• 7 min read (EN)
What Is the Dividend Discount Model: DDM Explained
The dividend discount model (DDM) values a stock by adding up every future dividend it will pay, then discounting those payments to what they are worth today. The total is what the model says one share is fairly worth.
It is one of the oldest valuation methods in finance, built on a simple idea: the only cash a shareholder ever directly receives from a stock is the dividend. Everything else is hoping someone else pays a higher price later.
Why the Dividend Discount Model Matters
DDM gives investors a way to put a price tag on a stock without depending on multiples like P/E that only compare against other stocks. Instead of asking "is this cheaper than its peers", it asks the harder question: what is this stock actually worth on its own.
The model is based on present value. A dollar paid in 10 years is worth less than a dollar paid today because today's dollar can be invested in the meantime. DDM applies that logic to every future dividend a company is expected to pay.
For mature dividend payers like consumer staples and utilities, DDM works because the payouts are predictable. Coca-Cola has raised its dividend for over 60 consecutive years. Johnson & Johnson has done the same. With that kind of track record, projecting future payments becomes a reasonable exercise rather than a guess.
The model also forces an investor to think explicitly about three things: how fast dividends will grow, how risky the business is, and what return is required to hold it. Multiples hide all of those inside one number. DDM puts them on the table.
How the Dividend Discount Model Works
The full formula sums every future dividend, discounted back to today:
In practice nobody projects dividends to infinity. Most analysts use the constant-growth shortcut known as the Gordon Growth Model, after economist Myron Gordon:
D₁ is next year's expected dividend, r is the required rate of return (often the cost of equity), and g is the long-run dividend growth rate.
Take a stock paying $2.00 in dividends next year, growing payouts at 4% per year, with a required return of 8%:
If the stock trades at $45, the model calls it undervalued. At $60, overvalued. Nudge the growth rate up by one percentage point to 5% and the value jumps to $66.67 ($2.00 / 0.03), a 33% swing from a single-point change. That input sensitivity is both DDM's defining feature and its biggest weakness.
DDM Variants Compared
The Gordon Growth Model is the simplest version, but it assumes a single growth rate forever. Real companies rarely behave that cleanly. Multi-stage variants relax the assumption at the cost of more inputs.
| Variant | Growth pattern | Best fit |
|---|---|---|
| Gordon Growth (one-stage) | One constant rate, forever | Mature payers in slow-growing markets |
| Two-stage | High growth for N years, then constant | Companies maturing out of an expansion phase |
| Three-stage | High, declining, then stable | Banks, insurers, long-life dividend payers |
| H-model | Linearly declining, then stable | Smooth transitions without an arbitrary break |
Each layer of complexity buys realism and adds inputs. A two-stage model trades one growth assumption for two plus a transition year. A three-stage model adds more. The harder the model, the more degrees of freedom an analyst has to nudge the result toward the answer they wanted in the first place.
Where the Dividend Discount Model Falls Short
DDM only works for companies that pay dividends, and pay them on a predictable schedule. That excludes most of the technology sector, almost all biotech, and many growth firms that return cash through buybacks instead. For those, a discounted cash flow on free cash flow is a better fit, because it values the cash the business produces rather than the slice that happens to be paid out.
The biggest pitfall is input sensitivity. A 1 percentage point shift in growth or required return can swing the valuation by 30% or more. When r and g sit close together, the math gets unstable: as g approaches r, the denominator shrinks toward zero and the value explodes. The model becomes useless for high-growth dividend payers.
DDM also assumes the dividend policy holds. A company that cuts its payout during a recession breaks the model retroactively, no matter how well the past was forecast. Pair DDM with a reverse DCF to see what growth the market is already pricing in, then ask whether your own assumption is more or less aggressive than the consensus.
Use DDM where dividends are reliable and growth is steady. Find candidate companies on the dividend yield heatmap and run the numbers with conservative assumptions before trusting any single result.
Frequently Asked Questions
Who invented the dividend discount model?
The idea has roots in John Burr Williams's 1938 book "The Theory of Investment Value", which argued that a stock's worth equals the present value of its future dividends. Myron Gordon and Eli Shapiro popularised the constant-growth simplification in 1956, which became known as the Gordon Growth Model and remains the most commonly used DDM variant.
Can the dividend discount model value stocks that don't pay dividends?
Not directly. DDM needs a dividend to discount. For non-payers like most tech firms, investors use a discounted cash flow model on free cash flow instead, or a free cash flow to equity model that treats the cash a company could pay as if it were already being paid out.
What discount rate should I use in the dividend discount model?
The discount rate is the required rate of return for holding the stock, which most analysts approximate with the cost of equity from the capital asset pricing model. The discount rate must be greater than the dividend growth rate, otherwise the constant-growth formula produces a negative or infinite result and the model breaks.
Why does the dividend discount model give such different answers for small input changes?
The Gordon Growth formula divides by (r − g), which is usually a small number. Any shift in r or g changes that denominator by a meaningful percentage, and the valuation moves with it. A single percentage point on g can swing the result by 30% or more, which is why DDM is most reliable when r and g are well separated and the inputs are based on conservative, defensible assumptions.