Basic Prerequisites: 01 · From Zero (price vs. value, the margin of safety).
1. Intuition & Practical Objective
Benjamin Graham invented the discipline of systematic fundamental analysis, and his criteria are worth learning literally because they are the ancestor of every screen you will ever run. This page lays out the two Graham rule-sets - defensive and enterprising - and the deeper valuation tools from Security Analysis (earnings power, asset-value floors, net-nets) that turn his heuristics into an analyzable method.
Three ideas organize everything Graham wrote:
- Two kinds of investor, two bars. The defensive (passive) investor wants safety and simplicity: large, financially strong, dividend-paying, stable-earning firms bought at moderate multiples. The enterprising (active) investor is willing to do the work and accepts less-liquid or temporarily unpopular situations in exchange for a higher required margin of safety (e.g. P/E ≤ 10, or a price at two-thirds of net current assets). The same principles, two thresholds.
- Safety is in the balance sheet and the dividend record, not the forecast. Graham distrusted growth projections (“the miraculous workings of compound interest … a new kind of philosopher’s stone which can produce or justify any desired valuation”). His defensive tests lean on what has already happened - ten years of earnings, twenty years of dividends - precisely because those are auditable.
- Every criterion is a floor, and the floors combine. Cheap and strong and stable and growing is far more restrictive than any one test; Graham designed the set so its intersection leaves only genuinely defensive candidates. (This is the screening philosophy that 03 · Screening Metrics modernizes.)
2. Mathematical Ground Truth & Derivations
The defensive criteria (The Intelligent Investor, Ch 14). A defensive stock satisfies all of:
The last two are Graham’s paired price caps; their product gives the Graham Number ceiling derived on 01.
The enterprising criteria (Ch 15) - the harder bar. A representative enterprising mix: a P/E below about 10, a P/B no greater than 1.5 (or price no more than two-thirds of net current assets), and a dividend yield at least two-thirds of the AA-bond yield - with the understanding that the investor does the extra analytical work.
Asset-value floor - the net-current-asset (net-net) rule (Security Analysis). Define net current asset value per share as current assets minus all liabilities (including long-term debt), over shares:
The logic is a liquidation floor: if the company were wound up, current assets alone might cover the price several times over, so the operating business comes “for free” - the purest margin of safety Graham ever wrote down.
Earnings-power value (EPV). When there is no asset floor, capitalise normalized (mid-cycle) earnings at the required return:
where is the cost of capital. EPV is the value of the business as it is now, with no growth assumed - the conservative anchor Greenwald later formalized.
3. Computational Implementation - net-nets, EPV, and the enterprising test
Runs on the standard library only. It evaluates “Deep Value Co” against Graham’s three deeper tools: the net-net liquidation floor, earnings-power value, and the enterprising P/E and P/B caps.
The three answers disagree, and that is the lesson: the stock passes the enterprising multiple caps (P/E 7.5, P/B 1.29) but is not a net-net (price $9.00 equals NCAV, above the two-thirds floor of $6.00), while EPV of $18/share says the business - no growth assumed - is worth twice the price. A mechanical net-net screen would reject it; an earnings-power analysis would flag it as cheap. Graham’s mature view (Greenwald’s “three buckets”) is exactly this: asset value, earnings power, and franchise value are separate estimates, and you take the most conservative one that applies.
4. Failure Modes & First-Principles Breakdowns
- The criteria were written for 1970s industrials. The $100M sales floor and the utilities’ asset test are period-specific; a modern screen must rescale them (the form of the tests survives, the constants do not). Apply the shape, not the dollar figure.
- Net-nets barely exist now, and when they do it is usually for a reason. A stock trading below two-thirds of liquidation value frequently has a shrinking business, a governance problem, or a value-destroying manager who will burn the asset floor - the value trap in its purest form (05 · Failure Modes).
- Earnings-power value inherits the accounting. “Normalized earnings” is a judgment: capitalising a peak-cycle year overstates EPV; capitalising a trough year understates it. EPV is only as honest as the earnings you feed it (the accrual problem of Accounting Quality & Red Flags).
- Criteria as a checklist give false precision. Passing all seven Graham tests does not make a stock safe - it makes it superficially defensive. Graham himself insisted the mechanical tests were a starting point, and that the analyst’s reading of the business decides. This is why the systematic translation (Piotroski’s F-score on cheap stocks) is more robust than the raw 1930s constants.
5. References
- Graham, Benjamin: The Intelligent Investor (rev. 1973; HarperBusiness annotated 4th ed. 2003)
- Graham, Benjamin & Dodd, David: Security Analysis (McGraw-Hill, 6th ed. 2008)
- Greenwald, Kahn, Sonkin & van Biema: Value Investing: From Graham to Buffett and Beyond (Wiley, 2001)
- Fisher, Philip A.: Common Stocks and Uncommon Profits (Wiley reissue)
- Piotroski, Joseph D.: “Value Investing…” (JAR, 2000)
6. Connected Graph Bridges
- Back: 01 · From Zero · Index Hub
- Continue: 03 · Screening Metrics · 04 · The Research Workflow
- Value anchor: Equity Valuation - DCF, Comps & Value Logic (EPV as a no-growth DCF) · Core Financial Ratios (every criterion is a ratio)
- Systematic cousin: Quantitative Fundamental Investing (Graham’s criteria as a backtested factor)