Net worth grew $613.6 million in 1985, or 48.2%, and Buffett's first move is to tell his owners to forget it. Halley's Comet passed the same year, and he does not expect to see either again.

Last year's letter carried what he now calls a wildly upbeat message: nothing much was in the works, but something big occasionally popped up. It did, four times, and book value per share has now compounded at 23.2% a year for twenty-one years, "another percentage that will not be repeated." The first anchor is the market. About a decade ago the only question was which bargain to choose. Today the same exceptional businesses sit on the shelf at full price, and Berkshire cannot find significantly undervalued stocks to buy (Exhibit 1).

The second anchor is far more telling. "And an iron law of business is that growth eventually dampens exceptional economics." No company he knows of has amassed $1 billion of equity and then earned 20% or more on it for ten years while reinvesting substantially all of its profits. To average 15% over the next decade Berkshire needs $5.7 billion of profit, up 48% from the $3.9 billion he quoted a year ago, as it must be, since the capital base grew 48%. Leaving aside oil companies, only about 15 American businesses earned that much in the past ten years. "Their problem is our problem."

One more warning, for recent buyers. Berkshire's shares have historically sold modestly below intrinsic value, so a buyer's own result could at least match the business's. Lately the discount has gone and a modest premium has sometimes appeared (Exhibit 2). In aggregate "investor gains must equal business gains," so a newcomer who paid a premium needs it maintained merely to keep pace. His preference is a price that tracks value, and he credits the shareholders for it: virtually all are individuals, not institutions, and institutions, with their large staffs of experienced professionals, are not a force for stability. Ben Graham's prospector explained why forty years ago. Told the oil men's compound was full, he shouted "Oil discovered in hell," watched the oil men stampede out, and then followed them, in case there was some truth to the rumor.

Berkshire reported $488 million of pre-tax gains from selling securities, most of it from General Foods. Buffett wants you to know that the number says nothing about how 1985 went.

"Security profits in a given year bear similarities to a college graduation ceremony in which the knowledge gained over four years is recognized on a day when nothing further is learned." General Foods is the case in point (Exhibit 3). Berkshire held most of its shares from 1980, bought far below business value. Jim Ferguson and Phil Smith raised that value year by year, and last fall Philip Morris made an offer that reflected it. The sale produced $338 million of the year's $488 million pre-tax gains, and the year of sale gets credit for all of it.

Four factors made the money: a bargain price, fine economics, able managers working for shareholders, and a buyer willing to pay full value. Only the last produces reported earnings, and he spends no time looking for it. "In selecting common stocks, we devote our attention to attractive purchases, not to the possibility of attractive sales." Berkshire also again sold shares back to the Post and General Foods in proportion to their buybacks, leaving its stakes unchanged, which the IRS taxes as a dividend and the accounts, as last year, show as a gain.

The basket is smaller and worth far more per dollar paid: $275 million of cost worth $1.20 billion at year-end, $4.36 for every dollar paid, with GEICO alone nearly half of it.

General Foods is gone, sold to Philip Morris, and the list now shows only positions worth over $25 million. Beatrice is a parking place: a short-term arbitrage holding in an announced sale, taken "when we have more money than ideas," and not a totally safe one. The insurers also hold about $400 million of tax-exempt bonds, including $194 million at cost of the WPPSS issues described last year, now carrying an unrealized gain of $62 million and paying about $30 million a year tax-free.

In July Buffett decided to close the textile mills that were all of Berkshire when he bought it, and in early 1986 sold their looms for scrap. The history, he says, is instructive, and the instruction is about which business you are in.

Buffett Partnership bought control of Berkshire Hathaway 21 years ago with $22 million of book value, all in textiles, and a nine-year record of $530 million of sales producing a $10 million loss: "one step forward, two steps back." Ken Chace, made president at once, ran the mills excellently, and in early 1967 their cash bought National Indemnity, Berkshire's entry into insurance. The mills themselves never became a good earner. He stayed in for the reasons the 1978 letter gave, including that the business should average modest cash returns. "It turned out that I was very wrong about (4)." After a moderately profitable 1979 the mills consumed major amounts of cash, and by mid-1985 that was almost sure to continue. Buyers' interest was nil.

Why not invest his way out? Every proposal to modernize looked like a winner, often better on paper than spending in the candy or newspaper businesses. But rivals at home and abroad made the same investments, and once enough did, lower costs became lower prices for everyone (Exhibit 5). Burlington Industries, the largest American textile company then and now, chose to stay and spent about $3 billion over 21 years (Exhibit 6). Its shareholders got a stock worth a little over its 1964 price while the cost of living more than tripled. "[A] good managerial record (measured by economic returns) is far more a function of what business boat you get into than it is of how effectively you row." Intelligence and effort help considerably in any boat, he allows, but they do not change the boat. He faults himself for not quitting sooner: 250 mills had closed since 1980, whose owners simply processed the same facts more objectively. The union, poorly paid and cooperative throughout, deserves no blame, and the workers performed superbly even during the liquidation.

The postscript is an auction (Exhibit 7). Machinery that cost about $13 million, stood on the books at $866,000, and would cost perhaps $30 million to $50 million new brought $163,122, less than the costs of selling it. Looms bought for $5,000 in 1981 found no takers at $50 and went for scrap at $26. Two paper routes in Buffalo, or a single See's shop, carry more economic goodwill than the whole lot fetched.

See's, the News, and the Mart earned $72 million pre-tax in 1985, against about $8 million fifteen years ago. Buffett says even that gain must pass a test before it earns applause, and most corporate gains fail it.

The three businesses pass. Their earnings are $64 million higher and they needed only about $40 million more invested capital to get there, so most of the cash they generated was free to go elsewhere (Exhibit 8). The average American business needs about $5 of added capital for each $1 of added pre-tax earnings, and would have needed over $300 million from its owners to match the group. This is the power of economic goodwill during inflation, as the 1983 appendix explained. He checks the base year too: the $8 million was excellent against the capital then employed, so the comparison is honest.

Quadruple the money in a savings account and you quadruple its interest. "You would hardly expect hosannas for that particular accomplishment." Yet retirement notices praise chief executives for quadrupling earnings with nobody asking whether the capital quadrupled too (Exhibit 9). At 8% interest, a savings account with the interest left in quadruples its earnings in 18 years. Praise may be deserved when returns stayed superior or capital merely doubled, and withheld when capital rose in step with earnings on lackluster returns.

The businesses themselves are much as they were. The Mart set a sales record in a worsening farm economy, and Mrs. B, at 92, still sets a pace nobody can keep up with. See's store volumes remain far beyond any competitor's he knows of, but same-store poundage keeps slipping and margins will narrow unless it stabilizes. At the News, advertising shifted from the paper's own pages into less profitable preprints, costs stayed controlled, and penetration stayed exceptional. Chuck Huggins, Stan Lipsey, and the three generations of Blumkins run them, and Buffett counts himself extraordinarily lucky.

From the savings account Buffett builds a case against the ten-year fixed-price stock option, the pay device companies grant routinely, and explains the different system Berkshire uses instead.

Give your $100,000 savings account to a trustee who pays out a quarter of the 8% interest and keeps the rest (Exhibit 10). After ten years the account is worth $179,084, earnings are up about 70%, and every chart marches skyward. Now grant him a ten-year option at year-one value: he profits handsomely for having held on to your money. Many stock options have worked exactly that way. Managers apply a double standard: warrants aside, he believes, nowhere in business are ten-year fixed-price options granted to outsiders. "Negotiating with one's self seldom produces a barroom brawl."

Options are also blind. Once granted, "the sluggard receives rewards from his options precisely as does the star," and a managerial Rip Van Winkle could not wish for a better incentive system. The rhetoric says options put managers and owners in the same boat. Exhibit 11 shows the two boats. An owner bears the cost of capital and the risk of loss. An option holder bears neither, which is why Buffett would accept a lottery ticket as a gift and never buy one. His three points: tie options only to managers with overall responsibility, since a .350 hitter on a last-place team still deserves his payoff. Structure them with a charge for retained earnings and price them at true business value, not at the depressed market prices managers call unrealistic when a bidder appears. And some managers he admires enormously disagree, and have built cultures where options work: "If it ain't broke, don't fix it."

Berkshire pays bonuses for results in each manager's own bailiwick, defined by that business's economics, and never looks at the stock price when writing the checks. The bonuses can run to five times salary or more, with no cap and no regard to seniority, and one manager's could top $2 million in 1986. "(I hope so.)" Managers who want to own Berkshire buy it in the market with their own money, and many have. Bearing the risks and the carrying costs, they truly walk in the shoes of owners.

The industry had its worst combined ratio on the table, an estimated 118, and Berkshire's own fell from 134 to 111 while its premiums more than tripled. Buffett explains why the better prices arrived, and why they will not last.

Premiums grew nearly 21% in 1985, which should have cut the combined ratio. Instead losses grew 17%, after 16% in 1984, more than 13 points above inflation, "a modern record." Catastrophes were not the cause, at about 2% of premiums. Reserve corrections were part of it, confessed at what he calls a revival meeting. The larger worry is judicial inflation: the deep pocket "is being sought and found, no matter what the policy wording, the facts, or the precedents." The ratio should still fall sharply in 1986, barring a supercatastrophe, and then two principles take over (Exhibit 12). A commodity earns good profits only when prices are fixed or capacity is short, the price-fixing is gone for good, and "capacity is an attitudinal concept, not a physical fact." About 15 insurers raised well over $3 billion in 1985 to write all the business they could at today's prices, and the pace has accelerated.

Berkshire's day, promised in earlier letters, arrived with the tight market. Buyers who learned in 1985 that "a policy is only an IOU" fled to its capital, and brokers come to it for something rarer still: the willingness to keep a large risk. Many insurers write big policies and lay off most of each one (Exhibit 13). For years that paid, because issuers kept a far larger share of the premiums than of the risk, until the reinsurers, like Mark Twain's cat, learned to avoid all stoves. Berkshire never played that game and is now willing to risk $10 million of its own money on a single event, if the price is right and the risk is not correlated with others it carries. National Indemnity's advertisement seeking policies with premiums of $1 million or more drew 600 replies and about $50 million of business. "Hold the applause," he adds. It is all long-tail, and it will be at least five years before anyone knows whether it was priced right.

The past misdeeds are still surfacing: the loss figures again revealed significant underreserving at the end of 1984, virtually all of it in reinsurance, and most of that in a few contracts discontinued years ago. He refuses the excuse, recalling the General Reinsurance chairman who, tired of hearing "except for," proposed an Except-For Insurance Company to hold the losses his managers would rather not count. "[T]he real mistake is not the act, but the actor." Elsewhere, Jack Byrne left GEICO, 38% Berkshire-owned and still far ahead of virtually all its major rivals, to run Fireman's Fund, and Berkshire followed him: from September it takes 7% of Fireman's Fund's own business (its reinsurance for outside companies excepted) for four years, holding the premiums until the claims are paid. "Never one to let go of a meal ticket," he explains.

Right after year-end Berkshire paid $517.5 million for three million shares of Capital Cities/ABC and about $320 million for Scott & Fetzer. Buffett explains why he paid full price for the first, and what he handed its managers along with the check.

The Washington Post shows the market that is gone (Exhibit 14). In 1973 anyone could see that a company worth $400 million to $500 million was selling for $100 million. "Our advantage, rather, was attitude," learned from Ben Graham, while the institutions, taught by business school academics that the market was efficient and thinking a waste of energy, traded without looking. He is enormously indebted to those academics. The stake cost $10.6 million, fell to $8 million by the end of 1974, and with the shares sold back this year is worth $221 million, most of the extra $160 million being the work of Kay Graham, who bought back stock at bargain prices and built the business. The Post now sells for over $1.8 billion, and no holding can grow from there the way it grew from $100 million.

Cap Cities is bought at a full price, amid an enthusiasm for media that for some properties has approached a mania, for its people: Tom Murphy and Dan Burke, whose management he thinks is "the best of any publicly-owned company in the country," and who are the sort of fellows you would want your daughter to marry. The purchase helps them buy ABC for $3.5 billion, whose economics are likely to be unexciting for a few years. That bothers him not an iota: "you can't produce a baby in one month by getting nine women pregnant." Then come the arrangements, drawn in Exhibit 15. Tom votes Berkshire's stock for an extended period, and the block cannot be sold to a would-be large holder without management's approval, as at GEICO and the Post before. A block that could go to a raider often commands a premium. He gives it up gladly: the managers he admires will never lose sleep over Berkshire's large stake, and "these agreements put Berkshire's signature where my mouth is." Shareholders wondering why he now pays $172.50 for shares he sold at $43 in 1978-80 get a one-line answer: "A little more time, please."

Scott & Fetzer came the way most things come to Berkshire: he read in the papers that an employee buyout had collapsed, wrote its chief executive, Ralph Schey, a short letter on October 10, met him for dinner on October 22, and signed the contract the following week. Seventeen businesses, about $700 million of sales, and World Book, which sells more encyclopedia sets in America than its four biggest rivals combined. There is no master strategy and no planning staff. "[W]e simply hope that something sensible comes along (and, when it does, we act)." The standing advertisement is back, with the minimum raised to $10 million of after-tax earnings.