Theories of dividend policy

Retained earnings are the largest source of finance most companies use. They carry no issue costs, they need nobody’s approval, they are available immediately, and they dilute nobody’s control.

So every kwacha paid out as a dividend is a kwacha of that finance given up. The dividend decision and the financing decision are the same decision seen from two ends: a company that raises its payout has to fund the shortfall from somewhere dearer or invest less.

Which raises the question this step is about. If the money can be paid out or retained, and shareholders own it either way, does the choice change what the company is worth?

• Relevance theory says yes — the dividend affects the share price, so the policy matters.
• Irrelevance theory says no — shareholders are indifferent, and only the investment decision creates value.
• Residual theory says the total matters but the timing does not — pay out whatever is left after the good projects are funded.

Relevance theory — the traditional view — holds that investors depend on dividends as income, so the amount paid influences the share price. It is the assumption the dividend valuation model is built on.

Read the formula and the claim is immediate: the share price is a function of the dividend, so raising the dividend raises the price. The theory is really a set of arguments about why market imperfections make that true, and the arguments are strongest in the direction of a cut.

Dividend signalling

Investors do not have perfect information about a company’s prospects, so they read the dividend as a statement about them. A cut conveys bad news whatever the board says in the announcement — the market reasons that a company confident about next year would have maintained the payment.

Investor liquidity

The standard reply to a cut is that shareholders can sell a few shares to manufacture their own dividend. That is not costless: there are dealing charges, the shares may be illiquid — very much the case on a small exchange — and the shareholder is left with a smaller holding.

Tax and the clientele effect

Shareholders are taxed differently on dividend income and on capital gains, so a change in policy can upset their tax planning. More than that, a company that has paid a steady, generous dividend for years has attracted shareholders who wanted exactly that — a clientele. Changing the policy does not persuade them; it makes them sell, and the buyers who replace them will pay a price that suits a different preference.

The practical conclusion companies draw from all of this is not to maximise the dividend but to stabilise it: adopt a policy, hold to it, and tell shareholders in advance about any change.

Modigliani and Miller argued that in a perfect capital market — no taxation, no transaction costs, no market imperfections — dividend policy is irrelevant. Shareholders care about the increase in their wealth and are indifferent to whether it arrives as a dividend or as capital growth.

This is the same argument they made about capital structure, applied to the other side of the balance sheet, and it turns on the same move: whatever the company can do, the investor can do for themselves.

• An investor who wants cash from a company that pays no dividend sells part of their holding — a home-made dividend.
• An investor who wants retention from a company that pays one buys more shares with the dividend received.

In a perfect market both are free and instant, so nobody will pay a premium for a company to do it on their behalf. A company can pay any level of dividend it likes and meet any shortfall with a new equity issue, provided it is investing in all available positive-NPV projects.

That proviso is where the theory’s real content sits. Value is created by investing in viable projects, not by the pattern of distributions — so a company that wants to be worth more should improve its investment decisions rather than adjust its payout.

The criticism is almost entirely of the assumptions. Transaction costs make home-made dividends expensive, especially where shares trade rarely. Tax treats income and gains differently. Information is not symmetric, which is what makes signalling possible at all. And issuing equity to fund a dividend involves issue costs a perfect market does not have.

Residual theory sits between the other two. Dividends are important, it says, but the pattern of them is not: provided the present value of the dividend stream stays the same, the timing of the payments is irrelevant.

From which a clear operating rule follows. Fund every positive-NPV project first, from retained earnings, and pay a dividend only from whatever is left.

The logic is impeccable. Paying a dividend while turning down a positive-NPV project means handing shareholders cash that would have been worth more inside the business — and if they need cash, they can sell shares.

The trouble is the table. A dividend that jumps from ZMW30 million to nil tells shareholders nothing they can use, and under relevance theory the nil year would be read as a disaster rather than as a good year for investment. Residual theory shares irrelevance theory’s assumptions — no tax, no market imperfections — and inherits the same weaknesses.

It nevertheless describes real behaviour well in two situations: young companies in a growth phase with more projects than cash, and companies without easy access to other sources of finance. A Zambian unquoted business that cannot readily issue shares or borrow more is, in effect, following a residual policy whether it has chosen one or not.

Three theories, three different instructions to a finance director. Set them out and the disagreement is easy to state.

The gap between them is entirely a gap about assumptions. Irrelevance and residual theory both need a perfect capital market. Relevance theory is the argument that the imperfections are the point — and taxes, dealing costs, and shareholders who know less than managers are all plainly real.

So the practical position most companies take borrows from all three. Accept irrelevance theory’s central claim, that value is created by investing well rather than by distributing cleverly. Accept relevance theory’s warning that the market reads the dividend as information, and that a cut is expensive whatever caused it. And use residual thinking to set the level, while smoothing the path so shareholders are not asked to interpret a figure that jumps around.

That is why real dividend policies look the way they do: a payout the board is confident it can sustain through a bad year, raised only when the improvement looks permanent. It is not what any single theory prescribes, and it is what all three, read together, point at.