Kevin Drum Smackdown Watch

I respect Salmon but his argument seems to be that if a Federal injection of $ 90 billion through AIG can keep the CDS market from melting down for a month (minus 3 days at the time AiG passed the 90 billion mark) then there is no problem.

My standard would be that the market can function without extraordinary public intervention. That sure isn't true of the CDS market.

The argument (which Salmon didn't really make) that CDS must be OK since they are in zero net supply has no basis in economic theory (so what) and is inconsistent with the evidence on huge contestable transfers (CDS contracts haven't been to court much yet but they will).

I suspect that part of the reason there is little public jubilation by in-the-money CDS traders is that they have no assurance of being paid. By virtue of the CDS's, a bundle of risk (say, the default risk on $100 bln of mortgages) gets duplicated and passed on from one entity to another, with the cumulative nominal obligation getting multiplied at each step of the way. Because of this process, there is massive exposure for many entities that are tied together like climbers on the side of a mountain. And if (as has happened) there is a massive triggering of obligations, any single entity's obligations may be so great that bankruptcy looms. While you'd much rather be in the money than out of the money, it's not comforting to know that the firms that owe money to you could very well be so overexposed that (a) you have to get money out of them through their bankruptcy trustee and (b) you'll only get dimes on the dollar. Additionally, however deep in the money you are, you know that your fortunes are largely tied up with those of the system as a whole. Some will get rich from a total systemic collapse, but most will be swamped.

That's my guess as to why the in-the-money guys are not grinning from ear to ear.

I dont know why it matters that CDS are zero sum. A particular institution or some type of institutions could have such a large exposure on one side of CDS that they may go bankrupt. This could cause a chain reaction of bankruptcies. As an empirical matter, this may not have happened in the current crisis as per Felix. But I dont see why the fact that CDS market is zero sum is germane.

Just to be clear...the Minn. Fed piece wasn't by real Fed staff economists (who are much more knowledgeable about the data), they were academics from U of Minn and Northwestern who could have easily have published that as a working paper without attaching the Fed's name to it.

(Obviously well over my head here but I'll comment anyway.)

The function of the CDS's was to provide a kind of quick-and-dirty insurance, that is, as a way of theoretically spreading risk so that the system as a whole could stay solvent even if individuals had problems. The plain fact is that whatever the mechanism is that was supposed to spread risk, IT DIDN'T WORK.

The lesson, from what I can see, is that if you are going to rely on insurance schemes, they're much worse that useless if they collapse when you need them to work. If we're going to justify massive leverage because we have an insurance scheme underlying it, then we need it to be real insurance and not some ersatz jerry-rigged thing that blows over in a stiff breeze. If we can't afford to put real insurance in place to justify our leverage, then maybe we shouldn't be trying to do so much leverage.

Felix should know better - yeah, it's a zero sum game except that some of the chips have been pulled from the stack and pocketed. This gets a little complex....

Many players in the CDS game "net" out their positions by going the other way at the point that their position is more valuable. In theory, this is OK. Where the theory fails is in the accounting. If you net out two CDS contracts, the profit (the difference in the transaction price) is totalled for the entire length of the contracts (let's say 5 years) and the profits are booked today. No cash, remember - but the fees and bonuses are paid now, not 5 years from now.

So let's say at Total Animal (which is where I work at my trading desk) the split is 50% to bonus pool and 50% to my boss, and I've booked two trades that net us a $15MM profit over the 5 years, I and my team get $7.5MM bonus this year - and those chips are gone from the table.

Multiply this by the zillions and you can see where some of the problem is - especially as I and my team have taken our personal wealth and leveraged it up to speculate in Irvine real estate - and we all know how well those investments have done.

I'm another guy who knows nothing about this but I agree that this does NOT appear to have been a zero sum game. Lehman and AIG sold insurance against defaults of mortgages. Say their premium was 2% per year. What did they do with the premiums? Were they reserved so that there was cash to pay off in the case of a default? Or did they just call it profit and distribute it as bonuses? I used to work for a Title Insurance Company in New York. At the time, there was a lawsuit pending by a local Indian tribe claiming that a good deal of the county I was in was wrongfully taken from them, and they wanted it back. Thousands of houses stood on the disputed land, and every time a house was sold the title policy contained an exception for the Indian Land Claim. But, we routinely insured over that risk for an additional premium of about 50 bucks. I always thought it was crazy to buy that insurance since, if the Indians won, there was no way in God's Green Earth that the title company could have covered the losses. State law required us to have a $500,000 loss reserve and just during the year I was there we easily insured 10 times that amount against Indian land claims. Everyone assumed that the Indians would eventually settle, (which they did.) The premiums were regarded as pure profit, and there was no contingency whatsoever in the event that the claims had to be paid. If the unexpected had happened (as it did with subprime foreclosures), the losers would have already spent their money and there would be no winners. Seems to my untrained eye that the only way to prevent the kind of debacle we have now is to regulate these instruments to insure that the people writing the insurance have the wherewithal to pay off if they lose.

... What that means is that everybody else, on net, was a buyer of credit protection, and therefore is likely to have MADE money during the current crisis, rather than lost it. And yes, that includes all the banks.

... the CDS market is actually much smaller than many people feared.

Posted by: Felix Salmon on 11/06/08

If someone bought "protection" CDSs to just cover any losses, then they would break even on defaults. If someone bought CDSs and didn't have defaults then they spent money (which is a loss).

It's only those who bough extra protection CDSs AND had defaults who have made money.

But, when these toxic assets aren't being sold at ridiculously low burst-balloon prices there isn't a default and the CDSs aren't executed, so there's a stasis brought on by sheer terror.

I think the assumption (is it an assumption?) that 'everybody made money' has to be questioned.

Anyway, the toxic mortgages remain and there are a lot of them.

Looking at the issue from another perspective:

What's the value of the dollar? It's changing, right? So, say 2 years ago, should one buy 'insurance' CDSs to have a future asset in dollars?

Could an insurance policy on toxic assets be somehow related to the future value of dollars? Hmmm. But then, what if the economy crashes? How much are those 'future dollars' going to be worth?

Seems the most valuable asset is the power to sit and watch events and control if or when the economy is pushed over the cliff.

Although Felix is logical and persuasive, I think he is still missing the bigger picture. I'd agree that CDS's in themselves do not need to be demonized, but the way they were used was a big contributor to banking problems.

In order to leverage their capital better to earn better returns, banks in recent years were being innovative in finding ways to get risks off their balance sheets. The biggest tool in doing this was a combination of securitizations and CDS which combine to make very marketable products. This scheme helped banks maintain their regulatory capital ratios even though their equity to asset ratios rose.

I think it is very clear at this point that the scheme has not worked and we have to go back to the drawing board. Flaws with the securitization process have been highlighted by many - unbundling the origination process and holding the credit has lead to poor credit origination, banking risks were not mitigated like banks expected, and in fact losses are being magnified.

Becuase of the loss in confidence of orgination, morgage back securities losses in the securities are far greater than the projected defaults of underlying loans. Due to the contagion affect nobody want to touch anything to do with these securities even though many of them will be fine. So what happens is that insurance companies, and other sellers of CDS's are taking huge losses as spreads go up. Banks or other holders of CDS's who theoretically benefit, have a risk position that is offset by a CDS. What is happening is that the benficiaries are not accounting for gains in the value of the CDS and won't unless it really defaults.

So what you have are losses on both sides that far outwiegh the underlying problems. Insurance companes taking massive losses on CDS spreads. Banks writing down mortgage back securities (which they might get some of it back in the case of defaults but they are having to take charges on even the stuff that will probably be ok).

Bottom line: While its true that CDS on their own are not a reason for economic collapse. But they have been used as an important tool that led to a system that has proven faulty. I think the banking system would be better off to start a square one and dismantle the derivatives that have helped them lever up.

I'd add that i'm sympathetic to Felix's arguements. As a financial analyst, I thought the securitization and CDS schemes were a smart way of matching capital with true risks. Seemed much better than the old Basle accord that was so crude. And while everyone knocks greedy bankers for taking these risks to increase profits, they should bear in mind that a more effient bank leads to better rates for borrowers. Its a shame these scheme has failed so badly. Now when the dust settles, banks will be less profitable and individuals will have to pay higher rates on their loans.

A simpler and more general way to ask the question, Kevin:

     Where did all that money GO?

Which leads to a follow-up question:

     Is there any way to get (some of) it back?

I believe that Tony's analysis (above) is correct.

Banks and insurance companies have regulatory requirements to maintain certain capital ratios. These ratios are are quite high, and are based on the assumption that there will not be widespread credit defaults.

When such defaults do occur (because of a housing bubble, for example) the ratios are violated, the lenders and/or insurance writers become undercapitalized, and they cannot lend any more -- or write more insurance.

There is also a positive (bad!) feedback effect: as more defaults occur, the capital of an institution (AIG) goes down while the risk rating of its portfolio (and therefore its capital ratio requirement) goes up -- which makes it even more under-capitalized.

And there is a domino effect, because of the interconnectedness of the financial system. As AIG goes down, its counter-parties have to write down AIG paper that they hold (as well as other paper that AIG had insured for them), and then they enter their own positive feedback loop. And then it spreads to their counter-parties, etc.

So the explosive behavior of the system is due not to any derivative multiplier effect, but to these two other phenomena (positive feedback re capital ratios and domino effect as counter-parties become infected, and the infection spreads).

To pile onto the zero-sum thing:

Derivatives are only zero-sum if they're priced correctly. Being priced correctly depends on a lot of things, and the more complex your derivative, the greater the probability that you'll get something wrong. The idea of a CDS might be relatively simple, but in practice actually constructing the probability universes that decide how these things are priced is incredibly difficult. I wouldn't be surprised if plenty of CDS owners got paid very little in exchange for insuring a very large amount of risk.

The finance industry's MBA's are like Microsoft engineers. They know how things work until their bugs are exploited by script kiddies.

link here:

http://www.berkshirehathaway.com/2002ar/2002ar.pdf