nueva página del texto (beta)
Inglés (pdf)
Artículo en XML
Referencias del artículo
Traducción automática
Enviar artículo por email
Citado por SciELO 
Similares en
SciELO 
Permalink
1. Introducción
The dividend policy has been one of the most debated topic in the financial literature. Although several researchers have developed theories and provided us with empirical evidence, there is still lack of a general consensus. In 1976, Fisher Black correctly predicted that the bigger the emphasis academics put on the topic the bigger the debate will appear. Different explanations to the problem have been proposed. According to Bhattacharya (1979) the payment of dividends is the instrument used by the companies to give signals of profitability’s perspectives to the investors. In other words, if a company stops paying dividends or reduces them, once a dividend policy has been declared, it will experience a drop in its share price. Later, the company could also experience higher costs in issuing new stocks.
The companies’ dividend policy is important for several reasons: first a company uses the dividends to give signals about its stability and growth opportunities to the investors; second dividend policy plays an important role in the firm’s capital structure because the latter could be modified by the former and finally dividends affect the company’s stock price (Ghosh & Woolridge, 1988 and 1991).
On the other side, there are several theories that relate the dividends’ payment to the agency costs (La Porta, Lopez-de-Silanes, Shleifer and Vishny, 2000). The agency theory points out that the dividend payout could be diminished because free cash flows might have been invested in non-profitable projects by the firm’s management (Jensen, 1986). Hence, there is a link between dividend distribution and the discretionary decision of what to do with available free cash flows.
According to Gompers, Ishii and Metrick (2003) agency costs are linked to the strength of shareholders’ rights, namely to the firm’s corporate governance. Therefore, it is interesting to study whether the adoption of good corporate governance practices is linked to the dividend distribution of companies. Good Corporate Governance practices aim to attract capitals to guarantee proper managerial practices and to protect investor’s rights with the objective of promoting trust in the financial markets and competitiveness.
Managers operating in Peru are not unaware of the discussions about good corporate governance practices. In fact, during the years, the local corporate governance regulation has been adapting to these new trends, focusing the efforts on the accomplishment of the international standards to protect minority investors given the high ownership concentration of the companies at the Lima Stock Exchange (LSE).
In July 2008 the LSE launched the Good Corporate Governance Index (GCGI) that is built upon a survey questionnaire on the 26 principles of good corporate governance for Peruvian firms made by the Securities Exchange Superintendence
Fuenzalida, Mongrut, Arteaga and Eurasquin (2013) find that the announcement of the inclusion of a listed company in the GCGI yields a positive abnormal return of about 1% on the day of the announcement. These authors also find that investing in a portfolio of the best performers (democratic portfolio) in the GCGI outperformed consistently in about 3% annually the portfolio of investing in the worst performers (autocratic portfolio) in the GCGI during the years 2004-2008.
Furthermore, Milton (2004) shows that the implementation of a Corporate Governance Code leads to a higher dividend payout ratio in emerging countries. Similarly, Kowalewski, Stetsyuk and Talavera (2008) demonstrate the existence of a strong correlation between the quality of the corporate governance code and the dividend payout ratio.
Given the above results, the main objective of this research is to analyze if there is any link between the adoption of a corporate governance code and the company’s dividend payout ratio in Peru. The study goes further into examine whether the quality of the adopted corporate governance code is linked to the dividend payout ratio. The main contribution of this research is to establish this link for companies operating in Peru.
The next section contains a short discussion of the relevant literature and the development of the two main hypotheses. The third section one describes the methodology applied and offers the main results and the last one examines the implications of the study and concludes the study.
2. Literature Review
La Porta, Lopez-de-Silanes, Shleifer and Vishny (2000) establish that managers prefer to retain dividends or to reduce the risk of human capital loss, implying that low dividend payout ratios are associated with low corporate governance standards and with a poor protection of the stockholders’ rights. These authors believe that those firms located in countries with high legal standards of protection of the minority investors pay higher dividends compared to countries where legal protection is weak. It is likely that the severity of the agency cost would be inversely proportional to the strength of the shareholders’ rights (Gompers, Ishii and Metrick, 2003). Companies that are exposed to the agency problem tend to experiment higher gaps between property and control where the rights of the minority shareholders are more repressed.
With respect to the dividend policy and its relationship with corporate governance, Bebczuk (2005) states that the better the corporate governance practices of the firm, the higher the dividend payment. This is consistent with the results of Kowalewski, Stetsyuk and Talavera (2008) that demonstrate that the adoption of the Corporate Governance Code is an important determinant in the dividend policy of Polish companies. To measure the quality of the corporate governance they build a Transparency Disclosure Index (TDI) for 110 companies and find out that there is a strong positive correlation between the index and the dividend payout ratio.
According to Burkart and Panunzi (2006) when minority stockholder’s protection is weak, agency problems are very serious and they suggest the separation between the control and the property of a company. The threat lies in the fact that managers keep a strong preference for control and they do not promote the payment of dividends. Jang-Chul and Young (2011) use the agency theory to explore how the corporate governance quality affects the dividend policy. Their results show a strong positive association between the corporate governance’s quality and the dividend payment meaning that those companies with a stronger corporate governance policy tend to pay higher dividends.
However, according to Sáez and Gutiérrez (2014) dividend payout ratios decrease when the structure of the ownership is highly concentrated, evidence that might be explained by the absence of supervision’s mechanisms so controlling shareholders could take this advantage to expropriate minority shareholders. It seems that this fact it is not always true because according to Pindado, Requejo and de la Torre (2012) family-controlled firms distribute more dividends in order to diminish the takeover apprehension of minority shareholders.
Gonzales, Molina, Pablo and Rosso (2016), through a study that includes six Latin-American countries, confirm that the higher the ownership concentration, the fewer the dividends, although they also find out that if the largest shareholder lives in a common-law country, the dividend is significantly higher. They also suggest that a second large shareholder might acquire a monitoring role and its presence would decrease even more the dividend payout ratio. Koo, Ramalingegowda and Yu (2016) showed that also the quality of the financial reporting might act as a corporate governance mechanism that leads managers to distribute more dividends because it mitigates the free cash flow problems.
Although it is possible to argue a double causality between corporate governance and dividend policy it seems that the causality between both goes from corporate governance to shareholder return’s measures, among them the dividend payout (Shabbir and Padgett, 2008). This remark leads us to establish the following hypotheses:
H1: The dividend payout ratio increases as a consequence of the adoption of the Corporate Governance Code.
By combining the model developed by La Porta, Lopez-de-Silanes, Shleifer and Vishny (2000) and the theories related to the free cash flows distribution (Jensen, 1986), under the assumption that markets offer a weak protection to shareholders, it is possible to establish two possible explanations for the relationship between corporate governance and dividend policy (Easterbrook, 1984):
a) The “Substitute Model” hypothesis according to which the higher the shareholders’ rights the higher the dividend payout will be, meaning that there is a positive relationship between corporate governance and dividend policy.
b) The “Opportunistic Management” hypothesis, which states that managers tend to retain dividends with the aim of investing them in projects that might benefit them but not the shareholders. Under this assumption, one should expect an inverse relationship between corporate governance and dividend payout.
Given the big disparity that one might encounter in emerging economies, even among companies within the same country, concerning corporate governance practices it is necessary to establish a second hypothesis in order to favor one of the two previous explanations for the first hypothesis (Black, 2001). The second hypothesis must be related to the quality of the corporate governance code:
H2: The higher the quality of the Corporate Governance Code the higher the dividend payout of the company.
These two hypotheses will be tested in the next section
3. Methodology and Results
3.1 Data
For this study one used a sample of 111 non-financial firms from 2007 to 2015 listed on the Lima Stock Exchange (LSE). The choice of the time span lies on the fact that in July 2008 Peru has launched the Good Corporate Governance Index (GCGI), so it is important to explore the consequences of this change for the shareholders.
Companies in the sample mostly belong to the manufacturing sector (37.84%), others to the mining sector (15.32%) and other sectors (18.92%). These three groups are followed by public companies (11.71%), agricultural companies (10.81%), pension funds and insurance companies (5.40%). Their financial statements, ownership composition and dividends declarations come from Economatica and Bloomberg databases. The information related to the adoption of a corporate governance code comes from the LSE while the one concerning the quality of the code comes from the Securities Exchange Superintendence (Superintendencia del Mercado de Valores - SMV) in Peru.
In order to obtain control groups that will make possible a robust conclusion, the sample includes companies that pay and do not pay dividends and with and without a Corporate Governance Code. Additionally, the sample has been adjusted in order to withdraw observations with negative utility and extreme positive dividends (outliers) to avoid biases.
From the total sample of 111 firms 63 companies (57%) do not have a Corporate Governance Code, 41 companies (65%) paid dividends while 22 companies (35%) did not. The other portion corresponds to 48 companies (43%) that adopted a Corporate Governance Code of which 36 companies (75%) paid dividends and 12 companies (25%) did not.
Alternatively, one classified the companies that have paid dividends and the ones that did not, and then split them into those that adopted a Corporate Governance Code and those that did not. One obtains the following results: 34 companies (31%) did not pay dividends, of which 22 companies (65%) did not establish any Corporate Governance Code while 12 companies (35%) did. Of the 77 companies that paid dividends (69%), 41 companies (53%) do not have a Corporate Governance Code while 36 companies (47%) do.
This leads to the definition and the analysis of three different samples and, therefore, three different models: a general model with 111 companies, a model which includes the 63 companies without a Corporate Governance Code and a model which includes the 77 companies that pay dividends. The first subsample of 63 firms is examined in order to find out the significance of ownership concentration as a proxy variable for not having a Corporate Governance Code. The second subsample will help to understand the interaction between the adoption of the code and its quality.
3.2 Variables
Table No 1 shows all the variables used in this study. The dependent variable is the dividend payout ratio (Dpay). The literature offers several alternatives to measure it such as dividends over cash flows, dividends over sales and dividends over net profit. Authors such as La Porta, Lopez-de-Silanes, Shleifer and Vishny (2000) recommend using the ratio dividends over sales since it does not depend on the accounting principles, which might vary among countries. However, this study is not a cross-country one so differences in accounting principles won’t play a role. Hence, the chosen ratio is dividends over net profit, which measures how the net profit is distributed and reinvested in the period.
Due to the fact that our goal is to determine whether the adoption of a Corporate Governance Code can influence the dividend payout ratio, one uses a dichotomous variable (GOB) that assumes value one (1) starting from the year in which the Code has been adopted and zero (0) before that year.
Another goal of this study is to check the impact of the quality of the Corporate Governance Code (CGOB) into the dividend payout ratio, so one uses the results of the survey questionnaire conducted by the SMV. The questionnaire is based on 26 principles related to different topics such as shareholders’ legal rights, communication and information transparency, Board of Directors’ transparency, and so on. A score between zero (0) and four (4) is given to each principle where zero represents no compliance and 4 means full compliance to the principle. Based on this survey questionnaire, every firm in the sample has a weighted average number for every year, not all principles have the same weight, and this number is a proxy of the quality of the Corporate Governance Code.
Several papers such as the one of Ahmed and Javid (2009) and Nuhu (2014) have tried to explain the determinants of the dividend policy in emerging economies and in particular they have used the dividend payout ratio. They have encountered that the following variables usually pay a role in explaining the dividend payout ratio of companies operating in emerging economies: the firm size (SIZE), the degree of ownership concentration (CA), the company’s financial leverage (LEV) and the operating income (EBIT). Hence in this study one must include them as control variables.
The size of the company, usually measured as the natural logarithm of sales, has a positive effect on the dividend payout ratio since high sales’ volumes are usually associated to larger companies and to larger dividend payments.
Ownership concentration is usually negatively related to the dividend payout ratio this is because when ownership concentration’s levels are high, the risk of wealth expropriation against minority shareholders increases. In this situation majority shareholders will rather keep free cash flows than distribute them among minority shareholders.
Ownership concentration is usually measured through the Herfindahl Index:
Where
The main advantage of using this index is that it can be easily reproduced (Cubbin and Leech, 1983). Let
It is reasonable to expect a negative relationship between the company’s leverage ratio and the dividend payout ratio. In other words, a high leverage ratio implies a smaller payment of dividends, due to the fact that the free cash flow must be used to repay the principal and the interests of debts.
The operating income reflects the availability of resources allocable to new investments, which should increase the dividend payout ratio. Hence, an increase in the company’s operating efficiency, measured through the EBIT, represent an increase in available free cash flows and this could lead to a higher dividend payout, so the expected relationship between both is a positive one.
Finally, it is important to control for the companies’ environment, in other words for the macroeconomics effects. In particular, it is expected a positive relationship between the growth of the Gross Domestic Product (GDP) and the dividend payout, which means that as long as the economy grows, dividends will grow as well.
3.3 Descriptive Statistics
Due to the fact that in 97% of our sample companies the ownership is concentrated in the first five shareholders, one uses equations (1) and (2) to calculate the Herfindahl Index. As reported by Fuenzalida, Mongrut, Nash and Benavides (2008) companies in Latin America are characterized by high concentration levels from the first to the fifth shareholder: 53% to the first shareholder, 73% to the third and 79% to the fifth one. Our findings are consistent to this pattern: 62% to the first shareholders, 80% to the third and 97% to the fifth one.
The Corporate Governance Quality is measured through a score that is calculated only for the 48 companies that have implemented the Corporate Governance Code since year 2007 and that have been evaluated by the SMV through a survey questionnaire. These companies achieved an average grade of 3.57 on a scale that goes from 0 (no compliance to the principle) to 4 (complete compliance to the principle), with 2.29 as the minimum grade and 4 as the maximum grade. It has been encountered that companies that paid dividends registered an average rating of 3.65 points against companies that did not pay dividends with an average grade of 2.91 points.
Out of the 48 evaluated companies, 28 improved meaningfully their rating between 2007 and 2015, 11 kept it more or less stable and the remaining ones suffered a meaningful decrease. It is important to note that out of the 36 firms that paid dividends 37% improved their ratings, while out of the 12 that did not pay dividends, only a 17% improved their ratings.
Table No 2 offers the descriptive statistics of the above variables; it shows the average dividend payout ratio for the 77 companies with a CGC that is equal to 0.33 and it is in turn higher than the average dividend payout ratio of the 63 companies without a CGC that is equal to -0.34. A simple average of the dividend payout ratio by economic sector reveals that the pension funds register the higher dividend payout to their shareholders during the sample period (77%), followed by public utilities (67%). Manufacturing companies, mining companies, insurance companies, and agricultural companies follow have averages of 42%, 22%, 13.5% and 0.95%, respectively.
Table 2 Descriptive Statistics.
The table reports descriptive statistics (average, coefficient of variation-CV, maximum and minimum values) for four subsamples: the first subsample (A) includes the total number of companies; the second subsample (B) corresponds to all the companies without a Corporate Governance Code; the third subsample (C) includes companies that have adopted the code, and the last subsample (D) includes all the companies that have paid dividends at least once during the period.
Source: Own elaboration
From Table No 2, companies with a Corporate Governance Code have, on average, a leverage of 60%, slightly less than those that did not have it (61.9%). These results show that the relationship between leverage and the Corporate Governance Code is not a clear-cut as stated by Benavides and Mongrut (2010) for the case of Colombian companies. However, using the sample of companies that pay and do not pay dividends there is a significant difference in the level of leverage. Companies that pay dividends have an average leverage equal to 58.8%, while companies that do not pay dividends have an average level of leverage equal to 65.3%. This can be explained because highly leveraged companies prefer to pay lower dividends in order to avoid insolvency risk.
Concerning company’s size, one finds out those companies that paid dividends are bigger than those that did not. In particular, the former group is 241% bigger than the latter. Moreover, those that established a Corporate Governance Code are 202% bigger than those that did not.
With respect to the EBIT, companies with a Corporate Governance Code have a better operating performance than the others. In fact, they have, on average, an EBIT of 107.46 MM Soles as compared to companies without a code with an average EBIT of 47.99 MM Soles.
3.4 Results
One has decided to use an unbalanced Panel Data Model to test for the first hypothesis (H1) because in this way one can take full advantage of the sample. Hence, one performed the Breusch and Pagan test to model the error, whose null hypothesis is that the variance of the error, which does not vary through time but only through companies, is equal to zero. The null hypothesis was rejected, so one must model the non-observable heterogeneity.
Therefore, one used the Hausman test in order to choose between fixed-effect or random-effect model for heterogeneity, whose null hypothesis is that the non-observable effects are not correlated with the explanatory variables. The null hypothesis was rejected, which means that was better to use the fixed-effect model. Then, it was important to test for the presence of heteroskedasticity in the residuals, so one used the modified Wald test for fixed-effect Models, whose null hypothesis is that the volatility is equal for all the individuals and whose main advantage is that it works properly for unbalanced panel data. The null hypothesis was strictly rejected.
One also tried to account for time effects using dummy variables for the year of adoption of the code in order to model unobservable characteristics that do not change between companies but over time. The test of jointly significance of these variables was rejected, meaning it was not necessary to include time variables in the model.
Finally, one verified the presence of contemporary and serial correlation; the null hypothesis was rejected, so the model should be corrected for the heteroskedasticity and contemporary correlation. There are two methodologies to correct for this kind of problem in panel data. The first one is to include a dichotomous variable for each cross-sectional observation that is present in the sample. However, following DeAngelo and DeAngelo (1990), Fama and French (2001), Bebczuck (2005) and taking into account the possibility of an “incidental parameters” problem, stated by Neyman and Scott (1948) it is better to model without controlling for each company’s fixed-effect.
The second one is the generalized least square (GLS) regression with panel corrected standard errors. Beck and Katz (1995) demonstrated that the GLS estimators are corrected only for samples with a great amount of time units (years in this case) and few cross observations (number of firms). Although there is not an appropriate minimum time span that every cross section must have, Beck (2001) argues that a good number would be 10. Concerning the number of cross section units, there is not a minimum amount required, meaning that there is no need for them to be a large quantity. Because of these considerations, one used the GLS regression with standard errors corrected for the panel in order to correct for serial correlation and heteroskedasticity.
Equation (3) shows the first model (all variables previously defined):
where:
First panel (Model 1) of Table No 3 shows positive relationship between the adoption of the corporate governance code
Table 3 Estimation of the Models.
The dependent variable is the dividend payout ratio (Dpay). The table shows the results of the panel data model with standard errors corrected for heteroskedasticity and contemporary correlation. The sample of Model 1 and 2 is composed by 111 non-financial companies. Model 1 measures the impact of the adoption of the Corporate Governance Code (GOB). Model 2 measures the impact of the quality of the Corporate Governance code (CGOB). Model 3 only includes companies that did not adopt a CGC regardless if they pay dividends or not. Model 4: only includes companies that pay dividends regardless if they have adopted a CGC. The definition of the control variables are the ones reported in Table 1.
Source: Own elaboration
As expected, the company’s size
In order to test for the second hypothesis, the variable adoption of the code
where:
The second panel of Table No 3 (Model 2) shows results that support the second hypothesis. Hence, the higher the quality of the corporate governance code
The third panel of Table No 3 (Model 3) shows the results for the subsample of companies (63) that didn’t adopt a corporate governance code, regardless if they pay dividends or not. Given the fact that these companies didn’t adopt a
This variable could be thought as a substitute for the corporate governance of the company. The variable ownership concentration
Finally, the fourth panel in Table No 3 shows the results for the sample of companies (77) that pay dividends, regardless if they have corporate governance code or not. It is interesting to note that for this subsample the statistical significance and sign of the variables remain the same with the exception of the annual growth rate of the
4. Conclusion
This research aims to identify whether there is a link between the adoption and the quality of the Corporate Governance Code and the dividend payout for Peruvian firms, a relationship that has not been explored before. Through an unbalanced panel data model and the use of control groups, it has been possible to address the initial question, concluding that both factors have a significant impact on the dividend payout of the companies quoted of the Lima Stock Exchange (LSE).
This study also highlights the fact that ownership concentration could be a good proxy for corporate governance in companies that didn’t adopt any corporate governance code because the more concentrated is their ownership is, the lower the dividends they are going to pay. Furthermore, there seems to be a well-established group of companies with a declared dividend policy that they have to honor regardless if they have or not a corporate governance code.
During the years of the study (2007-2015), the Peruvian economy has grown 6% annually well above the average of Latin American countries. Hence, good investments opportunities came along and companies became more indebted and invested more in new projects instead of distributing more dividends unless they had a declared dividend policy. Larger firms distributed more dividends than smaller ones, most of them because they have adopted a corporate governance code due to the higher awareness of the average Peruvian investor with respect to corporate governance practices since the launch of the Good Corporate Governance Index in 2008.
It is interesting to note that the economic sector where the vast minority of investors is investing now (i.e. pension funds) is precisely the one that pays higher dividends. In other words institutional investors, when considering where to invest their fund, are perhaps already taking into account the corporate governance of the company. In particular, this study implies that they do not have to pay attention only to the adoption of a code but also to its quality especially in the case of companies with a non-declared dividend policy. Future line of research might concentrate efforts at the industry level in order to find out what explains differences in dividend payout. In addition, it might focus in explaining the relationship between a declared dividend policy and corporate governance.
