How Goldman deals with the government

By Felix Salmon
June 8, 2010

Comment of the day comes from HBC, on the subject of the FCIC’s Goldman subpoena:

Aww, cut Goldman some slack willya… They’re just not used to cooperating with the government. For years, it’s always been the other way round.

Stephen Gandel tries to do just that, turning the tables on the FCIC:

Is anyone surprised that the FCIC got billions of documents? …

The real question is what is the FCIC looking for, and does it even know…

How about calling up the heads of CDO trading at every other bank? Goldman wasn’t the only bank pitching synthetic CDOs. There are still a lot of issues that have been barely touched, and Phil, baby, its June, there ain’t that much time left.

Firstly, yes, the enormity of the Goldman document dump is surprising. The FCIC asked in January for “documents and information concerning Goldman’s synthetic and hybrid collateralized debt obligations based on mortgage-backed securities”, with a deadline of February 26. Goldman asked for an extension, and was given until March 5. Then Goldman asked for a second extension, and was given until March 8. And then Goldman’s submission was inadequate, but the FCIC allowed Goldman some time off because it was dealing simultaneously with requests from the Senate.

After the Senate hearings were over, at the end of April, the FCIC started badgering Goldman again, and was eventually told that the information would arrive on May 3; on May 4, more incomplete information arrived. Lots more back-and-forth resumed, and far from trying to help out, Goldman simply said that they had already provided everything asked for back on March 8. Eventually, on May 18, the five-terabyte document dump began: that’s roughly 2.5 billion pages. More back-and-forth, including a further incomplete submission on May 21; eventually the subpoena was issued on June 4.

You can see the running-out-the-clock here. And you can also see how the data dump was not a good-faith attempt to comply with the FCIC’s request. Gandel writes:

They asked for all the documents for all the synthetic or hybrid deals Goldman has ever done on residential or commercial mortgages. Goldman did hundreds of those deals. And the offering documents alone for each of those deals would be hundreds of pages. The correspondence between bankers could be thousands of pages.

Fine. Let’s say 10,000 pages per deal, and 1,000 deals. That’s 10 million pages. Multiply by 10 for good measure, we’re up to 100 million pages. We’re still in the realm of a rounding error compared to the billions of pages that Goldman provided.

It’s worth remembering here that, pace Gandel’s broader point, Goldman is just one of many different lines of inquiry that the FCIC is pursuing. Yes, the FCIC is looking at other banks, and the ratings agencies, and all manner of other players in the crisis. It’s not like they’re concentrating on Goldman alone. And Goldman knows that, so they know also that if they’re obstructive and unhelpful, there’s not much that the FCIC is going to be able to do to them before its time runs out.

Except, you know, start pillorying them in public and serving them with subpoenas.


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

If you’re not paying, Goldman isn’t going to do the work for you. It’s ludicrous to expect them to hand over a twenty-page, smoking gun dossier.

Wired had an article a few months ago on how the SEC lawyers deal with these enormous data-dumps. It made it seem like this activity was commonplace for responding to government requests by large financial companies.

Posted by jdanielwright | Report as abusive

Do you know how many pages are actually in a typical CDO prospectus? How about the underlying tranches? See page 37 of Andrew G Haldane, Executive Director, Financial Stability, Bank of England’s speach titled, “RETHINKING THE FINANCIAL NETWORK” for a guideline.

The number of pages Goldman has provied is not surprising if there are some CDO^2′s in there. ons/speeches/2009/speech386.pdf

Posted by david3 | Report as abusive

This wouldn’t be the first time Goldman has upchucked a load of worthless paper for the taxpayer to pick through. It’s what they do for a living.

Posted by HBC | Report as abusive

You’re forgetting the relevance of information of underyling mortgages. It’s not possible to evaluate the synthetics or CDOS based on the offering documents themselves. Data, documentation and research on all the underlying securities would certainly be relevant.

If you asked for documents regarding a equity ETF, information and research regarding the constituent companies and securities would similarly be important.

It’s a giant exponential effect, especially considering some of these CDOs and synthetics have CDOs as their constituents.

Each CDO / Synthetic could possibly contain tens or hundreds if not thousands of mortgages, other CDOS. The inherent problem with these securities in the first place was that for most, if not all participants it was nearly impossible to wade through all the documents and properly evaluate the risks.

Posted by JonLW | Report as abusive

Note: The above isn’t to absolve Goldman. The data dump should give people an idea of just how screwed up the market for these securities really were. Most investors clearly did not perform proper diligence regarding their investments in CDOs. It would take significant amounts of time to evaluate even a single one of these synthetics, let alone create portfolios consisting of many many CDOs.

The takeaway for the FCIC, banks and investors should be that the obfuscation of details and reliance on faulty “diversification” in the face of complexity was a huge factor in the crisis.

Posted by JonLW | Report as abusive

When you lack any moral compass and, in fact, are legally duty bound to maximize profits, it makes sense to run out the clock, ruthlessly default, or take any number of morally dubious actions.

We don’t really expect them to actually cooperate do we? Americans are required to check their morals at the office door and we have done so willingly for decades.

Posted by silliness | Report as abusive

JonLW, what makes you think that investors were interested in doing proper due diligence? For them it is a one way bet – thing doesn’t blow up then they are geniuses for getting that extra yield and if it does then they are poor chumps who were fooled by the evil sellside who failed to disclose some spurious bit of information.

All this data dump does show is how regulations are forcing more and more useless boilerplate onto documentation and apparently the solution is to create the need for more boilerplate.

Posted by Danny_Black | Report as abusive

Oh I have no doubt that investors had no interest in actually doing proper due diligence. Though unstated, that was kind of my point.

I think the FCIC along with most people investigating the financial crisis are using seriously faulty logic. The idea that documentation, prospectuses and misleading information were the main culprit risks simply adding the wrong type of regulation. The problem isn’t just bad information, it was that there was simply TOO much information for people to willingly process.

Posted by JonLW | Report as abusive

Danny_Black and JonLW, the investors didn’t have to do proper due diligence. That was done by the ratings agencies when they stamped AAA, AAb, etc. on the different tranches of these CDOs. Really, the SEC should be tapping the people at those agencies, who obviously have the ability to evaluate these complex securities.

Posted by KenInIL | Report as abusive

GS delivered 5 TB of data. This is very little data compared to corporate bankruptcy or giant lawsuits. There are companies that specialize in indexing all those documents so lawyers can query them like a search engine. In fact, you can buy a Google Search Appliance and index all those documents overnight and then search the data. In fact, I could probably index all this using free tools on my desktop. It’s a trivial technical issue. Thanks for blowing it out of proportion.

Posted by projectshave | Report as abusive