A growing company spends every dollar it can find. A mature one eventually can't, it generates more cash than its best projects can absorb, and the question quietly shifts from "how do we fund growth?" to "what do we do with the surplus?" Returning capital to owners is the answer most boards reach for. The interesting part is how, and the choice between a dividend and a buyback is far less cosmetic than it looks.

The quick version

  • Returning capital is a residual decision. Hand back what's left after you've funded every project that beats your cost of capital, not before.
  • Dividends are a promise; buybacks are a one-off. Markets treat a regular dividend as a commitment and punish cuts hard, so managers raise it slowly and reluctantly. Buybacks flex with the cash you actually have.
  • Neither creates value by itself. In a frictionless world the payout method is irrelevant; in the real world it signals confidence, reshapes ownership and interacts with tax, but it doesn't conjure value from nothing.
  • A buyback only helps owners if the shares are cheap. Buying back over-valued stock destroys value as surely as overpaying for an acquisition.

The idea in depth

Start with the uncomfortable baseline. In 1961, Merton Miller and Franco Modigliani showed that, in a world without taxes, transaction costs or information gaps, dividend policy is irrelevant to a firm's value, what a company is worth depends on its investments and earning power, not on how it slices the cash between dividends and retained capital (Miller & Modigliani, 1961). An investor who wants cash can sell a few shares; one who doesn't can reinvest a dividend. The packaging shouldn't matter.

It's a deliberately unrealistic model, and that's the point: every argument for a particular payout policy has to live in the gap between that perfect world and the real one, taxes, signalling, discipline, ownership. So the useful question isn't "do dividends create value?" (mostly, no), it's "which friction am I actually trying to exploit or avoid?"

The honest limitation: M&M's irrelevance holds only under its assumptions. Once you add real-world frictions, and you always must, payout choices genuinely move outcomes. The theorem is a compass, not a conclusion.

Why a dividend is a promise, and a buyback isn't

The single most useful empirical fact about dividends is more than sixty years old. In 1956, John Lintner interviewed managers at 28 companies and found that they don't set dividends from a formula, they manage them. Firms target a long-run payout ratio and move toward it slowly, raising the dividend only when they're confident the higher level is sustainable, and cutting it only as a last resort (summarised in the Lintner model). A dividend, in other words, is a soft promise. The market reads an increase as "management is confident" and a cut as "something is wrong", which is exactly why cuts are punished so severely.

That behaviour has barely changed. When Brav, Graham, Harvey and Michaely surveyed 384 finance executives in 2002, they found managers still treat maintaining the dividend level as roughly as important as the investment decision itself, they would pass up projects or raise external funds before cutting the dividend, while buybacks are made out of the cash left over after investment (Brav et al., 2005). Managers like repurchases precisely because they carry no such commitment: you can do a big one this year and none next year without anyone reading it as a signal of distress.

A dividend is what you promise. A buyback is what you can afford.

The practical rule that falls out of this: match the instrument to the durability of the cash. Cash flow you're confident will recur, year after year, that can support a dividend. Lumpy, one-off, "we had a great year" cash, return it through a buyback, where stopping next year costs you nothing. Promising a dividend off a windfall is how boards trap themselves.

The honest limitation: this is partly self-fulfilling. Dividends are punishing to cut because everyone treats them as a signal; the signal exists because everyone believes it. That makes the convention strong, not absolute, a board with a credible, well-communicated story can sometimes reset expectations without the usual penalty.

The great shift to buybacks, and its dark side

Over the past few decades, large firms have quietly moved their money. Grullon and Michaely documented the substitution of repurchases for dividends: firms increasingly funded buybacks with cash that, a generation earlier, would have gone into raising the dividend, to the point where repurchases overtook dividends as the dominant form of payout among US industrial firms (Grullon & Michaely, 2002). In parallel, Fama and French found dividends themselves "disappearing": the share of US-listed firms paying a dividend fell from 66.5% in 1978 to 20.8% in 1999, partly because of a wave of young, growth-hungry listings and partly because firms of all kinds simply became less inclined to pay (Fama & French, 2001).

Buybacks earned their popularity honestly: they're flexible, they're more tax-efficient than dividends for many shareholders, and, when the stock is genuinely undervalued, they hand the remaining owners a larger slice of a good business at a discount. Treat a buyback like any other investment, then. You are, in effect, buying shares in your own company, so only do it when those shares are cheap relative to their intrinsic worth. A repurchase is value-creating at a low price and value-destroying at a high one.

Here's the dark side, and it's real. Because buybacks shrink the share count, they mechanically raise earnings-per-share, even when nothing about the underlying business improved. William Lazonick's Profits Without Prosperity showed how far this can run: of the S&P 500 firms public throughout 2003–2012, the 449 surviving companies spent 54% of their earnings, about $2.4 trillion, buying back their own stock, often while under-investing in R&D and wages, and while executive pay was tied to the very EPS the buybacks inflated (Lazonick, HBR, 2014). A buyback that exists to flatter a metric, fund nothing, or hit a bonus target is capital allocation working in reverse.

The honest limitation: Lazonick's account is contested. Critics note that cash returned via buybacks doesn't vanish, it flows to selling shareholders who can redeploy it into firms that do have good projects, which is roughly how capital is supposed to move. The defensible reading sits in between: buybacks are neither a scandal nor a free lunch. They're a tool that rewards discipline and punishes the lack of it.

flowchart TD
    A(["Cash generated by the business"]) --> B{"Projects that beat your
cost of capital?"}
    B -- "Yes" --> C(["Reinvest, that's the
highest-value use"])
    B -- "No / not all of it" --> D(["Surplus cash to return"])
    D --> E{"Is this cash
durable & recurring?"}
    E -- "Yes" --> F(["Dividend, a sustainable,
signalled commitment"])
    E -- "No / lumpy" --> G{"Are our shares
cheap vs intrinsic value?"}
    G -- "Yes" --> H(["Buyback, flexible,
and value-accretive"])
    G -- "No" --> I(["Special dividend, pay down
debt, or simply hold"])
A residual view of returning capital: reinvest first, then choose the instrument that fits the cash. Leaders Loop

A worked example

All figures below are illustrative, chosen for round arithmetic, not drawn from any real company.

Imagine Meridian Tools, a profitable mid-cap with 100 million shares trading at £10 (a £1 billion market value). This year it generates £80m of free cash flow. Its growth projects can soundly absorb £50m at returns above its cost of capital; beyond that, the remaining ideas don't clear the bar. So £30m is genuine surplus, the residual the board can return.

Option A, raise the dividend. Meridian already pays £20m a year (a 20p dividend per share). It could lift this to £30m (30p). But Lintner's lesson bites: this £30m looks unusually high, helped by a strong year. Promise a 30p dividend and a softer year next year forces an ugly choice, cut it and spook the market, or borrow to defend it. If the board isn't confident 30p is the new floor, over-committing is the trap.

Option B, buy back shares. Suppose the board believes Meridian is genuinely worth £12 a share, not £10. Spending the £30m surplus at £10 retires 3 million shares, leaving 97 million. The remaining owners now hold a slightly larger slice of a business the board thinks is under-priced, value created, because they bought low. Run the same buyback at £15 (above intrinsic value) and it would destroy value: Meridian would be paying £15 for something worth £12, handing a gift to the sellers at the expense of those who stay.

The decision. A disciplined board keeps the dividend at a level it's sure it can sustain through a bad year, say a modest, well-signalled rise to 22p (£22m), and returns the genuinely cyclical remainder (£8m) through a buyback executed only while the shares look cheap. The instrument follows the durability of the cash and the price of the stock.

flowchart LR
    A(["£80m free cash flow"]) --> B(["£50m reinvested
(beats cost of capital)"])
    A --> C(["£30m surplus"])
    C --> D(["£22m sustainable dividend
(modest, signalled rise)"])
    C --> E(["£8m buyback
(only while shares are cheap)"])
Meridian Tools (illustrative): reinvest first, dividend for the durable slice, buyback for the cyclical remainder. Leaders Loop

Frequently asked questions

Are dividends better than buybacks?

Neither is inherently better, they answer different questions. A dividend suits cash you're confident will recur and shareholders who value a predictable income stream; markets treat it as a standing commitment, so it's costly to cut. A buyback suits lumpy or uncertain cash and works only when the shares are under-valued. The right choice depends on the durability of your cash, your share price versus intrinsic value, and your shareholders' tax position, not on which one sounds more shareholder-friendly. The tax part especially varies by jurisdiction and changes over time, so treat anything here as the general logic rather than advice, and check the specifics with a qualified tax or finance professional for your market.

Do buybacks just inflate earnings per share?

They mechanically raise EPS by shrinking the share count, and that effect alone creates no value, the business is exactly as profitable as before. A buyback creates value only when the shares are bought below their intrinsic worth; otherwise the higher EPS is cosmetic. Treat any buyback whose main justification is "it lifts EPS" with suspicion, especially where executive pay is tied to that number.

Why do companies almost never cut their dividend?

Because Lintner's 1956 finding still holds: managers raise dividends slowly and cut them only as a last resort, since the market reads a cut as a confession that the business is in trouble. The penalty is reputational as much as financial, which is why disciplined boards set the dividend at a level they can defend through a downturn, and use buybacks, not dividend hikes, to return windfall cash.

Should a fast-growing company return capital at all?

Usually not. Returning capital is a residual decision: hand back only what you can't reinvest at returns above your cost of capital. A company with a deep pipeline of high-return projects should keep the cash and compound it, paying a dividend it could have reinvested at 25% to satisfy a convention is value-destroying. That's exactly why young, growth-rich firms so rarely pay dividends.

What's a special dividend, and when does it fit?

A special (one-off) dividend returns a chunk of cash without the implied promise of a regular one, useful for a genuine windfall (an asset sale, an exceptional year) when you want shareholders to receive cash directly rather than via a buyback, but don't want to raise the recurring dividend you'd then be expected to defend. It's the dividend world's answer to the buyback's flexibility.

Related in the Toolkit

Returning capital is the back end of a longer chain. The decision only makes sense once you've defined what "a good project" means, which is why it sits downstream of your broader capital allocation philosophy and your cost of capital, the hurdle every reinvestment (and every buyback) has to clear.

Where to go next