What I was trying to say

In my last post was that the meltdown of Bear Stearns threatens the entire financial system. This was the reason that the Fed had to intervene to arrange the sale of Bear.

The reason that any element can bring down the system now is the interdependence of the players. The rise of structured finance, which enables risk to be sliced into various pieces and traded between banks, funds, insurance firms, investors, the government, corporations, farmers, travellers - in short, everyone - means that the failure of any large player inevitably weakens everyone else.

Derivatives essentially function like insurance - enabling the buyer to receive protection from specific risks (like having one's house burn down or having to make payments in currencies other than those in which a firm earns its revenue) at a cost - an insurance premium. Likewise, the seller function as an insurer, taking on that risk in return for consideration (the premium paid). If the seller decides to lay off some or all of that risk, they seek reinsurance, by assigning the contract (selling it) or by purchasing an offsetting position. This can be done for all or part of a position.

The problem is that as the risk becomes dispersed throughout the financial system, those at teh center, the investment banks and their large clients - hedge funds and insurance firms, and some corporations - find themselves counterparties to everyone. If one fails, the insurance that they have offered becomes worthless, forcing the insured to recognize a loss from an exposure they had believed to be hedged. Once that loss appears, equity of the counterparties can be quickly wiped out and more selling and liquidation ensues.

In a very real sense, what structured finance has done is not only spread the risk widely, it has meant that when any sector has a major problem the effects are impossible to isolate, because, at the end of the day, through the banks, everyone becomes a counterparty to everyone else, and failure at one bank is enough to bring down the system.

I point out that this is not news, though many seem to be surprised by it. In 1998, LTCM, having suffered a $5 billion loss, nearly had its equity wiped out and it stood to default on over $100bn in trades. The effect of the write-downs on the markets would have been huge.

Why does this happen? Well, insurance is based on estimates of what is likely to happen. Most of the time, it's a good bet for the insurance company. They reliably collect premiums that are more than adequate to offset small losses on in the position taken. However, sometimes really big events happen and when they do there is a surprising amount of correlation among assets, particularly financial assets and particularly among assets of similar classes. This means that if you regularly issue insurance against a specific type of default - all of your claims are likely to come at the same time. This would be OK, were it the case that the insurer had the assets in place and that those assets were sufficiently liquid to ensure that conversion to cash to pay claims didn't lead to greater imbalance and higher losses.

Virtually no firm, and certainly no investment bank, which operate with massive leverage, has the balance sheet necessary. It requires huge stores of cash - and cash is not a big earner. High returns on equity and the big bonuses that come with them aren't earned with cash. If you own 1/3rd of the company and live in a nice house in Omaha, you might not care that much about holding cash (especially if you operate with an underwriting profit, in which case the cash is free), but if you work in NYC and are competing to have your kids get in the most prestigious pre-school and you rely on annual performance bonsuses to pay for your lifestyle then you don't want lots of cash sitting around. You need to keep dealing and generating fees.

This is not to knock the investment banks. Generating fees is a very good idea, since those profits can be used to offset the loses (if you can continue operations).

But what I want to caution against is the belief that the risk models used to predict the insurance payoffs are good. Even Alan Greenspan, a fan of structured finance - and asset bubbles - admits that overreliance on models is a problem. This is because of the Rumsfeldian "unknown unknowns" which are impossible to model, no matter how good we get at modeling "known unknowns". Worse, the moments that the models will be most useless will be precisely those moments when insurance is most needed, that is during crises when traditional risk measures, such as Value-at-Risk (VaR) that depend on predicting the correlations among a series of risks, will break down.

I posted the first, unedited, and considerably less clear explanation, because I wanted to share my thoughts at the time. There has been evidence that my thinking on this was pretty good - the bank is selling for much less than its close on Friday (even I didn't think it would essentially go for liquidation value) and because I predicted more problems. The WSJ already hinted at problems at Lehman (though, thus far, it's speculation - on my part as well).

The Times of London is also insinuating that several leveraged hedge funds are near to collapse. This would also not surprise me. Leverage, so great on the way up is even more deadly on the way down. Long periods of being rewarded for taking increasing risks and getting paid less and less (but still getting paid) to take them inevitably leave one naked. And the tide is rushing out.

As an investor, I remain patiently on the sidelines. Right now, there is panic selling, knives are coming out and I'm not real interested in trying to catch it - never was that coordinated. The deleveraging that is underway will dramatically change the investing climate that we have come to know. It will truly be unlike anything we have seen in my lifetime.

Most of the old rules will be out of the window. I think there is still a great deal of reason to expect deflation and not inflation. As an investor, preparing oneself to thrive regardless of which scenario occurs is very tricky, but as always - the first rule is Hippocrates: "do no harm".