09 November, 2026

Quora Answer: What are the Various Theories of Dividend Policy?

The following is my answer to a Quora question: “What are the various theories of dividend policy?”

A dividend is a cheque a company writes to its shareholders.  Except that behind that cheque sits one of the most contentious debates in corporate finance — whether paying it was the right decision at all.  Academics have been fighting about dividend policy since the 1960s.  Practitioners have largely ignored the academics, which has produced both excellent companies and spectacular disasters.  The theories themselves are worth understanding, not because they resolve the debate, but because they reveal what a dividend signals about a company’s confidence.  A dividend is not a gift.  It is a statement.  The question is whether the statement is true.

The Miller-Modigliani Position: Dividends are Irrelevant

In 1961, Dr Franco Modigliani, Institute Professor at the Massachusetts Institute of Technology, and Dr Merton Howard Miller, Robert R. McCormick Distinguished Service Professor at the University of Chicago Booth School of Business, published a paper arguing that, in a perfect market, dividend policy is irrelevant to firm value.  It is one of the most elegant and most useless propositions in the history of finance.  The argument is that if a firm pays a dividend, it must raise equivalent capital elsewhere to fund its operations — either by issuing new shares or by borrowing.  The increase in shareholder wealth from the dividend is offset by the dilution from the new shares.  Total value is unchanged.  The form of the return — dividend or capital gain — is immaterial.  In a world without taxes, transaction costs, or information asymmetry, investors can manufacture their own dividend policy by selling shares.  They do not need the company to do it for them.

The problem is that perfect markets do not exist.  Taxes exist.  Transaction costs exist.  Information asymmetry — where management knows more about the firm’s prospects than shareholders do — exists and matters enormously.  The Miller-Modigliani theorem is correct in its own frictionless universe.  It is a useful baseline because reality deviates from it, and the deviations tell you where dividend policy matters.  Modigliani won the Nobel Prize in Economic Sciences in 1985 and Miller in 1990, partly on the strength of this work.  The market proceeds to disagree with them every time a company cuts its dividend.

The Bird-in-Hand Theory: Take the Money Now

Professor Myron Jules Gordon, Professor Emeritus of Finance at the Rotman School of Management, University of Toronto, and Professor John Virgil Lintner Jr., George Gund Professor of Economics and Business Administration at Harvard University, proposed what became known as the bird-in-hand theory.  Investors, they argued, prefer a certain dividend today over the uncertain promise of capital gains tomorrow.  A dollar in hand is worth more than a dollar theoretically accruing in retained earnings, because retained earnings can be misdeployed, the company can underperform, and the future is not guaranteed.

Gordon’s model formalised this: the value of a stock is the present value of its expected future dividends.  Firms that pay higher dividends are therefore worth more, because they reduce shareholder uncertainty.  Miller and Modigliani called this the bird-in-hand fallacy.  The riskiness of the firm’s cash flows does not change because a portion is paid out as dividends, they argued.  Investors who want certain income can achieve it by selling shares.  The timing of the return does not alter its risk.  The practitioners sided with Gordon and Lintner.  Pension funds, insurance companies, and income-oriented retail investors built entire portfolios around dividend yield.  Whether their preference is rational in the Miller-Modigliani sense is irrelevant.  Their preference is real, and real preferences move prices.

Signalling Theory: A Dividend is a Declaration

The signalling hypothesis is the most psychologically interesting of the theories.  It says that dividends carry information.  Management knows things about the firm’s future earnings that outside investors do not.  A dividend announcement — especially an increase — signals that management believes the firm can sustain that payout from future cash flows.  Cutting a dividend signals the reverse.  Because maintaining a dividend that cannot be sustained is expensive and embarrassing, the signal is credible.  Cheap talk is ignored.  A dividend backed by real cash is not cheap talk.

The empirical evidence for signalling is strong.  Studies consistently find that dividend increases are followed by positive share price reactions and that cuts are followed by sharp price declines — often disproportionate to the size of the cut itself.  Between 1973 and 2025, dividend cutters and eliminators in the S&P 500 returned negative 0.96 per cent annually, against 10.22 per cent for growers and initiators.  One hundred dollars invested in dividend growers in 1973 became US$17,375 by 2025.  The same hundred in dividend cutters became US$60.  The signal matters more than the cash.

The Clientele Effect: Every Policy Attracts Its Investor

Different investors have different tax positions and different income needs.  Retired individuals living on investment income want dividends.  High-income investors in high tax brackets prefer capital gains, taxed at lower rates and deferred until realisation.  Pension funds and endowments, often tax-exempt, are largely indifferent.  The clientele effect theory says that firms attract investors who match their dividend policy.  A firm that pays high dividends attracts income-oriented investors.  A firm that retains earnings attracts growth-oriented investors.  Once a policy is established, changing it disrupts the clientele — triggering selling by the existing investors before the new clientele arrives, and producing a temporary share price dislocation in the transition period.

This is not a theory about whether dividends are good or bad.  It is a theory about consistency.  Choose a policy, hold it, and attract investors for whom it is optimal.  Change it arbitrarily, and you lose your clientele before you gain a new one.  The share price pays the price of the transition.

Agency Cost Theory: Dividends as a Leash

Professor Michael Cole Jensen’s free cash flow hypothesis, developed in 1986 while he held the Jesse Isidor Straus Professorship of Business Administration at Harvard University, takes a more cynical view.  Management, left to its own devices, will deploy free cash flow in ways that benefit management rather than shareholders.  Empire building.  Overpriced acquisitions.  Vanity headquarters.  R&D that generates publications rather than returns.  Dividends reduce the free cash flow available for management to misallocate.  They act as a discipline mechanism.  By committing to a dividend — especially one that requires debt financing when cash flow is insufficient — the board creates a credible constraint on management’s discretion.

The practical implication: firms in mature industries with limited reinvestment opportunities should pay high dividends, because the alternative is management deploying capital into projects with negative net present value.  Jensen’s empirical case was drawn heavily from the oil industry in the 1980s, where cash-rich companies spent billions on diversification into businesses they had no competence to run.  The theory has aged well.  Every conglomerate that destroyed value through acquisition binges in the 2000s and 2010s was, at some level, a Jensen problem in search of a dividend solution.

Settling the Hierarchy with Data

The Hartford Funds analysis of S&P 500 returns since 1960 is among the most cited in the dividend literature.  Reinvested dividends and their compounding effect accounted for about 85 per cent of the S&P 500’s cumulative total return over that period.  On an annualised basis, dividends contributed about 30 per cent of total return.  In the decade of the 2000s — when the index’s price return was slightly negative — dividends contributed about 1.8 per cent per year, the difference between losing money and not.  The dividend payer hierarchy is equally instructive.  From 1973 to 2025, in the S&P 500 universe:

-    Dividend growers and initiators returned 10.22 per cent annually, with a beta of 0.89 — lower risk than the market.

-    All dividend payers returned 9.20 per cent annually.

-    The equal-weighted index returned 7.74 per cent.

-    Non-payers returned 4.21 per cent, with a beta of 1.18 — more risk, less return.

-    Dividend cutters and eliminators returned negative 0.96 per cent annually, with a beta of 1.22.

Dividend growth compounds.  Dividend cuts destroy.  The 69 S&P 500 companies that have raised dividends for at least 25 consecutive years — the Dividend Aristocrats — returned 10.22 per cent annually from 2013 to September 2026, per ProShares, against an index yield that by October 2026 had fallen to about 1.05 per cent.

Proving the Theories

General Electric Company, 2017–2018

For over a century, the General Electric Company was the American industrial benchmark.  It was the only company that remained in the Dow Jones Industrial Average from its founding in 1896 until 2018.  Its dividend was a signal of corporate permanence.  Retirees held it.  Pension funds held it.  Widows and orphans, as the finance phrase goes, held it.  In November 2017, General Electric cut its quarterly dividend from US$0.24 to US$0.12 — a 50 per cent reduction, the first cut since the Great Depression save for the 2009 financial crisis adjustment.  The stock dropped 8.2 per cent on the day of the announcement.  In October 2018, the company cut again, from US$0.12 to US$0.01 per share — effectively eliminating the dividend.  The 2018 cut saved about US$3.9 billion annually.  It also confirmed what the market had suspected: the dividend had been a fiction.  General Electric had been paying out more in dividends than it generated in free cash flow for years, funding the gap with asset sales, debt, and financial engineering.  The signalling theory predicted this.  When a dividend is cut, the market does not merely price in the lost income.  It prices in everything the cut implies about the underlying business.  General Electric’s stock had already fallen over 50 per cent from its 2016 peak before the 2018 cut.  The cut was not the cause.  It was the admission.

Berkshire Hathaway

Warren Edward Buffett, Chairman and Chief Executive Officer of Berkshire Hathaway Incorporated, has not paid a dividend since 1967.  He addressed this in his 2012 annual letter to shareholders; the most coherent public argument ever made for a no-dividend policy.  The logic is this.  Every dollar Berkshire retains earns a return substantially above what a shareholder could achieve by reinvesting a dividend.  Berkshire’s compound annual growth in book value per share from 1965 to 2022 was 19.8 per cent.  The S&P 500’s total return over the same period was 9.9 per cent.  A shareholder who received dividends and reinvested them would have earned 9.9 per cent on the reinvested amount.  A shareholder whose earnings were retained and compounded inside Berkshire earned 19.8 per cent on the same capital.

Buffett’s framing is owner earnings: the cash a business generates after all capital expenditure required to maintain its competitive position, as distinct from accounting earnings.  A business that earns 15 per cent on retained capital every year does not need to pay dividends.  The capital retained is growing faster inside the business than it would outside it.  Paying it out would be, in Buffett’s formulation, a gift to the tax authorities.  The clientele effect is part of the equation.  Berkshire attracts investors who agree with this framework.  It has done so for six decades.  Investors who want income own other things.  Berkshire’s shareholder base is self-selected for patience and for confidence in management’s ability to allocate capital.  Gregory Edward Abel, who will succeed Buffett as chief executive, has indicated no intention of initiating a dividend.  The clientele will hold.

The Counterargument for Almost Everything Else

Berkshire’s model works because of one condition: management allocates capital at returns substantially above the market average.  Remove that condition, and Jensen’s agency cost argument takes over entirely.  A mediocre company with large free cash flows and no dividend is a Jensen problem waiting to happen.  The case for retained earnings rests entirely on management’s ability to compound them usefully.  Most management teams cannot match Buffett.  For companies that cannot demonstrate 15-plus per cent returns on reinvested capital over sustained periods, retaining earnings is not intelligent allocation.  It is management protecting its discretionary budget.  For those companies, dividends are not a concession.  They are a discipline.

The Buyback Problem: Jensen’s Theory, Perverted

Jensen’s agency cost argument is sound.  Its American implementation is not.  The theory says: drain free cash flow before management wastes it, and return it to shareholders.  American corporations heard return capital to shareholders.  They chose the method themselves.  They chose share buybacks.  Between the twelve months to September 2025, S&P 500 companies repurchased a record US$1.020 trillion of their own shares.  They paid US$664.9 billion in dividends.  The combined payout was US$1.685 trillion.  Apple spent US$26.2 billion on buybacks in the first quarter of 2025 alone — more than Toyota earned in the whole of that financial year.

Buybacks are not dividends.  A dividend returns cash to every shareholder, proportionally, as income.  A buyback reduces the share count, which inflates earnings per share arithmetically, without the underlying earnings growing at all.  The company’s total earnings are unchanged.  The number of shares claiming those earnings shrinks.  EPS rises.  Management, compensated on EPS targets, collects its bonus.  The share price follows EPS upward.  Management, holding options priced below the new level, collects again.  This is not capital discipline.  It is accounting manipulation in a suit.  Dr Milton Friedman, the Nobel Prize-winning economist, wrote in September 1970 that a corporation has one social responsibility: to increase its profits.  That sentence rewired the Western boardroom.  Jensen gave the theory.  Friedman gave the ideology.  The buyback gave management the tool to satisfy both while extracting maximum personal benefit from the arrangement.

Boeing is the definitive case.  Between 2013 and 2019, Boeing spent US$43 billion repurchasing its own shares.  It spent proportionally less on engineering, manufacturing safety, and supplier quality control.  Two Boeing 737 MAX 8 aircraft crashed — Lion Air Flight 610 in October 2018 and Ethiopian Airlines Flight 302 in March 2019 — killing 346 people.  Investigators found systematic failures in the MCAS flight control software and in Boeing’s own safety certification process.  The aircraft was grounded globally.  Boeing sought a government bailout in 2020.  The shareholders who received the buybacks kept their gains.  The families of 346 people received nothing back at all.

The buyback boom was fuelled further by one condition: a decade of near-zero interest rates after 2008 that made debt cheap.  Companies borrowed to buy back shares.  The debt sat on the balance sheet.  The shares were gone.  When rates rose from 2022 onward, those companies faced refinancing at three and four times the original cost, with the equity cushion already distributed.  The free cash flow Jensen wanted deployed as discipline had instead been deployed as leverage, and leverage, unlike a dividend, compounds in the wrong direction when conditions turn.  A dividend is a promise backed by operating cash flow.  A buyback is an engineering choice backed by whatever the CFO can arrange.  One tests the business.  The other tests the balance sheet.  For a company with real earnings durability, both can coexist.  For a company masking stagnant earnings behind financial engineering, the buyback is the last act before the curtain falls.

My Preferred Policy

The theories point in the same direction when properly read.  Miller-Modigliani establishes the baseline: in the absence of market frictions, dividend policy is neutral.  Reality is not frictionless, so the deviations from neutrality are where the analysis begins.  The signalling theory is the most practically important.  A dividend you cannot sustain is worse than no dividend at all.  It creates a clientele that trusts the signal, and when the signal proves false, the destruction of that trust is priced into the stock at a multiple of the dividend itself.  General Electric destroyed billions in market capitalisation not by cutting the dividend but by having paid it for years when the underlying cash generation did not support it.

The clientele effect argues for consistency, not for a particular policy.  Choose a policy suited to the business’s earnings profile and growth stage.  Early-stage growth companies with high returns on reinvested capital should retain earnings — paying dividends to fund operations through new share issuances, as Miller-Modigliani notes, is moving money in a circle with transaction costs at each end.  Mature businesses with limited reinvestment opportunities and strong cash generation should distribute — the Jensen argument applies, and the alternative is poorly-priced acquisitions.

The bird-in-hand theory matters for the client base I work with.  HNW individuals and family offices managing multigenerational wealth require predictable income streams.  For them, dividend yield is not irrational preference.  It is portfolio construction discipline.  The question is not whether to target dividend income, but which companies signal sustainable dividend growth rather than a yield that will be cut.

Dividend aristocrats — companies that have grown their dividends for 25 or more consecutive years — demonstrate one thing above all else: earnings durability.  A company cannot grow its dividend for 25 years without underlying cash flow that supports it.  The aristocrat screen is a proxy for Jensen-compliant capital allocation, signalling-theory credible signals, and clientele stability all at once.  The 10.22 per cent annualised return from 1973 to 2025 against the index’s 7.74 per cent is not an accident.  It is the premium the market pays for predictability.  For high-net-worth clients in jurisdictions with unfavourable dividend tax treatment — China’s new 20 per cent withholding rate being the clearest current example — structured insurance products provide tax-efficient income streams that replicate dividend cashflows without the tax drag.  The income is real.  The tax is not.  The vehicle matters as much as the asset.

When evaluating companies for income-oriented portfolios, I treat buyback intensity as a warning sign rather than a virtue signal.  A company that returns more capital through buybacks than through dividends is making an asymmetric bet: it benefits management through EPS inflation and option value while giving shareholders no recurring income they can rely upon.  When rates rise, and the cheap debt that funded those buybacks must be refinanced, the first casualty is the buyback programme, then the dividend, then the share price — in that order.  Dividend aristocrats by definition cannot sustain their status through buybacks alone.  The dividend must grow.  That constraint is the discipline.

The core principle is this.  A dividend policy is a promise.  Make promises you can keep.  Keep them publicly and consistently.  Raise them when the underlying business earns it.  Never cut them because you should never have made a promise the business could not sustain.  A company that follows this discipline attracts the right investors, keeps them, and signals real confidence in its own future.  A company that pays dividends its cash flow cannot support is not signalling confidence.  It is writing cheques against an overdraft.  Eventually, the bank notices.


Terence Nunis | Executive Chairman, Equinox Zenith | Author, The 1% Playbook: The Billionaire Cheat Code



No comments:

Post a Comment

Thank you for taking the time to share our thoughts. Once approved, your comments will be poster.