All the Devils Are Here [78]
There were a few critical differences between CDOs composed of securitized mortgages and CDOs composed of corporate loans. The former contained not two but three levels of debt. Instead of “merely” using debt to buy the debt of a company, CDOs were using debt to buy the debt from a pool of mortgages, which was itself homeowner’s debt. A second critical difference was that bonds backed by mortgages generally had higher yields than similarly rated corporate bonds. Defenders of mortgage-backed securities tended to explain away this anomaly, once again, by claiming that investors didn’t understand mortgage-backed bonds as well as corporate bonds, and thus demanded a higher yield for what was really a very safe asset. And to be sure, that was one possibility. Another possibility, though, was that the market understood quite well that mortgage-backed securities were riskier than corporate bonds and was compensating by insisting on a higher yield.
Wall Street didn’t really care which explanation was correct. All it cared about was that it had discovered an anomaly it could take advantage of. And, oh, did it ever. Firms bought mortgage-backed bonds with the very highest yields they could find and reassembled them into new CDOs. The original bonds didn’t even have to be triple-A! They could be lower-rated securities that once reassembled into a new CDO would wind up with as much as 70 percent of the tranches rated triple-A. Ratings arbitrage, Wall Street called this practice. A more accurate term would have been ratings laundering.
Soon, CDO managers were buying the lowest investment-grade tranches of mortgage-backed securities they could find and then putting them in new CDOs. Once this started to happen, CDOs became a self-perpetuating machine, like cells that won’t stop dividing. From the very beginnings of the mortgage-backed securities business, marketers had always had to work hard to find enough investors to buy the lower-rated tranches. The triple-As were easy to sell because investors around the globe that were legally confined to conservative investments, or didn’t want to hold the capital against a higher-risk investment, embraced their higher yield relative to their super-safe rating. The triple-B and -B-minus tranches were a harder sell, with a much smaller universe of potential investors. But once the CDO machinery itself became the buyer of the triple-Bs, there were suddenly no limits to how big the business could get. CDOs could absorb an infinite supply of triple-B-rated bonds and then repackage them into triple-A securities. Which everybody could then buy—banks and pension funds included. It really was alchemy, though of a deeply perverse sort.
In time, CDOs became by far the biggest buyers of triple-B tranches of mortgage-backed securities, purchasing and reassembling an astonishing 85 to 95 percent of them at the peak, according to a presentation by Karan P. S. Chabba, Bear Stearns’s structured credit strategist. Among other consequences, this practice helped perpetuate the worst, most dangerous securities, because they were the ones that had the highest yield relative to their rating. One Wall Street executive would later liken CDOs to “purifying uranium until you get to the stuff that’s the most toxic.”
Lang Gibson, a former Merrill Lynch CDO research analyst, wrote a novel after the crisis in which a character describes the CDO market as a Ponzi scheme. You can see his point. As the triple-Bs were endlessly recycled, CDOs begat CDO squareds (in which triple-B portions of CDOs were reassembled into a new CDO) and even CDO cubeds (reassembed triple-B tranches of CDO squareds). The rise of ratings arbitrage helped push sales of CDOs from $69 billion in 2000 to around $500 billion in 2006. It was an endless cycle of madness.
The rating agencies were at the very heart of the madness. The entire edifice would have collapsed without their participation. “Get the rating out the door—that was it,” says a former S&P executive. Once a tranche of a mortgage-backed security was stamped