Thursday, January 29, 2009

CDS are good, Naked CDS are bad, OK?

Following Barry K comment on my previous post I went and read what the FT blog had to say about the debate. The only thing bone-headed here is the coverage and the knee-jerk reaction to anything legislative.

Let me see if I can give a 10 sec crash course on CDS. A credit default swap is a protection against a default of debt. If you hold a bond, you can buy "protection" against default. You pay a premium, and if the bond defaults you get the principal back. If you don't hold the bond then it is called a naked CDS. You pay the premium on a fictitious bond (you never paid the principal) but if it defaults you get paid the principal, pooof! out of bad debt, comes more bad debt.

CDS are good, good in the theoretical economic value sense as it moves risk around from a party unwilling or unable to assume the risk to a party that is willing and ,again in theory, able to assume the risk (AIG anyone?). When the bond defaults, the system remains strong. This is good from a systemic standpoint.

A naked CDS, however, will create a leverage on the default. People never really lent the money but if the bond goes bad they get the principal back. Multiply this by the number of people taking naked bets, 4 to 1 according to the article.

The effect of 'real' debt going bad is bad enough, sub-prime was responsible for this crisis, but the idea of turning 'virtual' debt into real debt only on default is really scary. Bad debt gets multiplied.

When all is said and done, the added value of naked CDS, if the bond does not default is marginal (better pricing of CDS, lots of money for the traders taking undue risks), but when the bonds default, they become weapons of mass destruction.

The call to ban naked CDS is right on. Don't be too quick to throw the 'bone-headed' moniker just because it is legislation and it comes from a elected representative. This ideology is partially to blame for the situation we are in today. So here is to you, Collin Peterson, chairman of the agriculture committee of the US House of Representatives: godspeed.

8 comments:

Andrew C. Oliver said...
This comment has been removed by the author.
Andrew C. Oliver said...

But don't you think they should be more tightly regulated with regards to reporting and transparency. My understanding was that the reporting requirements on these kinds of things loosened up to the point where no one knew what was in them. "Standard practice" then takes root and no one tells you what is in them because no one else does (since it isn't in any sellers best interest there is a natural collusion). I don't think they should be banned either, but labeled like a pack of cigarettes just in case you weren't sure that they might kill you (or at least your bank) and that they are full of tar (or in this case McMansions mortgaged to actual McEmployees).

Marcf said...

Andy,
Yes, more transparency would come by virtue of being traded on a exchange.
No, the analogy of a pack of cigarettes does not hold. In this case naked CDs are good for the smoker and bad for the rest of us. In browsing the news this morning I see they are already watering down the ban. That is bad.

FBA said...

That's a great misunderstanding.

CDS are ruled by specific contracts (refer to ISDA).

Considering that naked CDS are a bad thing, this wouls also include...
Future contract
Options (whatever the underlying)
and all derivatives, including also stock lending and borrowing.

This is also a very strong misunderstanding of the terms of a CDS.
A CDS has not to be settled exclusively by the delivery of the reference entity underlying (which by the way may not exists) but could be cash or any debt that apply to the reference obligation clause of the contract.

Saying this would be also saying "an option buyer should hold the equity". That is a complete non sense, since the goal of holding a "naked CDS" is ofrmelry to gain exposure on a credit when you can't for any reason (no cash but diversification requirements, not liquid).

So most of your assumptions are completely false, since most of CDS are settled in cash (difference of value).

Again, if you wan't to manage a physical settlement, then the mecanism is different :
You deliver the defaulted debt to the protection seller,
The protection seller will then ..
Deliver you a equivalent bond according the contract
Pay you the correct notionnal amount.

A naked CDS is when you have to find a bond to deliver as protection buyer, but it is not made within a minute, all CDS contract have a 30 days with equivalent roll over period to settle (grace period). If none of the parties can deliver, then the settlement is made out of pure cash or mixed (x% cash/y% debt).

The true reality of this mass explosion is n derivatives that exists on a bond
CDS of bond a, aggregated in CDO x, issuing CDS on tranche x, packaged on bond e; and so go on.
At this time, a single event or a sufficient number of default could lead to a Naked situation because you can't find debt and one event generate many CDS execution.

It is not a question of a single hedged or collateralized positions, It is the question of one to many company exposed through one to many way due to the possession of one to many security, exposed at any level to the risk of default.

FBA said...

And by the way, yes it is complex; yes a lot of people have considered that maths were sufficient to make the job. But it is not.

Finance are not maths, market is a everytime moving entity, what fails exactly is:
Too much people talking on a subject they don't know.

As far as I'm concerned, my 'mentor' was right :

If you put money in a debt and don't precisely know how the flows are payed (What is generating enough money to pay the flow ?), you are in risk
If you have a price, that's a good thing, but the price is not the signification of a risk. It is just the price someone is ready to pay for buying, or is ready to accept for selling.

That's precisely were the quant approach is inefficient : it is computing prices, it is considering cash flows as a formal element of a calculation.
BUT it can't determine the human part

Do someone know when buying a GE Bond were the money come from ?
I don't know. And that's the primal form of credit risk. And that's a basic rule applicable to any form of market that a generation of super quant forget :
Buying and selling is not a question of pure math, whatever the complexity or the approach. Pricing is a question of giving a price. But do we know someone selling for just the pleasure of selling at the cost price ? No.
Pricing does not consider that a human is by definition greed and looking for margin.

Finance is behavorial not computational.

Marcf said...

FBA,

I was awaiting the argument that it this was commensurate to calling for a ban on derivatives in general. It is not. Not by a long shot. The reason is nominal amounts. The CDS market by representing defaults on debt is a MUCH BIGGER market than derivatives on stocks.

The bond market is much bigger than the stock market. So no, this is not like saying "a holder of a derivative should hold the stock" because I didn't say it and the orders of magnitude mean everything.

The main point I made, which you corroborate is that the principal has to be materialized. I did not make any assumptions of cash vs physical for the settlement, you do and get all worked up about it. I merely made the point that materializing bad debt and multiplying is a bad idea.

Then your points about quants vs non-quants are not well received. Mathematics still helps me understand what is going on and call bullshit to guys like you who huff and puff and mirrors and smoke.

Look FBA, at the end of the day MATERIALIZING AND MULTIPLYING BAD IDEA IS A BAD BAD IDEA. Period. Levels of debt and therefore bad debt are heavily regulated. All the 'innovation' lately has been made so that banks could circumvent legislation. Self regulation my ass, misunderstanding my ass.

Marcf said...

That being said. I do agree with the point that market based prices are not rational in extreme times. You really have 3 ways to approach asset prices,
1/ cost base, make a profit, based on accounting of building it
2/ revenue base, discounted cash flow analysis
3/ bullshit base, the emotional component of market expectations.

Apply to housing: 1/ what does the house cost to build 2/ what revenue does it bring in 3/ the crazyness of debt created bubbles and the expectations associated to it.

The problem is that these expectations work in reverse and with a revenge on monetary levels as is going on right now. If the bubble is big enough then there is nothing the FED can do at a monetary level because the levels of money are linked to demand which is based on expectations of return which go negative. Market based pricing can be the devil on the way down.

I am discovering minsky's work through Keen, the dynamics of market based chaos due to endogenous money created by the banking system. Debt sometimes is created for the wrong reasons, to feed the banks and not to be serviced yet the market based economy mostly based on greed and ignorance will still clear these products.

So, yes and no, there are no equations in what I just described (and by the way the trigger WAS naked CDS in aug 07 in subprime) yet some models already are trying to capture the dynamics of the breakdown as opposed to the clear sailing equilibrium.

I would enjoy more conversation with you on the topic but in a more civilized way.

Anonymous said...

Excellent discussion. Just wanted to ask FBA about the technical detail of whether even the buyer of a naked CDS has to deliver to the CDS issuer the reference underlying bond in case of default even he does not have it in its portfolio ... does the CDS seller get the recovery rate in cash settled in case of default?