The repeated appeals to the market-maker excuse for Goldman’s CDO sales merits a rant. (Note: inspired by zerobeta tweets).
The secondary market for CDOs has always been very thin. Basically the bank that issued the CDO stood ready theoretically to buy the CDO back, but, well, it almost never happened.
So when people claim that banks were making markets in CDOs, I think the question is: “Well, then, where was your CDO trading inventory?” CDO trading inventory — as I am using the term — can only include CDOs that were placed by the issuing investment bank with an investor and were subsequently repurchased by the same or another investment bank. Such trading inventory is entirely distinct from the CDO inventory that was created by issuing new CDO securities and failing to sell them. (Citigroup, Merrill Lynch and UBS were chock of new issue CDO inventory).
Now, maybe somebody will correct me, but it’s my understanding that there really wasn’t any secondary market to speak of in CDOs and that the investment banks held minuscule quantities, if any, of second-hand CDOs in their trading inventories. If this understanding of the market is correct, I would like to know how anybody can claim that investment banks “made markets” in CDOs. They may have originated CDOs, issued CDOs and placed CDOs, but unless they carried trading inventories in second-hand CDOs, the investment banks can not claim to have made markets in CDOs in any meaningful sense of the word.
Using models to give theoretic prices to clients who need to mark their CDOs to market is not market making — for the simple reason that these prices are not tested by the market unless transactions are actually taking place at these prices. Unless we have the evidence of a transaction to demonstrate that the market maker was willing to take the CDO onto it’s books at the price in question, the price quoted has very, very limited meaning in a market economy.
In short, one of the biggest failures of the investment banks in the CDO market was precisely the failure to make markets in CDOs. The creation of an illiquid, untradeable product that the issuers themselves did not want to hold in trading inventory on their books was a disaster. And for these same investment banks to turn around now and claim market making as a shield in their defense is almost beyond belief.
Synthetic Assets 
Depends what you mean by secondary CDOs. Citibank nearly went bankrupt because it had an arrangement to buy back its CDOs via liquidity puts. MS wrote a load of CDSes on CDOs – which have the same economic effect as buying them – and lost 9billion USD on that market making. Bear bought out their CDOs which pretty much bankrupted them. GS were more selective and are clearly are now paying in the court of public opinion for not being incompetent.
I think I was clear that in order to claim to be a “market maker” in CDOs, an investment bank has to carry a trading inventory of second-hand CDOs. None of the cases you cite apply.
I agree that the IBs are market makers in CDS — but that derivative market is distinct from the CDO (securities) market.
OK, you don’t understand how these things work – possibly wilfully misunderstand. Debt obligation = risk free bond – CDS. Buy CDS you no longer own a debt obligation you own a risk free bond – modulo basis, model and counterparty risk. They are economically exactly the same and there are technical reasons you trade the CDS not the CDO.
Danny, please be polite. We are talking about whether or not there is a market in CDOs, not whether or not there is a way to synthesize exposure to them. These are two clearly distinct questions.
Over at Interfluidity you were kind enough to point me to the fact that the prospectus for Abacus states outright that there is no market in the CDO. I’m writing up a post on that right now — because the idea that there can be a market maker where there is no market is really just a gross abuse of the English language.
Furthermore your comment contradicts itself. Obviously CDOs and CDS are not economically equivalent precisely because of basis, model and counterparty risk. For the same reason making a market in CDS is a distinct act from making a market in commercial debt, sovereign bonds or CDOs.
Perhaps your first sentence is really just a reflection of your own state of mind.
Ok let me try and state it more slowly. CDO is basically a risky bond. A risky bond is economically a risk-free bond with a short position on the CDS on that risky bond. Hence If I want to sell a CDO I can get the same effect by selling a treasury and writing protection on the CDO. There are a number of reasons why I would do this rather than buy and sell the CDO – no cash up front, no hassle of registering securities, wider universe of buyers, tax issues etc. The risks I mention are generally considered relatively minor risks – except when they are not… – counterparty risk. Obviously if you outright sell a CDO then you get a lump of cash and then don’t care if your counterparty can pay premiums or not – thats your counterparty risk. Basis risk is that the maturity of the CDO might not have an exact match on the treasury or CDS side. But it is exactly the same and CDOs did change hands and helped to bankrupt the banks who were not rigourous in their risk management. These statements don’t contradict themselves, it is just the mechanics of the trading.
The fact is that if the CDO investor wanted to offload the risk on his books then he could – and many did and more didn’t. The fact is that most of the investment banks suffered massive losses as a result of trading these products. UBS, ML, MS, CS and yes GS all lost money trading these products. DB lost less than most almost single-handedly due to one guy. Your basic premise is false. Sorry.
For the sake of comparison, if I said that the senate hearing was nonsense because none of the banks went short CDOes they just went short CDSes on the CDOs then I would like to think I was playing semantics.
If GS said that Paulson didn’t short the bonds he bought a product that replicated the economic effect of shorting the bonds I would also like to think you’d be tearing GS a new one.
PS I am sorry if I sound rude, I don’t mean to be.
No offense taken. I can state my case a little strongly too sometimes.
For the sake of argument let’s assume that your premise is correct and that CDS are an imperfect substitute for CDOs, but all the imperfections are trivial.
Then clearly CDS are also substitutes for corporate and sovereign bonds. According to your view, we might as well just close down all the bond markets and just trade CDS. I think there’s a reason the market chose to keep corporate and sovereign markets open even after CDS were invented.
In my view, small imperfections in financial markets regularly become huge problems — because of the common use of leveraged arbitrage that exaggerates tiny differences. For example part of LTCMs losses in 1998 were due to subtle differences in 10 year Treasury bonds.
In short, the lack of a market in CDOs is a problem that is not solved by the existence of a CDS market — precisely because of those “trivial” imperfections.
Well CDSes ARE replacements for certain bonds. For instance until CDSes came along you pretty much couldn’t short ABSes as borrowing the bonds was impossible. So if you insist on being strict about talking about trading then none of the banks ever shorted CDOs in the sense of borrowing those CDOs and selling them. Was pretty much all done via CDSes. Shame GS didn’t think of this argument!!
I think we are talking at cross purposes here.
Just as the options market is a derivative market that can be used to hedge equity market risks without eliminating the equity market, so the CDS market could be used to hedge bond and CDO risk without eliminating the underlying market.
However, you appear to be claiming that the derivative market is a replacement for the underlying market — which seems to me an extreme position, that I genuinely don’t understand.
I’m trying to make a very simple point here: In order to claim to be a market maker in the CDO market, a firm has to actively trade in secondary markets in CDOs. Whether or not an almost equivalent position can be taken in a derivative market really doesn’t seem very relevant to me.
Just because a firm makes markets in options on Apple does not mean that the firm can claim to be a market maker in Apple stock. Is the difference between derivative and security markets really so confusing?
Maybe we are talking cross-purposes. It is not a matter of hedging, it is just you normally don’t trade the CDO or the bonds you trade the CDS. There are a number of technical reasons why this is, liquidity, tax reasons, funding reasons, legal and regulatory reasons.
The option vs underlying is not relevent here because there is no a direct arithmetic relationship in the price. With a CDS there is… again modulo those differences – CDS price = risky bond – risk-free bond. There is not such a direct link with options. If want to short a CDO I would not go out and try and borrow the CDO and sell it and hope to buy it back later. I would buy a CDS on it. Ditto most RMBSes. Maybe I am too close to this but it seems perfect obvious that it is merely a technical difference, a market microstructure difference.
I understood your underlying point to be that the underlyings weren’t traded due to some IB risk management reasons which isn’t true or because the IB didn’t want exposure which i think post 2008 has to clearly be untrue. Also the CDS prices affect the CDO price and vice versa. If I buy protection on a bond and sell an equivalent maturity risk-free bond then I essentially have sold that CDO. Obviously in times of stress these things are more complicated – ie if i am cash constrained and the income and expenses of these positions are not matched then i can get screwed, whereas if i had actually sold the CDO I would be sitting on a pile of cash – but i don’t think this technical difference makes your point.
Sorry if I am making a piss-poor job of explaining this, it seems kinda obvious to me….
Maybe the problem here is that you are looking for an “underlying point” when I don’t have one. I am simply making an observation about the fact that secondary markets in CDOs (for all practical purposes) have never existed. The Abacus prospectus supports this view.
You aren’t actually disputing my point. You are simply arguing that there is an “equivalent” market via CDS. Which seems to me completely irrelevant. The fact that the CDS market was used as a substitute for the non-existant CDO market is no more interesting than the fact that one can simulate stock exposure on the options market. In neither case are the prices perfectly correlated, but in both cases there is (almost always) a strong relationship between movements on the two markets.
You are claiming that the differences between the CDS and CDO market don’t matter, but the fact is that in a crisis these differences suddenly become very important — just think how different 2008 would have been if AIG had bought the underlying CDOs rather than selling CDS on them.
As I said before derivative and security markets are distinct and the differences cannot be ignored, by simply assuming that “there is a direct arithmetic relationship in the price” or that they are “merely technical”.
No offense meant, but the fact that you work in the industry makes the appeal to “technical differences” especially dubious. There’s nothing more terrifying these days than a financier who thinks he understands markets well enough to write the regulations himself. TED expresses this problem very well here: http://epicureandealmaker.blogspot.com/2010/01/im-dancing-as-fast-as-i-can.html