click here />

Prompted by this letter, I decided to share a short essay on a large-scale fraud. Unable to visualize the complicated "credit default swap transactions that have hamstrung A.I.G.," I created a hypothetical easy-to-understand scenario. Does it help to understand real situations? Probably not. But it probably sheds some light on what is going on. I was not aware that most of those responsible for the AIG fraud "have left the company and have conspicuously escaped the public outrage."

Suppose a fire insurance company is created for homeowners in a town where houses are very similar. In what follows I am going to simplify things as much as possible, in order to focus on what is essential. Suppose the average cost connected with a fire is $300,000. This includes all associated expenses the insurance company must pay, such as repairs, reconstruction, reimbursement for damage, temporary hotel fees, etc.- Assume that the probability of fire, for each house, is 1/1,000 per house. The town has 100,000 of insured houses. What should the yearly cost of insurance be for each house?

The expected number of fires per year, in the town, is

n= (1/1,000)*100,000 = 100

and the associated cost is 100*300,000=$ 30,000.000, or $300 per house each year. But this is not the total yearly cost. The company pays salaries to 10 people, totaling $1,000,000. Taking this under consideration, the associated cost becomes $31,000,000, or $310 per house per year. What is a reasonable profit? The owner of the company believes that his profit should be 5% of the total associated cost; this amounts to 0.05*31,000,000=$1,550,000. I suppose that 5% is reasonably acceptable in our society. Adding this to $31,000.000 one has $32,550,000, or $325.50 per house each year. The situation is probably reviewed every ten years, to account for changing costs of everything, using the same basic rules.

So much for trivial arithmetic--in this hypothetical illustration. Let us turn to possibilities of fraud. Who is checking figures on which calculations are based? I suppose that government regulations call for transparent bookkeeping. But small "mistakes" are not easy to recognize. Suppose that the first average cost, $300,000 in my illustration, becomes- $ 303,000. This is only 1% higher than before. Multiplying the extra $3,000 by 100,000 houses, one gets the extra of $300,000,000 dollars each year. Small "mistakes" can be very profitable.

But this is not the worse possible scenario. Suppose the owner of the company develops a taste for gambling. Having $10,000,000 of cash available, he decides to "borrow" it, and to play roulette (or to conduct a risky market operation). This is not a theft, in his mind. If I win, he reasons, then I will put this money back and keep the difference. But he does not win. The result would be obvious, after a certain number of fire victims are reimbursed. The company would declare bankruptcy. Should this bankruptcy be accepted by the local government or should it try to rescue the company, at the expense of homeowners (each paying an extra $100 for one year)? The government could say, for example, that "not having insurance is much worse than bankruptcy." Is this situation analogous to what is happening now to AIG, on a much larger scale?