Graham Fair Value: How to Estimate a Stock's Intrinsic Worth

Graham's most important contribution to investing may not be the formula itself but the concept of margin of safety: the idea that you should only buy a stock when its market price is substantially below your estimate of intrinsic value.

This buffer protects you against three kinds of risk: errors in your assumptions, deterioration in the business after you buy, and general market downturns. Graham did not fix one percentage; the point is to demand a discount large enough to absorb being wrong.

On AlphaStocks, every fair value comes with a low–base–high range rather than a single point, and the Buffett model includes a margin-of-safety test of its own: price ≤ 75% of intrinsic value (owner earnings per share, 10 years at the company's own growth held to -5%…10%, then 3% forever, discounted at 10%).

The Graham model on AlphaStocks is not the Graham Number. It checks the seven criteria Graham set for the defensive investor in chapter 14 of The Intelligent Investor, computed as written: adequate size (annual sales ≥ $100M); strong financial condition (current ratio ≥ 2 and long-term debt ≤ net current assets); positive earnings every year (10 years); uninterrupted dividends (20 years); per-share earnings growth over ten years (≥ +33.3% (3-year averages, ten years apart)); moderate p/e on 3-year average earnings (≤ 15); moderate price to book value (P/B ≤ 1.5 or P/E × P/B ≤ 22.5). The score is the share passed out of the criteria that can be evaluated.

Graham himself made two sector exceptions, and the model makes only those:

One fair value per stock, from one method chosen by sector profile. The market price is never an input: no blending toward it, no hiding a value that disagrees with it. Each value comes with its method and a low–base–high range, and is withheld only when the method's inputs are invalid, with the reason shown. The fair value is displayed; it is not an input to the composite score.

For an ordinary company the method is: Owner earnings per share (net income + D&A − capex, 3-year average) grown for 5 years at the company's 5-year revenue growth (0 to 13.95%), faded over 5 years to 3%, discounted at the Damodaran industry cost of equity, with a terminal value at the stable-growth cost of equity (8.41%). Banks, insurers and holding companies use a justified price-to-book, utilities and asset managers a justified P/E, and REITs FFO times their property group's P/FFO. Every method and parameter, with its source and date, is on the methodology page.

The Graham model carries 45% of the Value axis in the general profile, and the Value axis carries 10%of the composite — plus it sets the Timing axis (15%), which is the lower of Value and Momentum.

It is deliberately weighted alongside two other valuation approaches — Lynch PEG analysis (25%) and Greenblatt Magic Formula (30%) — so that no single valuation method dominates. Only the Value models that apply to a company are averaged. A stock that appears cheap by Graham's standards but expensive by earnings yield and PEG will receive a moderate Value score, not a high one.

The Graham Number requires positive EPS and positive BVPS. For companies with negative earnings or negative book value (some high-growth tech firms, for example), the formula cannot produce a meaningful result, and AlphaStocks shows no Graham Number. In the Graham criteria, a P/E or P/B over negative earnings or book is undefined, and undefined counts as a fail — never as a cheap reading.

Book value itself can be misleading. Companies with large intangible assets (brand value, intellectual property, customer relationships) often have book values that significantly understate their economic worth. Conversely, companies carrying legacy assets at inflated book values may appear cheaper than they are.

These are not flaws in Graham's logic — they are limitations of any formula that relies on reported financial data. The most effective use of the Graham Number is as one data point within a broader analytical framework.