The Benjamin Graham Formula: Still Relevant After 75 Years

Benjamin Graham once published an intrinsic value formula that fits on a napkin. Take a base P/E of 8.5, add twice the expected annual growth rate, and multiply by current earnings. Next to the 40-tab DCF models and Monte Carlo simulations people build today, it looks almost too simple to bother with. So why does anyone still use it?

Because it works better as a sanity check than most complicated models work as primary valuation tools. The formula won't hand you a fair value to the penny. It'll tell you whether you're in the right ballpark or way off. And plenty of investors genuinely can't tell the difference between a stock that's a little rich and one that's priced for a fantasy, so a fast way to sort those apart is worth having.

How the formula works

The original version from The Intelligent Investor is: Intrinsic Value = EPS x (8.5 + 2g). EPS is trailing twelve-month earnings per share, and g is the expected annual growth rate over the next 7 to 10 years.

The 8.5 is the P/E Graham thought was fair for a company with no growth at all. If earnings are never going to climb, he figured 8.5 times current earnings was the right price. The 2g part pays up for growth, adding two points to the acceptable multiple for each percentage point of expected growth.

So a company earning $5 a share with expected growth of 10% comes out at $5 x (8.5 + 20) = $5 x 28.5 = $142.50. Trade below that and the stock is potentially cheap. Trade well above it and the market is betting on either faster growth or lower risk than the formula assumes.

Graham later adjusted the formula for interest rates: Intrinsic Value = [EPS x (8.5 + 2g) x 4.4] / Y, where Y is the current yield on AAA corporate bonds and 4.4 was the average AAA yield when he wrote it. The tweak makes the formula stingier when rates are high and more generous when they're low, which matches the basic idea that higher rates shrink the present value of future cash flows. Run it through the earlier example: that $5-a-share, 10%-growth company came out at $142.50 in the original version. If AAA yields sat at 6%, the 4.4/Y factor is about 0.73, and the intrinsic value drops to roughly $104. Same earnings, same growth, but a higher discount rate takes a real bite out of what you should be willing to pay.

Where it actually shines

The formula is at its best as a screen and a reality check. Run it across a big list of stocks and it quickly separates the obviously overpriced from the plausibly cheap. That matters, because a lot of people burn hours dissecting companies that are clearly expensive and spend almost none on the ones trading at reasonable prices.

It's also great for exposing the growth assumptions baked into a price. Say a stock trades at $200 and earns $4 a share. You can run the formula backward to find the implied growth rate: $200 = $4 x (8.5 + 2g) gives 2g = 41.5, so the market is implying roughly 20.75% annual earnings growth for the next decade. That's a steep bar. If you don't honestly believe the company can compound earnings north of 20% for ten straight years, it's overvalued by Graham's math.

I like this reverse-engineering trick because it makes the market's assumptions explicit. Instead of arguing in the abstract about whether something is cheap or dear, you get a concrete question to answer: can this company realistically grow earnings at X% for the next decade? Often the answer is plainly no, and you can move on without building a full model.

Where it falls short

Graham built this in an era when most companies were industrial, asset-heavy, and fairly steady. A few things about modern markets chip away at its accuracy.

The base P/E of 8.5 for a no-growth company can feel low depending on where rates sit. Some analysts bump it to 10 or 12 to reflect current conditions, though that reintroduces exactly the subjectivity Graham was trying to strip out.

The formula also treats the link between growth and value as linear, when in reality it curves. A company compounding at 30% is worth more than twice one compounding at 15%, because the effect of compounding accelerates as growth rises. The 2g multiplier understates very high growth and overstates the merely decent kind.

It says nothing about the quality of that growth either. A company posting 15% earnings growth by piling on debt to buy other companies is a very different animal from one growing 15% organically with high returns on capital. Graham dealt with quality all over his broader writing, just not inside this formula.

And it ignores the balance sheet. A business with $10 a share in net cash is worth more than an otherwise identical one carrying $10 a share in net debt, yet the formula scores them the same as long as earnings and growth match.

Making it useful today

You can patch most of those gaps with a few practical habits.

Use normalized earnings instead of raw trailing EPS. Cyclical businesses can post earnings that are temporarily inflated or crushed, and either one distorts the output. Averaging earnings over three to five years, or using a mid-cycle estimate, gives you a cleaner input.

Be stingy with growth. Graham himself pushed for estimates below consensus. If the street expects 15%, plug in 10%. That builds a margin of safety into the assumption itself, not just into the price you pay.

Adjust for rates using the revised formula. When AAA yields are elevated, the 4.4/Y factor pulls the value down and makes you more conservative. When yields sat near zero, that same factor made the formula far more generous, which was reasonable since competing returns were low. The point is to let the rate environment move the number rather than eyeballing it.

Deal with the balance sheet by hand. Since the formula ignores net cash and net debt, adjust after the fact. Compute the Graham value on earnings and growth, then add net cash per share or subtract net debt per share to get a truer picture of what a buyer of the whole business is actually paying for. On a cash-rich company that one step can meaningfully change whether it clears your discount hurdle.

Layer it with other approaches. Run it alongside a DCF, a comps analysis, and the Piotroski F-Score, the nine-point checklist from Joseph Piotroski's 2000 paper that separates financially improving value stocks from deteriorating ones. When Graham says cheap, the DCF agrees, and the F-Score is high, you've got several independent methods converging, and that's a much stronger signal than any one of them alone.

Using it as a screen

Here's the workflow I'd run. Pull every stock in your investable universe. For each one, compute the Graham value from trailing EPS and a conservative growth estimate. Consensus minus three to five points is a fine starting point. Then compare that value to the current price.

Keep the names trading at 66% or less of Graham value. That roughly one-third discount is the margin of safety Graham argued for. You're paying about 66 cents for every dollar of estimated worth, which cushions you against the errors you're guaranteed to make in the estimate.

From what survives, drop anything with negative earnings, since the formula breaks on money-losers. Drop the heavily indebted names too. A common convention is debt-to-equity above 1.5x, but pick a threshold that fits your universe. And cut companies with clearly declining revenue. Those filters clear out the most obvious value traps.

What's left is a starting point, not a portfolio. You've turned a whole market into a short list that probably hides a few real bargains, and now you can dig into those by hand.

What the track record tells you

Value approaches like this have a long history of working, and an equally long history of testing your patience. Studies of margin-of-safety and low-multiple strategies going back decades generally show them beating the broad market over full cycles, with the edge concentrated after downturns and the worst stretches during speculative bull runs.

The drawdowns are real. During the late-1990s tech bubble and the growth mania of 2020 and 2021, disciplined value approaches lagged badly. The people who bailed at the bottom locked in the pain and missed the snapback that followed. In a systematic value strategy, staying disciplined tends to matter more than being clever.

Why simple is a feature

Finance has a bias that treats complexity as sophistication. A 200-line DCF with 50 assumptions feels more rigorous than a one-line formula. But every one of those 50 assumptions is a place for error to sneak in, and small errors compound into an answer that's precisely wrong instead of roughly right.

The Graham formula has two inputs, earnings and expected growth. You can be off on both and still get useful direction, because there just aren't many places for mistakes to hide. A simple tool used consistently beats a complex tool used badly, and most people use the complex tools badly.

Graham understood this himself. He never claimed the formula was precise, only that it was useful, and that's still the right way to think about it. When you're drowning in data and analytical machinery, a decent approximation you can run in half a minute has real value. It won't replace deep work on your best ideas, but it'll make your first pass far faster and keep you from wasting afternoons on companies that were never cheap to begin with.