In dollars, 1979 looked like a step up: more profit than 1978, and earnings per share about 20% higher. Buffett opens by asking you not to be impressed. By the measure he trusts, the year was a touch worse than the last.

Here are the two numbers, side by side. The businesses earned 18.6 cents on every dollar of the owners' money in 1979, down a little from 19.4% the year before. Yet because Berkshire had more capital to work with, the profit sliced per share still rose about 20%. Buffett calls that per-share figure "improper" to focus on, and gives the reason in a single barbed sentence: leave money untouched in a savings bond and its "earnings" pile onto themselves year after year, so even a stopped clock can look like a growth stock.

So instead of a victory lap, he restates the rule. The real test of a manager is a high return on the capital employed, earned honestly, without piling on debt or accounting tricks, and not a string of rising per-share numbers. It is the kind of distinction nobody makes in a good year and everybody wishes they had made in a bad one.

Then comes the heart of the letter. A few years back, Buffett says, a business compounding its owners' money at 20% a year would have made them comfortably rich. By 1979 he is no longer sure it would make them any richer at all.

Buffett does the grim arithmetic out loud. Suppose Berkshire keeps compounding at 20% a year, and the share price keeps pace. At 14% inflation, your real gain before tax is only six points. Hand most of those six to the tax collector when you cash in, and your after-tax purchasing power gain is, in his words, very close to zero. He gives the squeeze a name, the "investor's misery index": add the inflation rate to the slice of your gain that tax claims, and whenever that total tops what the business earns on equity, you grow poorer in real terms while consuming nothing at all.

To make it vivid, he passes along a needle from a sharp-eyed reader. Berkshire's book value at the end of 1964 would have bought about half an ounce of gold. Fifteen years later, after plowing back every dollar of profit "along with much blood, sweat and tears," the far larger book value would buy about the same half ounce. The trouble, he notes, is that governments have proved wonderfully good at printing money and conspicuously bad at printing gold (or, he adds, creating oil).

Buffett then owns up to a mistake, and draws from it one of his most durable rules. A few years earlier he had bought a textile mill because the price was a steal. The price, it turned out, was the only good thing about it.

The mill was Waumbec, in Manchester, New Hampshire. On paper it was a giveaway: Buffett bought it for less than the working capital alone, which meant the buildings and machines came free. He had acquired machinery and real estate, by his own account, for less than nothing, and it was still too expensive. New problems arose as fast as the old ones were tamed. Set against that, he points to a different kind of business entirely: a network television station earns lavish returns on its assets almost no matter who runs it, which is exactly why such a station sells for ten times its tangible worth or more. Despite the fancy price, he concludes, the easy business is usually the better road.

He is careful not to call Waumbec a disaster. Some of it strengthened Berkshire's decorator-fabric line, the one corner of textiles where it had a real edge, and the Manchester plant might yet run at a slimmed-down profit. But the reason he bought it, the bargain price, never paid off. The energy would have done more good almost anywhere else.

Insurance was again Berkshire's biggest engine, and 1979 sharpened a lesson from last year's letter. The whole industry was sliding into a loss on the policies it wrote. Berkshire's best insurer went the other way, and it did so by selling less, not more.

Last year Buffett predicted the industry's combined ratio would drift up, maybe far enough to push the whole business into a loss. It did exactly that, climbing past the break-even line from about 97.4% to 100.7%. Berkshire moved the opposite way, nudging its own ratio down to 97.1%. The standout was the slice of National Indemnity run by Phil Liesche, which earned an $8.4 million underwriting profit on about $82 million of premiums, a result almost no one in the industry matched. And the telling detail: those premiums were down from the year before. Liesche had written less business on purpose. If a policy made sense he wrote it; if it did not, he turned it away.

Buffett does not expect the easy years back soon. With interest rates high, rivals will happily write policies at a loss just to get hold of the premiums to invest, and he thinks the industry's tolerance for underwriting losses has permanently widened. His forecast is for combined ratios averaging around 105 over the next five years: a loss on the insurance itself, year after year. In that climate, the temperament to say no is worth more than any clever pricing model.

All that insurance money has to go somewhere, and where it goes is the stock market. By the end of 1979 Berkshire's insurers held common stocks that had cost about $185 million and were worth around $337 million.

Buffett adds a quiet point that becomes a thread through later letters. Most of the companies he owns pieces of keep the bulk of their profits rather than paying them out, and that kept-back money never shows up in Berkshire's reported earnings. But it is working all the same. He expects these managers to turn each dollar they retain into a dollar or more of future market value, and trusts that, over time, the share prices will reflect it. Here is the portfolio behind that claim.

The other corner of the portfolio is bonds, and here Buffett confesses a second mistake. The whole industry had been lending money long and cheap into the teeth of inflation, and he had joined in, only half awake.

Buffett spots a contradiction the whole industry was living with. Insurers had decided that a one-year auto policy was too risky in an inflationary world, since they could not guess a year ahead at the cost of repairs and hospital stays, so they switched to six-month policies. Then they took the premiums and lent the money out at a fixed rate for thirty or forty years. They would not name a price for car insurance twelve months out, yet happily named a price for money to be used in 2010 or 2020. The long bond, he says, was the last great fixed-price contract still being written as if inflation did not exist.

He is hard on himself here, harder than on the rest. The mild caution he had shown, sticking to shorter maturities and special features, was "an improper response to the world unfolding about us." As he puts it: "You do not adequately protect yourself by being half awake while others are sleeping." Worse than buying the fifteen-year bonds was failing to sell them, at a loss if need be, once he saw the danger clearly. It is a rare thing to watch an investor mark his own paper down in public.

Buffett closes the letter with a theme that is almost personal: who owns Berkshire, and why he writes to them the way he does. A company, he argues, chooses its shareholders as surely as a restaurant chooses its diners.

The numbers behind this are striking. Year to year, about 98% of Berkshire's shares stay in the same hands, and perhaps 90% of its shares are owned by investors for whom Berkshire is their largest single holding, often by far. That is why Buffett builds each letter on the last instead of repeating himself, and why the report sounds like one owner writing to another. He passes along an analogy from the investor Phil Fisher: a restaurant can win a loyal crowd as a steakhouse or as a French bistro, but not by serving steak one week and take-out chicken the next. A company that keeps changing its story ends up with a revolving door of confused and disappointed owners.

The same plain-dealing runs through how Buffett manages. Head office is barely a basketball team in 1,500 square feet, with financial decisions made at the very top and the running of each business pushed right down to its managers, the kind, he notes, who treat the job as if it were their own. He also says a quiet goodbye: 1979 is the last year he can report on the Illinois National Bank, which earned a remarkable 2.3% on its assets under 82-year-old Gene Abegg before new rules force Berkshire to let it go. He would rather hand owners a candid account of all of it, the bargains and the blunders alike, than a polished one.