You can insure against the tragedy of a washing machine malfunction, but not against an air crash

I feel obliged to write about Warren Buffett once more. Remember him? World's most successful investor, worth about $35bn; also the world's most annoying 71-year-old, retailing homespun aphorisms from his Omaha lair.

What is particularly irksome about Buffett is that he is usually right. So once again, with gritted teeth, I must digest the letter he sends to shareholders in the annual report of his investment company, Berkshire Hathaway. One passage stands out: "Though our performance last year was satisfactory, my performance was anything but. I manage most of Berkshire's equity portfolio and my results were poor, just as they have been for several years."

When have you ever heard the head of a big company tell his investors that he or she has been something of a dud? There are whole industries whose sole purpose is to prove that LameDuck plc is in fact FirstSwallow Inc.

Buffett does not bother with any of this. He has the confidence that comes from more than 25 years of superior corporate performance. A couple of mediocre years, of the kind that Berkshire Hathaway has experienced recently, can be tolerated. But the disarming honesty raises an interesting question. When does Buffett's personal failure become so acute that he concludes he is a liability to Berkshire Hathaway and should be replaced? This would be unthinkable to most of the company's shareholders, but cannot be eschewed on the basis of his own credo.

Buffett's biggest mistake, by his own admission, came in the management of Berkshire Hathaway's enormous insurance businesses. He was conscious of the implications for insurance payouts of terrorist atrocities, but failed to ensure that the group's premiums reflected them. So 11 September cost Berkshire Hathaway more than half of a $4.3bn underwriting loss from insuring other insurers (so-called "reinsurance"). "I violated the Noah rule," he says: "predicting rain doesn't count; building arks does."

As it happens, almost the entire insurance industry is clinging to a fragile makeshift raft, rather than sheltering inside a craft of biblical robustness. The costs of 11 September continue to rise. But there are also egregious losses stemming from decades of insurance against: a) asbestos-related illness, b) putative poisoning caused by the use of lead paint in US schools and public buildings, c) an epidemic of toxic "mould" in US energy-efficient office blocks, and d) ill health linked to smoking.

In the US, medical bills just keep on rising, and judges just keep on passing the costs to big manufacturers and their insurers. Nor is this simply a North American problem. In the UK, Royal Sun Alliance and CGNU, too - to a lesser extent - are exposed.

Most insurers broke the cardinal rule of insurance. They underwrote risks that they could not properly assess. Which would be all very well if this were simply a one-off problem. However, there is reason to believe that the economics of the industry have collapsed in a fundamental and enduring way.

Insurers, in effect, borrow money from the insured in the form of premiums. They then invest these premiums in the hope of generating a sufficient return so that, when the claims eventually come in, they will be able to cover these outflows with a bit left over.

The problem is that, just at the moment when insurance risks have become so huge and difficult to pin down, investment returns have collapsed. The performance of stock markets has been lousy; and the return on fixed-interest securities, at a time of deflationary pressure, remains low in cash terms.

Buffett is particularly ghoulish: "No one knows the probability of a nuclear detonation in a major metropolis this year . . . Nor can anyone, with assurance, assess the probability in this year or another, of deadly biological or chemical agents being introduced simultaneously (say through ventilation systems) into multiple office buildings and manufacturing plants." He reckons a medium-size outrage would cost insurers roughly $1trn, which would be enough to wipe them out (though it is a moot point whether their plight would be at the forefront of our thoughts in such circumstances).

Even without disaster on that scale, the insurance industry in its traditional form seems to me to be more or less defunct. Insurers happily insure us against those niggly misfortunes that befall us in home and car - but even for this bog-standard cover, premiums have greatly increased. We are all going to be offered vastly more protection against the tragedy of washing-machine malfunction.

But the big risks - the life-threatening ones, the ones that really matter - are being eschewed. The US airline industry is forming its own mutual company to provide air travel cover, because insurers will no longer do it. And most insurers are desperately trying to limit their exposure to terrorism, wherever it may lurk.

So if you are struggling to understand why George Bush is hunting down alleged terrorists from the Philippines to Somalia, the answer, at least in part, is that it is a battle to preserve a private sector insurance industry. The struggle to protect the American way of life is an attempt to restore the status quo in the Anglo-American financial markets. The unthinkable alternative for Bush is that the state becomes guardian and underwriter of our vitals and chattels.

Buffett, however, warns that it is a lost cause: "Fear may recede with time, but the danger won't. The war against terrorism can never be won . . . There can be no checkmate against hydra-headed foes." On that basis, you buy Buffett and sell Bush.

Robert Peston is editorial director of QUEST(TM); www.csquest.com; e-mail rpeston@csquest.com