Berkshire bought no company outright in 1981. It came quite close to one major purchase, then walked away when the seller's final price would have left Berkshire's owners poorer. Around that non-event Buffett builds his most famous explanation of why other companies, in the same spot, charge ahead anyway.

His starting rule is blunt: regardless of what it does to next quarter's reported numbers, he would rather buy 10% of Wonderful Business T at X per share than 100% of it at 2X. Most corporate managers prefer exactly the reverse, and, he notes, have no shortage of stated reasons. The real reasons, he suspects, are three, and usually unspoken. Leaders relish activity and challenge (even at Berkshire, he admits, "the corporate pulse never beats faster than when an acquisition is in prospect"). Managers are ranked and paid by size, not profit: ask a Fortune 500 chief where his company stands on that famous list, and the answer will come from the ranking by sales. Where it stands on the profitability list Fortune compiles just as faithfully, he may not know. And many managements, Buffett writes, seem to have been overexposed in childhood to the tale of the handsome prince freed from a toad's body by a princess's kiss, and are certain their kiss will do wonders for the profitability of Company T(arget).

That optimism is the whole game. Without it, why would the buyer's shareholders want to own the business at 2X when they could buy it themselves at X? As he puts it: "investors can always buy toads at the going price for toads." If they bankroll princesses who insist on paying double for the right to kiss one, those kisses "had better pack some real dynamite." The record says they rarely do. Many managerial princesses remain serenely confident about their kisses, he observes, "even after their corporate backyards are knee-deep in unresponsive toads."

Berkshire's own 1981 test case ended quietly. It came close to a major purchase involving a business and a manager it liked very much, but the price finally demanded would have left its owners worse off than before. "The empire would have been larger, but the citizenry would have been poorer." He is happy to skip small deals too, passing along a maxim he likes: "If something's not worth doing at all, it's not worth doing well."

In fairness, Buffett writes, some acquisition records have been dazzling. He finds exactly two patterns behind them, names the four executives he most admires, and then grades his own record against both standards. He does not award himself a pass.

Category one is about what you buy. Companies that, "through design or accident," purchased only businesses adapted to inflation (able to raise prices easily and to grow dollar volume with only minor added capital) have done excellently, even under managers of ordinary ability. The catch is that very few such businesses exist, and the competition to buy them has become, in his words, fierce to the point of being self-defeating.

Category two is about who does the buying: the managerial superstars, people who can recognize "that rare prince who is disguised as a toad" and actually peel away the disguise. He salutes four by name: Ben Heineman at Northwest Industries, Henry Singleton at Teledyne, Erwin Zaban at National Service Industries, and especially Tom Murphy at Capital Cities Communications, a "twofer" whose deals sit in category one and whose operating talent puts him atop category two. Such records, Buffett notes, are rare, and the men who own them know it: these champs have made very few deals lately, and often decided the best use of their capital was buying back their own shares (readers of last year's letter will recognize the move).

Then the self-grading, carried out in public. "Your Chairman, unfortunately, does not qualify for Category 2." As for category one, he understood the economics but acted too little: "Our preaching was better than our performance. (We neglected the Noah principle: predicting rain doesn't count, building arks does.)" The confession comes with a fresh example: only last year he had volunteered his expert opinion on the rosy future of the aluminum business, an opinion since revised by adjustments "aggregating approximately 180 degrees." Where Berkshire has done well, he says, is princes that were princes when purchased, and above all in buying fractional interests in easily identifiable princes at toad-like prices: the stock portfolio, in a sentence.

Last year's letter taught owners to count the earnings their part-owned companies keep and reinvest. This year the idea returns with bigger numbers and a sharper conclusion: the profit on Berkshire's own books is now the smaller part of the truth.

The invisible earnings, he reports, have grown to exceed the reported kind, and he expects that to continue. Just four holdings (GEICO, General Foods, R. J. Reynolds and The Washington Post) should pile up well over $35 million of undistributed, unrecorded earnings on Berkshire's behalf in 1982. Set that against the $39.7 million of operating earnings the 1981 books actually show, and last year's iceberg image still holds: the mass below the waterline outweighs the tip above it. He is candid that the reward arrives on no schedule: market values track business values well over long periods, but in any given year the relationship "can gyrate capriciously."

Why keep so much of the company's future in businesses it doesn't control? Choice, mostly. The market will sell you small pieces of exceptional businesses at reasonable prices. Whole ones come up rarely, and almost always at high prices (Lesson 1, from the other side of the counter). And mistakes mend faster: a mistaken stock can be sold, while a mistaken subsidiary is a long project. "Lack of control, in effect, often has turned out to be an economic plus." Accountants do not record such heresies. Buffett banks them anyway. And when a purchase does disappoint, his diagnosis sheet has only three possible entries.

1981 obligingly supplied the proof. Berkshire's net worth grew by $124 million, about 31%, and more than half of that came from the market performance of a single holding, GEICO. He is quick to add that securities prices outran business values that year, and that "Such market variations will not always be on the pleasant side." The long record now reads: book value $19.46 per share when the present management arrived seventeen years ago, $526.02 at year-end 1981, a compounding of 21.1% a year.

Here is the actual list. By the end of 1981 Berkshire's net interest in common stocks had cost $351.7 million and was worth $639.2 million. GEICO alone accounted for $199.8 million of the worth, nearly a third of the whole portfolio.

The list is tighter than a year ago: several of 1980's smaller names have dropped off it, two newcomers (Arcata and GATX) appear, and the R. J. Reynolds position has grown to about nine times its year-ago cost, making it the largest holding by money put in. Aluminum tells its own story: Kaiser Aluminum, 3% of which Berkshire's insurers held a year earlier, is no longer on the list, and the Alcoa stake shown is smaller at cost than it was. The 180-degree adjustment of Lesson 2, in other words, was not merely rhetorical.

The letter's central economics lesson starts from first principles. The whole case for owning a business is that it earns more than passive money, parked in a bond, earns by itself. By late 1981, Buffett shows, that case had quietly collapsed for the average American company.

Several decades back, the bar was low. Long-term taxable bonds paid 5%, tax-exempts 3%, and American business earned about 11% on its equity, so even a 10% earner was a "good" business, worth more than the capital inside it. Stocks in aggregate sold for over 150 cents per dollar of book value. By late 1981 the picture had inverted. Taxable bond yields ran past 16% and tax-exempts past 14%, while business earned about 14% on equity, before the owner's own taxes. The tax-exempt bondholder keeps every cent. The shareholder does not. For a person in the 50% bracket, a company earning 14% and paying it all out delivers the equivalent of a 7% tax-exempt bond, and a perpetual 7% bond, he notes, "might be worth fifty cents on the dollar as this is written."

Retaining the earnings does not escape the trap: growth of 14% a year still trails what passive money pays once the capital-gains tax (then a maximum of 20%) takes its cut on the way out. "Unless passive rates fall," he concludes, "companies achieving 14% per year gains in earnings per share while paying no cash dividend are an economic failure for their individual shareholders." By that standard most American businesses had become, in his uncomfortable word, "bad" businesses, and American equity capital in aggregate produced no value-added for individual investors. The failure is not athletic: companies were jumping a few points higher than a decade before. The crossbar of passive return had simply been raised much faster than anyone improved their jump.

He declines to re-run the full inflation sermon (copies of previous discussions, he offers, "are available for masochists") but repeats the conclusion that matters: Berkshire's efforts will keep doing "a much better job of filling your wallet than of filling your stomach." Berkshire itself retains earnings for offensive reasons, and its historic 21% return still clears the crossbar of after-tax passive return, "but barely." He puts the stakes plainly: "It would be a bit humiliating to have our corporate value-added turn negative." Then he applauds Chairman Volcker's fight at the Federal Reserve, and gives long-term inflation the one-line verdict that made the quotes section below.

If most businesses now earn too little to justify keeping the money, reason says they should hand their earnings back to the owners. Inflation, Buffett explains, forces the weakest of them to do precisely the opposite, and he gives the mechanism a memorably unpleasant name.

Start with the logic no bondholder would ever violate. A person stuck with a 5% bond does not take its interest coupons and pay full price for more 5% bonds while similar bonds sell for, say, forty cents on the dollar. He reinvests wherever the return is best. "Good money is not thrown after bad." What makes sense for the bondholder makes sense for the shareholder: a company earning splendid returns should keep its earnings and compound them, and a company earning poor returns should pay everything out and let its owners find better soil. Buffett notes that Scripture agrees, citing the parable of the talents: the high-earning servants are rewarded with 100% retention, while the unproductive one is chastised ("wicked and slothful") and has his capital redirected to the top performer.

Then inflation walks the logic through the looking glass. The bad business must retain every nickel, not despite being unattractive but precisely because it is so unattractive. To operate next year at this year's size, it needs more dollars for receivables, inventory and plant, whatever its reported profits say. "For inflation acts as a gigantic corporate tapeworm," consuming its required daily diet of investment dollars regardless of the host's health. Under 1981 conditions, a business earning 8% or 10% on equity often has nothing left for expansion, debt reduction or real dividends. "The tapeworm of inflation simply cleans the plate."

The inability to pay is often dressed up rather than admitted. Companies push dividend reinvestment plans, sometimes discounted to the point of all but forcing owners to hand the dividend straight back. Others "sell newly issued shares to Peter in order to pay dividends to Paul." His warning is one sentence: "Beware of 'dividends' that can be paid out only if someone promises to replace the capital distributed." Berkshire, for its part, retains earnings "for offensive, not defensive or obligatory, reasons": because it believes it can compound them, not because the tapeworm left it no choice.

The insurance section opens with a borrowed warning about forecasts, then makes one anyway: 1982 will be the worst year in recent history for insurance underwriting. A forecast with no danger in it, Buffett says, because the result "already has been guaranteed" by the prices written in 1981.

The mechanism is the frozen price. Prices are fixed for the life of the contract, so "if you make a mistake in pricing, you have to live with it for an uncomfortable period of time." In 1981 industry premiums written grew just 3.6%, the slowest of the decade, while the combined ratio reached 105.7, the worst since 1975. Buffett's benchmark makes the future automatic: with claims costs expected to rise at least 10% a year, premium growth of about 10% is needed merely to hold the loss level steady, and every point below that pace quickens the rot. Quarterly figures through 1981, he adds, showed the deterioration accelerating.

The bitter twist is that the record was achieved "in spite of good luck, not because of bad luck." Hurricanes had politely stayed at sea, and motorists were driving less. "They won't always be so obliging." Meanwhile the bond trap of last year's letter kept squeezing: insurers carrying bonds "valued, for accounting purposes, at nonsensically high prices" had little choice but to keep the cash revolving by "selling large numbers of policies at nonsensically low prices," fearing a drop in volume more than an underwriting loss. Since it is difficult, as he notes, to price much differently than your most threatened competitor, and virtually none would shrink to the point of significantly negative cash flow, prices stayed bad for everyone. Commentators kept talking of the underwriting "cycle," with its comforting implication of a rhythm. Buffett's view was darker: large losses as the norm, with the best years of the coming decade possibly beneath the average of the last.

Berkshire had no magic formula, only discipline. Managers Phil Liesche, Bill Lyons, Roland Miller, Floyd Taylor and Milt Thornton had done "a magnificent job of swimming against the tide," sacrificing much volume while keeping a substantial underwriting edge on the industry. Berkshire's own underwriting profit thinned to $1.5 million before tax, from $6.7 million. The outlook: continued low volume, maximum financial flexibility (a very rare condition in the industry), and patience, since "should fear ever prevail throughout the industry, our financial strength could become an operational asset of immense value." As for GEICO, its extreme and improving efficiency left it, in his view, better protected than almost any major insurer: "a brilliantly run implementation of a very important business idea."

The letter's last stretch is not about markets at all. In October 1981 Berkshire began letting each shareholder direct where the company's charitable money goes, an idea conceived by Charlie Munger, and the owners answered with a response Buffett says he had never seen matched.

The numbers were remarkable for a voluntary reply: of 932,206 eligible registered shares, 95.6% responded, over 90% even setting aside Buffett-related shares, and the response came, he notes, "without even the nudge of a company-provided return envelope." More than 3% of shareholders went further and wrote letters, all but one approving. In all, $1,783,655 went to about 675 charities of the owners' choosing. The "father-knows-best" school of corporate governance, he adds, will be surprised to learn that not one shareholder asked Berkshire's officers to choose for them, in their superior wisdom, and none suggested matching the directors' pet charities, a practice he describes, deadpan, as popular, proliferating and non-publicized.

The one sour note came from the middlemen. The Treasury's tax ruling arrived in early October and covered only shares registered in the owner's actual name, so Berkshire wrote to everyone on October 14, urging brokers to forward the letter to owners holding through them in time to re-register by the November 13 record date. Many never did. "The results from our urgings," Buffett reports, "would not strengthen the case for private ownership of the U.S. Postal Service." One of the largest brokerage houses, which claimed to hold stock for sixty clients (about 4% of all Berkshire's shareholders), apparently forwarded the letter roughly three weeks after receiving it, too late for all sixty. Its billing department, Buffett notes, suffered no such delays: the invoice for mailing services arrived within six days.

His advice stands for any shareholder of anything: keep at least one share registered in your own name, so important news reaches you directly. And then the letter's warm close: whatever the titles, he and Munger work as partners, enjoying it "to almost a sinful degree," and enjoying the shareholders as financial partners too.