Financial innovation is one of the favourite targets of those who take the demonological view of Wall Street.
Everything from ordinary securitization to more complex CDOs and credit default swaps gets put in the bag and beaten, often ignoring the immense good that much of the innovation has accomplished.
Simon Johnson and James Kwak have an uneven article in Democracy: A Journal of Ideas that tries to divide good financial innovation from bad financial innovation. That’s an important task, especially given our tendency to over-react to crises and throw the proverbial baby out with the proverbial bathwater.
Unfortunately, they don’t accomplish this task very effectively. They lay a blanket condemnation on collateralized debt obligations and credit default swaps when a more nuanced approach is required. Part of the problem with their critique is that they just don’t seem to understand the details of the financial innovations they are critiquing.
Take, for instance, this description of CDOs as “magic.”
The magic of a CDO, as explained in the research paper “The Economics of Structured Finance” by Joshua Coval, Jakub Jurek, and Erik Stafford, lies in how CDOs can be used to manufacture “safe” bonds (according to credit rating agencies) out of risky ones. Investors as a group were willing to buy CDOs when they would not have been willing to buy all the assets that went into those CDOs. We don’t have to decide who is to blame for this situation–structurers, credit rating agencies, or investors. The fact remains that at least some CDOs boosted financial intermediation by tricking investors into making investments they would not otherwise have made-because they destroyed value.
The correct reaction to this is captured over at the Economics of Contempt:
This is usually how CDOs are portrayed these days: they’re obviously voodoo finance, because–get this!–they claimed to take a bunch of risky bonds and transform them into a super safe bond. What a ridiculous idea, right? Now do you see how useless financial innovation is?
Of course, this isn’t a remotely accurate description of CDO.
Economics of Contempt does a good job of describing tranching. So you might want to click here and read what he says. Basically, he concludes that the problem wasn’t the structure of the CDOs. It was the fact that the investment banks that put them together, the ratings agencies, and investors didn’t realise that the underlying loans were of far worse quality than they thought.
In truth, however, there was a problem with the structure of CDOs. But in order to see what it was, you have to get past demonizing structured finance and understand how it works.
The idea behind a CDO is that bundling thousands of different credit obligations into a common pool should produce a reliable stream of income because exposure to unlikely events on the downside and upside should be limited. Geographic diversity limits exposure to local market downturns. Mixing mortgages with low prepayment risk and those with low default risk can further ameliorate interest rate risk. Mixing classes of credit can further reduce risk from being over-exposed to one type of credit.
Of course, all this breaks down if you get a correlated downturn across the country, across various types of debt and by a huge variety of debtors. And this is where tranching comes in.
The tranched nature of CDOs confuses a lot of people. How can some tranches have more risk than others when the risk of default on the underlying loans is the same for entire pool? This can seem like voodoo.
But what tranching does is make it possible for one part of the CDO—the “senior” part—to continue paying income even if a large portion of the underlying loans default. The senior tranches only absorbed losses after the lower tranches. In effect, the senior tranches were built on the leverage of the lower tranches, who received higher yields in exchange for greater risk.
So what went wrong? What happened was that the CDO senior tranches were too big. The senior tranches were far more popular than the subordinate tranches. There are lots of explanations for this but legislative prohibitions on certain institutional investors owning non-triple A rated investment played a large role. This put pressure on investment banks—who then pressured ratings agencies to go a long with the scheme—to enlarge the size of the senior tranches. And when the senior tranches grew too big, they were actually becoming far riskier than many understood.
To understand how the width of the senior tranche creates risk it helps to start with a very simple mental experiment. Imagine a situation where 90 per cent of mortgages in a pool default. Now if there were a super-senior tranche consisting of only 10 per cent of the total investment in the pool, that money would be safe. But as the senior tranche grows, its exposure to losses increases.
Historical data about borrowers defaults and prepayment risk from before 2002 seemed to justify wide senior tranches. Unfortunately, lending of all sorts—mortgages, credit cards, student loans–after 2002 had deviated from historical norms, which turned out to make historical data misleading. Until very recently, the ratings agencies and investment banks were not using data that reflected this change. This meant that the senior tranches were too broad, they included a greater percentage of investments in the overall debt pool than turned out to be justified by the risk.
In short, it wasn’t the complexity CDOs that led to their meltdown. It was the fact that demand for safety of senior tranches was so strong—and historical data from before 2002 misleading—that the pools were widened beyond warrant. The question we really should be asking is if its possible not to make this kind of mistake in the future, or if we’ll always get the size of the tranches wrong.