Benjamin Graham Formula Calculator

V = EPS × (8.5 + 2g), where EPS is earnings per share and g is the expected annual growth rate over the next seven to ten years entered as a whole number. Graham presented it in The Intelligent Investor as a rule of thumb, not a valuation method.

The price-to-earnings multiple Graham associated with a company expected to grow not at all. Everything above it is what the formula is willing to pay for growth.

It was Graham's empirical shorthand for how much extra multiple a point of growth deserved. It is a calibration from observation, not a derivation from theory, which is part of why it behaves badly outside the range he had in mind.

Graham meant expected growth over the next seven to ten years. Analyst consensus, the company's own historical rate, or your own estimate all work — but note that the answer is extremely sensitive to this input, so it is worth running two or three values rather than one.

Trailing twelve months diluted EPS is the usual choice. If the last year was distorted by a one-off, use a normalized figure — the formula multiplies whatever you give it, so a bad input produces a bad valuation with no warning.

With EPS of $5.00 and 7% expected growth: 5.00 × (8.5 + 2 × 7) = 5.00 × 22.5 = $112.50. Applying the bond adjustment at a 5.2% yield gives 112.50 × 4.4 ÷ 5.2, or about $95.19.

People usually mean one of two things: the growth formula above, or the Graham Number, √(22.5 × EPS × BVPS). They are different measures answering different questions, which is why this page computes both.

V = EPS × (8.5 + 2g) × 4.4 ÷ Y. It multiplies the original by the ratio of the bond yield Graham observed to the yield available now, so the valuation reflects the return on offer elsewhere.

It is the yield on high-grade AA corporate bonds around the time Graham wrote. It is a historical constant, not a parameter to tune — changing it means you are no longer using Graham's formula.

The current yield on AAA or AA-rated corporate bonds. Moody's series is the conventional source and is widely quoted; it has typically been in the 4-6% range in recent years.

The revised version is more defensible, because a valuation multiple that ignores interest rates ignores the alternative use of the money. The original is what most people quote, so this page shows both and lets you switch.

Intrinsic value falls, because 4.4 ÷ Y shrinks. That is the formula behaving correctly: when safe assets pay more, a stream of uncertain future earnings is worth less today.

√(22.5 × EPS × BVPS) — the highest price at which a stock would satisfy both of Graham's defensive limits, no more than 15 times earnings and no more than 1.5 times book value. The 22.5 is simply 15 × 1.5.

It uses no growth estimate at all, relying on earnings and book value instead. That makes it more conservative and less sensitive to your assumptions, but blind to genuine growth.

Because it needs positive earnings and positive book value. A negative product has no real square root, so the calculator reports it as unavailable rather than inventing a figure.

Rarely. Asset-light businesses carry little book value relative to their earning power, so the Graham Number will call almost any software company expensive. That is a limitation of the measure, not an insight about the sector.

None on its own. Their value is in the comparison: when the growth formula runs far ahead of the Graham Number, the valuation is resting entirely on your growth assumption, and that is worth knowing before you act.

Graham's central idea was to buy well below your estimate of value, precisely because the estimate could be wrong. Given how sensitive this formula is to the growth input, a wide margin is more defensible here than in a model with more moving parts. The buy-below figure applies whatever you choose.

As a fast sanity check, yes — it tells you quickly what growth a price implies. As a valuation method for a real position, it is too crude and too sensitive to one input. Use it to generate a question, then answer that question with a proper model.

A DCF makes every assumption explicit and lets you separate growth stages, terminal value and discount rate. The Graham formula compresses all of that into one linear term. The DCF is more work and far more honest about what you are actually assuming.

Because 8.5 + 2g implies a P/E of 38.5 at 15% and keeps rising with no ceiling. Graham never sanctioned multiples like that, and the formula silently stops being sensible. The warning exists so a large number is not mistaken for a confident one.

A separate idea about asset allocation, not valuation: never hold less than 25% or more than 75% of your portfolio in stocks, with the balance in bonds, rebalancing as valuations move. It appears in the same book but has nothing to do with this formula.

Nobody can answer that honestly, and this site will not pretend to. Graham's defensive criteria — adequate size, strong financial condition, earnings stability, a dividend record, moderate P/E and P/B — are documented in The Intelligent Investor and can be applied yourself. Note also that late in life he doubted whether that level of individual analysis was worthwhile for most investors.

No. The calculation runs entirely in your browser. Nothing is transmitted, logged or stored.

Because the formula multiplies one EPS figure by one multiplier, so a distorted input produces a distorted valuation with no warning. A single impairment, legal settlement, disposal or exceptional year passes straight through. Normalizing asks what the business earns in an ordinary year instead.

The average is simplest but one exceptional year drags it. The median ignores a single distortion but is blind to trend. The trend fit is a least-squares line evaluated at the most recent year and does not understate a company whose earnings genuinely grew. A forward estimate is the most relevant of the four and the least verifiable. The block shows all of them — if they disagree sharply, the history is too unstable for this formula.

Long enough to span most of a business cycle and short enough to still describe the current company. Ten is the convention for CAPE; seven is a common compromise for individual company normalization and keeps the input manageable.

Graham's multiplier is 8.5 + 2g, with g as a whole-number percentage. A normalized 10% growth rate gives 8.5 + 20 = 28.5, so the formula considers 28.5 times normalized earnings a fair price. Applying a margin of safety lowers the multiple you would actually pay.

A compound rate needs two positive endpoints. If the first or last year in your series is a loss, the rate is undefined rather than zero — enter your own estimate in the override field instead.

Because no company sustains an exceptional growth rate for a decade, and Graham's g is a decade-long expectation. Competition, scale and the law of large numbers pull growth toward the economy's rate. Holding a high first-year rate flat across ten years is the single most common way this formula is misused.

The constant annual rate that compounds to the same cumulative earnings as your fading path. It is a geometric result, not the average of the percentages — averaging them would overstate the outcome. That equivalent rate is the single number fed into 8.5 + 2g.

A quality compounder growing 15% today, easing to 6% by year seven, follows roughly 15, 13.5, 12, 10.5, 9, 7.5, 6 per cent. Earnings roughly double over the period, and the equivalent constant rate is about 10.5% — implying a P/E near 29.4. Held flat at 15% instead, the formula would imply 38.5, which is the same company with a third more value bolted on by an assumption.

Somewhere around 5-7% is the usual landing point. Beyond a decade very few businesses outgrow nominal GDP by much, and an ending rate above that implies the company eventually becomes an implausible share of the economy.

It is a reality check, not an input. The compound rate from your seven EPS figures is shown beside the forecast so you can see whether you are projecting faster than the business has ever managed. If you are, that is allowed — but you should be able to say why.

You are buying future earnings, so a credible forward estimate is the more relevant figure. The catch is that it is a forecast, and forecasts are optimistic on average and least reliable when the business is turning. Normalized history is verifiable but backward-looking. The block offers both so the choice is explicit rather than assumed.