Upgrade Your Graham Calculator: Worked Examples, Excel Formulas, and Key Limitations

Disclaimer: This article is for educational and informational purposes only. Nothing here constitutes financial advice or a recommendation to buy or sell any security. Always conduct your own research and consult a licensed financial advisor before making investment decisions. All data reflects figures current as of March 2026 and may have changed since publication.

Most Benjamin Graham calculators do one thing: spit out a number. You type in EPS and book value, you get a Graham Number. You either feel good about the stock or you don't. That's it.

That's not enough.

The Graham Number is a starting point for analysis — not an endpoint. To actually use it, you need to understand what goes into it, how to replicate it in a spreadsheet, what it looks like on real stocks with real data, and critically, when it gives you completely wrong answers.

This guide closes all of those gaps. By the end, you'll have Excel formulas you can copy, worked calculations on five real 2026 stocks, and a clear-eyed view of the model's failure modes. Then we'll show you how to go further with the free stock screener at valueofstock.com/screener, which applies the Graham Number across thousands of tickers automatically.

Quick Recap: What Is the Graham Number?

Before the Excel formulas, a brief grounding. Benjamin Graham — Warren Buffett's mentor and author of The Intelligent Investor — developed a formula to set a ceiling price for any stock. The Graham Number represents the most you should rationally pay based on two fundamentals:

Earnings Per Share (EPS) — Is the company making money?
Book Value Per Share (BVPS) — What are the company's assets actually worth?

The Formula:

Graham Number = √(22.5 × EPS × Book Value Per Share)

The constant 22.5 comes from Graham's twin rules:

Never pay more than 15× earnings (P/E ≤ 15)
Never pay more than 1.5× book value (P/B ≤ 1.5)
15 × 1.5 = 22.5

Any stock trading below its Graham Number is, by Graham's definition, potentially undervalued. The discount is called the margin of safety — the buffer that protects you against errors in your inputs or surprises in the business.

Margin of Safety = (Graham Number − Current Price) ÷ Graham Number × 100

Simple. Elegant. And often misapplied.

The Excel Formula Reference

Here's everything you need to build a Graham Number spreadsheet from scratch.

Setup: Your Column Structure

| Column | Cell | Data | |--------|------|------| | A | A2 | Ticker Symbol | | B | B2 | EPS (trailing twelve months) | | C | C2 | Book Value Per Share (BVPS) | | D | D2 | Current Stock Price | | E | E2 | Graham Number (formula) | | F | F2 | Margin of Safety % (formula) | | G | G2 | Verdict (formula) |

Cell E2 — Graham Number

=IF(AND(B2>0, C2>0), SQRT(22.5*B2*C2), "N/A")

Why the IF guard? The SQRT function throws an error if either EPS or BVPS is negative. The IF check returns "N/A" instead of crashing — useful when you're pasting in a list of 50 tickers and some have negative earnings.

Cell F2 — Margin of Safety %

=IF(E2="N/A", "N/A", ROUND((E2-D2)/E2*100, 1))

A positive result means the stock trades below its Graham Number (undervalued). A negative result means it trades above (overvalued by this metric).

Cell G2 — Conditional Verdict

=IF(F2="N/A", "⚠ Formula fails", IF(F2>=30, "✅ Deep value", IF(F2>=10, "🟡 Fair value zone", "🔴 Above Graham Number")))

This gives you an instant traffic-light read across your watchlist:

≥30% margin of safety → Deep value territory (Graham would approve)
10–29% → In the reasonable range, but not a screaming bargain
Negative → Trading above intrinsic value by this metric

Cell H2 — Graham Number as P/E × P/B Check (Verification Formula)

=IF(AND(B2>0, C2>0), SQRT(15*B2 * 1.5*C2), "N/A")

This is mathematically identical to the main formula (15 × 1.5 = 22.5), but written out in a way that makes the P/E and P/B logic visible. Use it to double-check your main formula or to explain the math to someone who's new to it.

Bonus: Current P/E and P/B

=IF(B2>0, ROUND(D2/B2, 2), "N/A")   '← P/E ratio
=IF(C2>0, ROUND(D2/C2, 2), "N/A")   '← P/B ratio

Graham's rules say P/E should be ≤ 15 and P/B should be ≤ 1.5 and their product should be ≤ 22.5. A stock can pass the Graham Number test while still having a P/E of 20 if its P/B is especially low — which is why checking both individually alongside the composite formula matters.

Worked Examples: 5 Real Stocks in March 2026

Let's run the formula on actual tickers from our screener database. All figures are sourced from Yahoo Finance (trailing twelve months, March 2026) and are verified independently.

1. Comcast Corporation (CMCSA) — The Textbook Case

| Metric | Value | |--------|-------| | EPS (TTM) | $5.39 | | Book Value Per Share | $26.89 | | Current Price | $30.16 | | Graham Number | $57.10 | | Margin of Safety | 47.2% | | P/E Ratio | 5.6× | | P/B Ratio | 1.12× | | Dividend Yield | 4.32% |

The calculation:

Graham Number = √(22.5 × $5.39 × $26.89)
             = √($3,261.4)
             = $57.10

Comcast trades at $30.16 against a Graham Number of $57.10 — a 47% margin of safety. Both P/E (5.6×) and P/B (1.12×) individually pass Graham's individual tests. This is one of the cleaner Graham-qualifying examples in the current market.

The caveat: Comcast is in structural cable decline. The formula measures value; it doesn't measure business quality or trajectory. Graham himself would have looked at earnings trend before pulling the trigger.

2. Lincoln National Corporation (LNC) — Financial Services, Deep Value

| Metric | Value | |--------|-------| | EPS (TTM) | $5.83 | | Book Value Per Share | $52.20 | | Current Price | $32.59 | | Graham Number | $82.75 | | Margin of Safety | 60.6% | | P/E Ratio | 5.59× | | P/B Ratio | 0.62× | | Dividend Yield | 5.43% |

Graham Number = √(22.5 × $5.83 × $52.20)
             = √($6,848.0)
             = $82.75

A 60% margin of safety is extraordinary. But financial companies like insurance firms carry complexity Graham acknowledged: their book values are sensitive to interest rate assumptions and actuarial estimates. Lincoln National recently went through a complex restructuring period.

Rule of thumb: For financial companies, treat Graham Number output as one signal among several — not a standalone buy signal.

3. Huntington Bancshares (HBAN) — Regional Bank, Moderate Discount

| Metric | Value | |--------|-------| | EPS (TTM) | $1.39 | | Book Value Per Share | $13.79 | | Current Price | $15.48 | | Graham Number | $20.76 | | Margin of Safety | 25.4% | | P/E Ratio | 11.1× | | P/B Ratio | 1.12× | | Dividend Yield | 3.96% |

Graham Number = √(22.5 × $1.39 × $13.79)
             = √(431.4)
             = $20.76

HBAN shows a cleaner picture than LNC — it's a straightforward regional bank, and the 25% margin of safety plus below-1.5 P/B ratio makes this a conservative Graham-style candidate. Not spectacular upside, but a reasonable margin of protection.

4. Regions Financial Corporation (RF) — Similar Profile, Lower Discount

| Metric | Value | |--------|-------| | EPS (TTM) | $2.30 | | Book Value Per Share | $20.39 | | Current Price | $25.32 | | Graham Number | $32.48 | | Margin of Safety | 22.0% | | P/E Ratio | 11.0× | | P/B Ratio | 1.24× | | Dividend Yield | 3.93% |

A 22% margin of safety puts RF in "acceptable" Graham territory — below his preferred 33% buffer but still offering meaningful downside protection. Comparing HBAN and RF side-by-side illustrates how two similar regional banks can offer different levels of Graham Number value at the same moment in time.

5. The Coca-Cola Company (KO) — A Famous Failure of the Formula

| Metric | Value | |--------|-------| | EPS (TTM) | $3.04 | | Book Value Per Share | $7.48 | | Current Price | $77.61 | | Graham Number | $22.62 | | Margin of Safety | −243% | | P/E Ratio | 25.5× | | P/B Ratio | 10.4× |

Graham Number = √(22.5 × $3.04 × $7.48)
             = √($511.6)
             = $22.62

The formula says Coca-Cola is worth $22.62. It trades at $77.61. By strict Graham Number logic, KO is massively overvalued.

Is it? Not necessarily. Warren Buffett — Graham's most famous student — has held Coca-Cola since 1988 and calls it one of his best investments ever. The problem isn't KO; it's the formula applied to the wrong type of business. KO's true value lies in its brand, pricing power, and consistent cash flow growth — none of which appear in the BVPS number.

This brings us directly to the limitations section.

When the Graham Number Breaks Down: 6 Critical Limitations

The Graham Number was designed for a specific type of company. Applied outside those bounds, it produces unreliable — sometimes absurd — results.

1. Negative EPS → Formula Returns No Answer

If a company is losing money (negative EPS), √(22.5 × negative × BVPS) requires taking the square root of a negative number. The math is undefined. The formula simply doesn't work.

In Excel: Your IF(B2>0, ...) guard returns "N/A" and protects the spreadsheet from crashing.
In practice: This disqualifies hundreds of growth companies, early-stage businesses, and companies in cyclical downturns.

2. Negative Book Value → Same Problem

Some high-quality businesses carry negative book value because they've returned so much capital to shareholders over decades. McDonald's (book value: −$2.52/share), Home Depot, and many other Dividend Aristocrats fail Graham Number math entirely. A negative BVPS makes the calculation undefined — yet these businesses are arguably some of the safest dividend payers on the market.

3. Asset-Light Business Models (Brand Power, SaaS, Tech)

Graham's formula implicitly assumes that book value matters — that if the business were liquidated, its assets would be worth something meaningful relative to earnings. This was true of 1950s industrial companies. It is not true of Apple (which would be worth far more than its balance sheet in any scenario), Coca-Cola (brand worth billions not on the balance sheet), or Salesforce.

For asset-light, moat-driven businesses: use a DCF model with earnings growth assumptions. The Graham Number will always make these stocks look dramatically overvalued.

4. Financial Companies: Book Value Is an Estimate

For banks and insurance companies, book value on the balance sheet is management's estimate of what the loan book and investment portfolio are worth. It depends on assumptions about credit losses, interest rate movements, and actuarial tables. A bank's "book value" can shift materially in a recession.

This doesn't make the formula useless for banks — it's actually widely used in bank valuation — but you should apply a larger margin of safety and cross-check with return on equity and net interest margin trends.

5. High-Growth Tech: EPS Is Too Low (Today)

For a company like NVIDIA in 2024, the Graham Number based on trailing EPS significantly understated the stock's fundamental value because earnings were accelerating. The formula looks backward (TTM EPS) while the stock's price reflects future earnings power. This isn't a flaw in the stock — it's a limitation of backward-looking metrics.

The fix: Graham himself acknowledged this in his revised formula, which multiplies EPS by (8.5 + 2g) where g is the expected long-term growth rate. But forecasting growth rates introduces subjective judgment that the simple Graham Number deliberately avoids.

6. The Constant 22.5 Assumes a Specific Interest Rate Environment

Graham derived his 15× P/E assumption during an era of roughly 4–5% AAA bond yields. When risk-free rates are higher (as in 2022–2023), equities deserve lower P/E multiples and the 15× ceiling makes sense or becomes generous. When rates are near zero (2020–2021), equities can rationally command higher multiples. The magic number 22.5 is not etched in stone.

Some practitioners adjust the P/E ceiling based on the current 10-year AAA yield: Adjusted P/E ceiling = 4.4 ÷ Current AAA yield. At a 5% yield, that gives you a ceiling of 8.8× — significantly more conservative than 15×.

From Formula to Action: Using the Screener

Working through Graham Numbers manually is educational, but impractical at scale. The valueofstock.com screener applies the Graham Number formula across thousands of tickers continuously — showing you margin of safety, EPS, BVPS, and a safety-weighted composite score.

The stocks in the worked examples above — CMCSA, HBAN, RF, and others — surface in the screener automatically when you filter by margin of safety. You can sort by sector, filter by yield, or combine Graham Number criteria with dividend metrics to find income-producing value stocks.

Where to start:

Filter by Margin of Safety ≥ 20% to see stocks trading meaningfully below Graham Number
Add a Dividend Yield ≥ 2% filter to narrow to income-paying value stocks
Check the Payout Ratio column — a high yield with a 100%+ payout ratio is a warning sign, not a feature
Sort by P/E within the results to quickly spot the lowest-multiple names

No Excel spreadsheet required — but now you understand exactly what the screener is calculating.

Summary: The Graham Number Toolkit

| Use Case | Tool | |----------|------| | Quick single-stock check | Free calculator → valueofstock.com/screener | | Personal watchlist analysis | Excel formula: =IF(AND(B2>0,C2>0),SQRT(22.5*B2*C2),"N/A") | | Margin of safety | =IF(E2="N/A","N/A",ROUND((E2-D2)/E2*100,1)) | | P/E individual check | =ROUND(D2/B2,2) | | P/B individual check | =ROUND(D2/C2,2) | | Growth stocks | Use DCF or Graham's revised formula instead | | Banks / insurers | Apply Graham Number + ROE trend + capital ratios | | Negative EPS/BVPS stocks | Formula not applicable — use another method |

The Graham Number is one of the sharpest tools in value investing — precise, fast, and grounded in first principles. Its limitations are real, but so is its track record when applied to the right types of businesses. Know what it measures, know where it fails, and you've upgraded from calculator user to genuine value analyst.

All stock data sourced from Yahoo Finance, trailing twelve months through March 2026. Data is provided for educational illustration only and does not constitute investment advice. See valueofstock.com/screener for current live figures.