The emperor's clothes

April 29, 2010 in Finance

One of the problems with the latest mess is that the financial press and more specifically financial bloggers have built up a considerable amount of “[wall] street cred” through accurate and intelligent reporting on the financial crisis. In one sense, it’s amazing that they were able to gain such a foothold (I humbly include TGR) by doing nothing more than explaining what was happening on Wall Street without holding any punches – it really speaks to how limited/timid traditional financial reporting was. Most of the discussion centered on new but relatively simple concepts like “capitalization”, “liquidity”, “debt” and – for the most daring – “CDOs”.

But now the ABACUS deal has thrown a wrench in the process. We have reached a point where the nuances of these structures are so fine that the casual observer/blogger/reporter is oblivious to their significance. But we have also established an oligarchy of financial journalists/bloggers who form the authority on the ongoing recession. The problem arises when we collectively rely on those authorities for information they are not qualified to distribute. In that sense the internet is serving very well as a means of collecting and aggregating information (such as ABACUS pitchbooks) – and very poorly as an echo chamber of bad commentary and groupthink (such as ABACUS analysis). It’s not the first time this has happened – I’ve gone to task with the community a few times before. But now it is reaching a critical mass.

I began to think this more specifically last week when the NYT ran a bunch of silly stories claiming to expose Goldman Sachs’ fraudulent ways to the tune of just $2mm in some cases (none of which were, in fact, fraud and none of which were, in fact, remotely outside normal or healthy business practices). A friend in the know confirmed that he thought the stories were pure linkbait, and was surprised they did not anticipate the traffic caused by the GS hearings (the site went down).

Most recently, we have this post from James Kwak, who is quite esteemed in the economic blogging community. In it, he declares that ABACUS was not a CDO nor even a synthetic CDO. No, it’s a synthetic synthetic CDO. The distinction is more than redundant – it is wrong. But it has not stopped Kwak from using it as the basis for his ongoing argument, as here.

But Kwak was himself basing his argument on this post from Steve Randy Waldman, who hasn't got his head around this ABACUS thing. How can be there be no equity tranche? The term “synthetic” should be the tip off, but no – this must be a synthetic synthetic, since in his experience no synthetic tranche works this way.

Unfortunately for their attempts to obfuscate this deal, in my experience, all synthetics work that way. The reason the synthetic market arose in the first place was so that banks could avoid holding un-sellable parts of the CDO capital structure on their balance sheets – through synthetics, they can choose to hold/sell only the tranches they want exposure to. And most importantly, the distinction between “funded” and “unfunded” is silly – it is merely a choice between purchasing the tranche as a bond or as a derivative, with slight risk implications (different convexity profiles) and large balance sheet implications (namely holding risk on or off the balance sheet).

To be fair, there is a difference between the ABACUS deal and the classic textbook synthetic CDO. In the latter, CDS is sold on all of the issuers in the reference pool and noteholders agree to take responsibility for certain tranches of that insurance. In the ABACUS framework, CDS is sold on certain tranches of the reference pool and noteholders take responsibility for their entire insurance swap. The distinction is largely arbitrary; the textbook case is used because it has a more direct analogy to cash CDOs.

So the “synthetic synthetic” label really bothers me. It means that a) we have sufficient understanding of what’s going on to see that the whole capital structure isn’t represented but b) we lack sufficient understanding to communicate that observation without introducing another layer of abstraction. One wonders how many pages it would take these bloggers to explain the CDO squared civil lawsuits that are surely on the horizon. Is it some basic need to re-establish the blogging hierarchy by claiming that “my post will explain this confusing product” but actually cloaking it in further mystery? I would have thought such tactics belonged solely to Goldman Sachs...

I’m getting increasingly nervous about the stock we put in the ability of these “experts” to disseminate real information when it comes to non-macro factors. I dislike the extent to which the public has outsourced reasonable discourse to this hegemony. It’s time to notice that the emperor might not be wearing any clothes.

Here ends the rant. Please note this is not an argument in favor of any specific position on ABACUS (my opinions lie pretty in line with the consensus); merely a call for clarity.

{ 6 comments… read them below or add one }

Steve Waldman April 30, 2010 at 7:26 am

Unfunded super-senior tranches were common, and easy to understand. The last losses were allocated to issuers who didn’t put up collateral. The risks were spoken for although the tranches weren’t funded.

Unfunded equity and first loss tranches imply there cannot be a waterfall structure. It’s a different thing, and as you point out, implies a very different kind of product than just a CDO whose reference portfolio is synthesized. Neither James nor I had encountered so-called “single tranche CDOs”, which is a qualitatively very different product than a traditional synthetic CDO. We’ve encountered them now, and described them to readers, since they have suddenly become relevant to the public debate.

I’m not sure why the anti-love. For someone closer to the CDO business, single tranche CDOs (and more complex structures based upon them) might be old hat, and calling them “synthetic synthetic” or whatever might seem breathless and annoyingly dramatic. But as far as fishing out facts that the general public might not understand, understanding them ourselves, and presenting them, I don’t see any substantive critique here.


J April 30, 2010 at 10:16 am

I agree with you, the blogosphere has performed beyond any expectation as far as locating and presenting information. I become concerned when that information is laced with opinion which is passed, unvetted, into the next post. In an example of groupthink, the blogosphere has already formed its opinion of the case; any new information is used to uphold it.

For the record, I don’t even think that opinion is wrong. I’m concerned with the way it is propagating, which seems contrary to an academic understanding.

The defining characteristic of a synthetic CDO is the absence of a waterfall. Waterfalls exist because cash CDOs pool funds into a single reserve account and therefore the procedure for distributing those funds from one the single cash account to many noteholders must be stated explicitly. Synthetic CDOs get rid of these often expensive and cumbersome devices by simply agreeing ahead of time to distribute funds in a segregated manner.

Synthetic CDOs are premised on an agreement that can stand alone, without the need for other tranches. Say you had an insurance company which sold a $10mm policy that has a $1mm deductible – or, equivalently, a hedge fund who wants to speculate on senior risk. That firm could create a vehicle and place $9mm inside. It would then construct a reference portfolio with $10mm of credit exposure. The SPV enters into a swap which pays par less any losses over $1mm. The SPV then issues notes which are simply passthrough securities to the swap proceeds. The original firm purchases those notes and has thus synthesized senior exposure without any equity investor.

I think the distinction between synthetics and cash is particularly important in today’s political climate. Synthetic products – derivatives – are merely agreements that replicate economic payoffs that could have been created with cash products (though they can be arbitrarily complex). The interest rate, FX, and commodity markets are premised on this idea. Equity markets have seen dramatic increases in option use in the last three decades and ETN’s more recently. The less liquid a market, the harder it is to gain consensus on a derivative format – hence credit lags the rest. But the idea behind synthesized credit products is no different than that behind synthesized rate products: a standardized format for acquiring and selling various exposures.

We are spoiled by the stock market, in which there is no ambiguity about a company’s identity. There is only one class of common share (mostly): Apple is AAPL; Microsoft is MSFT. In credit, even that level of standardization disappears. What’s the yield on Hewlett-Packard bonds? Hard to say, because there are twenty of them, all with different maturities, coupons and other idiosyncrasies. There’s nothing wrong with that at all, but it speaks volumes to why the credit markets have embraced CDS and other synthetic products: they standardize exposures. There’s only one HP CDS contract (well, there’s two. Standardization is less than perfect.). Choose your maturity and you can have HP credit exposure. If bonds were uniform, than selling CDS would be exactly equivalent to buying a bond. As it is, it is almost the same because savvy investors will arbitrage the difference – and that efficiency is the engine for all of our financial markets, not just in credit and not just in synthetics.

But financial markets don’t serve much purpose for corporations except at the very moment they need to interact with the public (new issues, dividends, voting) – and even then it’s hard to justify continuous secondary trading. The private equity markets accomplish the same with only a few discrete moments of liquidity. Moreover, we don’t require secondary markets in order to disclose financials or public information. Instead, secondary markets exist because people *want* to trade exposures. And synthetics, in my mind, merely provide a clean way of doing so. As long as I can enter an agreement regarding a cash product – “I’ll sell you shares at $90 if you decide, for a fee”; “I’ll pay you the LIBOR rate in exchange for 2.5% a year”; “I’ll insure the top 90% of losses in return for a 1% coupon” – then I will be able to transact via synthetic or derivative.

When synthetics start to scare us because they aren’t “real” or because they are “gambling”, we need to start to take a close look at the cash transactions we take for granted. When I use my credit card to buy a cup of coffee, I enter a synthetic loan agreement with my credit card company. Moreover, if I prepay my card then I provide them with credit enhancement and synthetically create a senior credit card tranche for that monthly period. It sounds a lot scarier with the buzzwords du jour — but it’s a very simple transaction.


Sandrew April 30, 2010 at 10:25 am

I agree with you that Kwak is a bit far afield to the extent that he’s using his perceived level of the abstractness of the ABACUS deal as support for any wrongdoing. I don’t recall reading him as saying that, but perhaps I wasn’t reading carefully enough to catch it.

But I disagree that it is evidence of some failure of understanding on SRW’s or Kwak’s part to introduce new or non-industry language (“synthetic-synthetic”, “unfunded”, “non-existent equity tranches”) or to belabor the mechanics of the structure. Steve and James may not be in the biz and hip to its lingo, but I don’t think that discredits their insights and opinions.


Steve Waldman April 30, 2010 at 4:36 pm

J — I’m going to quibble with you on two details, but then agree with you almost entirely. The absence of a waterfall is not the defining characteristic of a synthetic CDO. Early synthetic CDOs, at least, did have the usual CDO waterfall structure. They simply replaced the cash debt portfolio with a debt portfolio synthesized from CDS positions, and potentially left the super senior tranches unfunded (which meant that the originator bore the super senior risk but did not post collateral). Synthetic CDOs without the waterfall structure are newer, and more complicated: they have to be dynamically hedged by the issuer (assuming the issuer does not want the equity/junior portfolio risk, and assuming, as I am told is usually the case, there is no single party interested in simply taking a short position on synthetic tranches). What makes a CDO synthetic is that the debt is synthesized. ABACUS-style CDOs are (as you point out) a level of abstraction beyond.

I’m also going to quibble with your equation of synthetic products and derivatives. Synthetic products are built from derivatives, but should not be called derivatives themselves. Suppose you build a fund that synthesizes the total returns from holding corn in efficient storage while holding sufficient debt cover a loss of value all the way to zero. You then issue claims against it as an ETF. Those claims are synthetic securities, and not usefully characterized as derivatives. Yes, their value is derived from the value of an underlying (corn), but they are liquid, tradable, limited-liability instruments that to the user are not different in kind than holding a stake in the company that most efficiently stores corn. They are securities, in most meaningful senses, including their nondescript “dangerousness” to more novice investors who are warned away from derivatives. If your mom doesn’t know finance but does know corn, and thinks corn is going to the sky, by all means she should purchase shares in a competently managed synthetic corn ETF.

Derivatives are the necessary building blocks for synthetic products, but a synthetic security behaves very differently from a straight unfunded bet with various collateral arrangements on the side to manage counterparty risk.

I agree with everything else you say. As an investor, I trade derivatives (public futures and options), and am very glad to have them. I think synthetic ETFs (when well-managed) are great. I think the economic rationale for derivatives and synthetics is sound. The problem is not derivative-ness or synthetic-ness, but complexity and asymmetric information. Because derivatives can offer very fine-grained exposure (a theoretical benefit!) and can be combined in very complex ways, one can tailor design products that maximally exploit ones informational advantages. Further, to the degree your derivative deals and public markets are not well-arbitraged, you can escape paying the usual and socially useful price of exploiting your informational advantage — informing market prices. Difficult to understand agglomerations of fine-grained bets that exploit informational advantage without ultimately informing market prices really are not socially useful, unless you consider penalizing the ignorant and overconfident useful. Even there, unfortunately, the ignorant and overconfident are usually playing with other peoples’ money, people who have no meaningful say in the deals or way to discipline the managers. We can’t have both a caveat emptor market and a delegated-to-professionals model of investing. We need either to regulate how we trade derivative and synthetic products so that constraints on professional money (its need for AAA and other forms of CYA branding) are sufficient to limit really stupid or predatory behavior by managers, or we need to stop delegating money so much and move to a model where risk-investing is performed much more by the economic owners of the returns.

By the way, I agree with you entirely that the distinction between synthetic and “natural” is wholly overstated and arbitrary. I view all financial contracts as synthetic — they compel cash flows at certain times and in certain states of the world. We are just most comfortable with the ones that compel a single cash flow once at the beginning, with variable cash flows coming back later. Organizing things that way (and not undoing it with leverage) does offer a certain degree of safety to the least sophisticated buyer possible: “I can’t lose more than I spend on this”, and it’s pretty sensible that we give this kind of investment a “G” rating and make them most easily available to retail investors. (No idea why binds are so much harder than stocks to trade, though.) Nevertheless, I would hate to see a world where we limit the investable menu to what’s suitable for children. But that creates an onus on those who (like me!) think derivatives and synthetics and stuff are useful to organize markets in such a way that supposed-adults don’t keep losing other people’s money to the smartest guys in the room.


J May 1, 2010 at 12:51 pm

Steve – very well put, and I agree on the whole. It seems to me that our difference on the term “derivative” is primarily semantic rather than substantive. I apologize as well; looking at my original post, it does seem unnecessarily harsh in coming down on you; it was not my intent to single you out, it was just those posts that triggered the thought in my mind.


Steve Waldman May 1, 2010 at 11:45 pm

J — Respect right backatcha. I’m glad to have discovered “the green room”.


Leave a Comment

Previous post:

Next post: