Banks - The Next Dot-Com Crash?

Accounting was invented by a Franciscan monk in the 15th century, but its First Rule - called "matching" - is still good today. It says that to compute true profit, you must match revenues with all costs that gave rise to these revenues. If you neglect even some costs, intentionally or otherwise, the resulting "profit" is incorrect, and may even be imaginary.

As an example, in the 1990s many tech companies wrote off cost of capital (mainly research and development) instead of amortizing it - and so did not include its portion in the price of the product. As a result they underpriced their gizmos, which, because of their low prices, then sold in large volume until the R&D "ran out" (became stale), and the companies had no money for new products. The products died, and the stocks melted.

When you sell below true cost, sooner or later you come to grief, whether you are a tech company, or a bank.

We already saw what happened when banks sold subprime mortgages below their real cost (which should have included the high risk of default): The imaginary "earnings" that banks booked over many years suddenly had to be wiped out in one year.

But what of bank loans to legitimate corporations? And what of derivatives - daisy-chain obligations to other banks or financial institutions? Are they priced any better? The answer is, most likely not. You see, most bank loans and derivatives today are based on a benchmark called Libor - the London interbank overnight rate. That's the cost of money that banks charge each other for overnight loans. Every morning, banks report the interbank rates they charged the previous night, then compute an average after tossing out the highest and the lowest rates. (This is done so no one bank can manipulate the average by reporting an extreme number.) Normally this works well. But recently, with the near meltdown of the financial system and the failure of Bear Stearns, banks have often been reluctant to lend to each other, even overnight. Because of this, the U.S. Federal Reserve Board had to lend to banks even against junky assets - subprime mortgages that may never be paid off in full - and the Bank of England even took over a failing bank. But banks still need to borrow from each other. Only now they do it at higher rates than they report, apparently.

How do I know? Because the BIS (Bank of International Settlements - a sort of central bank for central banks) said so. There is an apparent fudging going on, the BIS said, and the result is that Libor is likely higher than reported.

Why is this important? Well, Libor is the base rate on which almost any serious money being lent today is priced. A large industrial mortgage, say, may be priced at 2 per cent more than Libor, a corporate bond at an extra 1.5 per cent, and some derivatives at 3 per cent more.

Thus when Libor goes up, so does the cost of borrowing - as does the revenue of the banks, based on interest payments. But what if Libor goes down? Or even worse: What if Libor actually went up, but banks reported it as going down? What happens to true bank profits then?

Let's focus on the profit. Because what if banks, too ashamed to admit they have to borrow at high costs, understate these costs? What if they collectively, allegedly, start fudging Libor? Then, like dot-com companies that ignored cost of capital when they computed the cost of their gizmos and deluded themselves into thinking they were making money, banks, too, might be deluding themselves into thinking they are making money on Libor-based loans and derivatives, when in truth they are making much less, or perhaps even losing.

And if this is indeed the case - if the BIS is right and Libor is understated - then more bank writeoffs may be coming.

Just how much?

Well, here's a quick calculation: There are about $600-trillion (U.S.) in derivatives outstanding. The BIS thinks that Libor may be understated by about 0.3 per cent. Let's see now: 0.3 per cent of $600-trillion is about $1.8-trillion, or about 13 per cent of U.S. gross domestic product. This equals about four to five years of U.S. economic growth, or about twice the cost of subprime writeoffs to date. If Libor is repriced properly, surely there will be an impact.

Just how soon will this impact be felt? Perhaps not too far off. First, a BIS committee is now looking into the alleged fudging issue, so this may cause reported Libor to go up. But second, since it's unhealthy for North American loans to be based on European rates, something called NYibor has been suggested - the New York interbank overnight rate, an average based on overnight lending costs of North American banks. I suspect, however, that instead of alleviating the problem, this would only expose it: Once true overnight North American lending costs are shown, the base rate for all loans may go up, and more writeoffs may follow.

By that time the next leg of the bear market will very likely resume.

So continue to be wary of dicey financial stocks, stay liquid and bide your time. The real panic is probably yet to come.