DCF vs Graham Formula vs Gordon Growth: Which Valuation Method Should You Use?
Every intrinsic-value formula is a different lens on the same question: what is this business actually worth? Here's how the three most-used methods compare, when each shines, and when each misleads.
The three methods in one paragraph each
Discounted Cash Flow (DCF) projects 10 years of free cash flow, adds a terminal value, and discounts everything back to today. It's the most flexible method — and the most sensitive to your assumptions.
The Graham Formula — V = EPS × (8.5 + 2g) × 4.4 / Y — turns EPS, expected growth, and the AAA bond yield into a single fair-value number. Fast, blunt, and surprisingly durable as a sanity check.
The Gordon Growth Model — V = D₁ / (r − g) — values a stock as a stream of dividends growing forever at a constant rate. Elegant for mature dividend payers, useless for everyone else.
Side-by-side comparison
| Dimension | DCF | Graham | Gordon Growth |
|---|---|---|---|
| Core input | Free cash flow | Earnings per share | Dividend per share |
| Best for | Any business with predictable FCF | Stable earners, quick screens | Mature dividend payers |
| Worst for | Early-stage, cyclical, capital-intensive | Growth stocks, no-earnings companies | Non-dividend payers, high-growth firms |
| Time to calculate | 10–20 minutes | 30 seconds | 1 minute |
| Sensitivity risk | High (discount + growth) | Medium (growth assumption) | Very high (r − g denominator) |
| Buffett's take | Conceptually correct method | Useful sanity check | Only when dividends ≈ owner earnings |
Where DCF wins
DCF is the only method that lets you model a business as it actually behaves — different growth in years 1–5 vs 6–10, changing margins, buybacks, reinvestment. For a wide-moat compounder like Visa, Microsoft, or Costco, a well-built DCF is the closest you can get to Buffett's textbook definition of intrinsic value: the discounted cash a business will produce over its remaining life.
Where it breaks: garbage in, garbage out. Change your discount rate from 9% to 11% and fair value drops 20–30%. Assume 12% growth instead of 8% and it swings the other way. Always run a sensitivity table.
Where the Graham Formula wins
Graham's formula was designed before spreadsheets — for humans doing math on the train home. That's its edge. In 30 seconds you can screen a watchlist of 50 stocks and flag the ones worth a real DCF. It's also remarkably resilient: because it anchors to a base P/E of 8.5 and adjusts for the AAA yield, it self-corrects across interest-rate regimes better than most people expect.
Where it breaks: anything without stable EPS. Cyclicals with a bad year look wildly underpriced. High-growth firms with low current earnings look overpriced. And the linear 2g term overstates the value of very high growth rates.
Where the Gordon Growth Model wins
For a mature dividend aristocrat — think Coca-Cola, Procter & Gamble, Johnson & Johnson — where dividends have grown steadily for decades and payout policy is unlikely to change, Gordon Growth gives you a clean answer with three inputs. It also produces an implied dividend yield at fair value, which is a useful yardstick for income investors.
Where it breaks: the model explodes as growth approaches your required return. If g = 8% and r = 9%, a small revision to either variable can double or halve fair value. And any company that reinvests instead of paying dividends is invisible to it.
How Buffett actually uses them
Buffett has said the mental model is DCF — but he doesn't build spreadsheets. He estimates a rough 10-year owner-earnings trajectory, discounts at the long Treasury rate plus a premium, and demands a margin of safety wide enough to absorb the errors baked into every input. The Graham Formula and Gordon Growth are useful as second and third opinions — if all three methods land within 20% of each other, you probably have a reasonable value range. If they diverge wildly, the business is either too complex, too cyclical, or outside your circle of competence.
A practical workflow
- Screen with Graham. Run the formula on any name that catches your eye. If the current price is above the Graham value, you probably don't need to go further unless there's a growth story the formula misses.
- Build a DCF for the survivors. Use conservative inputs. Model at least two scenarios — base case and bear case. The bear case is what your margin of safety needs to protect.
- Cross-check with Gordon if the business pays a reliable dividend. It's a fast way to see whether the DCF's terminal value assumption is sane.
- Look at the range, not the number. If DCF says $110, Graham says $95, and Gordon says $102, your fair value is "around $100 with meaningful uncertainty." Buy well below the low end.
Try each method on a real stock
We built a free calculator for each: DCF calculator, Graham Formula calculator, and Gordon Growth calculator. Or run all three at once — plus a moat score, financial-health check, and CAGR projection — with the full intrinsic-value calculator.
Frequently asked questions
Which valuation method is most accurate?
None of them is 'accurate' in an absolute sense — every method makes assumptions about the future. DCF is the most conceptually correct because it directly models what a business is worth (the discounted value of its future cash flows). But its output is only as good as your inputs. Cross-checking with the Graham Formula and Gordon Growth Model gives you a range, which is more honest than any single number.
When should I use the Graham Formula vs a DCF?
Use the Graham Formula as a 30-second screen when you're evaluating many stocks — it filters out obviously overpriced names. Use DCF when you've narrowed to a specific business worth analyzing carefully, especially wide-moat compounders where free cash flow is predictable.
Does the Gordon Growth Model still work today?
Yes, for mature dividend-paying businesses with stable payout policies — utilities, consumer staples, and dividend aristocrats. It's useless for anything that reinvests cash instead of paying it out, or where growth is unstable. Because the (r − g) denominator is fragile, always test how the answer changes when you tweak either variable by 1%.
What if the three methods give very different answers?
That's a signal, not a problem. Wide divergence usually means the business is too cyclical, too complex, or too early-stage for any of these methods to work well. In practice, that's a good time to skip the stock — Buffett's rule is to put anything you can't confidently value in the 'too hard' pile.
Continue reading
Put this into practice
Run any stock through a Buffett-style analysis — intrinsic value, moat scoring, financial health, and CAGR projections in seconds.
Analyze a stockDisclaimer: This is not financial advice. All analyses are for educational purposes only. Always do your own research (DYOR) and consult a licensed financial advisor before making investment decisions. Past performance does not guarantee future results.