How smaller banks would have helped shrink the CDO market

By Felix Salmon
December 13, 2009
new details on how banks would pass CDO risk between each other in an improbably long chain:

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">

The WSJ has new details on how banks would pass CDO risk between each other in an improbably long chain:

One of Goldman’s trades with AIG involved a financial vehicle called South Coast Funding VIII. South Coast was one of many pools of bonds backed by individual homeowners’ mortgage payments that Wall Street turned into collateralized debt obligations or CDOs.

Merrill Lynch, now part of Bank of America Corp., underwrote the South Coast CDO in January 2006 by stuffing it with packages of home loans originated by firms such as Countrywide Financial Corp., the big California lender.

Once a CDO debt pool is assembled, it is sliced into layers based on risk and return. Merrill sold the safest, or top layer, of deals like South Coast to large banks, including in Europe and Canada.

The banks wanted protection in case the housing market tanked. Many turned to Goldman, which effectively insured the securities against losses. Then, to cover its own potential losses, Goldman bought protection from AIG, in the form of credit-default swaps.

The risk, here, originated with Countrywide. It then got moved to South Coast Funding, whence it moved to Merrill Lynch, and thence to European banks, who sold it on to Goldman Sachs, who in turn passed it on to AIG. When the music stopped, AIG bore the brunt of the losses.

The obvious question here is why the chain was so long: why couldn’t Merrill (or even Countrywide) just insure the mortgages with AIG itself, instead of sending them off on a long and winding road to end up in the same place?

There’s actually a good answer to the question. Countrywide considered itself to be in the originate-to-distribute business, it didn’t want an exploding balance sheet. So it was very happy to sell its mortgages to the likes of South Coast Funding and book the profits immediately. South Coast Funding, in turn, had no real equity of its own, it was just a special-purpose vehicle set up by Merrill Lynch. And Merrill considered itself to be in the moving business rather than the storage business, so it wanted to sell as much of the debt as it could.

Eventually, the bonds ended up with big European and Canadian banks, who were attracted by their triple-A credit ratings: under Basel II, that meant they needed to hold very little capital against them. Crucially, it was these banks, who had no size limits, which were happy expanding their balance sheet as much as they could, so long as that meant extra profits. They even managed to bring their risk down near zero by getting Goldman to insure the credit.

Finally, Goldman was basically in the trading business, insuring CDOs only if and when it knew that it could get someone else (in this case, AIGFP) to reinsure the risk at a lower rate; eventually the European banks got wise to that trade, and started dealing directly with AIG themselves.

There’s a lot of blame to go around here, but right at the heart of it is the fact that there were no limits on the size of banks’ balance sheets. The European banks (and, later, once AIG stopped insuring this stuff, Merrill Lynch itself) would happily balloon up as large as they could with stuff like this, because sheer size was one of the profitable attributes that regulators didn’t mind in the slightest. If there were a cap on size, this chain would have been much harder to construct.

More From Felix Salmon
Post Felix
The Piketty pessimist
The most expensive lottery ticket in the world
The problems of HFT, Joe Stiglitz edition
Private equity math, Nuveen edition
Five explanations for Greece’s bond yield
7 comments so far

Smaller banks are one of the least relevant proposed solutions to this crisis. Note that the U.S. has an enormous number of banks and that the little ones are failing all over the place (over 130 this year). The same was true in the S&L crisis and the Depression. The same would be true in Japan if Japan let banks fail. The losses of 100 bad banks with $1 billion each in bad assets are going to look pretty similar to the losses of 1 bad bank with $100 billion in bad assets. Talk about underwriting, talk about leverage, talk about regulatory arbitrage, talk about incentives for irrational risk-taking, talk about executive over-compensation, talk about bailouts – these topics all make sense – but why would we think that having more/smaller banks would have any bearing on these other problems?

The real problem with all the transactions in this CDO chain is that every participant thought that they were being paid real money to take a non-existent risk, and therefore they took as much of this “risk” (which they didn’t really believe in) as possible. While this may be feasable on very small scales (e.g. for 5 minutes there is a risk-free arb that you can put a few million into to make a tiny spread), it’s unlikely that our financial system is such that people can earn spreads over Treasuries for years on billions of assets without taking any risk. The smarter players eventually recognized the risk and did everything they could to get rid of it. The rest got bailed out. I would have much preferred that my government teach people the valuable lesson that there’s no free spread by putting a haircut on every bit of debt they bailed out, but some people still think they’re geniuses for buying agency mortgages and Merill Lynch bonds last year. Everyone who bought Lehman bonds was an idiot though. What’s the difference between Merill and Lehman? About two additional weeks of independent solvency and a cooler logo.

I have no problem with leverage and size for financial speculators – I just don’t want to bail them out. If someone wants to bet at the bucket shop or play foreign exchange levered 100-1, it’s fine with me – they add liquidity to the market and maybe their speculative excess provides me with a nice entry point.The problem is that our government has bailed out so many speculative entities, which means that people will continue to underestimate risk and provide easy capital to insufficently-skilled speculators while pretending that what they’re doing is investing. Without bailouts and the continued faith of the sheeple that they create, a company like GS would have a significantly higher cost of capital and cost of leverage (see Buffett’s investment). Then we wouldn’t have to whine about their traders being overpaid because they wouldn’t have the money with which to overpay them. It’s only because people are willing to hold GS equity and debt at pitiful returns that they can do what they do. If you invest in index funds you’re an enabler :)

Posted by najdorf | Report as abusive

Was the problem the size of the balance sheets, as you say, or rather the ratios the banks were allowed to maintain because of the flawed ratings? What if there were 5x as many banks, each with 20% of the size of CDOs? More banks would have been in trouble then, and with the same amount of bad loans. That doesn’t sounds any better.

If the ratings can not be trusted, then the debt to capital ratios must be lowered. This would have a better effect on reducing risk than just limiting the size of banks. I don’t want to have to bail out any bank, regardless of their size, and only regulating size wouldn’t have prevented the collapse we experienced.

Posted by OnTheTimes | Report as abusive

I’ll second OnTheTimes. For me, the key sentence is this: “Eventually, the bonds ended up with big European and Canadian banks, who were attracted by their triple-A credit ratings.” How did all this risk not get factored into the bond ratings? And the recently passed House financial industry reforms did bupkiss to address any of this.

Posted by jte | Report as abusive

I don’t see how a cap on absolute size would have done anything. A lot of the banks bought the CDOs, directly or through SIVs, were pretty small with balance sheets in the tens, rather than the hundreds of billions – IKB, HSH, SachsenLB. A gross leverage ratio which took into account off balance sheet vehicles might well have done something to slow it, however.

Posted by GingerYellow | Report as abusive

I agree with the other commenters that a cap on bank balance sheets would not have unambiguously prevented these risks from being distributed through this chain. Furthermore, it’s not clear to me why the chain itself is a bad thing. I suppose there are extra transaction costs (intermediary profit) at each step that could have been avoided had a few steps been short-circuited. Note that what we ended up with was dodgy loans originated by Countrywide, funded by European banks, and insured by AIGFP (who, it turns out, was back-stopped by US taxpayers). I take it your bigger gripe is with find that end state, not how we got there. That is, even if we arrived at this particular risk distribution more efficiently (with fewer steps and less intermediary profit), the risks still would have been borne by taxpayers.

I’m not sure I agree that capping the balance sheets of financial firms would help anyone. By what mechanism would this curb systemic risk and/or the need for taxpayers to bear it? (Or is that not the idea?) I’ve yet to hear this spelled out clearly. The only thing I can come up with is that if we shrink all the banks, then no one bank will be TBTF. Well, maybe; but so what? How is the bankruptcy of one big bank materially different from the simultaneous bankruptcy of a bunch of little banks? What makes you think that the balance sheets of a bunch of little banks wouldn’t be just as correlated as the combined balance sheet of one big bank?

Posted by Sandrew | Report as abusive

It seems more than one factor matters and financial reform should look into ALL of them:

- size of the balance sheet
- debt to capital ratio/mandatory cash reserve level
- the type of financial institutions that are allowed to do any kinds of “financial alchemy” as opposed to plain old lending (re-institute Glass Steagall or at least some form of it)
- change tax code to encourage risk-aversion and discourage excessive risk-taking

Posted by jian1312 | Report as abusive

jian1312: Why should we discourage risk-taking in the tax code? Risk-takers have been very, very good to the IRS with their tendency to generate growth, profitability, and corporate income/capital gains taxes. I would rather work on modifying our social contract so that no one expects that the government will save risk-takers when downside materializes. Size has been addressed. I agree with you on the other two.

Posted by najdorf | Report as abusive
Post Your Comment

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see