This governs the two decisions that sit underneath every other financial decision: what a business is really worth, and where the next available dollar goes. The core claim is that a business is worth the cash it will hand its owners over its remaining life, discounted, and that almost every popular measure (earnings per share, growth, price history, reported net income) is a substitute someone invented because the real question is hard. Because these letters run from 1977 to 2024, they also record a man revising his own method in public, which is the most useful part of them.

Do

Judge performance by what a business earns on the capital already inside it. Rising earnings per share proves nothing, because a dormant savings account produces rising earnings through compounding alone, and a company that grows equity 10% while growing earnings 5% has actually gotten worse. Buffett replaced the popular measure with return on beginning equity capital, 19% in 1977. [Source: "1977-1978: The Owner-Oriented View"]

Count the earnings retained on your behalf, not just the cash mailed to you. In 1980 Berkshire's roughly one-third stake in GEICO produced about $3 million of reported dividend income while Buffett estimated Berkshire's real share of GEICO's earning power at about $20 million. His image for it: a tree growing in a forest you partly own is still growing, whether or not your statements record it. [Source: "1979-1980: The Earnings Iceberg"]

Require every dollar a business keeps to become at least a dollar of value. This is the only sound basis for judging whether a company should retain earnings, reinvest them, or hand them back, and it replaces any fixed payout rule. A business that must keep consuming capital at low returns is destroying value while reporting growth. [Source: "1982: The Dollar-for-a-Dollar Test"]

Subtract the capital spending required just to stand still before you call anything profit. Buffett's "owner earnings" is reported earnings plus non-cash charges minus the capital expenditure a business genuinely needs to hold its competitive position and unit volume. He concedes that last figure is an estimate and defends the imprecision by preferring to be roughly right rather than exactly wrong. [Source: "1986: Owner Earnings Defined"]

Write down where your understanding actually ends, and stay inside the line. Buffett tells individual investors that intelligent investing is not complex but is far from easy, that they do not need to understand beta or option pricing, and that only two subjects matter: how to value a business, and how to think about market prices. What is required is honest self-knowledge about the boundary, not broad expertise. [Source: "1996: The Owner's Manual and the Circle of Competence"]

Want the price to fall on anything you are still buying. Someone who eats hamburgers but does not raise cattle should want beef cheaper, and an investor who will be a net buyer for years should want a falling market. Buffett rewrites the standard headline about investors losing when markets fall: disinvestors lose, buyers gain, and every transaction has both sides. [Source: "1997: Market Prices and Patience"]

Hold enough spare cash that nobody else's panic can force your hand. Berkshire's standing pledge is to never depend on the kindness of strangers, and that discipline is exactly why it could put $14.5 billion into Wrigley, Goldman Sachs and General Electric securities in 2008 while the credit market was frozen. Buffett states he will not trade a night's sleep for the chance of extra profit. [Source: "The 2008 Letter: Buying Into the Panic"]

Put money into your best idea rather than your twentieth. Buffett argues concentration can lower risk, because it forces harder thinking and a higher bar before buying, and asks why anyone who has found five to ten sensibly priced businesses with real advantages would fund their twentieth favorite instead of adding to their first. Berkshire settled for needing to be right very few times, even one good idea a year. [Source: "1993: Risk, Focus, and Diversification"]

If you cannot tell one business from another, buy a cheap slice of all of them and then stop. Buffett's instruction in his own will is 10% short-term government bonds and 90% a very low-cost S&P 500 index fund, and his stated goal for a non-professional is not to pick winners, since neither they nor their helpers can, but to own a cross-section that will do well in aggregate. He adds that the chatter urging activity is a cost, not a service. [Source: "2012-2013: Index Funds and the Case for Simplicity"]

Don't

Don't buy a mediocre business because the price is cheap. Buffett calls this the cigar-butt approach: one last free puff and no real smoke. He bought control of Berkshire itself this way in 1965 and a Baltimore department store in 1966, sold the store three years later for roughly what he paid, and drew the rule that time is the friend of the wonderful business and the enemy of the mediocre one. [Source: "1989: Mistakes and the Institutional Imperative"]

Don't keep funding a business whose losses have no visible end. Berkshire's textile mills got twenty years of patience and equipment that had cost about $13 million sold at liquidation auction for gross proceeds of $163,122, with looms bought at $5,000 apiece going for $26 as scrap. Buffett's own verdict on the delay is that 250 mills had closed since 1980 and their owners had no information he lacked, they simply processed it more objectively. [Source: "1985: The Textile Mill Closes"]

Don't buy anything with stock you believe is worth more than its price. A seller can hold out for full value, but a buyer issuing undervalued shares has no such leverage, so the deal quietly trades away two dollars of your own value to receive one of theirs. Buffett also reframes the announcement language: "Company A to acquire Company B" is more honestly read as part of A being sold to buy B. [Source: "1982: The Dollar-for-a-Dollar Test"]

Don't accept price movement as a definition of risk. Under beta-based theory Washington Post stock became riskier after it fell sharply in the early 1970s, precisely when Buffett was buying, even though a lower price for an unchanged business is plainly safer. His own definition is whether the after-tax proceeds will preserve your purchasing power plus a modest return, judged on the business, the management, the price, and expected tax and inflation. [Source: "1993: Risk, Focus, and Diversification"]

Don't let the daily quote tell you whether you were right. Graham's Mr. Market is a partner who shows up every day with a price, and the whole point is that he is there to serve you and not to guide you. In 1987 the Dow moved 2.3% for the year while portfolio-insurance programs mechanically dumped stock as prices fell, which Buffett compares to a farmer ordering his land sold every time a neighbor's plot changes price. [Source: "1987: Mr. Market"]

Don't run a risk you have not priced, and don't sit on one you have already spotted. Berkshire's General Re had priced property coverage off past windstorm, fire and earthquake experience while charging nothing for large-scale terrorism, and September 11th produced the industry's largest loss ever. Buffett names his own failure directly as violating the Noah rule: predicting rain does not count, building arks does. [Source: "2000-2001: The Bubble Bursts, and September 11th"]

Don't sign a long-dated contract whose risk you cannot read. After finishing the derivatives footnotes of major banks, Buffett said the only thing he understood was that he did not understand how much risk the institution was running. Berkshire inherited a derivatives dealer it never wanted, and ten months of active liquidation still left 14,384 contracts with 672 counterparties. [Source: "2002: Derivatives as Weapons of Mass Destruction"]

Don't treat growth as evidence of value. Growth only helps when a business can put incremental capital to work above the rate that returns a dollar of value per dollar deployed, and decades of growth financed at poor returns destroyed shareholder capital across the U.S. airline industry. Buffett also calls the "value versus growth" split fuzzy thinking, since growth is always one variable inside the calculation of value. [Source: "1992: Intrinsic Value Defined"]

Where they disagree

Is the thing you are buying a low price, or a good business? Buffett arrived trained by Ben Graham to buy statistically cheap companies below sell-out value, and did exactly that with Berkshire, Hochschild Kohn and Dexter Shoe, whose $433 million purchase paid in stock ended up costing shareholders billions once the stock appreciated and Dexter's advantage evaporated. Munger pushed the opposite blueprint, buy wonderful businesses at fair prices, told Buffett in 1965 that taking control of Berkshire had been a dumb decision, and was proven right by See's Candy, bought in 1972 for $25 million and earning $1.9 billion pretax on only $40 million of further investment. What settles it is the size of the money and the length of the hold: cigar butts worked on small sums in the 1950s and stopped scaling, and Buffett calls himself the slow learner of the two. [Source: "1989: Mistakes and the Institutional Imperative"] [Source: "The 50-Year Letter: Buffett's Retrospective and Munger's Essay"]

Which mistakes actually cost you, the ones you made or the ones you skipped? Buffett's public accounting is dominated by errors of commission, and he itemizes them: USAir preferred written down to a quarter of cost after sloppy analysis of deregulation, ConocoPhillips bought near peak oil prices, Irish bank stock down 89% in a year, Precision Castparts written down $11 billion because he was simply too optimistic about normalized profit. Munger, given his own section of the fiftieth-anniversary letter, says nearly all the big errors were omissions instead, names Walmart as one that was sure to work out, and estimates the missed opportunities cost at least $50 billion, far more than any visible blunder. Buffett does concede the point under a different name, calling it thumb-sucking as early as 1989, and Munger's last recorded use of the word makes it the one unforgivable failure: delaying the correction of a mistake you already know about. [Source: "The 50-Year Letter: Buffett's Retrospective and Munger's Essay"] [Source: "The Munger Tribute and Succession"]

Which single number should you grade yourself on? Buffett spent the first letters insisting on return on beginning equity capital, then abandoned that yardstick in 1982 when a growing equity base and earnings locked inside partly-owned companies made it misleading. He replaced it with per-share book value for nearly three decades, defending it as understated but stable enough to track intrinsic value year to year, and then retired that too in 2018, on the grounds that operating subsidiaries carried at depreciated cost, marked-to-market equities and buybacks had pulled book value away from what the business was worth. Through all of it he kept one rule constant, which is the transferable part: watch operating earnings, not GAAP net income, and change the yardstick the moment it stops tracking the thing you actually care about. [Source: "1982: The Dollar-for-a-Dollar Test"] [Source: "Buybacks and Operating Earnings"]

The one line

Work out what the business will hand its owners over its life, pay less than that, keep enough cash that nobody can force you to sell, and then leave it alone.