A year ago, a paper written by a student at Harvard College -- Anna Katherine Barnett-Hart -- received a fair amount of media attention. The paper examines the collateral performance and credit rating of asset-backed security collateralized debt obligations, so-called ABS CDOs. The author's findings are not surprising, in light of what is already known about the mortgage meltdown, but they do provide a better delineation of the parameters of the CDO meltdown.
Given the intense interest in CDO deals following the Securities and Exchange Commission's charges of civil fraud against Goldman Sachs for a deal known as Abacus 2007 AC1, it is worth revisiting the findings in this research and analysis by a college senior seeking to graduate with honors.
The author looks at both broad data on CDO performance and the ratings of credit rating agencies, but she also closely examines 735 ABS CDO deals.
The author's main finding is that there was far too much in the way of low quality assets in CDOs -- and given the continued deterioration since the report, updated information would find the conditions of those assets even weaker.
Relying on data from Lehman Live, Barnett-Hart found that the share of floating-rate collateral (adjustable-rate mortgages or ARMs) in ABS CDO deals, which had been a small share of the deals, rose sharply to become a majority of collateral in 2003. Synthetic collateral began to take off in 2005 and rose to about a third of the deals in 2006 before declining slightly in 2007.
Barnett-Hart explored a number of hypotheses and measured them by looking at a range out outcomes, from defaults to CDO downgrades.
In terms of collateral, Barnett-Hart examined three effects over the period from 2002 to 2007 with the following outcomes:
The Housing Effect: Increasing exposure to residential mortgages, specifically subprime and Alt-A residential mortgage-backed securities (RMBS), is associated with worse CDO performance as measured by defaults.
The Vintage Effect: Increasing exposure to 2006 and 2007 vintage collateral, particularly assets with floating interest rates, is associated with worse CDO performance as measured by defaults.
The Complexity Effect: Increasing the amount of synthetic collateral, the amount of pre-securitized CDO collateral, and the overal number of collateral assets is associated with worse CDO performance as measured by defaults.
CDO Underwriters and Originators
The paper also ranks the biggest CDO orginator clients of the credit rating agencies. And its findings agree with much of the anecdotal evidence accumulated in a number of press reports and an increasing number of books on the meltdown.
Merrill Lynch, with $76.9 billion in rated deals, was at the top of the list for Moody's, followed in order by Citigroup, UBS, Wachovia, Calyon, Goldman Sachs, Deutsche Bank, Credit Suisse, RBS, Lehman Brothers, Bear Stearns, Bank of America, West LB, Dresdner Bank, Morgan Stanley, Barclays Capital, and JP Morgan.
After reviewing underwriters and originators, Barnett-Hart concluded the following:
The Underwriter Effect: Holding constant general CDO characteristics, CDO performance varies based on the underwriting bank.
The Size Effect: The performance of an underwriter's CDOs varies according to the size of their CDO business, with overly-aggressive or very inexperienced banks issuing worse CDOs, as measured by their ex-post defaults and rating downgrades.
The Originator Effect: Controlling for the type of mortgages issued, as measured by average FICO, the combined loan-to-value, and debt-to-income scores, the performance of a CDO depends on the specific entities that originated the assets that became collateral in the CDOs.
Underwriters as Originators
Barnett-Hart also looked at how much the credit quality of CDOs is affected when banks are both CDO underwriters and collateral originators.
Here, she found some interesting correlations.
The Assymetic Information Effect: CDO performance will be affected if it contains collateral originated by its underwiter, although the performance might improve or decline, depending on the important of reputation vs. adverse selection and moral hazard.
To read the entire paper, click this link: