Abstract
Purpose
This study aims to analyze the influence of taxes and corporate governance on the dividend policy of Brazilian companies.
Design/methodology/approach
The authors identify the changes of the tax legislation in Brazil in the period 1986-2011 and check their effect on corporate dividend policies for preferred and common shares. The authors use panel data Probit and Tobit estimation to verify the probability of companies to pay dividends under different tax regimes. The final sample comprises 672 companies, 1,159 traded stocks and 30,134 observations
Findings
The authors’ results suggest that changes in the tax legislation have a significant influence on dividend payments. Also, firms do not follow target payout ratios, but dividends are moderately dependent on past payments. Dividend payouts are affected by stock voting rights, privatization and dividend deductibility. Changes in regulation that reduce the agency problems among shareholders affect positively payout ratios.
Practical implications
For managers, maximizing shareholders’ value requires taking into account the consequences of the taxation when designing financial policies for the firm. For investors, stock portfolio selection should take into account payout behavior and how changes in dividend taxation affect stocks’ value. For policymakers, the effects of changes in the tax code on corporate behavior are of utmost importance to stimulate private investment and economic growth.
Originality/value
There are several tax law changes in Brazil within the period analyzed, creating a good opportunity to study the effect of taxation on dividend policy and its dynamics over time.
Keywords
Citation
Zagonel, T., Terra, P.R.S. and Pasuch, D.F. (2018), "Taxation, corporate governance and dividend policy in Brazil", RAUSP Management Journal, Vol. 53 No. 3, pp. 304-323. https://doi.org/10.1108/RAUSP-04-2018-006
Publisher
:Emerald Publishing Limited
Copyright © 2018, Timoteo Zagonel, Paulo Renato Soares Terra and Diogo Favero Pasuch.
License
Published in RAUSP Management Journal. Published by Emerald Publishing Limited. This article is published under the Creative Commons Attribution (CC BY 4.0) licence. Anyone may reproduce, distribute, translate and create derivative works of this article (for both commercial and non-commercial purposes), subject to full attribution to the original publication and authors. The full terms of this licence may be seen at http://creativecommons.org/licences/by/4.0/legalcode
1. Introduction
One of the main and more complex corporate financial decisions deals with the definition of the firm’s dividend policy, that is, the trade-off between distributing funds to shareholders and retaining the profits within the company. Marginal corporate and individual tax rates are one of the main factors that corporations ponder in such decision. Among emerging markets, Brazil is known as a country where such marginal tax rates are high, which suggests that companies pursue active tax planning strategies, which include changes in their dividend policies in response to changes in the tax legislation. This paper tests the influence of the tax preference theory on public Brazilian companies and is motivated, fundamentally, by the many changes in the tax legislation that Brazil has been through in this period, which makes for a good laboratory to study this problem.
Another relevant aspect of this paper is the legal framework of corporate law in Brazil, which allows a firm to issue up to two-thirds of its total equity in nonvoting preferred shares. Therefore, to hold complete sway over a public corporation in Brazil, the controlling shareholders may have as little as 16.7 per cent of firm’s total equity. Such disproportional power raises obvious agency problems between majority and minority shareholders.
Such legislation was originally established in 1976 by Act 6404/1076. It has been changed in 2001 by Act 10303/2001, which limited the maximum proportion of nonvoting shares to one-half. However, the new legislation preserved the status of existing companies. Thus, the new requirement applies only to firms, who were either created or went public after 2001, preserving the statutory privileges of existing ones.
To investigate such problem, we collect data on 672 firms listed in the São Paulo Stock Exchange (Bovespa) between 1986 and 2011, making up a total of 30,134 observations. We identify the pertinent changes of the tax legislation and check the effect of such exogenous changes on corporate dividend policies of preferred and common shares of listed firms.
The paper proceeds as follows: Section 2 presents a brief history of taxation law applied to dividend payments in Brazil. Section 3 summarizes previous empirical literature on this subject in Brazil. Section 4 presents the research method and describes the sample. Section 5 presents the results of our analysis. Section 6 concludes the paper.
2. Tax legislation on dividends and capital gains in Brazil (1986-2011)
Brazil has gone through several changes in its taxation regarding dividend payments in the past couple of decades. Between 1986 and 2004, there were four major changes in such legislation. Also, taxation on capital gains from stock trading has changed once for individual investors and has not changed for corporations. In addition, two other legislation changes that indirectly affect firm’s dividend policies have been implemented in this period. Here, we briefly summarize such legislation.
2.1 Tax treatment of dividends and capital gains in Brazil
In the period between January 1, 1980 and December 31, 1988, dividend payments to individuals were taxed according to three different tax rates (according to Decree 1790/1980, Articles 1 and 2; Decree 2065/1983, Article 1-I and Decree 2303/1986, Article 7): 23 per cent if the distributing company was publicly listed (except for firms from the agriculture industry whose profits originated from these activities); 15 per cent if the distributing company was from the agriculture industry (whose profits originated from these activities); and 25 per cent for all remaining cases.
If the beneficiary was a firm, then there were two tax rates: 23 per cent if the beneficiary was a publicly traded company, a tax-exempt firm (except for pension funds), a subsidiary of a publicly traded corporation or when the distributing firm was from the agriculture sector (whose profits originated from these activities); and 25 per cent for all remaining cases.
From January 1, 1989 to January 1, 1996, several changes have occurred in the tax legislation. Table I summarizes such legislation changes in the case of an individual or corporate beneficiary, including legal basis.
Regarding taxes on capital gains in Brazil, corporations have been taxed on capital gains also at the flat rate of 15 per cent since 1977, according to Decree 1598/1977, Article 31 (Brazil, 1977). Such legislation has not changed since. Individuals, on the other hand, were exempt from capital gains taxes until 1988. Since 1988, however, individuals have been taxed also at the flat rate of 15 per cent, according to Act 7713/1988, Articles 2 and 3 (Brazil, 1988) and Act 8981/1995, Article 21 (Brazil, 1995a, 1995b). Table II summarizes the legislation on capital gains taxes.
The reader should keep in mind that this paper refers to periods of taxation as follows: TAX2 – exempt, except for the net income; TAX3 – exempt; TAX 4 – taxation; TAX5 – exempt. The Brazilian legislation becomes complex in the case where shares are transferred to the heirs of a deceased individual. If the heirs opt for the transference according to market value, then they are subject to tax incidence, according to Act 9532/1997, Article 23, § 1 (Brazil, 1997). However, if the transference is made according to the historic value recorded in the previous tax returns of the deceased, then there is no taxation according to Act 9532/1997, Article 23 (Brazil, 1997).
Compared to the USA, the Brazilian taxation may be considered generous, as its rates are well below the ones practiced in the USA. Table III, based on Elton, Gruber and Blake (2005), revised with recent Internal Revenue Service (IRS) data, reports the changes in dividends and capital gains tax rates in the USA in the period of our study. Brazilian tax rates are roughly half of those in the USA.
2.2 Dividends for nonvoting preferred stocks
Preferred shares in Brazil were introduced by Decree 21536/1932 and modified by Act 6404/1976. According to this piece of legislation, preferred nonvoting shares are entitled to either of the following preferences:
preference in dividend payouts;
preference in share repurchases with or without a premium; and
both of the previous preferences.
In 2001, Act 10303/2001, Article 17 (Brazil, 2001) rewrote the dividend preference, explicitly calling for a fixed or minimum dividend instead of the vague “preference” of the previous legislation. Moreover, its first paragraph went further in restricting the trading of preferred nonvoting shares in case they did not include at least one of the following preferences:
a dividend of a least 25 per cent of net earnings or 3 per cent of book equity and equality of treatment to common stocks regarding dividends;
a dividend at least 10 per cent higher than the one to common stocks; or
equality of treatment to common stocks regarding dividends, share repurchases and takeover offers with or without a premium (mandatory bid rules or tagalong rights).
Which such preference each company attributes to its preferred shares must be explicitly stated in the company’s acts of incorporation. Therefore, the new legislation aimed to reduce the discretion of majority (voting) shareholders in the expropriation of minority (nonvoting) shareholders through the dividend policy. Nonetheless, the legislation is still timid to fully address the agency problem between majority and minority shareholders in Brazil.
2.3 Deductible “interest on equity capital”
In 1996, Act 9249/1995 (Brazil, 1995) came into effect introducing the concept of “Interest on Equity Capital” (Juros Sobre o Capital Próprio – henceforth IOEC). This legislation in its Article 9, §7, allows firms to partially deduct payments of dividends as operating expenses. Article 9 of the Act, incorporating Articles 347 and 668 of the Income Tax Regulation of 1999 (Brazil, 1999a), limits the deductibility to a maximum of twice the amount of interests paid or received by the firm, as well as to the amount resulting from the computation of the official long-term interest rate on book equity. IOEC paid out, according to the Internal Revenue Service’s Instruction 11/1996 (Brazil, 1996, Article 30), is subject to a flat tax rate of 15 per cent retained at the source. Such legislation in effect created a tax incentive to equity capital – however limited – equivalent to the tax benefit of debt that is usual elsewhere in the world. To the best of our knowledge, such legal provision is unique to Brazil.
Therefore, Brazil offers a peculiar legal frame to test the tax preference theory of dividends. As dividends have become gradually tax-free while capital gains have become more taxable over time, with different tax rates in different periods, we expect that dividend payouts increase when the tax rate on dividends decreases and/or the tax rate on capital gains increases. Also, until recently, dividend policies for common (voting) and preferred (nonvoting) stocks have been subject to the discretion of controlling shareholders at the expense of minority shareholders.
2.4 Stock voting rights and corporate governance regulations
Stock voting rights in Brazil have been established by Act 6404/1976 (Brazil, 1976) in its Article 15, §2, which allowed firms to issue up to two-thirds of its shares in nonvoting preferred classes. Such legislation allowed shareholders to hold absolute controlling interest in any public firm with as little as one-sixth of its shares plus one share. Of course, such disparity between voting and cash flows rights increased the agency problem between majority and minority shareholders, where the latter may be subject to wealth expropriation by the former.
In 2001, Act 10303/2001, Article 2 (Brazil, 2001), modified the previous legislation reducing the maximum proportion of nonvoting shares to 50 per cent. However, such change applies only to firms that go public (Initial Public Offering) and to those public firms that prepare seasoned equity offerings after the promulgation of the Law (Brazil, 2001, Article 8, §1). Moreover, the minimum capital that any shareholder must commit to a public firm to have absolute control over it is still only 25 per cent.
Also in 2001, Bovespa launched its “Novo Mercado” (New Market), a special listing segment designed for shares issued by companies that voluntarily undertake to abide by corporate governance practices and transparency requirements in additional to those already requested by the Brazilian Law and the Brazilian Securities and Exchange Commission, from 2004 to 2011, 103 companies entered the “Novo Mercado”. Such special segment requires, among other things, one vote–one share rule. In that year, Bovespa also launched two intermediary listing segments for firms wishing to gradually improve its corporate governance standards. These levels (Level I and Level II), however, tolerate the existence of nonvoting shares even though they introduce new disclosure and floating requirements. Despite being an improvement over the previous regulation, the new rules are still timid steps toward the international corporate governance standards.
3. Previous empirical evidence on dividend policy in Brazil
Several empirical studies on dividend policy have been conducted in Brazil. The pioneer study by Brito and Rietti (1981) replicated the well-known study by Elton and Gruber (1970), testing the Clientele Effect in the Brazilian market, between years 1973 and 1976. Their study did not find any evidence of the existence of this effect in the Brazilian stock market. The authors concluded that, in opposition to the North American stock market, in Brazil, it does not seem to have any association between marginal levels of investors’ taxation and the preference for retention or distribution of dividends.
Procianoy and Poli (1993) discussed the possibility of a dividend policy that could simultaneously produce large tax savings to investors and encourage more liquidity in the stock market and, as a consequence, higher share values. The authors concluded that companies that would adopt their suggested policy would get a substantial increase in share prices. The investors would look for shares that could give them the larger possible net earnings, as a result of tax savings. The clientele effect was central to this new outfit, and, in such case, the dividend policy was relevant and maximized the value of the company.
In another study, Procianoy and Snider (1994) observed that changes in the tax legislation on dividends and capital gains, in 1989, offered an opportunity to test the maximization of shareholder wealth through dividend payouts. At the time of the legislation change in 1989, there was an inversion of tax burden: up to that year, dividends were taxed and capital gains were not, and, from 1990 on, capital gains were taxed and dividends were tax-free. The study tried to determine if an increase in payout ratios by companies listed in Bovespa has really happened after the tax legislation changed. The authors identified that, for the period from 1987 to 1988, the average dividend payout ratio was 25 per cent when dividends were taxed. Meanwhile, from 1990 to 1992, when the dividends were tax-free, the average payout ratio was 40 per cent. Therefore, the authors concluded that after the tax legislation has changed the dividend payouts have generally increased. However, the increase has not been verified in all companies, as suggested by the observed decrease of the minimum dividend. The authors also cautioned that their sample was small. In conclusion, Procianoy and Snider (1994) observed that, after the Brazilian tax legislation change, there was an increase of dividend payouts, but they were not increased to the 100 per cent level, which, according to the authors, would be the ideal rate. They also stated that companies that had more controlling shareholders had a more significant increase in dividend payout ratios than the companies that had just a single controlling shareholder.
Another study about dividend policy and taxes in Brazil is the paper of Procianoy (1996), motivated by the change in the Brazilian legislation with respect to dividends that happened in 1988 and 1989. This change stopped taxing the dividend distribution by the companies (23 per cent on dividend value retained at the source) and instead started taxing capital gains in 25 per cent, which should be calculated from the difference between the sale and purchase prices (including adjustment for inflation), a tax rule that was effective from 1990 to 1993. Using data from Bovespa from the years 1987 to 1992, Procianoy (1996) tested two hypotheses: the changes in dividend policies by the companies listed in Bovespa; and the maintenance of a new dividend policy, in the period between 1990 and 1992. The author concluded that taxation is one of the factors that influence the companies’ dividend policy. In his conclusion, the author also highlighted that the agency conflict between managers and shareholders may be responsible for the adoption of non-maximizing shareholders’ wealth dividend policies.
In a recent paper, Vancin and Procianoy (2016) find that the legislation has an important role in the determinants of dividend policy. They find that there is strong empirical evidence that companies paying dividends above the legal mandatory level present different determinants from those that only pay the minimum level. Thus, the samples that mix companies that pay both the mandatory minimum dividend and above the minimum influence regression coefficients, resulting in significant biases in dividend determinants in Brazil. As companies that pay above the mandatory levels are the real decision makers of dividend policies (given that other companies simply comply with the legislation), this aspect must be taken into account when studying dividend policy in countries where minimum mandatory dividend legislation is in place.
Ramos (1997) studied the influence of taxes on dividends in the Brazilian stock market, adjusting the returns for systematic risk levels. Following Litzenberger and Ramaswamy’s (1979) method of augmented capital asset pricing model, the author focused on two distinct periods: from 1984 to 1987 and from 1988 to 1992. The model states that, in equilibrium, the expected net return of a stock is linearly related to its systematic risk and its dividend–price ratio. Ramos’s (1997) results indicated a significant and positive coefficient for the two subperiods. His conclusions support the hypotheses that dividend taxation influences the stock returns in Bovespa.
Nossa and Nossa (2007) investigated the economic performance of Bovespa’s listed companies that distributed dividends or repurchased their shares in the period 1995-2004. The authors found that firms that paid dividends had comparatively increased positive returns in every single year of the sample, while firms that repurchased shares increased their returns only in the years 1997, 2002, 2003 and 2004.
Mota and Eid (2007) analyzed the choice of the method of payout policy of companies listed in Bovespa from 2002 to 2005. From the three distribution options (dividends, IOEC and shares repurchases), they found that dividends are preferred to IOEC, in spite of the latter’s tax deduction advantage. Also, share repurchases are usually used as a complement to dividends and IOEC, mainly because of its higher transaction costs and lack of legal obligation.
Decourt et al. (2007) test if the changes of the payout ratios point to variations on net earnings in the year prior to and subsequent of dividend payment. Using a sample of listed companies from 1997 to 2005, the authors document that an increase of dividend payout does not signal future earnings, but reductions are correlated to the increasing in contemporaneous net earnings and also signal increase in the future earnings. Apparently, earnings retentions signal good investment opportunities.
Cioffi and Famá (2010) analyzed if dividends could be used as a proxy for other information that contribute for the prediction of future returns, to test the effect of signaling on the market value of companies. The authors found that dividends affect positively the value of the companies, in accordance with other studies from Brazil and other countries.
Forti and Schiozer (2011) found a positive relationship between dividends and past earnings, when looking at dividend policy of Brazilian banks. This evidence reinforces the use of dividends to inform the quality of assets when sending a sign that the bank is capable of making profits with assets that perform well. Furthermore, privately held banks pay more dividends than publicly traded banks, which complies with the purpose of dividend signaling.
In contrast, Fiorati et al. (2007) studied whether the variation in earnings distribution policy (dividends and IOEC) had any relationship to the firm’s profitability in subsequent years. Using a sample of public firms from 1994 to 2004, the authors could not document a significant relationship between earnings payouts and future firm profitability. The authors suggest that such result may be because of market conditions and the legal and institutional politic–economic organization.
Futema et al. (2007) have performed a solid analysis of the capital structure, dividends and IOEC of Brazilian companies in the period 1995-2004, following the Fama and French (2002) model. They aimed to contribute to the joint understanding of capital structure and dividend policy in Brazil. They concluded that, as predicted in the static trade-off and pecking order theories, the profitability and growth options are the most influential variables on earnings distribution.
Pohlmann and Iudícibus (2010) show that there is a positive relationship between the tax level on income and debt level. This was also observed for companies with high debt and low profit taxation. These results support the theory of trade-off regarding the impact of taxation on profit over debt decision, and consequently over the capital structure.
Nakamura et al. (2007) investigate the basis of dividend policy definition by Brazilian companies. Their results indicate that companies with better growth opportunities tended to pay lower dividends, and larger companies tended to pay larger dividends, under the hypotheses that they would face less future restrictions to banking loans, as well as lower transaction costs in case of an eventual issue. The authors conclude that dividends are deemed relevant by Brazilian firms along the lines of the Lintner’s (1956) and agency theories.
Futema et al. (2009) confirm the predictions of pecking order, that is, the distribution of earnings varies positively with profitability and negatively with investment, but that do not happens with grow opportunities, which have a negative relationship with earnings distribution. Another positive factor for pecking order is the negative relationship observed between leverage and profitability, confirming that companies use retained earnings to finance their investments and only use debt after running out of internal sources.
As it has been shown above, the study of dividend policy in Brazil raises interesting questions regarding the relationship among dividends, capital structure, taxation, corporate governance and institutions. Moreover, Benetti et al. (2007) document that 62.9 per cent of Brazilian firms surveyed affirm to pay dividends, a larger sample share than Graham and Harvey (2001) report in their sample of North American firms (53.9 per cent). Such is the motivation of this study.
3.1 Previous empirical evidence on corporate governance and dividends in Brazil
Agency problems have a central role in the definition of dividends policy (Ferreira et al., 2010). The commitment to corporate governance practices can reduce the risk for investors because of a raise in shares value caused by best practices, ultimately increasing the liquidity and trading volume of the shares because of increased demand.
With the increase in the value of a company stock, there may be, consequently, a reduction in their cost of capital. For the company, it means fundraising at lower costs, with the likelihood of higher returns on investments, and for shareholders, higher dividends (Rabelo et al., 2007).
Silva (2004) shows that there is a relationship, often statistically significant, between governance structure, market value, leverage and dividend policy of the Brazilian companies. Gonzaga and Da Costa (2009) show that there is a relationship between accounting conservatism and conflicts over dividend policies between controllers and minority shareholders in Brazilian companies.
The financial corporate governance model seems to address the greatest number of features that apply to the Brazilian reality, given the importance it assumes the agency conflict (even if not always explicit) between controlling and noncontrolling shareholders and the importance attached to returns on investments (Bertucci et al., 2006).
Companies who have American depositary receipts Level II or participation in the “Novo Mercado” segment of Bovespa have higher returns in periods of crisis and distribute more dividends (Srour, 2005).
Bellato et al. (2006) show that there is a negative and significant relation between excess vote power in possession of the controllers and the dividend rate of the companies. Their study reinforces the arguments of the codes of good governance practices that the distance from the one share–one vote principle is detrimental to minority shareholders.
Through these studies, we can see that there may be a relationship between the levels of corporate governance and dividend policy. In companies where governance levels are higher, the tendency is that the dividend policy is more aggressive to distribute more dividends to its shareholders.
4. Research methods
We test the hypothesis of the influence of the Brazilian tax changes on the dividend policy of its listed companies. The sample is made up of companies whose shares are traded in São Paulo Stock Exchange (Bovespa), in the period 1986-2011. The data source is from Economática® (2012) database. The final sample comprises 672 companies in the period 1986-2011, making up a total of 1,159 traded stocks of both common and preferred classes. We have also included companies with multiple classes of preferred stock, but kept in the sample only those whose dividend rights are similar, for homogeneity. As we used annual data frequency, the total sample contains 30,134 observations over the sample period.
Firms whose government or governmental bodies have a controlling stake are classified as state-owned. Privatized firms are coded as state-owned up to (and including) the privatization year, and as privately controlled henceforth. Some companies of the sample had a relevant ownership – but not the controlling stake – of the national development bank (Banco Nacional de Desenvolvimento Econômico e Social – BNDES), through its holding subsidiary BNDESPAR. In spite of having such a relevant stake from a government institution, these companies are treated as private companies.[1]
Firm-level data are the dividends per share and earnings per share, both adjusted for inflation using the Augmented Consumer Price Index, which is the default inflation index of Economática® (2012) database. Dividend per share values are year sums obtained in December 31st of the respective year from 1986 to 2011 and include the values of IOEC, paid in the same year. The variables are standardized by their book value of assets, to remove any systematic size effect.
A panel data analysis is performed according to the following model:
Where DPSit is the firm’s i dividend per share in period t; EPSit is the firm’s i earnings per share in period t; COMMONi is a dummy variable for common stocks; STATi is a dummy variable for state-owned companies (privatization dummy); IOECit is a dummy variable when firm i paid IOEC in year t; GOVit is a dummy variable when firm i was listed on any corporate governance listing level in year t; TAXk is the dummy variable for each taxation period (1986-1988; 1989-1992; 1993; 1994-1995; 1996-2011); βj are the coefficients to be estimated and εit is the regression residual.
Equation (1) is estimated as pooled panel data model (simple stacking of time series and cross-section data) using Probit and Tobit estimation. To verify the probability of companies to pay dividends, we use probit, which transforms the dependent variable in a dummy variable, being equal to 1 for firms that distribute dividends in that year or equal to 0 for firms that did not distribute. In the Tobit model, which is applied to variables strictly positive, with the accumulation of values at 0, the dependent variable can take any positive value in its probability distribution, being used in a manner similar to the probit. In this study, Tobit will provide the determination of the amount of distributed dividends, not just the propensity pointed by probit. For treatment of outliers, we applied the winsorization process on DPS and EPS. Winsorization has the advantage of reducing the extreme values of the sample without reducing the number of observations as other alternative outlier treatments such as simple trimming for instance. Descriptive statistics are presented in Table IV.
A panel data specification is useful in this case because it allows the use of a larger number of observations, therefore increasing the degrees of freedom, reducing the collinearity between exogenous variables and reducing the missing variable bias (Hsiao, 1986). We choose the Tobit specification based on Kim and Maddala’s (1992) findings that suggest the importance of accounting for the large number of zero observations in empirical dividend studies. Also, the authors conclude that Tobit estimates, including lagged dividend dependent variables and earnings per share independent variables as well as industry, firm-specific and time effects, seem better equipped to deal with heteroskedastic errors and other specification problems in models of dividend behavior.
We expect that firms would increase their dividend payments in relation to earnings whenever the taxation on dividends decreased (1989-1993 – TAX2 – and 1996-2011 – TAX5) and/or taxation on capital gains increased (1989-2011). The coefficients βj of EPSit and any interaction terms of this variable in equation (1) represent payout ratios under different taxation regimes. We also expect an increase in payout ratios for the companies that are under corporate governance rules introduced by the new legislation and special listing segments, indicating whether corporate governance has any impact on the agency problem between majority and minority shareholders.
5. Results
5.1 Estimation results
The estimation results using Probit and Tobit methods for equation (1) are presented in Tables V and VI, respectively. Analyzing probit first, we have similar results to other empirical works on the performance of portfolios comprising companies with high levels of corporate governance (Rabelo et al., 2007) and on dividend payment (Srour, 2005), which brings evidence that companies with higher levels of corporate governance are more prone to pay dividends. The coefficients of the variable GOV were statistically significant and robust in all regressions.
Companies that distribute IOEC are more likely to distribute dividends too, as shown by the variable IOEC, which remained statistically significant and positive in all regressions. The variable STAT suggests that there is a statistically significant and negative relationship between state-owned companies and dividends payment, i.e. the probability of state-owned companies to pay dividends is lower.
We find a lower probability of common stocks to pay dividends, as we observe a negative and significant coefficient for the COMMON variable. This does not come as a total surprise given that the legislation change of 2001 (Act 10303/2001) required that firms with preferred nonvoting stock should choose one among three possible preferences: a minimum dividend, a dividend at least 10 per cent higher than the commons stock dividend or tagalong rights. Nevertheless, as our sample covers a large period before this regulatory change (1986-2011), this result is an indication of the behavior of firms toward nonvoting stocks in a more ample perspective[2].
Although the results for the variable EPS are not statistically significant for the probit model, they are significant for the Tobit model, as we will see further. For now, it is worth only to observe the positive sign of the coefficients, pointing to a positive relationship between dividends and earnings per share. Lag Dep, which is the lagged dependent variable (DPS), shows that if the company paid dividends in the previous period, then the probability of paying them again increases.
TAX3 period suggests that the likelihood of paying dividends increases in 1993, perhaps because it is exempt from taxation. The interactions TAX3-EPS and TAX5-EPS suggest that the effect of earnings per share in the likelihood of dividend payments is higher in these two tax periods, i.e. as predicted, an increase in earnings per share in periods which are exempt from taxation also increases the probability of dividend payments.
The interaction between the variables GOV-EPS is statistically significant and positive in all regressions, strongly suggesting that firms in special corporate governance listing segments that had increases in their earnings per share increase the likelihood to pay dividends. This result complies with other empirical work on corporate governance and performance of companies listed in Bovespa (Freire et al., 2010).
The results of Tobit regressions (Table VI) show that the EPS is positive and significant, unlike the probit model, suggesting that an increase in earnings per share of a company increases the amount of dividends distributed. The lagged dependent variable (Lag Dep) showed the same results of the probit model, suggesting that if the company paid dividends in the earlier period, then it would pay larger amounts too. That is an indication that firms pursue a persistent dividend policy.
The interactions TAX2-EPS and TAX3-EPS for Tobit regressions are also aligned with the probit, showing that, in periods of less taxation TAX2 and TAX3, an increase in earnings per share would increase the amount distributed as dividends. The interaction TAX4-EPS shows that during TAX4, there would be a reduction of the amount distributed, which makes sense because such period was one of higher taxation. For the period TAX5, when there was no incidence of taxes, the interaction suggests that an increase in earnings per share would also increase the amount distributed, which is consistent with the tax preference theory.
The variables COMMON, STAT, GOV and IOEC for Tobit regressions were statistically significant, as was the probit model, suggesting that common stocks and state-owned enterprises pay less dividends, companies that distribute IOEC also have higher total payouts, as do companies with higher levels of corporate governance. Also consistent with our probit results, the interaction GOV-EPS for Tobit regressions suggests that the effect of corporate governance increases the amount distributed by companies that had an increase in their earnings per share.
5.2 Robustness tests
To check for the robustness of the findings, we reestimate equation (1) using three different dependent variables as follows: DBV (the ratio between dividend/book value of the share), Payout and Yield (results omitted for the sake of concision but available upon contact with the authors).
Using DBV as the dependent variable, we observe the same results as the original model for almost all independent variables, for both methods probit and Tobit, only with a few changes in the significance level. The dependent variable Payout also kept the same results as the original model, presenting only a few changes in the significance level, and the interaction TAX4-EPS presented signal changing. The dependent variable Yield also kept the results practically constant, presenting only a few changes in the significance level, and signal changes for the interaction TAX2-EPS and for TAX4 and TAX5 in the probit method with fixed effects.
6. Conclusions
This study aims to investigate the tax preference theory of dividends in a sample of 672 Brazilian public firms for the period 1986-2011. We documented several changes in the tax legislation of dividends over the sample period and tested their effect on dividend payments using probit and Tobit regression analysis. Our findings suggest that the Brazilian firms do not follow target payout ratios, but they do try to pay dividends that are moderately dependent on past payments. The level of dividend payment is affected by stock voting rights, privatization, dividend payments deductibility provisions and changes of corporate governance rules. Changes in the tax legislation have a significant influence on dividend payout ratios, corresponding to the theoretical prediction. Payout ratios are positively affected by changes in regulation that reduce the agency problems among shareholders of the firm.
We have documented that the tax preference theory found empirical support in the Brazilian environment, given the numerous changes in the tax regulations over the year. Our results highlight the influence of taxation in the payout policies pursued by publicly listed firms. These results have a range of implications for managers, investors and policymakers. For managers, it is clear that maximizing shareholders’ value requires taking the consequences of the taxation – at the corporate and personal levels – into account when designing financial policies for the firm. For investors, the choice of which stocks to include in their portfolios should take into account their payout behavior and how it is affected by changes in dividend taxation. Finally, for policymakers, it becomes clear that taxes drive firms to change their payout policies and, possibly, their investment and financing strategies as well. Therefore, a careful study of the effects of changes in the tax code on corporate behavior is of utmost importance if the goal to stimulate private investment and economic growth is in the agenda of the authorities.
Another important aspect raised by our results is that of the role of agency conflicts in dividend policies. We found significant positive effects in dividend payouts for firms that belong to special corporate governance segments of the stock exchange – firms that voluntarily adopt better practices toward its outside shareholders, thus reducing agency problems. Also, state-owned companies seem to systematically and significantly less likely to pay dividends. These firms are known to be generally more opaque and less accountable to outside investors than privately held enterprises – thus more prone to exacerbate agency conflicts. Therefore, we find evidence that agency problems are also important determinants of dividend decisions in Brazil, along the lines of Easterbrook (1984).
In sum, we conclude that taxation does affect dividend policy, as shown by the coefficients of the regressions, and also that agency problems are an important issue when it comes to dividend payments. Further studies in the field of corporate governance of Brazilian listed corporations may provide a better understanding of this subject and guide future regulation reforms.
Summary of changes in the tax legislation on dividends in Brazil
| Period of earnings accrual | Applied fiscal regimen | Tax periods used in the study | ||||||
|---|---|---|---|---|---|---|---|---|
| Beneficiaries: individuals | Beneficiaries: corporations | |||||||
| Tax retained on distribuition? | Treatment on the beneficiary’s tax returns | Tax retained on distributuion? | Treatment on the beneficiary | Legal basis | ||||
| Earnings | Withholding income tax | |||||||
| Firms taxed under the real profit method | Other firms | |||||||
| Up to December 31, 1988 | Yes | Exclusive taxation on the income source (non-compensable) | Yes | Nontaxable by the corporate income tax | Compensation of the corporate income tax retained over earnings or distributed dividends | – | Decrees 1.790/1980 Articles 1 and 2; 2.065/1983 Article1, I; and 2.303/1986 Article 7 | (1) 1986-1988 |
| From January 1, 1989 to December 31, 1992 | No | Exclusive taxation on the net income at a flat tax rate (8%) | No | Nontaxable by the corporate income tax | – | – | Income Tax Regulation 1999, Article 660 | (2) 1989-1992 |
| From January 1, 1993 to December 31, 1993 | No | Nontaxable | No | Nontaxable by the Corporate Income Tax | – | – | Income Tax Regulation 1999, Article 659 | (3) 1993 |
| From January 1, 1994 to December 31, 1995 | Yes (15% Flat Rate) | Addition of dividends to taxable income and compensation of the tax retained or, optionally, declaration of dividends as subjected to exclusive taxation on the income source | Yes (15% flat rate) | Nontaxable by the corporate income tax | Compensation of the corporate income tax retained upon redistribution | Non-compensable | Income Tax Regulation 1999, Articles 655 and 656 | (4) 1994-1995 |
| From January 1, 1996 onward | No | Nontaxable | No | Nontaxable by the corporate income tax | – | – | Income Tax Regulation 1999, Article 654 | (5) 1996-2011 |
Source: adapted from IOB Thomson (2005), p. 8
Taxes on capital gains in Brazil, 1977-2011
| Taxpayer | Period | Rate (%) | Legal basis |
|---|---|---|---|
| Individuals | 1976-1988 | Exempt | Decree 1510/1976, Article 4 |
| Individuals | 1989-2011 | 15 | Act 7713/1988, Articles 2 and 3 |
| Corporations | 1977-2011 | 15 | Decree 1598/1977, Article 31 |
Source: Authors of the paper
Maximum tax rates on dividends and capital gains in the USA
| Year | Tax rate on dividends (%) | Tax rate on capital gains (%) |
|---|---|---|
| 1988-1990 | 28.00 | 28.00 |
| 1991-1992 | 31.00 | 28.00 |
| January 1996 to June 1997 | 39.60 | 28.00 |
| July 1997 to December 2000 | 39.60 | 20.00 |
| January 2001 to December 2001 | 39.10 | 20.00 |
| January 2002 to May 2003 | 38.60 | 20.00 |
| May 2003 to 2011 | 15.00 | 15.00 |
Source: Elton et al. (2005) and Internal Revenue Service (IRS), USA
Descriptive statistics of firm-level data
| Without winsorization | With winsorization (p = 0.01) | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Variable | N | Mean | Median | SD | Minimum | Maximum | N | Mean | Median | SD | Minimum | Maximum |
| DPS | 30,129 | 4.60 | 0.00 | 180.67 | 0.00 | 17,751.42 | 30129 | 0.48 | 0.00 | 2.24 | 0.00 | 18.65 |
| EPS | 16,429 | −4.89 | 0.19 | 3,029.30 | −125,797.90 | 310,714.00 | 16429 | 0.53 | 0.19 | 33.37 | −191.17 | 186.19 |
Source: Authors of the paper
Probit regressions
| Variables | Expected sign | REG1 | REG2 | REG3 | REG4 |
|---|---|---|---|---|---|
| EPS | + | 0.00136 (1.50) | 0.00137 (1.50) | 0.00132 (1.35) | 0.00131 (1.35) |
| Lag Dep | + | 1.836*** (72.75) | 1.848*** (71.91) | 1.356*** (46.21) | 1.378*** (46.18) |
| TAX2 (1989-1992) | + | –0.266*** (–4.10) | –0.194*** (–2.93) | ||
| TAX3 (1993) | + | 0.183** (2.10) | 0.157* (1.72) | ||
| TAX4 (1994-1995) | – | 0.0495 (0.70) | 0.102 (1.38) | ||
| TAX5 (1996-2011) | + | –0.260*** (–4.61) | –0.0617 (–1.00) | ||
| TAX2 × EPS (1989-1992) | + | 0.00203 (1.63) | 0.00122 (0.97) | 0.00134 (1.01) | 0.000343 (0.25) |
| TAX3 × EPS (1993) | + | 0.0106*** (3.26) | 0.0106*** (3.27) | 0.00779** (2.40) | 0.00781** (2.42) |
| TAX4 × EPS (1994-1995) | – | –0.000817 (–0.48) | 0.000595 (0.34) | –0.00267 (–1.39) | –0.00119 (–0.61) |
| TAX5 × EPS (1996-2011) | + | 0.00257** (2.48) | 0.00278*** (2.68) | 0.00256** (2.26) | 0.00274** (2.41) |
| COMMON | – | –0.00690 (–0.27) | –0.00173 (–0.07) | –0.133*** (–4.35) | –0.139*** (–4.54) |
| STAT | – | –0.150*** (–3.25) | –0.203*** (–4.31) | –0.665*** (–6.60) | –0.721*** (–7.02) |
| IOEC | + | 1.652*** (20.29) | 1.763*** (21.63) | 1.806*** (18.44) | 1.885*** (19.32) |
| GOV | + | 0.334*** (6.58) | 0.514*** (9.38) | 0.586*** (7.78) | 0.717*** (8.99) |
| GOV × EPS | + | 0.0628*** (3.13) | 0.0625*** (3.10) | 0.0551*** (2.60) | 0.0566*** (2.62) |
| N | 16,099 | 16,099 | 14,534 | 14,534 | |
| Pseudo R2 (%) | 41 | 42 | 45 | 46 | |
| Chi2 | 9,209.9 | 9,439.8 | 9,025.0 | 9,208.9 | |
| Year effect | No | Yes | No | Yes | |
| Industry effect | Yes | Yes | – | – |
This table presents the results for probit regressions with grouped data (pooled) and fixed effects, for the dependent variable DPS. This variable assumes value “1” if the company paid dividends and “0” otherwise. Regressions REG1 and REG2 represent grouped data (pooled), and regressions REG3 and REG4 represent the fixed effects. We used dummy variables for each year (for regressions without the variable TAXk) and industry (not for fixed effects) to capture these effects. The coefficients are shown for each variable, and t statistics follow in brackets.
,
and * represent the levels of statistical significance at 1, 5 and 10%, respectively
Source: Authors of the paper
Tobit regressions
| Variables | Expected sign | REG1 | REG2 | REG3 | REG4 |
|---|---|---|---|---|---|
| EPS | + | 0.0145*** (6.88) | 0.0146*** (6.95) | 0.0141*** (6.80) | 0.0140*** (6.79) |
| Lag Dep | + | 0.716*** (81.90) | 0.718*** (82.17) | 0.537*** (52.33) | 0.536*** (52.29) |
| TAX2 (1989-1992) | + | –0.259* (–1.78) | –0.436*** (–3.13) | ||
| TAX3 (1993) | + | –0.0575 (–0.29) | –0.335* (–1.74) | ||
| TAX4 (1994-1995) | – | 0.201 (1.27) | 0.00122 (0.01) | ||
| TAX5 (1996-2011) | + | –0.246* (–1.94) | –0.316** (–2.48) | ||
| TAX2 × EPS (1989-1992) | + | 0.00913*** (3.08) | 0.00752** (2.51) | 0.00916*** (3.18) | 0.00723** (2.48) |
| TAX3 × EPS (1993) | + | 0.0447*** (6.52) | 0.0446*** (6.53) | 0.0248*** (3.94) | 0.0248*** (3.96) |
| TAX4 × EPS (1994-1995) | – | –0.0189*** (–4.76) | –0.0182*** (–4.56) | –0.0249*** (–6.21) | –0.0239*** (–5.92) |
| TAX5 × EPS (1996-2011) | + | 0.0166*** (6.82) | 0.0170*** (7.03) | 0.0116*** (4.81) | 0.0119*** (4.97) |
| COMMON | – | –0.125** (–2.34) | –0.107** (–2.01) | –0.278*** (–4.90) | –0.284*** (–5.02) |
| STAT | – | –0.395*** (–4.12) | –0.473*** (–4.91) | –1.482*** (–7.36) | –1.555*** (–7.68) |
| IOEC | + | 1.598*** (19.67) | 1.857*** (21.55) | 1.162*** (13.10) | 1.345*** (13.84) |
| GOV | + | 0.164* (1.70) | 0.394*** (3.89) | 0.501*** (4.10) | 0.604*** (4.73) |
| GOV × EPS | + | 0.100*** (2.98) | 0.0890*** (2.69) | 0.110*** (3.02) | 0.109*** (2.99) |
| N | 16099 | 16099 | 16099 | 16099 | |
| Pseudo R2 | 14% | 14% | 19% | 19% | |
| Chi2 | 7735.9 | 7911.2 | 10753.3 | 10900.4 | |
| Year effect | No | Yes | No | Yes | |
| Industry effect | Yes | Yes | – | – |
This table presents the results for Tobit regressions with grouped data (pooled) and fixed effects, for the dependent variable DPS. This variable represents the amount of distributed dividends. Regressions REG1 and REG2 represent grouped data (pooled), and regressions REG3 and REG4 represent the fixed effects. We used dummy variables for each year (for regressions without the variable TAXk) and industry (not for fixed effects) to capture this effects. The coefficients are shown for each variable, and t statistics follow in brackets.
,
and
represent the levels of statistical significance at 1, 5 and 10%, respectively
Source: Authors of the paper
Notes
Although relevant, such ownership is clearly of a minority nature (between 5 and 40 per cent of voting shares – 22.45 per cent on average – in only 11 out of 672 firms in our sample). Therefore, we choose to treat such firms as privately owned. We thank an anonymous referee for pointing this out.
We are thankful to an anonymous referee for pointing this out.
References
Bellato, L.L.N., Silveira, A.D.M. and Savoia, J.R.F. (2006), “Influência da estrutura de propriedade sobre a taxa de pagamento de dividendos das companhias abertas brasileiras”, Encontro Nacional da Associação Nacional dos Programas de Pós-Graduação Em Administração. Salvador, p. 30.
Benetti, C., Decourt, R.F. and Terra, P.R.S. (2007), “The practice of corporate finance in an emerging market: preliminary evidence from the Brazilian survey”, Annual Meeting of the Financial Management Association, Orlando, FL, p. 37.
Bertucci, J.L.D.O., Bernardes, P. and Brandao, M.M. (2006), “Políticas e práticas de governança corporativa em empresas brasileiras de Capital aberto”, Revista de Administração, Vol. 41 No. 2, pp. 183-196.
Brazil (1976), “Congresso Nacional. Lei N° 6.404 de 15 de dezembro de 1976. Dispõe sobre as sociedades por ações”, Diário Oficial da União. Brasília, Art. 1° ao 300, 12/15/1976.
Brazil (1977), “Presidência da República. Casa Civil. Subchefia para Assuntos Jurídicos. Decreto-Lei N° 1598 de 26 de dezembro de 1977. Altera a legislação do imposto sobre a renda”, Diário Oficial da República Federativa do Brasil, Brasília, Art. 31, 12/26/1977.
Brazil (1988), “Congresso Nacional. Lei N° 7.713 de 22 de dezembro de 1988. Altera a legislação do imposto de renda e dá outras providências”, Diário Oficial da República Federativa do Brasil. Brasília, Art. 2° e 3°, 12/22/1988.
Brazil (1995a), “Congresso Nacional. Lei N° 9.249 de 26 de dezembro de 1995. Altera a legislação do imposto de renda das pessoas jurídicas, bem como da contribuição social sobre o lucro líquido, e dá outras providências”, Diário Oficial da República Federativa do Brasil. Brasília, Art. 1° a 36, 12/26/1995.
Brazil (1995b), “Presidência da República. Subchefia para Assuntos Jurídicos. Lei N° 8.981, de 20 de janeiro de 1995. Altera a legislação tributária Federal e dá outras providências”, Diário Oficial da União. Brasília, Art. 21, 1/20/1995.
Brazil (1996), “Secretaria da Receita Federal. Instrução Normativa SRF n° 011, de 21 de fevereiro de 1996. Dispõe sobre a apuração do imposto de renda e da contribuição social sobre o lucro das pessoas jurídicas a partir do ano-calendário de 1996”, Diário Oficial da União. Brasília, 2/22/1996.
Brazil (1997), “Presidência da República. Subchefia para Assuntos Jurídicos. Lei N° 9.532, de 10 de dezembro de 1997. Altera a legislação tributária Federal e dá outras providências”, Diário Oficial da União. Brasília, Art. 23, §1° 12/10/1997.
Brazil (1999a), “Presidência da República. Decreto n° 3.000, de 26 de março de 1999 (Regulamento do Imposto de Renda – RIR/1999). Regulamenta a tributação, fiscalização, arrecadação e administração do Imposto sobre a Renda e Proventos de Qualquer Natureza”, Diário Oficial da União. Brasília, 3/29/1999.
Brazil (2001), “Congresso Nacional. Lei N° 10.303 de 31 de outubro de 2001. Dispõe sobre as sociedades por ações”, Diário Oficial da União. Brasília, Art. 1° a 10, 31 out. 2001.
Brito, N. and Rietti, R. (1981), “Efeito clientela, níveis marginais de taxação e eficiência: o caso dos dividendos no mercado acionário brasileiro”, Revista de Administração, Vol. 16, pp. 33-46.
Cioffi, P.L.D.M. and Famá, R. (2010), “O modelo de Ohlson e a sinalização de dividendos no mercado de capitais brasileiro”, XIII SemeAd – Seminários em Administração, São Paulo, p. 8.
Decourt, R.F., Procianoy, J.L. and Pietro, N.J. (2007), “As variações nas distribuições dos proventos em dinheiro sinalizam variações nos lucros futuros?”, XXXI Encontro Nacional dos Programas de Pós-Graduação em Administração. Rio de Janeiro, p. 31.
Easterbrook, F.H. (1984), “Two agency-cost explanations of dividends”, American Economic Review, Vol. 74 No. 4, pp. 221-230.
Economática (2005), “Banco de Dados: Cotações das ações e Ibovespa”, Version: out.
Elton, E.J. and Gruber, M.J. (1970), “Marginal stockholder tax rates and the clientele effect”, Review of Economics & Statistics, Vol. 52 No. 1, pp. 68-74.
Elton, E.J., Gruber, M. and Blake, C. (2005), “Marginal stockholder tax effects and ex-dividend-day price behavior: evidence from taxable versus nontaxable closed-end funds”, Review of Economics and Statistics, Vol. 87 No. 3, pp. 579-586.
Fama, E.F. and French, K.R. (2002), “Testing trade-off and pecking order predictions about dividends and debt”, Review of Financial Studies, Vol. 15 No. 1, pp. 1-33.
Ferreira, W.O. Jr, Martin, D.M.L. Nakamura, W.T. and Bastos, D.D. (2010), “Evidências empíricas dos fatores determinantes das políticas de dividendos das firmas listadas na Bovespa”, Facef Pesquisa, available at: http://legacy.unifacef.com.br/facefpesquisa/2010/nr2/vol13_nr02_art03.pdf
Fiorati, A.R.S., Garcia, F.G. and Filho, E.T. (2007), “Dividendos e juros sobre capital próprio: sinalização de lucratividade futura? Estudo no mercado brasileiro 1999/2004”, XXXI Encontro Nacional dos Programas de Pós-Graduação em Administração, Rio de Janeiro, p. 31.
Forti, C.A.B. and Schiozer, R.F. (2011), “Do Brazilian banks use dividends for signaling?”, XI Encontro Brasileiro de Finanças, available at: http://bibliotecadigital.fgv.br/ocs/index.php/ebf/11EBF/paper/viewFile/2658/1242
Freire, D.R., Rodrigues, R.N., Lagioia, U.C.T., Miranda, L.C. and Araújo, D.M.D.C. (2010), “Governança corporativa e a crise financeira: um estudo sobre o comportamento de empresas do setor de energia que possuem boas práticas de governança corporativa em momentos de turbulência”, Registro Contábil, Vol. 1 No. 1, available at: www.seer.ufal.br/index.php/registrocontabil/article/view/167/150
Futema, M.S., Basso, L.F.C. and Kayo, E.K. (2007), “A estrutura de capital, os dividendos e os juros sobre capital próprio: evidências empíricas no caso brasileiro de 1995 a 2004”, XXXI Encontro Nacional dos Programas de Pós-Graduação em Administração, Rio de Janeiro, p. 31.
Futema, M.S. Basso, L.F.C. and Kayo, E.K. (2009), “Estrutura de capital, dividendos e juros sobre o capital próprio: testes no Brasil”, Revista Contabilidade & Finanças – USP, Vol. 20 No. 49, available at: www.scielo.br/pdf/rcf/v20n49/04.pdf
Gonzaga, R.P. and Da Costa, F.M. (2009), “A relação entre o conservadorismo contábil e os conflitos entre acionistas controladores e minoritários sobre as políticas de dividendos nas empresas brasileiras listadas na Bovespa”, Revista Contabilidade & Finanças - Finanças, Vol. 20 No. 50.
Graham, J.R. and Harvey, C.R. (2001), “The theory and practice of corporate finance: evidence from the field”, Journal of Financial Economics, Vol. 60 Nos 2/3, pp. 187-243.
Hsiao, C. (1986), Analysis of Panel Data, Cambridge University Press, Cambridge.
IOB Thomson (2005), Anuário Mapa Fiscal, Tabelas Práticas e Instruções, IOB, São Paulo.
Kim, B.S. and Maddala, G.S. (1992), “Estimation and specification analysis of models of dividend behavior based on censored panel data”, Empirical Economics, Vol. 17 No. 1, pp. 111-124.
Lintner, J. (1956), “Distribution of incomes of corporations among dividends retained earnings, and taxes”, American Economic Review. Menasha: The American Economic Association, Vol. 46 No. 2, pp. 97-113.
Litzenberger, R.H. and Ramaswamy, K. (1979), “The effect of personal taxes and dividends on capital asset price: theory and evidence”, Journal of Financial Economics, Vol. 7 No. 2, pp. 163-195.
Mota, D.C. and Eid, W. Jr (2007), “Dividendos, juros sobre capital próprio e recompra de ações: um estudo empírico sobre a política de distribuição no Brasil”, XXXI Encontro Nacional dos Programas de Pós-Graduação em Administração, Rio de Janeiro, p. 31.
Nakamura, W.T., Martin, D.M.L. and Ferreira, W.O. Jr (2007), “Evidências empíricas dos fatores determinantes das políticas de dividendos das firmas listadas na Bolsa de Valores de São Paulo”, VII Encontro Brasileiro de Finanças, São Paulo, p. 7.
Nossa, V. and Nossa, S.N. (2007), “O impacto da distribuição de dividendos e da recompra de ações na eficiência das empresas listadas na Bovespa (1995 a 2004)”, XXXI Encontro Nacional dos Programas de Pós-Graduação em Administração, Rio de Janeiro, p. 31.
Pohlmann, M.C. and Iudícibus, S. (2010), “Relação entre a tributação do lucro e a estrutura de Capital das grandes empresas no Brasil”, Revista Contabilidade & Finanças – Finanças, Vol. 21 No. 53, available at: www.scielo.br/pdf/rcf/v21n53/02.pdf
Procianoy, J.L. (1996), “Dividendos e tributação: o que aconteceu após 1988-1989”, Revista De Administração da Universidade de São Paulo, Vol. 31 No. 2, pp. 7-18.
Procianoy, J.L. and Poli, B.T.C. (1993), “A política de dividendos como geradora de economia fiscal e do desenvolvimento do mercado de capitais: uma proposta criativa”, Revista de Administração de Empresas, Vol. 33 No. 4.
Procianoy, J.L. and Snider, H.K. (1994), “Tax changes and dividend payouts: is shareholders’ wealth maximized in Brazil?”, The European Financial Management Association’s 4th Annual Conference, London, p. 4.
Rabelo, S.S.T., Rogers, P., Ribeiro, K.C.D.S. and Securato, J.R. (2007), “Performance das melhores práticas de governança corporativa no Brasil: um estudo de carteiras”, Facef Pesquisa, Vol. 10 No. 2, available at: http://legacy.unifacef.com.br/revistas/index.php/facefpesquisa/article/viewFile/96/161
Ramos, C.F. (1997), “A influência da tributação sobre o retorno das ações em função da distribuição de dividendos feita pelas companhias negociadas na Bovespa: um estudo do mercado brasileiro”, XXI Encontro Nacional dos Programas de Pós-Graduação em Administração, Rio das Pedras, p. 21.
Silva, A.L.C. (2004), “Governança corporativa, valor, alavancagem, e política de dividendos das empresas brasileiras”, Revista de Administração- USP, Vol. 39 No. 4.
Srour, G. (2005), “Práticas diferenciadas de governança corporativa: um estudo sobre a conduta e a performance das firmas brasileiras”, Revista Brasileira De Economia, Vol. 59 No. 4, pp. 635-674, available at: http://bibliotecadigital.fgv.br/ojs/index.php/rbe/article/view/982/528
Vancin, D.L. and Procianoy, J.L. (2016), “Os fatores determinantes do pagamento de dividendos: o efeito do obrigatório mínimo legal e contratual nas empresas Brasileiras”, Revista Brasileira De Finanças, Vol. 14 No. 1, pp. 89-123, available at: http://bibliotecadigital.fgv.br/ojs/index.php/rbfin/article/download/53448/59139
Further reading
Brazil (1991), “Congresso Nacional. Lei N° 8.383, de 30 de dezembro de 1991. Institui a Unidade Fiscal de Referência, altera a legislação do imposto de renda e dá outras providências”, Diário Oficial da União. Brasília, 12/03/1991.
Brazil (1998), “Presidência da República. Subchefia para Assuntos Jurídicos. Decreto n° 2.673, de 16 de julho de 1998. Dispõe sobre o pagamento, pelas empresas estatais, federais, de dividendos ou de juros sobre o capital próprio”, Diário Oficial da União. Brasília, Art. 1 a 8, 7/16/1998.
Brazil (1999b), “Secretaria da Receita Federal. Instrução Normativa SRF n° 012, de 10 de fevereiro de 1999. Dispõe sobre os juros pagos ou creditados a título de remuneração do capital próprio, e dá outras providências”, Diário Oficial da União. Brasília, 2/12/1999.
Acknowledgements
The authors are grateful to CAPES, CNPQ and FAPERGS for the financial support in the development of this research.