Dividend Policy and Fiduciary Duty

In 1990, Merton Miller was awarded with the Nobel Prize, together with Franco Modigliani (winner of the prize in 1985), for the development of a counterintuitive theory. The so-called Modigliani-Miller Proposition states that the value of a company does not depend on its capital structure. Among other consequences, it means that the dividend policy is irrelevant to the company.
In view of that, it’s curious to note that many of the recent initiatives to improve the corporate governance practices recommend the adoption of a dividend policy.
Firstly, all Brazilian listed companies must disclose their dividend payout practices in their Reference Form (item 3.4), including information (item e) about whether they “have a formally approved policy on the distribution of profits, the body responsible for its approval, date of approval and, in the event the issuer discloses the policy, the World Wide Web address where the document is accessible.”
It’s known that most companies simply repeat the legal text that deals with the legally required payout of 25% of profits, deducted from reserves. There are exceptions, including well-thought-out policies, such as Vale’s and Copasa’s. Other ones, such as Grendene’s, are specific to the company’s situation (allocate 100% of earnings, excluding tax incentives). Curiously, B3, the Brazilian stock exchange itself does not have such policy.
Leaving the legal nudge aside and talking about dividends, the Code of the Brazilian Institute of Corporate Governance brings, in its item 1.9, the following recommendation:
The company should disclose its Dividend Payout Policy and how often this document is revised. This policy, established by the Board of Directors and approved by the General Assembly, must contain, among other things: the frequency of payments, the benchmark used to define the amount (percentage of adjusted net income and free cash flow, among others); the process and the parties responsible for dividend payouts; the circumstances and factors that may affect a payout.
Along the same lines, the Brazilian Corporate Governance Code – Listed Companies also addresses the topic. Product of the perception of 13 entities related to the capital markets gathered in the so-called Interagentes Working Group, the Code was incorporated into the regulations of the Brazilian Securities and Exchange Commission – CVM. From October 2018 on, listed companies should disclose a Report (Informe) about the adoption of the code’s practices through the “comply and explain” system. One of the recommended practices is the adoption of a Profit Allocation Policy (Item 1.7). The recommendation is:
The company should prepare and disclose a profit allocation policy established by the board of directors. Among other aspects, such policy should provide the frequency of payments and the reference parameter to be used for definition of the related amount (percentage of the adjusted net profit and of the free cash flow, among others).
Why is there so much concern about dividends if Modigliani and Miller said they are irrelevant? Because, despite the Noble Prize, they are not so irrelevant as they may seem.
First of all, the laureates themselves introduced additional hypotheses into their models in an attempt to adjust them to the real world and show why dividends, in fact, make a difference. The first one is related to taxes.
There are only two ways through which shareholders receive a portion of a company’s profits: dividends or capital gains. Put simply, profits can be paid to shareholders or kept by the company, increasing its value in the market or in the in case of an acquisition. In most countries, the way these two options are taxed at very different rates; historically, in the United States, the taxes on dividends have always been higher than those levied on capital gains.
Additionally, the decision to pay dividends is tied to the company’s capital structure. When a company pays dividends, all else equal, its financial leverage increases through borrowing. The interests on these borrowings can be deducted from the income tax (unlike dividends), creating an incentive for companies to pay more dividends to increase leverage and reduce the average taxation.
Other reasons were added to the model. We can talk about the “clientele” effect: a company attracts a specific type of shareholder because of its dividend payout history. Assuming we are dealing with rational agents, the shareholder’s preference for the company is also related to taxes. In Brazil, for example, individual investors are exempt from paying taxes on dividends, that is, they have reasons to prefer companies with relevant and consistent payout histories. Accordingly, changes to the policy can either please or not this group of shareholders and, therefore, affect share prices.
Up to here, we are assuming companies and investors have access to the same information about the company – an easily questionable assumption. If we consider that this hypothesis is not entirely true (and get closer to reality), the dividend policy can play at least two additional roles. The first one is that dividends are a way of informing shareholders of the company’s operating perspectives. This is the so-called dividend signaling effect. If the company foresees bright future prospects, it increases dividend payments. On the other hand, when the perspectives are too risky, they reduce the payments. A consequence of this effect is that companies have difficulties in making changes to its short-term dividends (sticky dividends) as the market will try to understand what’s behind the decision.
Another situation in which dividends can mitigate the asymmetric information is related to the agency conflict, that is, when shareholders and directors – or, in markets like the Brazilian one, controlling and minority shareholders – have different objectives.  According to this view, an excessive retention of dividends could be the result of decisions that provide private benefits for the insiders, including from strategic decisions (entrenchment and gaining of power through a larger operation) to mundane decisions (more possibilities to obtain benefits, for example, acquiring a corporate jet). There are academic studies that show the relation between higher payouts[1] and countries with better investor protection mechanisms.[2]
In Brazil, although we are one of the few countries where the law establishes the payment of a minimum compulsory dividend (25% of the adjusted net profit), our companies’ payout is historically lower than the global average.[3] This reflects a trend among companies of retaining profits despite the legal and tax incentives (unlike the USA, the distribution of profits in Brazil is highly encouraged through attractive tax rates). If we consider this situation together with the academic findings, we can conclude that the possibility of minority shareholder expropriation encourages the retention of earnings. If we consider the cases of control premium in mergers and acquisitions (explicit or “hidden” in complex operations), it is easy to figure out the reasons for the retention. From this point on, we start to understand why the regulator and other entities that work to promote good practices are concerned about dividend decisions.
Finally, it is important to remember that even the assumption that the agents are rational can be questioned. Therefore, although investors can always “simulate” the dividend policy they prefer (for example, selling part of their stock in a company that does not pay dividends to create an income flow), many of them work with a “mental accounting system” that separates their main investment (shares) from the money made with dividends. All these factors should be considered by managers when they decide to make changes to their dividend policy.
The main thing that should be said about dividends is therefore that there is no single rule. It is fully justifiable that a high-growth and/or capital-intensive company does not pay any dividend. Shareholders benefit from the possibility that the operating cash generated is reinvested and, in fact, that is what they want. On the other hand, a mature company with a robust cash flow and few investment opportunities is expected to pay the biggest possible dividend to its shareholders.
From one extreme to the other, there are all types of situations. Any attempt to establish a formula for a dividend policy will have adverse impacts. And this is also true also for the Brazilian reality. Some companies may have no problem in paying the compulsory 25% of their earnings, while for other ones this percentage can be too high or even unfeasible.
Considering we are dealing with rational agents, the most accepted concept in the academic studies we analyzed is that the dividend is a residual variable in a capital structure decision. Based on long-term operating and investment plans (including contingency plans), officers should establish an ideal leverage level that takes into consideration tax aspects (tax rates, benefits, etc.), financial aspects (interest rates, risk spread, rating…), merger and acquisition opportunities, and the characteristics and volatility of the company’s market, including the capital intensity. The dividend can be the “adjustment” variable of this equation.
A legitimate decision-making process protected from asymmetric information and that aligns the interests of all stakeholders (management, controlling shareholder and minority shareholders [see above]) should be transparent and based on a formal policy – not on a dividend policy –, that is, on a capital structure policy that has the distribution of profits as one of its variables.
Regardless the existence or not of a formal policy to guide the decision to pay dividends, it’s important that the management considers the above-mentioned factors when taking the decision. It is not recommended to opt for the legal compulsory percentage without a further analysis as it can be good or bad for the company.
It is even more important to consider the implications of the directors’ duties of diligence (Article 154 of the Law 6,404/76) and loyalty (Article 155) with respect to the dividend decision. Put simply, directors should pay dividends that, after a deep understanding (diligence), meet the company’s best interests (loyalty). Accordingly, the value of the dividend should not be too low nor too high: it should be the ideal value.
Both extremes bring problems. A company that pays too few dividends can be incurring into agency problems: retaining earnings so that the controlling shareholder gets a disproportional portion of them through a control premium.
On the other hand, if the company pays excessive dividends, it can impair its financial health. The reader may be wondering: but why would a company do that? The answer is simple: to meet the interests of a controlling shareholder with high debts or that simply wants to have money available. There are two examples of extreme cases in the Brazilian market: Oi, which paid billions in dividends to its controlling shareholders and went bankrupt and Light, which almost went bankrupt after its privatization and whose operations were significantly affected.
Unfortunately, it seems that the lessons have not been learned: recently, it was announced that a heavily indebted company paid a significant extraordinary dividend to its even more heavily indebted controlling shareholder.
There are also cases of abuse on account of the signaling theory. Recently, Smiles, a non-capital-intensive company with a very strong cash generation capacity, has changed its dividend policy. The company, which used to distribute 100% of its earnings, started to retain a significant portion of them. Its shareholders, which valued the dividends not only as a signal of future activities, but also because they were a guarantee of control over conflicts of interests with controlling shareholders, were deeply disappointed and sold most of their shares. With that, Smiles’ stock dropped by around 15% in the days after the announcement.
Coincidence or not, 7 months later Smiles announced a corporate restructuring, through which, in practical terms, the controlling shareholders seek to increase their participation in the company and, therefore, a decrease in its value could be interesting for them. It is questionable whether the cut in dividends was part of a strategy of the controlling shareholder to decrease the company’s value.
Stories like this can teach us important lessons. For companies, they teach the advantages of having a well-thought-out policy based on the company’s reality and capital structure and that does not simply meet legal requirements. For directors – mainly members of Boards of Directors –, they show they should reflect on how payout decisions relate directly to their duties of diligence and loyalty. For investors, they teach the importance of understanding the companies’ governance practices and the quality of their management boards so that they can be trusted to take the best decision in behalf of all shareholders.
And for the market’s regulators, it is warning about how a company can actually go bankrupt because of an apparently legal decision that, if not based on the duties of loyalty and diligence, hides a transfer of values that has a deeply negative impact on the market. Managers involved in cases like that should be punished.
[1] Amount of dividend divided by the fiscal year earnings.
[2] See mainly Lopes de Silanes, Shleifer and Vishny (1998), available in https://papers.ssrn.com/sol3/papers.cfm?abstract_id=52871
[3] It is important to mention that this trend has shifted in recent years.