000) qualifies issue of two classes of shares as one of the most apparent and frequent ways of managers and shareholders "entrenchment".

In companies with two classes of shares cash flow and voting rights are distributed unequally. Thus, in companies with two classes of shares the conflict of interests is more acute than in those with but a single class of shares. Co-existence of two classes of shares within a company enables the largest holder to maintain control over the company through holding a large portion of voting shares without the need to have hold of shares conferring rights to cash flow only. Thus, the largest shareholder is capable to control decision-taking within the company without assuming the costs burden. This enables a greater extent of benefiting at the expense of minority shareholders as compares to largest shareholders of companies with but a single class of shares.

In Russia the two classes of shares were introduced in 1992 parallel to the privatization program launch. The program envisaged three variants of state-owned enterprises privatization and stipulated obligatory introduction of a corporate charter for all large stateowned enterprises proposed as privatization targets. One of privatization ways was transformation of state-owned enterprises into companies with chartered capital wherein preferred or non-voting shares freely distributed among the active and retired employees of the company could account for as much as 25%.

The legal status of the two classes of shares was defined in the charters of all privatized companies. Notably, the status of ordinary shares was similar to that of ordinary shares in most developed countries (grant of the right to vote at a general meeting of shareholders and the right to dividends the amount whereof was undefined) while preferred shares had certain specificity.

Since enactment of the federal law "On joint-stock companies" in 1996 the rights granted by preferred shares became variable, depending on changes in companies' charters. Thus, holders of preferred shares ceased to enjoy the veto right but were occasionally granted the right to vote at a general meeting of shareholders. But since the portion of preferred shares in the chartered capital was not in excess of 25%, their holders could not ban any decision. Additionally, the law ceased to attach to holder of preferred shares the right to dividends amounting to 10% of net profit, confining itself to indication that companies be obliged to define in their charters the amount of dividends on preferred shares in the form of a fixed percentage of the company's net profit or in any other clearly defined form. With a view of enhancing the level of minority shareholders' rights protection several important amendments were introduced to the law in 2001. Thus, the veto right was returned to preferred shares holders.

Preferred shares have certain advantages over ordinary ones. However, absence of the right of vote and the company being entitled to partly define the rights under preferred shares at its own discretion demonstrate essential inequality between the two classes of shareholders and the opportunity for preferred shares holders being expropriated by holders of ordinary shares.

3. Hypotheses

For a long time empirical studies in corporate governance dealt mostly with companies with dispersed ownership structure (Grossman, Hart, 1980). Investigators into dividend policy focused mostly on the impact of the owner holding the controlling block of shares on decisions taken on dividend payment (see, For example, (Jensen, Meckling, 1976; Shleifer and Vishny, 1986)). However, the ownership structures prevailing in many countries involve presence of multiple major shareholders within the company. For example, in Germany about a quarter of publicly traded companies have two or more shareholders with a portion of shares amounting to at least 20% (Gomes and Novaes, 2005). According to works available (Maury and Pajuste, 2002; Bebczuk, 2005), such companies' dividend policy is the result of largest shareholders interaction: formation of coalitions or largest shareholders fight for influence on decision-taking. Only recently researches started to study the categories and amount of largest shareholders' ownership portions as well as such controlling group's impact on the fact of minority shareholders expropriation.

According to the studies of S.Grossman and O.Hart (Grossman nad Hart, 1980) as well as A.Shleifer and R.Vishny (Shleifer and Vishny, 1986), it is the major shareholders that should carry out management monitoring. Presence of a major shareholder within the company mitigates "the free rider problem", accordingly reducing agency costs. Shareholders in possession of a major portion of shares are more incentivized to carry out management monitoring since the benefits from such monitoring considerably exceed the implementation costs. Notably, the more recent works by R.La Porta (La Porta et al., 2000) and A. Gomes (Gomes, 2000) state that in countries where the legal protection of shareholders is weak it is through high concentration of ownership that the agency problem is to be reduced.

At the same time, existence of shareholders in possession of a major portion of shares or a controlling shareholder among the company owners may be unfavourable for less influential stakeholders. A.Shleifer and R.Vishny (Shleifer and Vishny, 1997) assert that in a situation when major shareholders obtain almost complete control over the firm they begin to derive private benefits minority shareholders fail to participate in. There are multiple ways of minority shareholders' rights impairment, M.Faccio, L.Lang and L.Young (Faccio, Lang and Young, 2001) specially emphasizing low dividend payments.

Hypothesis 1. Increase of percentage of ordinary shares held by a major shareholder will lead to decrease of dividend payments.

Hypothesis 2. Increase of percentage of ordinary shares held by three major shareholders will lead to decrease of dividend payments.

Hypothesis 3. Increase of percentage of ordinary shares held by the second largest shareholder will lead to increase of dividend payments.

Hypothesis 4. Decrease of the difference between the amounts of ordinary shares held by the first and the second largest shareholders will lead to increase of dividend payments.

Hypothesis 5. Presence of a controlling shareholder within the company will lead to decrease of dividend payments.

Hypothesis 6. Largest shareholders having a portion of ordinary shares exceeding that of preferred shares will prefer dividend payments to decrease.

It is worthy of note that researchers' opinions differ with regard to practically every of the listed suppositions concerning the described relations character. This fact is quite understandable. Legislation peculiarities, level of financial markets development, corporation's evolution history and legal status of holders of different types of shares entail variable impact of ownership concentration and a specific shareholder type on dividend payments policy.

4. Methodology

The study aimed to reveal the character of ownership structure impact on the dividend policy of Russian companies with two classes of shares is based on a regression model (1):

... (1)

Dependent variable Div_Payoutit characterizes the dividend payout ratio of company i at moment t. In equation (1) Xit is the vector of variables characterizing concentration of ordinary shares in the hands of the company's shareholders (dimension (m ×'1)) ; Yit is the vector of variables characterizing the type of the company's major shareholders (dimension k×1)); Zit is the variable determining the largest shareholder's portfolio structure; x^sub it^ is the vector of variables "in charge of" the company's financial and economic standing indicators (dimension (l×1)); u^sub it^ is a random disturbance. All the vectors and variable have index it showing that the data are measured for each company i at time moment t. Note that the model is linear in terms of parameters but the variables vectors include non-linear components.

5. Data and sample

Included into the study sample were companies that were traded on RTS stock exchange during 2003-2009 and had two classes of shares. All in all, during 2003 - 2009 there were 145 such companies. The final panel is composed of 598 observations. The conditions for companies to be included in the sample were as follows: both classes of shares simultaneously traded on RTS stock exchange; dividends paid in cash form. The required data on the companies' ownership structure and dividend payments were obtained from the issuers' quarterly reports. For collection of data on the results of financial and business activities, ownership concentration, the largest owners identity, amount of dividends paid and other details of the issuers we used SKRIN and SPARK public databases. See Table 1 for description of the variables used in regression models.

For the results of descriptive statistics of the variables used in the econometric analysis see Table 2.

From Table 2 it follows that the average value of the aggregate dividend payout ratio for the whole of the period is 0.311 which means that the sampled companies paid out in the form of dividends, on the average, approximately 31% of their net profit. Preliminary statistical analysis showed that the minimum value of the dividend payout ratio was -0.401 since nine of the companies under observation paid dividends having a negative value of net profit while the maximum dividend payout ratio value was 2.393. Such situations when the dividend payout ratio value is negative or in excess of 1 are possible in cases when the company pays out dividends from reserves18. In the course of further statistical and econometric analysis the outliers were excluded.

Analysis of the average dividend payments dynamics evidences considerable fluctuation of the dividend payout ratio from year to year. Notably (as we already remarked above), the largest fluctuation of dividend payout ratio occurred in 2006-2009.

Thus, from the descriptive statistics, behaviour of dividend payout ratio as well as the main ownership concentration characteristics it follows that the amount of dividends paid by the companies underwent a considerable variation during the period under consideration. Notably, a variation also occurred in the companies' ownership concentration and the ratio of the amounts of the largest shareholders' portions, and consequently - in the character of such shareholders' interaction as well. In order to test the suppositions concerning existence and character of the relation between the dividend amount paid and the ownership concentration we applied regression analysis.

6. Regression analysis results

See Table 3 for the results of the regression analysis wherein the following three types of dividend payout ratio were used as the dependent variables: aggregate ratio of dividend payout on the both classes of shares (Div_Payout), for ordinary shares (Ord_Payout) and preferred shares (Pref_Payout). Consequent testing of the models demonstrated that the fixed effect model describes the empirical data most adequately.

All the models apart from that presented in Column 9 are statistically significant. Variable Share_3, Conc_3, (Conc_3)2 turned significant in all the models. Variable Control(d) is significant in the models where the dependent variables are: ratio of dividend payout on the both classes of shares Div_Payout ? dividend payout ratio for ordinary shares Ord_Payout. In the model (see Column 11) wherein the dependent variable is represented by the ratio of dividend payout on preferred shares Pref_Payout, in contrast to the other model, it is the variable characterizing the largest shareholder's portion of ordinary shar es Share_1 that is significant. In some parameter assessments the coefficients on the significant variables are different from those supposed. Thus, the coefficient on the variable Control(d) has a positive sign in spite of the hypothesis put forward alleging existence of a reverse relation between the dividend payments amount and existence of a controlling shareholder within the company.

Let us proceed to analysis of the variables characterizing the impact of ownership concentration on the dividend payout ratio (Table 3). Contrary to the supposition concerning existence of the link between the major shareholder's portions of ordinary shares and Div_Payout and Ord_Payout, variables Share_1 (largest shareholder's ownership portion) and Share_2 (second largest shareholder's ownership stake) are insignificant in the models (Columns 1, 2, 6 and 7). However, in the model with Pref_Payout as the dependent variable (Column 11) the coefficient on the variable characterizing the largest shareholder's portion of ordinary shares Share_1 is significant. Consequently, one can admit the supposition on existence of a reverse relationship between concentration of ordinary shares held by the largest shareholder and the amount of dividends paid on preferred shares.

Additionally, one revealed statistically significant relation between the third largest shareholder's portion of ordinary shares Share_3 and the dividend payout ratios. The negative value of the coefficient (Columns 1, 6, 11) shows that this variable is inversely related to the three types of the dividend payout ratio. With regard to this variable no hypothesis was put forward, the variable having been introduced with a view of analyzing the relation revealed in the paper (Maury and Pajuste, 2002).

We found statistically significant non-linear relation between the aggregate portion of ordinary shares held by the three largest shareholders Conc_3 and the value of the dividend payout ratio for the both classes of shares as well as for the ordinary and the preferred shares viewed separately (Columns 3, 8, 13). Analysis of the quadratic function (Column 3) shows that growth of ordinary shares concentration in possession of the three largest shareholders from 7.3 to 73% brings about decrease of the ratio of dividend payout on the both classes of shares Div_Payout, while further concentration growth within the range from 73 to 99.5% will entail increase of the dividend payout ratio. Similar analysis of the quadratic function (Column 8) that in case of variable Conc_3 value change from 7.3 to 70.5% the amount of dividend payments on ordinary shares Ord_Payout is reduced, further change of the variable value within the interval from 70.5 to 99.5%, entailing increase of the ratio of dividend payout on ordinary shares. Additionally, the results of analysis of the non-linear function (Column 13) enable one to conclude that at a 7.3-81.7% concentration of ordinary shares growth in concentration of ordinary shares in possession of the three largest shareholders will entail decrease of dividend payments on preferred shares Pref_Payout. Notably, increase of ordinary shares concentration within the interval from 81.7 to 99.5% will entail increase of the volume of dividend payments on preferred shares.

According to the results obtained (Columns 5 and 10), there is a statistically significant relation between the ratio of dividend payout on the both classes of shares Div_Payout and on ordinary shares Ord_Payout and existence of a controlling shareholder within the company. Relying on the coefficients signs, one may conclude that the amount of dividend payments on the both classes of shares and on ordinary shares with companies having a controlling shareholder are accordingly 17.5% and 14.7% higher as compared to companies without a controlling shareholder.

The other ownership concentration indicators (difference between the two largest shareholders' portions of ordinary shares Spread and fact of presence of a second largest shareholder with a block stake described by the binary variable Share2(d) turned insignificant in all the models considered.

Analysis of the models wherein variable Power_1 (characterizing the largest shareholder's portfolio structure) is used (Columns 4 and 5, 9 and 10, 14 and 15) fails to allow of a conclusion on the possibility of the largest shareholder using the portion of ordinary shares in excess of that of preferred ones to derive private benefits of control and reduction of the level of dividends paid. From the regression analysis it follows that this variable is statistically insignificant.

One may conclude that dividend policy on preferred shares considerably differs from that on ordinary shares and is determined predominantly by the company's performance measures and ownership concentration.

7. Conclusions

Taking into account high concentration of ownership in Russian companies with two classes of shares as well as imperfection of the Russian legislation concerning protection of minority shareholders' rights, we supposed that increase of ordinary shares concentration in the hands of the largest shareholders is associated with lower dividend payments because of the largest shareholders deriving private benefits of control. However, we did not reveal the relation between the amount of aggregate dividend payments on the both classes of shares and that of the ordinary shares portions held by the largest and the second largest shareholders. Notably, there is a nonlinear relationship between dividend payout ratio and concentration of ordinary shares in possession of the three largest shareholders. We found out that with companies where ordinary shares concentration in possession of the three largest shareholders grows from 7.3 to 73% the dividend payout ratio on the both classes of shares aggregately will decrease while further ordinary shares concentration growth (within the range from 73 to 99.5%) will lead to increase of the dividend payout ratio. One of the hypotheses was that existence of a controlling shareholder within the company will lead to decrease of dividend payments. It found confirmation, for example, in the market of Finland (Maury and Pajuste, 2002). However, according to our study, with companies having a shareholder with a controlling portion of ordinary shares, the average value of the dividend payout ratio is 17.5% higher as compared to companies without a controlling owner. Thus, the hypothesis previously put forward was not supported.

Relying on the results obtained, we may conclude that the character of the relation between the ownership concentration factors and the ratio of dividend payout on ordinary shares considerably differs from the said factors relation with dividend payments on preferred shares. We found a similar link between the ratio of dividend payout on ordinary shares and the company's financial and economic indicators, a non-linear relationship with concentration of ownership in the hands of three largest shareholders, a direct relationship with the factor characterizing the fact of a controlling shareholder presence. At the same time, we revealed that most ownership concentration factors fail to significantly affect the amount of dividend payments on preferred shares.

However, we confirmed the hypothesis on existence of a reverse relationship between concentration of ordinary shares in the hands of the largest shareholder and the amount of dividends paid. Such a result may be demonstrating a manifestation of the agency problem between the major and the minority shareholders in accordance wherewith the largest shareholder strives to reduce the portion of net profit paid in the form of dividends on preferred shares and to use this part of free cash flow for deriving private benefits. The reverse relationship between the ratio of dividend payout on preferred shares and the largest shareholder's portion of ordinary shares may be attributed to largest shareholder, as a rule, failing to strive at possession of preferred shares or having held of but a small portion thereof. According to the descriptive statistics data, 77% of the largest owners fail to have preferred shares in their portfolios, the average amount of the largest shareholder's portion of preferred shares being 6.8%. The reverse relation between the amount of dividends on preferred shares and concentration of ordinary shares in possession of the largest shareholder, at first sight, contradicts the result on existence of direct relation between the amounts of both dividend payments on ordinary shares and aggregate dividend payments on the both classes of shares and existence of a shareholder with a control stake within the company. This result may be attributed to controlling owners having hold of small portions of the company's preferred shares: only 20 % of controlling shareholders hold preferred shares, the average amount of preferred shares portion with controlling owners being 6.1% of the total amount of the company's preferred shares. Thus, largest shareholders, apparently, may be not interested in high dividend payments on preferred shares.

Thus, dividend policy on preferred shares considerably differs from that on ordinary shares and is determined predominantly by the company's financial and economic indicators and ownership concentration. However, the amount of dividend payments on preferred shares, unlike those of dividend payments on ordinary shares and aggregate dividend payments, is unaffected by existence of a controlling shareholder that, according to the results of the analysis, is represented by the state or state corporations to an extent of 74%. As one has previously noted, controlling shareholders, on the average, hold a small portion of preferred shares and thus may be not interested in increase of dividends on preferred shares, the same way they are interested in dividends on ordinary shares.

Based on the study results one may conclude that ownership concentration factors more apparently affect the amount of dividend payments on ordinary shares while dividend payments on preferred shares (determined predominantly by the company's financial results and decisions of the largest shareholder) are an obligation of the company similar to debt obligations on the one hand and means to manipulate distribution of voting shares among the largest shareholders - on the other.

All the aforesaid serves to raise questions regarding legal protection of preferred shareholders' rights in general and such rights abuse by large shareholders of the company that are holders or ordinary shares. It appears that even further steps for enhancement of investors' rights protection level are not likely to dismantle the acute problem. Only unification of shares classes meaning equalization of shareholders' control and cash flow rights as per one share may promote mitigation of the agency problem with regard to the conflict of interests between holders of different types of shares.

Footnote

18 See [Russian Federal Law on Joint-Stock companies, 1995, p. 2, page 42].

References

8. Bibliography

1. Bebczuk R. Asymmetric Information in Financial Markets. Cambridge, UK: Cambridge University Press, 2003.

2. Bebczuk R. Corporate Governance and Ownership: Measurement and Impact on Corporate Performance and Dividend Policies in Argentina Inter-American Development Bank Research Network. Working Paper N R-516. 2005.

3. Easterbrook F. Two Agency-Cost Explanations of Dividends American Economic Review. 1984., 74. pp. 650-659.

4. Faccio M., Lang L., Young L. Dividends and Expropriation. American Economic Review. 2001., 91(1). pp. 54-78.

5. Gomes A. Going Public without Governance: Managerial Reputation Effects Journal of Finance. 2000, 55(2). pp. 615-646.

6. Gomes A., Novaes W. Sharing of Control versus Monitoring as Corporate Governance Mechanisms . Working Paper N 01-029. Pennsilvania Institute for Economic Research, 2005.

7. Grossman S., Hart O. Takeover Bids, The Free-Rider Problem, and the Theory of the Corporation Bell Journal of Economics. 1980., 11(1) pp. 42-64.

8. Grossman S., Hart O. One Share, One Vote, and the Market for Corporate Control Journal of Financial Economics. 1988. 20(1/2) pp. 175-202.

9. Jensen M., Meckling W. Theory of the Firm: Managerial Behavior, Agency Costs, and Ownership Structure Journal of Financial Economics. 1976., 3(4 )pp. 305-360.

10. La Porta R., Lopez-de-Silanes F., Shleifer A., Vishny R. Agency Problems and Dividend Policies around the World Journal of Finance. 2000., 55(1). pp. 1-33.

11. Maury C. B., Pajuste A. Controlling Shareholders, Agency Problems, and Dividend Policy in Finland. Swedish School of Economics and Business Administration. Working Paper: pp. 15-45. 2002.

12. Maury C. B., Pajuste A. Private Benefits of Control and Dual - Class Share Unification. Swedish School of Economics and Business Administration. Working Paper. 2007.

13. Shleifer A., Vishny R. Large Shareholders and Corporate Control The Journal of Political Economy. 1986., 94(3) pp. 461-488.

14. Shleifer A., Vishny R. A Survey on Corporate Governance Journal of Finance. 1997., 52(2). pp. 737-783.