Dividend Discount Model Calculator (DDM)

CFA DDM DIVIDEND FINANCE VALUATION
Share:
By Published Updated Last reviewed

Dividend Discount Model calculator. Gordon growth (single-stage perpetuity) and two-stage DDM (high growth → stable growth). Intrinsic share price from dividends, growth, and required return.

RT-FIN-227 · Finance & Money
Your browser is too old for this tool’s interactive features. Please update to Safari 13.1, Chrome or Edge 80, Firefox 74 or newer.

Dividend Discount Model Calculator

Model
Gordon for mature firms; two-stage for growth dividend payers
to compare intrinsic vs market
Common inputs
most recent annual dividend
cost of equity from CAPM
Gordon growth parameters
must be < r forever
Gordon diverges if g ≥ r
Advertisement
After results · AD-W1Responsive · Post-tool — peak engagement

How to use the dividend discount model calculator

Choose Gordon or two-stage DDM

Gordon (single-stage): best for mature dividend payers with stable growth (utilities, consumer staples, REITs, dividend aristocrats). Assumes constant growth forever. Two-stage: best for growth dividend payers (J&J, Microsoft historically, mature tech) where growth is high now but will eventually stabilise. Two-stage is more realistic; Gordon is simpler.

Enter current dividend (D₀) + required return (r)

D₀: most recent annual dividend per share — sum of last four quarterly dividends, in dollars. From Yahoo Finance "Dividends" column, your broker statement, or company investor relations. r: required return on equity — your hurdle rate. Use the CAPM calculator to compute: r = Rf + β · (Rm − Rf). For mature dividend stocks, r typically 7-10%; for growth dividend payers 10-13%.

Enter growth rates carefully

For Gordon: one growth rate (g) used forever — must be < r and ≤ long-run GDP growth (~2-4% developed markets). For two-stage: high growth rate (g_high) for first N years, then stable growth (g_stable) forever. Critical: g_stable must be < r or terminal value diverges. Also g_stable should be ≤ GDP growth — a company can't grow faster than the economy forever.

(Optional) Enter current market price

If you enter the current market price, the calculator shows whether the stock appears undervalued (intrinsic > market = buy candidate), fairly valued (within ±5%), or overvalued (intrinsic < market = avoid). The fair-value verdict is only as good as your inputs — sensitivity-test r ±1% and g ±1% to see how robust the conclusion is.

Read intrinsic value + decomposition

For Gordon: intrinsic value, next-year dividend, forward yield, and r-g spread (a small spread = highly sensitive to inputs). For two-stage: year-by-year dividend projection table, terminal value, PV of terminal value, and what % of intrinsic value comes from the terminal stage. If terminal share > 80%, you're essentially making a Gordon-growth bet on year-N+1 — extend N or revisit g_stable.

Advertisement
After how-to · AD-W2Responsive

DDM — valuing a share as the present value of the dividends it will pay

The dividend discount model values a share as the present value of all the dividends it will pay. With dividends growing at a constant rate g forever and investors requiring a return r, the sum collapses to the Gordon growth model: P₀ = D₁ ÷ (r − g), where D₁ is next year's dividend. OpenStax's Principles of Finance spells out its two conditions: r must be greater than g, or the formula gives a meaningless negative price, and constant growth forever is unrealistic for many companies, which is why it also sets out two-stage and variable-growth models. This calculator offers the Gordon model and a two-stage version: a high-growth period, then stable growth from a terminal value.

When DDM works — and when it doesn't

DDM works best for: (1) Mature dividend payers with consistent payout policies — utilities (Duke Energy, Exelon), consumer staples (Procter & Gamble, Coca-Cola), banks (JPMorgan, Bank of America in normal cycles), pharma giants (J&J, Pfizer). (2) Dividend aristocrats — companies with 25+ consecutive years of dividend increases. (3) REITs — which, to keep their tax status, must pay out most of their taxable income as dividends. DDM fails for: (1) Growth stocks that pay no dividends (Amazon historically, Berkshire never, Tesla never, Meta until 2024). (2) Cyclicals with volatile payouts (oil majors during crashes, banks in distress). (3) Distressed firms that have suspended dividends. For non-dividend payers, use DCF or relative-multiple valuation (P/E, EV/EBITDA).

DDM tells you what a stock is worth if dividend policy is the entire story. For Coca-Cola, that's most of the story. For Amazon, it's none of it. Knowing which is which is half the value-investing battle.

The r-g sensitivity problem

The Gordon model's denominator is (r − g). When r and g are close — say r=9%, g=6%, spread of 3% — the model is reasonable. But small changes to either input swing the valuation dramatically. Drop g to 5%: spread widens to 4%, intrinsic value drops about 26%. Raise g to 7%: spread shrinks to 2%, intrinsic value rises about 51%. The model is structurally fragile to growth assumptions, which is why g_stable must be set conservatively — never above expected long-run GDP growth. A common mistake: using historical 10-year dividend CAGR as g_stable. Past growth is rarely sustainable forever; long-run g must equal economic growth or fall below it.

Where the inputs come from

D₀ is the dividend paid over the last twelve months. r is the return you require for the share's risk; it is a cost of equity, not a company's weighted cost of capital, because dividends go only to shareholders. g should be a rate the business can keep up indefinitely, so a stable rate above long-run economic growth is a warning sign. Damodaran's Dividend Fundamentals dataset at NYU Stern (January 2026) gives US payout ratios by industry, a useful check on the dividend you start from. Taxes on dividends depend on where you live and where the company is listed, and change, so this calculator values the dividend before tax.

P₀ = D₁ / (r − g): the g < r constraint, and why DDM fits REITs and not tech

01

P₀ = D₁ / (r − g) — Gordon's closed-form solution (Review of Economics and Statistics, 1959), valid only while g is below r.

02

OpenStax names it the Gordon growth model and requires r > g; with g ≥ r the formula gives no meaningful price.

03

Hard constraint: g < r always. If terminal growth equals or exceeds required return, the perpetuity diverges to infinity.

04

Two-stage DDM is the production workhorse: high growth for 3-10 years, then perpetual stable growth (≤ GDP growth).

05

Two-stage with 0 years of high growth is just the Gordon model at the stable rate; this page now treats it that way instead of failing.

06

Sensitivity: at r = 9%, cutting g from 5% to 4% lowers the Gordon value by about 20.8%, because r − g grows from 4 to 5 points.

07

Worked example: a USD 2.50 dividend growing 5% at r = 9% is worth USD 2.625 ÷ 0.04 = USD 65.63.

08

The model is structurally fragile: at r = 9% and D₀ = USD 2.50, g of 6% gives USD 88.33 and g of 7% gives USD 133.75, a 51% jump from one point of growth.

09

r-g spread matters more than absolute values. Small spreads = high sensitivity. Wide spreads = robust valuations.

10

Damodaran publishes US payout ratios, and dividends + buybacks, by industry (January 2026 data), a sanity check on DDM inputs.

Frequently asked questions

  • Use Gordon when growth is stable and sustainable — utilities, mature consumer staples, regulated banks, REITs with steady payouts. The constant-growth assumption isn't crazy because their businesses don't change much year to year. Use two-stage when growth is currently high but will eventually stabilise — typical for dividend-paying growth tech (Microsoft, Apple post-2012), early-stage REITs, dividend payers in cyclical recovery. The two-stage model produces a more defensible valuation because it explicitly separates the "current growth phase" from "mature steady-state".

  • Three approaches. (1) Historical CAGR: compute 5-10 year compound annual growth rate from the company's historical dividends. Easy but assumes the past repeats. (2) Earnings growth × payout ratio: long-run dividend growth equals retention rate × ROE = (1 − payout ratio) × ROE. More robust. (3) Management guidance: many dividend-aristocrat firms publish multi-year dividend growth targets. For Gordon (single-stage), use a blended estimate that won't exceed long-run GDP growth (~2-4% developed, ~4-5% emerging). For two-stage high-growth phase, near-term CAGR is reasonable; the stable phase must be ≤ GDP growth.

  • Healthy scepticism. Three checks. (1) Are your inputs aggressive? A small reduction in g (say 5% → 4%) often eliminates the apparent undervaluation. (2) Is r too low? If you used a Rf from years ago when rates were near zero, your r is stale; rates have risen materially. (3) Is the market pricing risk you haven't modelled? Many "undervalued" dividend stocks are pricing in regulatory risk (utilities), litigation risk (tobacco, asbestos exposure), or end-market decline (legacy print media). The market is approximately efficient most of the time; large divergences usually mean either (a) your model is wrong or (b) there's real risk not captured in growth/required-return assumptions.

  • Strictly no — if D₀ = 0, the model produces 0. For Amazon, Tesla, Berkshire, pre-2024 Meta, and most growth tech, DDM is the wrong tool. For these firms, use DCF on free cash flow to equity (FCFE), which captures cash returns via buybacks and reinvestment value, not just dividends. Some practitioners do "implied DDM" by assuming the firm will start paying dividends at some future point — but this is highly speculative and the math is sensitive to the assumed dividend-initiation date. For most analysts: use DCF for non-dividend payers, DDM only for actual dividend payers.

  • Buybacks are economically equivalent to dividends — both return cash to shareholders. Many firms (Apple, Microsoft, banks) prefer buybacks for tax efficiency. For DDM purposes, some practitioners use "total shareholder return" = dividends + buybacks per share, treating buybacks as dividend equivalents. Damodaran's dividend dataset reports dividends + buybacks by industry alongside dividends alone, which shows how much of the cash returned comes through buybacks. For Gordon: D₀ in the formula becomes (dividends + buybacks) per share over the last 12 months. Caution: buybacks are more discretionary than dividends — companies cut buybacks first when cash gets tight. If you include buybacks, sensitivity-test by running pure-dividend DDM separately.

  • A stock with a high current dividend yield (say 9%+) that looks "cheap" on DDM but where the dividend is unsustainable. Common causes: (a) cyclical earnings that won't support the dividend through the next downturn (energy companies in 2015-2020, banks in 2008-2009), (b) declining industry where revenue is shrinking faster than payout reductions (legacy print media, fixed-line telcos), (c) excessive payout ratio (>100% means the company is paying more than it earns — sustained only by debt or asset sales). Sanity-check: payout ratio (dividends / earnings) should be ≤ 80% for sustainable dividend growth. Above that, expect a cut.

  • No. DDM uses the cost of equity (re), not WACC. The model discounts dividends — which flow only to equity holders — at the return equity investors demand. Compute re via CAPM: re = Rf + β · (Rm − Rf). Don't confuse this with WACC, which blends cost of equity with after-tax cost of debt and is used for discounting firm-level cash flows (FCFF) to compute enterprise value. DDM gives equity value directly; no debt bridge needed.

  • Very. For Gordon at r=9%, g=5%: intrinsic value = D₁ / 4%. Drop g to 4%: intrinsic value = D₁ / 5%, a 20% reduction. Raise g to 6%: intrinsic value = D₁ / 3%, a 33% increase. The (r − g) denominator amplifies small input changes into large valuation changes. Always run sensitivity tables: g ±1% × r ±0.5% gives a 9-cell view of how robust your valuation is. If most cells produce similar values, your DDM is robust. If the range is 2-3× from low to high, you're really making a growth-rate bet, not a valuation.

  • No. Dividend, required return, growth rates — every input stays in your browser. The full DDM computation (Gordon perpetuity and two-stage cash-flow walk) runs as client-side JavaScript. Open DevTools → Network when you click Compute and you'll see zero outbound requests. Safe for confidential equity research work.

  • Original papers: Williams JB. The Theory of Investment Value. Harvard University Press, 1938. Gordon MJ. "Dividends, Earnings, and Stock Prices." Review of Economics and Statistics 1959;41(2):99-105. OpenStax Principles of Finance §11.2 sets out the constant-growth and multi-stage models with worked examples, and Damodaran's Dividend Fundamentals dataset gives industry payout ratios for checking inputs. Both are listed under Method & sources on this page.

Method & sources

How it computes

Gordon: P0 = D1 / (r - g), D1 = D0(1 + g), refused when g >= r. Two-stage: dividends grow at g_high for N years, each discounted at r, plus the terminal value D_N(1 + g_stable) / (r - g_stable) discounted N years; N = 0 is the Gordon model at g_stable. RecaCorpFin.gordon / twoStage, tested against hand-summed dividends.

What this tool implements

  • Bugs fixed: a high-growth stage of 0 years read cashFlows[-1] and threw, so the page failed instead of answering; 0 years now reduces to the Gordon model at the stable rate, which is what it means. When stable growth was at or above r, the error showed but the breakdown kept the previous result's numbers beside it; it is now hidden.
  • Gordon model and the r > g requirement follow OpenStax 11.2. Worked figures checked: USD 2.50 at 5% growth and r 9% is USD 65.63; at r 9%, g 6% gives USD 88.33 and 7% USD 133.75 (+51%), 5% gives -26%. Damodaran's Dividend Fundamentals (January 2026) is cited for industry payout ratios and dividends + buybacks; the page had said it published growth estimates, which it does not.
  • Removed as unsourced: the '~20% of US stocks pay meaningful dividends' and '~65 dividend aristocrats' counts, ASEAN yield and withholding-tax claims, a Damodaran 'augmented DDM' attribution, and textbook chapter numbers.

Sources

Advertisement
Pre-footer · AD-W3 728 × 90