In 30 seconds
- The Graham formula is V = EPS × (8.5 + 2g) × 4.4 / Y: a P/E of 8.5 with no growth, two more points for every point of growth, and an interest-rate adjustment.
- Kaplio shows it on the Valuation tab as a cross-check model, with five-year growth and a 5.3% bond yield, never as the verdict.
- The starting year and the growth rate decide almost the entire result: with cyclicals and high-growth companies it blows up.
- Read backwards, it tells you what growth the current P/E is pricing in.
- Its sister rule, P/E × P/B of 22.5 or less, is the one used in the Graham defensive investor list.
Introduction
Coca-Cola earned $3.04 per share in 2025. Plug that into the Graham formula with the growth rate of the past five years and you get a value of about $78. Measure the growth from 2019 instead of 2020 and you get $55. Same company, same earnings, twenty-three dollars apart just from picking a different starting year. Understanding why is half of what you need to know about this formula; the other half is knowing when not to use it.
Explanation
Benjamin Graham, Buffett's teacher, published the formula in the 1962 edition of "Security Analysis" and repeated it in "The Intelligent Investor" (1973 edition, chapter 11). The idea was simple: a company with no growth deserves a price-to-earnings ratio (P/E) of 8.5, and every point of expected annual growth adds two points to the P/E. In 1974 he added an interest-rate adjustment, because when bonds pay more, stocks have to offer more to compete.
Every piece has a reason behind it:
EPS: earnings per share for the trailing twelve months or the last fiscal year. Graham talked about "normal" earnings, not those of an exceptional year.
8.5: the P/E Graham considered fair for a company with no growth.
2g: expected annual earnings growth over the next seven to ten years, in percentage points, multiplied by two. At 5%, you add 10 points.
4.4: the average yield on top-quality (AAA) corporate bonds in 1962.
Y: the yield on those same bonds today. If Y rises above 4.4, the value falls.
Pay attention to what Graham himself said about his formula. He presented it as an approximation of what analysts of his day got from far more elaborate calculations, not as the true value of a stock, and he warned that projections lose reliability when growth rates are high. He didn't recommend it for companies with excessive debt either.
How Kaplio applies it. On the Valuation tab of every stock page it appears among the "cross-check models": EPS × (8.5 + 2 × growth) × 4.4 / bond yield. Growth is the annualized growth in earnings per share over the past five years, and the bond yield is fixed at 5.3%. It isn't calculated for banks, insurers, REITs or biotechs, where earnings are no use as a base. And it's never the verdict: the reference on the stock page is the multiples-based value range.
Graham has another rule, better known among defensive investors than the formula itself: don't pay a P/E and a price-to-book ratio (P/B) that, multiplied together, come to more than 22.5 (for example, a P/E of 15 and a P/B of 1.5). That's where the so-called Graham number comes from: the square root of 22.5 × EPS × book value per share. Kaplio applies that rule, along with six others from chapter 14, in its Graham defensive investor list.
Formula
V = EPS × (8.5 + 2g) × 4.4 / Y
EPS = Earnings per share (last fiscal year or trailing twelve months)
g = Expected annual EPS growth in %, over 7-10 years (Kaplio uses the past 5 years)
4.4 = AAA bond yield in 1962
Y = Current AAA bond yield (Kaplio uses 5.3)
Graham number = √(22.5 × EPS × Book value per share)
Example
Fill-in-the-blanks example: Coca-Cola, 2025 annual report.
Step 1, EPS. Diluted earnings per share for 2025: $3.04.
Step 2, growth. In 2020 it was $1.79. Annual growth over five years: (3.04 / 1.79)^(1/5) − 1 = 11.2%.
Step 3, the multiplier. 8.5 + 2 × 11.2 = 30.9.
Step 4, the rate adjustment. With Y = 5.3%: 4.4 / 5.3 = 0.830.
Step 5. V = 3.04 × 30.9 × 0.830 = about $78.
Your turn: 2020 was a pandemic year, and earnings slumped. Redo the calculation starting from 2019, when EPS was $2.07, which means six years of growth. (Answer: 6.6% annual growth, a multiplier of 21.7 and a value of about $55.)
See the problem? A depressed starting year inflates growth, and growth goes in multiplied by two.
Now try the formula on a growth company. Microsoft earned $17.95 per share in fiscal 2026 and $8.05 in fiscal 2021: 17.4% a year. V = 17.95 × (8.5 + 34.8) × 0.830 = about $645. With growth like that, any small change in g moves the result by hundreds of dollars. That's exactly what Graham warned about.
And to see how different the defensive investor's rule is, apply the Graham number to Coca-Cola. Equity attributable to Coca-Cola shareholders at the end of 2025 was $32,169 million; divided by about 4,313 million diluted shares, book value per share comes to around $7.46. Work out √(22.5 × 3.04 × 7.46) and compare it with the share price implied by the live table (P/E × EPS): the Graham number comes in at less than a third of the price. A brand that has bought back stock for decades has little book value and almost never passes that filter. That's why Kaplio uses it to screen for a diversified basket, not to pass verdict on a single company.
The table below shows the P/E of four big consumer staples companies. Before you look: on October 4, 2026, Coca-Cola traded at a P/E of 25.7 on trailing twelve-month earnings. With a bond yield of 5.3%, what annual growth do you think it's pricing in, according to Graham? Solve for g with the formula in "How to read it". (Answer: 11.2% a year, almost exactly the 2020-2025 growth that gave us the $78.)
Real-data example
| Company | Ticker | P/E |
|---|---|---|
| The Coca-Cola Company | KO | 25.7 |
| PepsiCo, Inc. | PEP | 16.5 |
| The Procter & Gamble Company | PG | 21.5 |
| Colgate-Palmolive Company | CL | 33.1 |
As of October 4, 2026, on trailing twelve-month earnings: PepsiCo trades at a P/E of 16.5, Procter & Gamble at 21.5, Coca-Cola at 25.7 and Colgate at 33.1. Read backwards through the Graham formula with a 5.3% bond yield, those P/Es price in annual growth of 5.7%, 8.7%, 11.2% and 15.7%. One caveat: the formula uses last fiscal year's earnings and the table uses the trailing twelve months.
How to read it
How to read it
The most useful way to read the formula is backwards: solve for g and see what growth the market is paying for. With Y = 5.3%, g = (P/E × 5.3 / 4.4 − 8.5) / 2:
| Current P/E | Growth priced in | Usual reading |
|---|---|---|
| 12 | 3.0% a year | A low bar: reasonable for consumer staples or utilities |
| 15 | 4.8% a year | The ceiling for Graham's defensive investor (chapter 14) |
| 20 | 7.8% a year | Calls for growth few mature companies sustain for a decade |
| 25 | 10.8% a year | Only fits companies with proven growth |
| 30 | 13.8% a year | Demanding even for high-quality tech |
Compare that implied growth with the company's own history and with its sector. If the market is asking for more than the company has delivered over ten years, your margin of safety is thin. Graham wanted to buy well below estimated value, never at full price.
By sector: the formula works reasonably well for stable companies with positive earnings (consumer goods, mature industrials, big pharma). It's no use for banks and insurers, which are valued on book value, or for loss-making companies, where negative EPS gives a negative value.
Pitfalls and limitations
- The starting year. You've already seen it with Coca-Cola: twenty-three dollars of difference from shifting one year.
- Growth runs away with the result. Above 15%, the formula produces values that depend almost entirely on g. Graham wanted projections over seven to ten years, and few companies keep up 15% for that long.
- EPS can be inflated. If earnings include an asset sale or a one-off gain, the formula multiplies it. The Summary on Kaplio's stock pages flags when trailing twelve-month earnings are inflated by items outside the business; if you see that warning, don't apply Graham to that EPS.
- It ignores debt. Two companies with the same EPS and growth are worth the same to the formula even if one owes three times as much.
- Y isn't a neutral input. Using 5.3% or some other rate changes the result, and in the opposite direction: the lower Y is, the higher the value comes out. With Y = 4.4, the Coca-Cola in the example would be worth about $94.
Case in point: Caterpillar breaks the formula
Caterpillar, a machinery maker tightly tied to the economic cycle, earned $5.46 per share in 2020 and $18.81 in 2025. With those two data points, the formula sees growth of 28.1% a year and a multiplier of 64.7: V = 18.81 × 64.7 × 0.830, more than $1,000 per share. But 2020 was the bottom of the cycle, not the start of a trend, and earnings had already slipped in 2025 from $22.05 in 2024. With a cyclical, the formula mistakes a recovery for growth. Graham talked about "normal" earnings, and his own defensive investor checklist uses a three-year average, not a single year. If you want a method that doesn't hinge on one EPS figure, compare it with the simple DCF or with the P/E against its own history.
Practice on Kaplio
Frequently asked questions
What is the Graham equation and how does it work?
Value = EPS × (8.5 + 2g) × 4.4 / Y. It starts from a P/E of 8.5 for a no-growth company, adds two points per point of expected annual growth (g), and scales the result by 4.4, the 1962 AAA bond yield, divided by today's yield (Y). Graham saw it as a rough approximation, not a stock's true value.
What is the formula for intrinsic value?
There isn't just one. The quick route is a multiples formula like Graham's, built on earnings per share and expected growth. The thorough route is a discounted cash flow model: project free cash flow, discount it, subtract net debt and divide by the number of shares. The sensible approach is to check one against the other.
How do I calculate the intrinsic value of a stock?
Estimate what the company is worth and divide by its shares. With Graham's formula, multiply EPS by 8.5 plus twice the expected growth rate and adjust for bond yields. With a DCF, discount future free cash flow and subtract net debt. Then compare with the share price: buying well below your estimate gives you a margin of safety.
Did Warren Buffett learn from Ben Graham?
Yes. Graham was Buffett's teacher, and his books "Security Analysis" and "The Intelligent Investor" shaped Buffett's approach: buy below estimated value with a margin of safety, and treat any formula as an approximation. Kaplio's Graham defensive investor list applies the rules Graham set out in chapter 14 of "The Intelligent Investor".