May 7, 2010
I would like to make the following comments and suggestions with respect to your proposed revisions to Reg AB as it applies to asset-backed securities.
Typically, a CDO collateral manager or equity holder could ask the trustee to auction the collateral after a fixed number of years (usually, 3, 5 or7) and pay bondholders off at par. Thus the buyer of a bond in a CDO effectively provides financing for an asset chosen by the collateral manager. If the asset defaults, the bond buyer suffers the writedown. If the asset appreciates and the collateral is auctioned, the most that the bond buyer can realize is par, and the collateral manager realizes any
further upside. Thus buying a CDO bond is the equivalent of selling the collateral manager a put option. CDOs which were "managed" gave collateral managers additional oppotunity to scam bondbuyers. The assets used to collateralize CDOs had optionality as well. If the
assets were mortgages or mortgage-related, as most were, the borrowers had the option to prepay or default. Any serious market participant paid much attention to the likelihood of prepayment and default in the
mortgage-related asset. By allowing CDOs to exist, you are allowing bondbuyers to buy options on options. This would not be allowed in the equity markets and as such, I see no reason why it should be allowed in the debt markets. As John Thain famously said about a CDO analyzed by Merrill Lynch: To model correctly one tranche of one CDO took about three hours on one of the fastest computers in the United States. There is no chance that pretty much anybody understood what they were doing with these securities. Creating things that you dont understand is really not a good idea no matter who owns it.
Although your revised regulations would prohibit synthetic collateral,nothing I saw would prohibit resecuritizations. The crux of the credit crisis was that debt could be originated which, thanks to the CDO, could
always be sold as AAA. In other words, if the ratings agencies rated 85% of your collateral as AAA, you could sell the 15% which wasn't AAA into a CDO of which a significant portion (70% or more) would be rated AAA.
Doing this ad infitum into infinite generations of CDOs meant that ultimately everything is AAA. As AIG found out the hard way by insuring these CDOs, squared and cubed, everything isn't. Unless you regulate
resecuritization, it is likely to happen again.
In addition, your proposed revisions would require an issuer to retain 5% of a pool. Previously, as it was difficult to sell anything not rated investment grade (BBB or higher), issuers already retained anywhere from
6-10% of a deal, depending on how much of the pool was rated BBB or higher (and, if they couldn't sell it into a CDO). Thus, as the ratio of sold versus retained would be higher than the ratio of investment to non-investment grade, your revisions might actually make things worse in
terms of the amount of leverage permitted in a debt structure. Wall St. makes money by increasing illiquidity turning something simple into something complex widens the bid/ask spread. Even to this day, they are buying securities which were once AAA-rated and resecuritizing them into the less-than-100% portion which ratings agencies consider to be still money good, a product for which the secondary market hardly exists. Nothing in your proposed revisions restrains them from doing so.
Finally, the idea of any revisions should be to increase transparency, not decrease transparency. Despite my respect for a company such as Intex's ability to model complex securities such as resecuritizations, it does not
mean that overly complex debt instruments should be permitted. Most people can not read Python or XML. If you need these to describe a deal's behavior, the deal should not be allowed to exist.