Benjamin Graham Number Formula: Uses and Limits

The Benjamin Graham Number is attractive because it is simple.

It does not require a five-year cash-flow forecast. It does not require a terminal growth rate. It does not ask the investor to estimate owner earnings, maintenance capital expenditure, or a discount rate. It takes two familiar inputs, earnings per share and book value per share, and turns them into a conservative valuation reference point.

That simplicity is the reason the formula is useful.

It is also the reason the formula can be misused.

The Graham Number should not be treated as a precise fair value estimate. It is better understood as a defensive screening tool. It asks whether a stock has support from current earnings power and balance-sheet backing. That is a narrower question than the ones answered by a DCF valuation example or a Buffett Intrinsic Value model.

This article explains what the Graham Number is, how StockGeniuses treats it, and why the formula can be useful for some businesses while breaking down for others. The real-world contrast uses Ford and Apple to show model fit, not to make a valuation call on either stock.

Benjamin Graham’s defensive investing idea

Benjamin Graham’s value-investing philosophy was built around defense.

The goal was not to forecast the most exciting growth story. The goal was to reduce the chance of overpaying. Graham-style analysis favored evidence that was already visible in the financial statements: earnings, assets, equity, and a meaningful margin of safety.

That philosophy made sense in Graham’s context, but it still matters today. Investors are constantly presented with stories about future growth, product cycles, market share, artificial intelligence, macro trends, and changing sentiment. Some of those stories may be valid. The Graham mindset starts from a stricter place: what can already be supported by the company’s reported earnings and balance sheet?

The Graham Number fits that defensive mindset.

It combines two questions:

Does the company have positive earnings power?
Does the company have balance-sheet value behind the shares?

If both answers are strong enough, the formula produces a conservative value reference. If either answer is weak, the model becomes less useful or not meaningful at all.

That is why this is a screening model. It is not meant to explain every good business. It is meant to filter for situations where visible earnings and book value provide enough support to justify more research.

The Graham Number formula

The public Graham Number formula is:

Graham Number = square root of (22.5 x EPS x Book Value Per Share)

The 22.5 constant comes from combining two conservative valuation limits commonly associated with Graham-style defensive stock selection:

maximum P/E ratio of 15
maximum P/B ratio of 1.5

Multiply 15 by 1.5 and you get 22.5.

In plain English, the formula says a stock should not receive a high valuation unless both earnings power and book value support it. The formula does not reward a company simply for having a strong story. It asks for earnings and equity.

That is very different from valuation methods that depend on future growth. The Graham Number does not ask how fast revenue might grow. It does not ask whether margins might expand. It does not ask whether the company has a powerful brand, ecosystem, network effect, or software-like economics. It asks whether earnings and book value provide enough conservative support.

This is why the formula should be read carefully. A high-quality company can look weak under the Graham Number if its economic value comes mostly from intangible assets rather than book equity. A mediocre company can look more supported by the formula if it has positive earnings and substantial accounting equity.

The formula is simple. The interpretation is not.

StockGeniuses method for the Graham Number

Inside StockGeniuses, the Graham Number is treated as a conservative value anchor, not as a complete valuation model.

The model uses:

3-year median TTM EPS
latest reported book value per share
market price
Value Gap
Graham Number score

The 3-year median EPS matters. A single year of earnings can be distorted by cyclicality, one-time charges, unusual demand, tax effects, or accounting items. Using a median smooths the input and reduces the risk of building the screen on one unusually strong or weak period.

The formula then computes:

Graham Number = square root of (22.5 x EPS median x BVPS)

The model becomes Not Meaningful if median EPS is less than or equal to zero, book value per share is less than or equal to zero, or the market price is missing. Those rules are important. The Graham Number cannot work properly if the formula is built on negative earnings or negative equity.

This is also why 10 Stock Analysis Metrics Serious Retail Investors Should Understand matters before using a model like this. EPS and book value are not just formula inputs. They are accounting signals that need context.

A real-world contrast: Ford and Apple

Ford and Apple help show why the Graham Number is really a model-fit test.

Ford is a traditional automaker with large manufacturing operations, financing assets, liabilities, physical capital, and meaningful accounting equity. Apple is a brand-and-ecosystem-driven business whose value is tied heavily to intangible strengths, services economics, product integration, cash generation, and shareholder returns.

Using Ford’s 2024 Form 10-K and Apple’s 2025 Form 10-K, the contrast looks like this:

Company	Filing data used	EPS input shown	Equity / shares context	What the contrast teaches
Ford	2024 Form 10-K	Diluted EPS of $1.46	Total equity attributable to Ford of $44.835B; about 3.963B common and Class B shares outstanding after year-end	Book value has more obvious relevance for an asset-heavy company, but cyclicality still matters
Apple	2025 Form 10-K	Diluted EPS of $7.46	Shareholders’ equity of $73.733B; 14.773B shares outstanding at fiscal year-end	Book value captures only part of economic value for an intangible-heavy company

These figures are not used here to make a current valuation judgment. They are used to show how the formula thinks.

For Ford, book value is a more natural part of the conversation because the business has factories, vehicles, financing assets, and tangible operating infrastructure. The Graham Number may be more relevant as a conservative check, although Ford’s cyclicality and capital intensity still make interpretation difficult.

For Apple, book value per share is much less likely to capture the full economic story. Apple’s brand, ecosystem, installed base, software and services economics, and buyback history do not show up in book value the same way a factory or financing asset might. A Graham Number screen can therefore make Apple look less supported by book-value logic even when the business itself is very high quality.

That is not a flaw in Apple. It is a limitation of the model.

The lesson is simple: the same formula can be sensible for one business type and misleading for another.

Where the Graham Number helps

The Graham Number can be useful when the investor needs a fast, conservative value check.

It helps in asset-backed businesses where book value is economically meaningful. Manufacturing companies, certain industrials, and other balance-sheet-heavy businesses may give the formula more useful information than asset-light companies do.

It helps when investors want to avoid speculative growth assumptions. The formula does not allow the analyst to justify a high price with an optimistic long-term story. If earnings and book value do not support the number, the screen stays conservative.

It helps as a first-pass filter. A simple formula can quickly separate situations that may deserve more research from situations where the defensive value case is weak.

It also helps discipline valuation language. Because the model is intentionally narrow, it reminds investors that a low-looking multiple is not the same as a complete thesis.

This fits the broader idea in Core Metrics in Stock Analysis: What to Understand Before Valuation: models should be built on clear, role-aware inputs. A valuation shortcut is only as useful as the accounting signals behind it.

Where the Graham Number fails

The Graham Number fails when investors treat it as universal.

It can fail for asset-light companies. Many modern businesses create value through software, brands, intellectual property, networks, customer relationships, and data. Those strengths may not appear fully in book value.

It can fail for high-quality growth companies. A business with strong reinvestment opportunities may look poorly supported under Graham-style book-value logic even if it has excellent economics.

It can fail when earnings are temporarily depressed or temporarily inflated. Cyclical businesses can swing sharply under the formula simply because EPS moves sharply.

It can fail when accounting equity is distorted. Large buybacks, write-downs, acquisitions, and accumulated losses can change book value in ways that do not cleanly reflect future economic value.

It can also fail for financial companies, real estate companies, and other specialized balance-sheet businesses if book value requires industry-specific interpretation. The formula is simple; the financial statements may not be.

That is why How to Evaluate Business Quality Before Valuing a Stock still matters. The Graham Number can screen for conservative value, but it cannot explain competitive advantage, pricing power, capital allocation, cash-flow quality, or durability.

How to read a Graham Number output

A Graham Number output should be read as a conservative screening signal.

It should not be read as:

a buy signal
a sell signal
a target price
a complete fair value estimate
proof that a stock deserves more capital
proof that a stock should be avoided

The useful questions are more disciplined:

Are EPS and BVPS both positive?
Are earnings stable enough for the formula to mean something?
Is book value economically relevant for this business?
Is the company asset-heavy or intangible-heavy?
Could buybacks, write-downs, or accounting distortions affect BVPS?
Does the Graham Number agree or conflict with other models?
What does the formula ignore?

Those questions prevent the model from becoming a shortcut to false confidence.

The Graham Number belongs inside a broader system. It can act as a conservative value anchor, but it should be weighed against business quality, financial strength, cash-flow models, and historical context. Which Signals Matter Most When Evaluating a Company? makes the same point at the signal level: evidence matters because of the question it answers.

How it fits with other value models

The Graham Number is much simpler than DCF, Buffett Intrinsic Value, or Earnings Power Value. The seven-model valuation guide places this defensive screen beside intrinsic-value and relative-ranking approaches without treating their outputs as interchangeable.

A DCF asks what future cash flows may be worth under explicit assumptions. Buffett Intrinsic Value asks what a business may be worth to a long-term owner based on owner earnings and conservative assumptions. The Graham Number asks whether current earnings and book value provide a defensive margin-of-safety anchor.

That makes it useful, but narrow.

If a company passes a Graham-style screen, the next step is not to stop. The next step is to ask why. Are the earnings durable? Is book value economically meaningful? Is the business in decline? Is the balance sheet carrying hidden risk? Is the market discounting something important that the formula cannot see?

If a company fails the Graham-style screen, the conclusion is also not automatic. The business may still be high quality, cash generative, and valuable under a different model. The Graham Number may simply be the wrong lens.

This is how analyzing a stock systematically should work. One model should sharpen the question, not replace the process.

What the Graham Number is best used for

The Benjamin Graham Number is useful because it is strict.

It forces the investor to look at earnings and book value before getting carried away by growth stories. It rewards conservatism. It makes a margin-of-safety question visible. It can be a helpful screen for traditional, asset-backed businesses where accounting equity carries real information.

But the same simplicity creates limits.

The formula does not understand brands, ecosystems, intangible capital, software economics, reinvestment quality, or future cash-flow durability. It can miss excellent businesses and overemphasize weaker businesses that happen to have accounting equity.

That is why the Graham Number should be treated as a screening tool rather than a thesis.

Used well, it helps investors ask a conservative question: do earnings and book value support enough further research? Used poorly, it turns a narrow formula into a false verdict.

Inside a StockGeniuses-style workflow, the Graham Number earns its place when it keeps the question narrow. It clarifies whether earnings and book value support further investigation; it should not pretend to answer quality, growth, or cash-flow questions by itself.