Methodology · v1.2.0
Replicly doesn't claim to know what Warren Buffett would buy. It takes what he has written and said, mostly in fifty-plus years of Berkshire Hathaway shareholder letters, and turns it into a consistent process you can inspect. Where Buffett gave a number, we use it. Where he didn't, we say what threshold we chose and why.
Three rules hold the whole thing together:
Buffett found most of his ideas by reading: Moody’s manuals page by page in the 1950s, annual reports ever since. The Desk does the same search every two weeks, in four stages. A sweep of SEC data for every US filer keeps the companies any screen below could still pass (about one in six). We then load each survivor’s full history and price and run the screens properly. Up to 80 of the best are read by Claude, who looks at the business description and risk factors for reasons to pass. Up to 35 of those it would look at further get the full analysis described on this page (fewer when the read rules more out, or an analysis can't be completed), and the Desk is ordered by what those analyses find.
Every screen is a set of rules over the same figures the reports show. A company passes when it meets all of them; one miss on a longer screen counts as a near miss. Companies that file a 20-F or 40-F are left out because their figures often aren’t in dollars per US share.
Berkshire's published acquisition test: large, consistent earnings, good returns on equity and little debt.
Sources: Acquisition criteria, printed in every Berkshire letter
High returns on capital, year after year, from a business that turns profit into cash and doesn't dilute its owners.
Sources: Berkshire Hathaway shareholder letters; Robert Hagstrom, The Warren Buffett Way (1994)
Steady returns on equity with no loss years, the test Buffett applied to Wells Fargo, GEICO and American Express.
Sources: Berkshire Hathaway shareholder letters
A business with proven returns that the market has marked down, as with American Express in 1964 and Coca-Cola in 1988.
Sources: Buffett Partnership letters, 1964; Berkshire letter, 1988
Benjamin Graham's seven tests for a cautious buyer: size, strength, an unbroken earnings record, growth and a moderate price.
Sources: Benjamin Graham, The Intelligent Investor (1973), chapter 14
Priced below two-thirds of its net current assets, Graham's bargain test that Buffett used in the 1950s. Often small and unloved.
Sources: Benjamin Graham, The Intelligent Investor, chapter 15; Buffett Partnership letters, 1957–1962
Small, consistently profitable companies with high returns, cheap on their cash earnings. Where Buffett said small sums could compound fastest.
Sources: Buffett, remarks to BusinessWeek, 1999
Joel Greenblatt's distillation of Buffett: a high return on capital bought at a high earnings yield.
Sources: Joel Greenblatt's magic formula (2005)
Buffett's principles stayed remarkably constant. What changed was what he could buy. A sum that fits in a small company's float behaves differently from Berkshire's hundreds of billions. We map each investor to the era whose constraints look most like theirs:
Small capital, a wide hunting ground. Graham-style bargains in small and overlooked companies, special situations, and deep margins of safety. Concentrated when conviction was high.
After See's Candies and Munger's influence: wonderful businesses at fair prices. Durable moats, high returns on capital, long holding periods, a handful of large positions.
Too big for small companies. Mega-cap franchises, whole-company acquisitions, and large cash reserves held patiently until prices make sense.
Your era sets the margin of safety we look for (40% for the partnership approach, 25% for quality compounders, 20% at Berkshire scale, adjusted for how you handle falls), the smallest company worth researching, and an illustrative range of holdings. These bands are our reading of his record, labelled as such, and they describe a profile type rather than advise you personally.
Buffett defines intrinsic value as the discounted value of the cash that can be taken out of a business during its remaining life. For most companies we discount owner earnings (net income plus depreciation and amortisation, minus the capital spending needed just to stand still) over ten years: five years at a growth rate drawn from the company's own history, cut by a quarter for caution, then fading to a terminal rate. The discount rate is the 10-year US Treasury yield plus a 3% hurdle.
Banks and insurers borrow as part of their business, so owner earnings don't describe them. For them we use a residual-income model: book value multiplied by (return on equity minus growth) over (discount rate minus growth).
We publish three scenarios and a sensitivity grid, never a single price target. Maintenance capex is estimated with Bruce Greenwald's method: capital spending beyond what revenue growth required, using the company's average ratio of plant and equipment to sales.
Invest only in businesses whose economics you can understand and whose future you can reasonably estimate. The size of the circle matters less than knowing its edges.
How we test it: Claude reads the latest 10-K business section and rates how predictable the company's economics are ten years out, citing passages. Low predictability means the engine stops short of a verdict.
Sources: Berkshire Hathaway shareholder letter, 1996; Berkshire Hathaway shareholder letter, 2014
A great business has an enduring moat that protects excellent returns on capital. The durability of the advantage matters more than its size today.
How we test it: Claude identifies the moat type (brand, switching costs, network, cost, regulation, scale) from filing evidence and judges durability. Every claim must quote the filing.
Sources: Berkshire Hathaway shareholder letter, 2007; “Mr. Buffett on the Stock Market”, Fortune, 22 November 1999
The single most important decision in evaluating a business is whether it can raise prices without losing business to a competitor.
How we test it: Judged from filings (price/mix commentary, gross margin stability) and checked against the gross-margin history.
Sources: Interview with the Financial Crisis Inquiry Commission, 26 May 2010; Berkshire Hathaway shareholder letter, 1983
Berkshire's acquisition criteria ask for demonstrated consistent earning power; future projections and turnarounds don't count.
How we test it: Our thresholds: no loss years in the last decade and operating-margin volatility under 5 percentage points. One loss year or 5–8 points is mixed.
The primary test of managerial economic performance is a high earnings rate on equity capital employed, without undue leverage or accounting gimmickry.
How we test it: Our threshold: average ROE of 15% or more with at least seven of the last ten years above 15%, and debt payback under five years so the returns aren't borrowed.
Sources: Berkshire Hathaway shareholder letter, 1979; Berkshire Hathaway shareholder letter, 1987
The best businesses need little incremental capital to grow and earn very high returns on the capital they do use.
How we test it: Our threshold: average after-tax return on invested capital of 12% or more (8–12% is mixed). Not applied to banks and insurers.
Berkshire looks for businesses earning good returns on equity while employing little or no debt.
How we test it: Our threshold: total debt repayable from four years or less of normalised owner earnings, and interest covered at least eight times. Banks and insurers are judged on capital instead.
Sources: Berkshire Hathaway shareholder letter, 2014; An Owner's Manual (Berkshire Hathaway)
What matters is cash an owner could take out: reported earnings plus depreciation and amortisation, less the capital spending needed to keep the business competitive.
How we test it: We compute owner earnings each year, estimating maintenance capex with Greenwald's method. Normalised owner earnings must be positive and free cash flow should average at least 80% of net income.
A gruesome business grows rapidly, requires significant capital to engender the growth, and then earns little or no money. Avoid it.
How we test it: Our threshold: capex averaging under 1.5× depreciation across the history.
Earnings should be retained only if each dollar kept creates at least a dollar of market value for owners over time.
How we test it: Market value created across the period divided by earnings retained (net income less dividends). Above 1× passes. Needs a price history, so it can be unknown.
Sources: Berkshire Hathaway shareholder letter, 1984; An Owner's Manual (Berkshire Hathaway)
Repurchases make sense only below intrinsic value; management should allocate capital the way an owner would, and say plainly when it can't use it well.
How we test it: Claude reviews MD&A and capital-return history. We pair it with the share-count trend: a shrinking count passes the deterministic half.
Sources: Berkshire Hathaway shareholder letter, 2011; Berkshire Hathaway shareholder letter, 1984
Buffett wants managers who report candidly, admit mistakes, and think like owners. Promotional language and constant adjustments are warning signs.
How we test it: Claude scores candour from the filing's own language: plain disclosure of problems versus promotional framing and non-GAAP emphasis.
Sources: An Owner's Manual (Berkshire Hathaway); Berkshire Hathaway shareholder letter, 2002
Intrinsic value is the discounted value of the cash that can be taken out of a business during its remaining life. It is an estimate. It is never a precise figure.
How we test it: A two-stage owner-earnings DCF (residual income for banks and insurers) under bear, base and bull assumptions, discounted at the 10-year Treasury plus a 3% hurdle. Every assumption is shown and adjustable.
Sources: An Owner's Manual (Berkshire Hathaway); Berkshire Hathaway shareholder letter, 1992
Buy only when price sits well below a conservative estimate of value, so errors in judgement don't become permanent losses.
How we test it: Price compared with the base-case value against the margin your profile requires (25% by default). The bear case is shown alongside.
Sources: Berkshire Hathaway shareholder letter, 1992; Benjamin Graham, The Intelligent Investor, ch. 20
Interest rates act on valuations the way gravity acts on matter: the higher the rate, the less any future cash is worth today.
How we test it: The owner-earnings yield is compared with the 10-year Treasury yield, and the Treasury anchors every discount rate.
Sources: “Mr. Buffett on the Stock Market”, Fortune, 22 November 1999
Market value relative to GDP is probably the best single measure of where valuations stand at any moment.
How we test it: We publish the Buffett indicator (Fed Z.1 equities ÷ GDP) and use it to frame how patient a Buffett-style investor would be.
Sources: “Warren Buffett on the Stock Market”, Fortune, 10 December 2001
Diversification is protection against ignorance. An investor who understands a few businesses well should own a few, in size.
How we test it: Your lens sets an illustrative position range from your experience and time budget.
If you aren't willing to own a stock for ten years, don't even think about owning it for ten minutes. Our favourite holding period is forever.
How we test it: Profiles with horizons under five years get a caution: the method assumes time for value to be recognised.
Sources: Berkshire Hathaway shareholder letter, 1996; Berkshire Hathaway shareholder letter, 1988
Small sums can go where big money can't. Buffett has said he could compound small capital far faster than Berkshire, by fishing among small, overlooked companies.
How we test it: Your capital decides which Buffett era your lens draws on, the smallest company size worth researching, and the margin of safety we look for.
Sources: Interview, BusinessWeek, 5 July 1999 (“I could make you 50% a year on $1 million”); Berkshire Hathaway shareholder letter, 2014; Buffett Partnership letters, 1957–1969 (generals, workouts and controls: January 1962 letter)
Buffett's own advice for investors without the time or inclination to analyse businesses is a low-cost S&P 500 index fund.
How we test it: Profiles with little time or experience see this note first. We say it because he did.
Sources: Berkshire Hathaway shareholder letter, 2013; Berkshire Hathaway shareholder letter, 1993
Warren Buffett and Berkshire Hathaway have no connection with Replicly. Prompt version [email protected]. Questions or corrections: see about us.