Ownership concentration, agency conflicts, and dividend policy
in Japan
1. Introduction
Agency conflicts play an important role in corporate decisions. In their seminal paper, Jensen and Meckling (1976) show that decision makers may select value-decreasing outcomes that are justified only because of the wealth transfer from other stakeholders. To give a well-known example, managers may invest in unprofitable projects whose costs are borne by shareholders if doing so enhances their own status and bring them private benefits. Likewise, shareholders may take excessive risk knowing that the downside is assumed by debt holders, while they benefit from the upside.
Agency conflicts can take many other forms. Myers and Majluf (1984) argue that firms acting in the interest of current shareholders should rationally pass up profitable investment projects if the benefits are captured by outside investors. Shleifer and Vishny(1997) emphasize the agency conflicts between majority and minority shareholders, prompted by recent evidence that dominant shareholders extract rents at the expense of small shareholders through the tunneling of assets and profits, such as the use of unfair transfer pricing between controlled entities. Johnson et al. (2000) provide a few examples of expropriation taking place in developed countries.
Payout policy is one area of corporate decisions that cannot escape the influence of agency conflicts. In fact, Easterbrook (1984) argues that dividends can be either the result or the solution to agency conflicts. Because managers prefer to retain earnings to increase private consumption or reduce the risk on their human capital, low governance standards and poor shareholder protection are likely to result in lower payout. This view is clearly established in the cross-country analysis of La Porta et al. (2000). Conversely, dividend payout can contribute to mitigate agency conflicts. Jensen (1986) advocates to lower the free cash flows available to managers in order to increase financial discipline. Paying high dividends precisely achieves this purpose, thus providing a cost effective substitute to shareholder monitoring.
More recently, Gugler and Yurtoglu (2003) highlight the importance of agency conflicts between shareholders by looking at the dividend policy of German firms. Since dominant shareholders can extract private benefits from the cash flows under their control, their preference translates into lower dividends. However, the presence of another large shareholder can restrain the rent extraction, leading eventually to a higher payout.
The purpose of this paper is to examine the dividend policy of Japanese firms from an agency perspective. If ownership concentration is consistent with the alignment of interest between management and shareholder, as studies of corporate performance have suggested, there
should be a higher dividend payout. However, ownership concentration can also facilitate rent extraction by dominant shareholders, resulting in lower payout.
Our results support the second hypothesis. Ownership concentration is associated with significantly lower dividend payments in proportion of operating earnings as in proportion of book equity. In effect, the difference between the high concentration and low concentration groups is found to be in the order of 10%.
We investigate the reasons for this difference, focusing on the role of profitability, growth opportunities and changes in leverage in explaining the decision to change dividends. Our analysis uncovers a number of agency conflicts. First, tightly controlled firms (i.e., firms with concentrated ownership) are less likely to increase dividends when profitability increases and when operating profits are negative. This pattern is consistent with their lower payout and the assumption that dominant shareholder extract private benefits from resources under their control.
We also find that tightly controlled firms are more likely to omit dividends when investment opportunities improve, which protects the interest of current shareholders. Clearly, this decision reduces the likelihood of requiring further funding that would benefit outside investors, hence preventing the under-investment problem present in more
likely to increase dividends when debt levels are high and less likely to omit dividends when debt increases, which is equivalent to a wealth transfer from debt holders to shareholders since it decreases the amount of collateral backing the firm’s debt.
Overall, the analysis of the dividend adjustment decision provides some reasons for the lower payout associated with ownership concentration. More importantly, perhaps, the results suggest that as in the case of many corporate decisions, dividend policy is heavily determined by agency conflicts between majority shareholders and other stakeholders. In particular, the emblematic debt holder-shareholder conflict appears to be exacerbated by the presence of dominant players able to coordinate the actions of shareholders.
The rest of the article is structured as follows. Section 2 articulates the hypotheses regarding the relationship between ownership structure and dividend policy. Section 3
Describes the sample and date sources. The empirical results are presented in Section 4.Section 5 concludes.
Relationship between ownership and dividends 2.1 Positive relationship
According to Shleifer and Vishny (1986), ownership concentration creates the incentives for large shareholders to monitor the firm’s management, which overcomes the free-rider problem associated with
dispersed ownership whereby small shareholders have no incentives to incur monitoring expenses for the benefit of other shareholders. Because of strict financial discipline, firms improve their capital allocation, reduce unprofitable investments and ultimately exhibit higher performance. Indeed, several studies (e.g., Claessens and Djankov, 1999) show that concentrated ownership contributes to higher profitability and market valuation.
The relevant consequence of financial discipline is that fewer resources are consumed in low return projects and more cash flows can thus be distributed as dividends. In support of this interpretation, Mitton (2005) show that firms with better corporate governance pay higher dividends in emerging markets. Likewise, La Porta et al. (2000) show that in countries with better shareholder protection, like the US, firms pay more dividends. Easterbrook (1984) suggests that dividend payments can be a substitute for monitoring.
Large shareholders have the power to constrain firms to disgorge excess cash flows in order to reduce monitoring expenses, resulting in the same positive relationship, but with reverse direction of causality. This interpretation is supported by the lower cash holdings of better-governed firms and firms with concentrated ownership (Dittmar et al., 2003).
Following these arguments, it is possible to formulate the hypothesis
that ownership concentration is associated with higher dividend payments.
2.1 Negative relationship
The opposite view is that well governed firms do not need to pay higher dividends to increase financial discipline or enhance the alignment of interest between managers and shareholders. In the absence of agency conflicts, shareholders can be confident enough that the firm’s cash flows are properly used. Hence, the higher dividend payout advocated by Easterbrook (1984) does not appear to be essential to disciplining management.
Consistent with this view, Jensen et al. (1992) show that insider ownership is associated with significantly lower dividend payout among US firms. Farinha (2003) documents a similar negative relationship in the UK. Chen et al. (2005) provide evidence that some indicators of governance quality (existence of audit committee and percentage of independent directors) negatively affect dividend payouts in Hong Kong. These studies strongly suggest that the higher alignment of interest between managers as agents and shareholders as principals should actually result in lower dividend payments.
What's more, agency theory has recently highlighted possible conflicts between large and small shareholders. Shleifer and Vishny (1997) emphasize that large shareholders prefer to generate private benefits of control that are not shared by minority shareholders. Johnson et al. (2000)
give several examples of controlling shareholders expropriating minority shareholders of profitable business opportunities. Claessens and Djankov (1999) explain downward-sloping firm value at high levels of ownership concentration by the potential risk of expropriation by controlling shareholders.
Gugler and Yurtoglu (2003) show that the lower dividend payout of majority-controlled firms in Germany is related to the probability that controlling shareholders extract private benefits at the expense of minority shareholders. Indeed, they find that increases in dividend payments are associated with significantly positive abnormal returns among firms where rent extraction is most likely given the discrepancy between cash flow rights and control rights. Furthermore, the presence of a second large shareholder contributes to increase the distribution of profits, as it decreases the scope of expropriation.
Maury and Pajuste (2002) document a similar negative association between ownership concentration and dividend payments in Finland, as well as evidence of the mitigating role of another large shareholder.
In a similar way, we can hypothesize that firms with concentrated ownership are associated with lower dividend payments.
Ownership concentration and dividend payout
In this section, we examine the effect of ownership concentration on dividend payout controlling for other firm characteristics that could
influence payout policy.
Table 3 show that the coefficients on the Herfindhal index (LHH) and its associated dummy (Q2H) are both significantly negative. The coefficient on Q2H for the dividend payout ratio is less negative than the 2.5% in university analysis; but the coefficient for dividend yield is more negative. The difference in dividend payment as proportion of equity is seen to be about 10% across the two groups.
These results do not support the role of dividends as substitute for shareholder monitoring suggested by Easterbrook (1984). In fact, firms with concentrated ownership, which are more likely to be closely monitored, actually distribute lower dividends. This pattern is more consistent with the claim by Shleifer and Vishny (1997) that dominant shareholders prefer to extract private benefits, such as synergies with other controlled entities, rather than receive dividends that equally benefit minority shareholders.
The results are consistent with Gugler and Yurtoglu (2003) who report that majority controlled in Germany have lower payouts. Maury and Pajuste (2002) also find that the cumulated ownership of the three largest shareholders has a negative effect on the dividend payout of Finnish firms.
Among control variables, firm size appears to reduce both the dividend payout and the dividend yield. A comparable effect, with a similar order
of magnitude, is reported by Gugler and Yurtoglu (2005) regarding German firms. Consistent with Gul (1999) the influence of profitability is positive for dividend yield, but negative for dividend payouts. This could be explained, following Lintner (1956), by the fact that dividends are sticky. Firms that are more profitable have higher dividend payments, although the dividends are lower in proportion of (their higher) earnings. Controlling for endogenously, Jensen et al (1992) find that ROA has a positive effect on the dividend payouts of US firms. Taken together, it is more likely that the true role of profitability is reflected by its effect on dividend yield.
Both measures of growth opportunities are seen to have a positive effect on dividend payout and yield. The result is in sharp contrast to studies regarding US firms. Fama and French (2002) and Jensen et al. (1992) report negative coefficients for growth proxies.Farinha (2003) also report a negative effect on the dividend payout of UK firms, although the results appear to be sensitive to the model specification. On the other hand, Gul (1999) shows that growth opportunities have a positive and significant effect on the dividend yield, but an insignificant effect on dividend payout.
Benito and Young (2003) offer an interesting explanation to this puzzle by controlling for firm fixed effects (unobservable characteristics). Examining a sample of UK firms, they show that
non-payers are most often high growth firms, which have never paid dividends, rather than troubled firms seeking to repair their balance sheet by cutting dividend payments. It is more likely that Japanese firms adjust their dividend policies to business conditions, consistent with Dewenter and Warther (1998).
Finally, leverage is seen to have a negative effect on dividend payouts as in Jensen et al.(1992). Gugler and Yurtoglu (2003) report a more negative sensitivity of about -0.46 for German firms compared with our result of -0.26 for Japanese firms. The positive coefficient on dividend yield seems to be due to the lower denominator (equity ratio) among highly-leveraged firms.
Although OLS is the commonly reported method for analyzing dividend payments, the fact that the dividend-to-equity ratio is bounded below zero can mean that OLS provides biased estimates. Following Barclay et al. (1995), we apply Tobit regressions to the censored dependent variable DYLD to check that we have consistent results.
Although 18.2% instances of observations are clustered at zero, the results appear qualitatively similar. Both proxies of ownership concentration Q2H and LHH are significantly negative with slightly more negative coefficients. Profitability (ROA) asset growth (GROW) and leverage (DEBT) retain their level of significance, but firm size and growth opportunities (Q) become less significant relative to OLS results.
Overall, the analysis confirms the negative influence of ownership concentration on dividend payouts.
Source: Kimie Harada, Pascal Nguyen, 2006“Ownership concentration, agency conflicts, and dividend policy in Japan” JEL Classifications: G35 .Working Paper Series
译文：
日本的股权集中度，代理冲突及股息政策
代理冲突在公司决策中发挥着重要作用。在他们的开创性论文中，延森和迈克林（1976）表明，决策者可以选择以价值递减的结果仅仅因为从其他利益相关者转移合理的财富。举一个众所周知的例子，管理者可以投资无益的由股东承担
的项目成本，如果这样做提高自己的地位并且带给他们私人利益。同样，股东可能采取由债权人承担过度的风险，而他们从上部受益。
代理冲突可以有许多其他形式。迈尔斯和麦吉罗夫（1984）认为，企业在对现有股东的利益行事时，应理性地放过有利可图的投资项目，如果他们被外部投资者所捕获。胥利夫和维什尼（1997）强调多数与少数股东之间的代理冲突，最近的证据表明，主导股东以牺牲小投资者的资产和利润来提取租金，如使用控制实体之间的不公平转让定价。约翰逊等人 （2000）提供一个在发达国家进行征用的几个例子。
股利政策是决定一个企业无法逃避的代理冲突影响的因素。事实上，伊斯特布鲁克（1984）认为，红利可以是结果也可以是代理冲突的解决方案。因为管理者倾向于保留盈利增加私人消费或减轻对他们的人力资本风险，低治理标准和对贫困股东的保护可能导致较低的支出。这种观点显然是建立在拉帕塔等人的跨国分析（2000年）。相反，派息有助于缓解代理冲突。詹森（1986）主以降低现金流量有效的提供给管理人员，以增加财政纪律。支付高额股息正是达到这一目的，从而给股东的监管提供了一种成本效益的替代品。
最近，古乐和郁塔路（2003）强调通过在德国的公司股利政策的股东之间寻找代理冲突的重要性。由于主导股东可以在其控制下从私人收益中提取现金流量，他们倾向于转换低股息。不过，另一大股东的存在可以抑制提取租金，最终导致更高的支出。
本文的目的是从机构的角度研究对日本企业的分红政策。如果股权集中度与管理层和股东之间的利益调整一致，作为企业绩效的研究建议，应该有更高的派息。但是，股权集中度也可以方便提取主要股东的租金，从而降低支出。
我们的结果支持了第二种假设。股权集中度是与经营中的股东权益账面价值的比例相关的。实际上，两者的高浓度，低浓度组差异在10％的顺序被发现的。 我们研究这种差异的原因，重点是对盈利能力的利用，成长机会和解释变化对股息改变决定的影响。我们的分析揭示了机构的一些冲突。首先，利润的增加和营业利润为负时严格控制企业（即与集中所有权公司）不太可能增加分红。这种模式是在与他们较低的支出和股东占主导地位的前提下提取其控制下的资源的私人利益是一致的。
我们还发现，严密控制的企业更容易忽略改善分红时的投资机会，这保护了现有股东的利益。显然，这一决定需要进一步降低资金，这将有利于提高外来投资者的可能性，因此，防止更多的分红可能增加的投资不足问题，当目前的债务水平很高并且不太可能省略股息时，这相当于把债权人的财富转移给股东，因为它降低了支持该公司的债务担保金额。
总体而言，股息调整提供了决策分析与股权集中度较低的支出相关联的一些理由。更重要的是，也许，结果表明，在许多公司决策的情况下，股利政策在很大程度上取决于大股东和其他利益相关者之间的代理冲突。特别是，象征债券股东持有人的冲突似乎会加剧，因为存在显著能协调股东利益的人。
这篇文章的其余部分的结构如下：第2条假设有关的产权结构和股利政策的关系。第3节描述样本和日期来源。实证结果显示在第4.第5部分的结论。 所有权和分红之间的关系 积极关系
根据胥利夫和维什尼（1986），股权集中度创造了大股东的监督激励公司的管理，克服了搭便车与股权分散的问题，使小股东没有激励去监测股东利益的其
他开支。由于严格的财务纪律，企业提高资本配置，减少无利可图的投资，最终表现出更高的性能。事实上，若干研究（如科莱森和达坚科，1999年）表明，集中的所有权能够获得较高的盈利能力和市场评价。
有关的财务纪律的结果是，在低回报的项目中消耗较少的资源和更多的现金流消耗，因此可以作为股利分配。在这种解释的支持下，米顿（2005年）表明，在新兴市场更好的治理公司与支付企业更高的股利。同样，拉帕塔等人（2000年）表明，在有些国家更好的保护股东，如美国，支付企业更多的股息。伊斯特布鲁克（1984）认为，分红可以成为监测的替代品。
大股东拥有限制公司交出多余的现金流量，以减少监测费用的权利，导致相同的正相关关系，但方向相反的因果关系。这种解释是支持更好的管治现金持有量的公司和集中所有权的企业（迪特马尔等，2003）。
按照这些论点，是有可能的制定所有权集中度较高的股息支付相关的假说。 消极关系
与此相反的观点是，治理良好的公司并不需要支付更高的股息，增加财政纪律或提高管理者与股东利益。在代理冲突的情况下，股东可以有足够的信心，使公司的现金流量用得其所。因此，伊斯特布鲁克（1984）所倡导的较高的股利支付是纪律管理必不可少的。
根据这一观点，延森等人（1992年）表明，部人持股与美国企业之间的较低的派息有关。法尼哈（2003）记载在英国存在着类似的负相关关系。等人（2005年）提供的证据表明，治理质量（审核委员会及独立董事比例的存在）的一些指标对派息产生负面影响。这些研究有力地表明，作为代理商之间的管理者和作为委托人的股东利益导致较低的分红。
更何况，代理理论最近强调大，小股东之间可能发生的冲突。胥利夫和维什尼（1997）强调，大股东倾向于生成对并不是由少数股东共享的私人利益的控制。约翰逊等人（2000年）提供了利用有利可图的商业机会剥夺少数控股股东的几个例子。科莱森和达坚科（1999）解释由控股股东侵占的潜在风险下，向下倾斜的所有权集中在高层次企业的价值中。
古乐和郁塔路（2003）表明，多数控股企业在德国的低股利支付是与控股股牺牲小股东的私人利益的可能性相关的。事实上，他们发现，在增加支付股息与提取租金之间，其中最有可能由于是由于现金流权和控制权的企业差异存在着正相关的异常报酬。此外，第二大股东的存在有助于提高利润的分配，因为它降低了征收围。
莫瑞和帕驹斯（2002）文档之间的负相关类似的所有权集中在芬兰和分红，以及另一大股东缓解作用的证据。以类似的方式，我们可以推测，集中所有权与较低的股息支付有关。 股权集中度和分红派息
在本节中，我们将探讨所有权集中度可能会影响公司派息政策的控制支出的影响。
表3表明，在赫希曼指数（LHH）及其相关道具（Q2H）的系数都存在显著的负相关。关于Q2H派息比率的系数小于单因素分析中的负2.5％，但对股息收益率系数是负面的。按股权比例分红，在这两个群体约10％被认为是作为支付差额。 这些结果不支持，作为供股东监督替代股息的作用的建议（伊斯特布鲁克（1984））。事实上，股权集中的公司，这是更容易受到密切监视，实际上降低了股利分配。这种模式是更符合由胥利夫和维什尼（1997）中提到的主要股东应该
倾向于私人利益的要求相一致，如与其他控制实体的协同作用，而不是平等地从少数股东获益分红。
其结果与古乐和郁塔路（2003）谁在德国控制了过半数控制权就可能有较低的支出一致。莫瑞和帕驹斯（2002）也发现，三个最大的股东累计所有权对芬兰公司派息有负面影响。在控制变量中，企业规模都出现减少派息和股息率的情况。一个类似的效果与一个数量级类似的命令，是由古乐和郁塔路（2005）关于德国公司提出的。
与居尔一致（1999年）的盈利能力在股息收益率的影响是正面的，但对股利发放是负面消极的。这可以解释林特纳（1956）提出的股息是有粘性的这一事实。因为公司有更多的利润，较高的股息支付，尽管股息（较高）占收入的比例较低。控制的生性，延森等人（1992年）发现，资产收益率对美国公司的派息有积极的影响。综合考虑，盈利的真正作用更可能是由它的股息收益率的影响反映出来的。
这两项措施的成长机会可以看成对股息支付和收益有积极影响。其结果是对于美国，法玛和法国（2002）形成了鲜明的对比，延森等人（1992年）报告代理负增长系数。法尼哈（2003）也报告了关于英国公司派息的负面影响，但是结果似乎是对模型设计是敏感的。另一方面，居尔（1999年）表明，增长机会对股息收益率有着积极显著的影响，但对派息影响不大。
贝尼托和青年（2003年）提供了一个有趣的解释，对于这个难题由企业固定效应（不可观察特征）控制。检查英国公司的样本，他们表明，非纳税人经常是高增长企业，它们并不是从未支付股利，而是试图通过削减分红弥补其资产负债表的困境。它更可能是根据日本企业的状况的分红政策调整，德文特和沃特
（1998年）提出的相一致。
最后，杠杆效应被延森等人认为在对股利发放上有负面影响（1992年）。古乐和郁塔路(2003)报告了一个更多负面的敏感性的比较德国企业为-0.46与此相比日本公司为-0.26。股息收益率上的积极系数似乎是由于在高杠杆企业中较低的分母（股权比例）。
尽管18.2％的观测实例在零聚集，结果会出现质的相似。所有权的集中Q2H和LHH代理是两个显着轻微的负面的系数。盈利率（资产收益率）资产成长（成长）和杠杆（债务）保留其显著性水平，但企业规模和成长的机会（Q）相对与最小平方法变得不那么重要。
总体而言，分析确认了派息股权集中的负面影响。
出处: 科姆原田, 帕斯卡尔阮,《日本的股权集中度，代理冲突及股息政策》, 中图分类号:G35工作论文系列.2006
因篇幅问题不能全部显示，请点此查看更多更全内容