Unlike almost every other named screening method on the internet, Benjamin Graham’s criteria really were published, in print, with numbers attached. That difference matters more than it sounds: a rules list you can read is a rules list you can disagree with, test, and – crucially – know where it will fail. His defensive-investor standard comes from The Intelligent Investor; the enterprising variant adds the more aggressive bargain hunt described in Security Analysis and covered under his own profile here.
The published defensive criteria, and what each one is for
Graham split investors into two types by temperament rather than ambition: the defensive investor wants freedom from watching portfolios closely; the enterprising one is willing to do unusual work for unusual results. The defensive filter asks for:
- Adequate size. Small issues were historically less liquid and less scrutinised; size was never a claim about quality, only about survivability and information.
- Sufficiently strong current assets – traditionally around two dollars of current assets for every dollar of current liabilities. A crude solvency sniff test, fully computable from the two lines the balance sheet explains.
- Earnings stability: positive reported earnings across roughly a decade. Not profitability excellence – survival.
- Dividend continuity: an uninterrupted record stretching back years, on the theory that a board unwilling to cut a dividend has been pressured to keep the underlying cash flow honest.
- Modest earnings growth, explicitly modest. Demanding rapid growth reintroduces exactly the forecasting he wanted the defensive investor to avoid.
- A moderate price relative to earnings, commonly summarised around fifteen times recent average earnings.
- A moderate price relative to book, commonly summarised around one-and-a-half times, with a composite shortcut: the earnings multiple times the book multiple not exceeding roughly twenty-two-and-a-half. The logic of the product is elegant – pay little measured against profit and against recorded capital, so weakness in one measure has to be covered by genuine strength in the other. The full argument for when book value means anything, and when it doesn’t, is on its own page.
For the enterprising investor there is also the famous deep-value case: buy when market capitalisation sits meaningfully below net current assets – current assets minus all liabilities – sometimes framed around two-thirds of that figure. Run the arithmetic on hypothetical figures: roughly $150 million of current assets and roughly $90 million of total liabilities leaves about $60 million of net working capital; two-thirds is around $40 million; ten million shares outstanding puts the action price near four dollars. Below that, liquidation economics rather than optimism decide whether you lose money. Such setups have become scarce, and when one appears on no news, the probability is high that something specific is wrong rather than that everyone missed it.
Why passing isn’t the signal people think it is
A screen encodes a hypothesis about what makes outcomes favourable. This one says: cheap relative to measured profit and recorded assets, plus boring evidence of durability, should produce better results over time than buying whatever is popular. Across long stretches and large baskets, that proposition has support. Per name, it produces three categories nobody advertises:
Cheap and improving – the objective, invisible in advance.
Cheap because it’s dying. Structural decline looks identical to temporary pessimism on a screen, and worse-looking numbers arrive later, when you’ve already bought. Margin of safety is supposed to protect against being wrong about the first group while guaranteeing nothing about the second – see the valuation discussion.
Cheap because the balance sheet lies. Receivables that never convert, inventory nobody wants at any price, goodwill from a deal gone wrong. Every screen trusts accounting inputs. Debt levels deserve their own scrutiny, since the difference between a bargain and a solvent-looking insolvency usually lives in the leverage column.
There’s a structural issue specific to our period. Rules built around price relative to book weight capital-intensive, asset-heavy businesses heavily, because those are the firms whose equity is large enough to make the denominator interesting. Economies that create value through contracts, code, brands and habits carry very little recorded capital, so they either fail the test outright or look expensive forever. Apply a vintage filter rigidly in a modern market and you’ll often end up owning yesterday’s business models, chosen precisely because their assets were easy to count. That is not an argument against the discipline – it’s an argument for knowing which decades it flatters.
The flip side deserves saying too, because this site covers both approaches: a quality screen hunting demonstrated high returns systematically skips everything a value screen likes, since durable performers get re-rated long ago. Ask the same question of each – does this rule describe a reason outcomes should improve, or does it merely describe what won recently?
Making it operational without paying anyone
Free screeners expose the necessary fields: trailing earnings multiple, price to book, current ratio, dividend history, long-term debt against equity, earnings record length. Set the criteria deliberately, export the candidates, then leave the tool entirely. The screening layer is coarse sieving; the judgement happens in filings, notes and last year’s letters, where you find out whether the cheap earnings included a tax settlement or a factory sale.
Two habits matter more than tuning thresholds. First, keep your own written record of why each candidate looked attractive, including the mechanism and the observation that would prove you wrong – otherwise the position becomes whatever story is most comfortable in hindsight. Second, hold the disposition Mr. Market teaches rather than adopting the crowd’s mood along with its prices: an offer to sell you a decent business cheaply is an event to evaluate, not a verdict to obey, and the same quote that tempts you to buy should be readable as temptation to sell in six months.
Treat every threshold as arguable, including his – interest-rate environments, share-buyback norms and asset mix have moved since these rules were written, and someone who cannot explain why a cutoff is two rather than five is repeating superstition with a citation. Read the original books where possible; secondary summaries drift, including this one, and published statistics quoted anywhere without a date attached expire faster than the paper they’re printed on.
This is educational explanation of a published investment method, not investment advice or a recommendation to buy anything. Screens select on reported figures that can be stale or misleading, cheap stocks can and do fall further, passing a filter is not a signal to purchase, and any holding can lose money regardless of how conservative its ratios looked on the day you bought it.