Multi-Stage DDM

A multi-stage dividend discount model refines the overly simple Gordon growth model by explicitly modeling different phases of dividend growth. A company might grow dividends at 15% while building market share, 8% as it matures, 4% as the industry stabilizes, and finally 2% in perpetuity. Each stage gets its own explicit forecast period; the final stage collapses into a perpetuity using Gordon growth.

The structure

A two-stage DDM forecasts dividends explicitly for, say, ten years, then assumes perpetual growth thereafter. A three-stage DDM divides the explicit period into two—high growth for five years, declining growth for five years—then perpetuity. A four-stage or five-stage model adds further granularity.

For each stage, you forecast dividend per share, discount it back to today at the cost of equity, and sum. The terminal value is calculated from the first year of perpetuity using Gordon growth, then discounted back to today.

The result is the intrinsic value per share.

Why multiple stages matter

The Gordon growth model requires you to choose a single perpetual growth rate immediately. For any real company, this is often indefensible. A high-growth software company might grow earnings and dividends at 30% while expanding internationally, but growth will eventually slow as the market matures or the company’s size limits the percentage gains.

Multi-stage modeling lets you be explicit about the decline. High growth for five years because you see a clear runway for market share gains. Lower growth for five more years as the tailwind fades and size becomes a constraint. Perpetual growth of 3% once the company is fully mature. Each assumption is contestable; the model’s narrative is transparent.

Two-stage is most common

In practice, two-stage DDM is the workhorse. Explicitly forecast dividends for 5–10 years based on business specifics—earnings forecasts, dividend policy, balance sheet health. Then assume dividends grow at a perpetual rate thereafter and apply Gordon growth to that.

This is equivalent to two-stage free cash flow valuation but using dividends instead of total free cash flow. It requires both a dividend forecast and conviction that the company will continue paying (or growing payouts) indefinitely.

Building the explicit period

Unlike free cash flow forecasting, which requires detailed operational projections, dividend forecasting depends on three things:

1. Earnings trajectory. Forecast earnings per share, either directly or by projecting revenue, margins, and share count.

2. Payout ratio. Decide what fraction of earnings the company will distribute as dividends. Utilities and REITs maintain high payout ratios—often 70–90%. Growth companies are stingier—20–40%. As the company matures, payout ratio often rises, releasing more cash as reinvestment needs decline.

3. Share buyback policy. Many companies supplement or replace dividends with buybacks. A total-payout DDM treats dividends and buybacks as equivalent. If you forecast earnings growing at 5% and assume the company pays out 50% in combined dividends and buybacks, the cash return per share compounds at roughly 5%.

Three-stage variants

Growth stage. Dividends growing at 15–25% annually for 3–5 years as the company reinvests heavily and expands.

Transition stage. Dividends growing at 8–15% for 5–10 years as growth rates decline and payout ratios rise.

Mature stage. Dividends growing at 2–4% in perpetuity as the company reaches steady state.

This structure is particularly useful for mature, high-dividend-paying businesses where you have confidence in the business model but uncertainty about precisely when and how growth will slow.

Limitations

Dividend policy risk. Companies change their payout policies. A consistent-dividend utility is more predictable than a technology company that pays no dividend today and might initiate one in five years. The further out your forecast, the less defensible the policy assumption.

Excludes buybacks or growth reinvestment. A company might retain earnings to fund buybacks or fund growth that eventually returns cash through higher future dividends. Multi-stage DDM captures this only if you explicitly model the payout policy changes.

Terminal value dominance. Like all perpetuity-based models, the bulk of value lives in the terminal perpetuity. Two-stage model with three-year explicit period? Probably 80%+ of value is in perpetuity. The explicit forecast matters less than the final-year Gordon assumption.

Limited for zero-payers. If a company pays no dividend and has no near-term plans to start, multi-stage DDM is unsuitable. Free cash flow models are better.

Comparison to DCF

Multi-stage DDM is a subset of free cash flow to equity valuation. The conceptual difference is whether you forecast the entire free cash flow (earnings plus depreciation minus capex minus working capital changes) or only the portion paid as dividends.

For capital-light, high-payout businesses like utilities, the difference is minimal. For growth companies or capital-intensive businesses, the difference is large. A software company with minimal capex and no dividend might generate free cash flow of 100 million while paying zero dividends. Multi-stage DDM assigns zero value to that 100 million; free cash flow models assign full value.

See also