Sunday, 21 August 2011

Down the Rabbit Hole on Information-Insensitive Debt: Inscrutable Complexity Is a Good Thing!

Part 5 of a detailed look at Gary Gorton's curious theory of information-insensitive debt in which we ask two key questions.



ONE: IS INFORMATION SENSITIVITY A USEFUL PRISM THROUGH WHICH TO VIEW THE WORLD OF DEBT IN THE FIRST PLACE?



Not really, it seems.



A more useful theoretical construct would steer away from the bi-phase nature of "information insensitive" and "information sensitive" and would at least posit a sliding scale between the two. But an even better theory would ditch information sensitivity completely. Risk is the key to understanding how the world of debt works and how securities are analyzed, not information sensitivity (note: it’s probably no accident, in fact, that certain bloggers have equated Gorton's "information insensitivity" phrasing with the quality of being "risk free" -- but a careful reading of Gorton shows he makes no such equivalence, so he appears to be aware that there's a critical distinction).



A better theory might assert that, with the financial sector's demands for collateral to back derivatives transactions and so on, there will be a need for less-risky securities to fill that role.



TWO: HOW DANGEROUS ARE THE WAYS IN WHICH ASSET-BACKED SECURITIES BECOME INFORMATION INSENSITIVE?



Gorton seems to like asset-backed securities as information-insensitive debt for all the wrong reasons.



He likes them partly because of the senior nature of the debt and the fact that it's backed by a portfolio. He doesn’t recognize that the worth of being senior is firmly attached to credit risk. As an investor, which would you rather hold, if you're anti-risk: the senior debt of Energy Future Holdings (the former TXU that’s freighted with debt after being bought in the biggest LBO in history) or some (not senior) debt of AAA rated Johnson & Johnson?



This isn't a gratuitously needling point, because structured debt likes "yieldy" (read: riskier) assets. Collateralized loan obligations, a type of CDO, are stuffed with leveraged loans -- the high-risk borrowing that private-equity firms take out to make an acquisition. Why? The structuring doesn't make sense using investment-grade debt; you can't wring out enough yield.



So to say a securitization is more "information insensitive" because it may be backed by a portfolio composed of senior debt -- and then to be agnostic about the contents of that portfolio -- is very wrong. And what's more, you should be looking at how correlated the movements within the portfolio are. Junk loans in CLOs will display high correlation if the economy double-dips; that's pretty much a given.



INSCRUTABLE COMPLEXITY IS A GOOD THING? SAY WHAT?



Then there's the really dangerous feature of asset-backed securities that Gorton, bizarrely, is attracted to: complexity. This should be a bug, not a feature, but we've gone down the rabbit hole, folks. Here's his rationale: complexity raises the cost of producing private information. It's too expensive to figure out the debt is mispriced. Ingenious, though the arrant screwiness of this is never acknowledged.



However, here's the catch: that same complexity will, at some point, confer a significant advantage for a dedicated investor (such as a Michael Burry type in the Big Short) to do enough research to determine the extent of the mispricing. This will only occur though, after the mispricing becomes significant enough.



So what you get in the trading of this complex debt is the equivalent of a tectonic shift, violent and jarring, instead of the smooth adjustments that are made by say a U.S. Treasury, which trades largely on public information -- millions of bits of it, clashing and conflicting and impressing various traders in various ways. The asset-backed security, however, manifests itself as stable and information-insensitive -- partly because of its impenetrability -- then, on reaching a certain tipping point of mispricing, lurches into “information sensitivity.” Also, because of its complexity, ratings services will be sluggish to downgrade the debt -- especially after they have been complicit in the initial misrating -- adding to the sudden volatility.



Note, however, that this volatility wouldn't have to be characteristic of a panic or widespread fire sales, as Gorton wants us to believe was the main problem during the financial crisis. This aspect of volatility is inherent in the very nature of complex debt -- a kind of debt that Gorton lauds because it raises the cost of producing private information.



And Gorton sees this as a feature, not a bug. Hmmm.



A BANKING SYSTEM SHOULDN’T BE BASED ON TRADING IN MAGIC PIGS



Information insensitivity is NOT what we need more of in our financial system. Magic pigs are information insensitive, until there is a revelation (the discovery that they are not magic), at which time they become dangerously information sensitive. We DON'T WANT a shadow banking system built on magic pigs (or on securities that want to become magic pig-like).



Next: What’s behind haircuts in the repo market, according to Gorton? (Surprise: It’s not what you think.)