The signalling theory of dividends: why management pays rather than talks

The signalling theory of dividends treats a payout not as a distribution of profit but as a costly message. Management knows more about future cash flow than an outside shareholder does, and anything it says about the strength of the business is cheap talk: a weak company can repeat the same words. Cash sent out of the balance sheet, by contrast, is expensive. It cannot be recalled, it reduces the liquidity cushion, and it creates an obligation to repeat the payment in the next period. That cost is exactly what makes a dividend credible: a board will not raise the payout unless it is confident it can sustain it. Hence the main practical consequence: the market reacts not to the size of the payout but to how it changes relative to expectations.

A costly signal versus a cheap statement

The logic is the same as in signalling models from labour economics: a signal works as long as faking it costs more than the gain from deception. In the dividend version (the work of Bhattacharya, and later of Miller and Rock in the mid-1980s), a weak company can in principle announce a generous payout, but it will then have to cut investment, borrow, or publicly slash the dividend before long — and any of these will cost it more than honest silence. A sustained increase in the payout is therefore read as a statement by management that the business can keep generating cash.

The same logic produces an asymmetry in the market's response. Investors greet an increase calmly, since it fits the expected trend. A cut draws a sharp reaction because it breaks a commitment that has been made. Empirically, cuts are a stronger signal than increases, which is consistent with Lintner's smoothing model (1956): management adjusts the dividend more slowly than earnings change precisely so as not to send false messages every time results fluctuate.

What a payout signals — and what it does not

The signal speaks to the sustainability of cash flow. It says nothing about the return on an investment and promises none. What is more, a generous payout has a directly opposite reading: Jensen's free cash flow hypothesis (1986) interprets a high dividend as a sign that the company has nowhere to invest its money at an acceptable return. The payout disciplines management and lowers agency costs, but it also tells investors that the pipeline of growth projects is thin.

The two readings differ not by the amount paid but by the context. A signal of strength is an increase while capital expenditure and leverage are maintained. A signal that growth is exhausted is an increase against a shrinking investment programme. The two can only be told apart from the financial statements, not from the announcement itself: look at issuers' financial reports alongside payout history.

The valuation framework in which the signal is turned into a price is the dividend discount model: a higher expected stream of payouts at an unchanged discount rate raises the estimated value of the share. Signalling theory answers the question of where the revision in expectations comes from; the model answers the question of what it translates into.

Why the signal is distorted on the Russian market

The classic signalling model assumes that the dividend decision is made by management that is free to choose. On the Russian market this assumption often does not hold, and the theory should be read with adjustments.

The first distortion is formula-based dividend policies. If a policy ties the payout to a share of adjusted net income or to free cash flow, the announcement ceases to be a statement of intent: it becomes arithmetic applied to results that have already been published. The information arrives with the financial statements, not with the board's decision.

The second is a controlling shareholder with needs of its own. When a large stake belongs to the state or to a holding company that needs an inflow of cash, the size of the payout says more about the recipient's budget than about the payer's prospects.

The third is regulatory and legal constraints. A company may not pay dividends when its net assets are negative, or in a number of other situations listed in Article 43 of the Law on Joint-Stock Companies; for details, see restrictions on dividend payments. Banks face an additional layer in the form of capital adequacy requirements: skipping a payout may reflect not weak cash flow but a regulatory requirement — this is covered in detail in our piece on bank capital adequacy. In such a case a zero dividend carries no information about earnings at all.

How to read the signal in practice

Step 1 — establish what was expected. The signal is the deviation, so you first need a baseline: the previous payout, the text of the dividend policy and the consensus forecast. Without a baseline, every announcement looks like news.

Step 2 — check whether the payout is funded by cash flow rather than by debt or asset sales. This can be read from the cash flow statement.

Step 3 — separate the decision date from the record date. The signal arrives on the day of the announcement; price moves around the cut-off date are a different mechanism, and confusing the two is costly. The dates are explained in our guide to the dividend calendar, and the dates themselves are listed in the payout calendar and the corporate events calendar.

Upcoming announced payouts:

The limits of the theory

Signalling theory describes a mechanism, not a pattern that can be traded for profit. Empirical tests produce mixed results: the link between a dividend increase and subsequent earnings growth is weaker than the model predicts, and part of the price reaction is explained not by the signal but by the shareholder structure, taxes and changes in the investor base. No specific numerical estimates for the Russian market have been calculated in the site's database for this article, so no percentages for the price reaction are given here — you will have to test the strength of the signal on your chosen stock yourself, using the equity market data and the issuer's announcements.

It is also worth remembering that, for a shareholder, the signal and the return are different things: receiving a dividend does not in itself generate a return unless the money is put back to work; see dividend reinvestment.