Felix Salmon

It’s impossible to price a CDO

By Felix Salmon
December 28, 2009

The discussion about Goldman’s synthetic CDOs just happens to coincide with a separate thread about a working paper by four researchers at Princeton, who seem to have a solid mathematical demonstration that CDOs simply can’t be priced: the amount of computing power necessary to do so just doesn’t exist.

Robert Oak has a good English-language explanation of the implications of the article: it boils down to the idea that if Goldman was loading up its synthetic CDOs with nuclear waste, there’s a good chance that its clients couldn’t have worked that out. Not because they were insufficiently sophisticated, but just because there aren’t enough computers in the world to do so.

It’s possible to overstate the connection between the paper and the events that happened at the height of the credit bubble. There’s a hidden implication in the paper that maybe with better models or faster computers we might have avoided some of this mess — but the solution to model risk isn’t more complex models, its less reliance on models altogether. And anybody who applied a simple smell test to the mortgages underlying the CDOs in question — rather than deciding instead to trust various quants both in-house and at the ratings agencies — would have come to the right conclusion without any computing power at all.

That said, there’s a case to be made that Goldman — along with much of the rest of Wall Street — was indeed arbitraging the models that investors and ratings agencies were using to price CDOs in general and synthetic CDOs in particular. The latter were particularly toxic, because synthetic CDOs are zero-sum, meaning that Goldman stood to gain in precise proportion to the degree to which the buyers of the paper were underestimating the risk involved.

The result was that Goldman had every incentive to structure CDOs which would blow up spectacularly, leaving investors with massive losses and Goldman with equal and opposite gains. The investors, meanwhile, might not have trusted Goldman (after all, they understood the zero-sum nature of these instruments), but did trust — far too much — mathematical models which turned out to be deeply flawed. And which, now, seem to be far too complex to get a grip on in any case, given the limitations of today’s computers.

In the final analysis, much of the problem here was a function of banks waving the magic letters “AAA” in front of the eyes of lazy institutional investors who thought they were getting free money by buying risk-free debt at hefty spreads over Treasuries. Some banks, like Merrill Lynch, did that and still contrived to lose billions; Goldman, by contrast, did that and managed to make a profit on the deals.

The whole sorry episode I think was fairly summed up by Goldman CEO Lloyd Blankfein when he said that “we participated in things that were clearly wrong and have reason to regret”. At the same time, however, Goldman wasn’t particularly evil or cunning here. It was simply acting like everybody else in the finance world, trying to take as much advantage of the credit boom as it possibly could. The main difference was that it had a bit more control over its own balance sheet than most of its rivals.

7 comments so far | RSS Comments RSS

All derivatives, futures, options and swaps are zero sum. One counterparty’s gain is the other party’s loss. CDOs behave as any other derivatives behave when it comes to winners and losers.

For many years, farmers have used agricultural derivatives purchased at planting time to protect themselves from the uncertainty about crop yields and prices at harvest time. Speculators, investors and food companies are the counterparties to the farmers.

All the computers in the world combined and all the supercomputers cannot accurately predict harvest time prices. Weather, droughts, floods, hurricanes, frost, crop diseases, foreign growers internal country politics, wars, tariffs, strikes, oil price spikes, e coli and salmonella outbreaks and other price effecting events are too unpredictable to know with any degree of uncertainty in advance.

Yet, the agricultural derivatives markets work very well and without the ability to compute future outcomes.

There is much more to the CDO problem than nuclear waste mortgages and computational difficulties.

Financial nuclear waste sells all the time without blowups as occurred in the mortgage market. Investors with high-risk tolerance buy, invest in and trade debt of bankrupt companies, junk bonds of very low credit rated companies, start-ups, and other high risk, high probability of default debt all the time.

Financial markets exist for all kinds of debt with high degrees of uncertainty about future events and with high likelihood of defaults without major problems.

If Goldman was selling snake oil, there was nothing compelling the buyers to buy and invest in the snake oil. Was Goldman holding a gun to each investor and buyer’s head of the CDOs?

Furthermore, and a more compelling argument, is that US Treasury bonds of all maturities from a few days to 20 years was available to the same investors who purchased Goldman’s so called nuclear waste. When there was a safe, transparent alternative investment that all the buyers knew about, why did they buy CDOs? US Treasuries are better than AAA rated!

Additionally, private equity, hedge funds and mutual funds were not the investors who ran into problems purchasing CDOs, (the mutual fund that broke the buck was a higher risk fund that attempted to increase its yield by heavily investing in higher risk securities like mortgage derivatives and just because it called itself a money market fund the Fed panicked).

The banks, who are sophisticated investors, suffered the most losses from the CDO market collapse.

What really went wrong for the financial institutions from the collapse of the CDO markets due to excessive mortgage defaults was that the financial institutions were overly invested in these kinds of debt.

Market prices collapse all the time for different kinds of investments, be it Gold, foreign currencies, commodity prices, etc. There is one simple rule of investing that has stood the test of time and does not need a lot of computer power. The rule is DO Not Put All Your Eggs In One Basket. The banks were not diversified. They bet the bank on mortgage related debt and they lost. They were gamblers that put everything on red on the roulette wheel and black came up.

Why banks did it is a whole other discussion, but Goldman was not the cause. Yes, it had the product that blew up, but it id not create the hunger in the buyers and investors for that product. Goldman met the banks cravings for these investments, but it id not make the banks the junkies that they were for this crack.

If the CDO outcomes were not computable by the buyers, then the CDO outcomes were not computable by Goldman. Goldman could not know in advance that it would see windfall gains.

Furthermore, the CDOs that ran into the most problems were the ones originated after the housing market started to collapse and that data was publicly available to the investors, including as to which areas of the US were seeing the sharpest price declines. Plus, the investors knew or could easily ask Goldman and Goldman would tell them which areas of the country had the mortgages in the CDOs they were buying.

Most of the responsibility for the problems should go the regulators. They overlooked the lack of diversification and the overinvestment in mortgage derivative debt. They promulgated rules that allowed banks to hold for each dollar of capital against a residential mortgage, two and one half times as much CDO mortgage debt. The regulators, through capital rules that promoted capital arbitrage, turned banks into crack heads for CDOs and then failed to supervise their addicted banks.

Goldman was like a bartender. It had a secondary and not a primary responsibility. It did not create the alcoholics, but at some point, it should have seen that the regulators were not stepping up to the plate and that their customers were in over their heads and too drunk to drive anymore. However, the regulators created the drunken banks and not Goldman. Goldman was at worst opportunistic.

Posted by MiltonRecht | Report as abusive

Milton: You managed to (deliberately?) overlook the simple fact that Goldman was also in the business of creating the job description and nominating their own regulators. Goldman is solely responsible for each and every one of the items mentioned in those last 2 paragraphs.

Regulators didn’t “overlook” the lack of proper capital ratios – Goldman, through Paulson who was then their CEO, demanded exemptions on leverage for the top 5 firms on Wall Street. This means that regulators COULD NOT make any moves because they were formally barred from doing so. It couldn’t be the regulators’ responsibility, Milton: regulators were absolved of any responsibility when all ability to exercise control was eliminated.

It’s all Goldman’s fault.

Posted by Unsympathetic | Report as abusive

This is a real Holy Cow result, IMO. The Efficient Market Hypothesis goes ten rounds with the Turing Machine, and loses. !!.

Posted by MattF | Report as abusive

Milton’s analogy of the bartender is basically correct, though the reality of what Goldman did is a little worse than just a bartender who keeps serving to an inebriated customer. In this case, Goldman brewed its own moonshine, which was stronger than the standard stuff, and served that to the tipsy customers, in order to juice its own profits. But what Goldman didn’t anticipate was that the drunks were going to pass out smoking and cause a fire which would nearly consume the whole establishment, until the Feds arrived to put it out.

Posted by o_nate | Report as abusive

“If the CDO outcomes were not computable by the buyers, then the CDO outcomes were not computable by Goldman. Goldman could not know in advance that it would see windfall gains.”

So, starting in 2004, they began setting up incredibly complicated synthetic mortgage-backed CDOs, selling them to customers, paying out premiums on a regular basis, not having a CLUE that they’d be making not only enough profit to cover their payouts, but to also make windfall profits.

Sounds like inept risk-management. It’s a good thing none of their counterparties on anything went bankrupt like, say, AIG.

Posted by nktrumpsall | Report as abusive

you argue that investors “did trust — far too much — mathematical models which turned out to be deeply flawed”.
My points are:
# Goldman (and other banks) used the flawed copula models
# Investors didn’t necessarily use any model (they just needed to mark to market the CDO)
# Investor just needed a triple-A asset yielding more than treasuries, that was enough
# All models are flawed, also the widely used Black model is deeply flawed. You can argue that the flaws of these models were not well known
# A better model would have helped those seeking to hedge the CDO risk (maybe); how many investors were hedging their CDO positions ? I guess not many. A better model would have saved some banks, not investors.

Posted by BozoTheGrey | Report as abusive

Felix, perhaps what you should say is that CDOs can’t be priced accurately. But what asset can be priced accurately? We don’t know the future. All CDOs got priced in an environment where people would make offers given the deal and pricing terms. Many offered to buy too much and lost. Systemic errors do happen. MOre people are buying aggressively at the peak of the market than the valley.

Posted by DavidMerkel | 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 http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/