The recently-released Senate financial reform package contains a number of provisions designed to prevent a recurrence of the performance problems that triggered the recent financial crisis. The most widely discussed initiative requires "securitizers" to retain a minimum of 5% of the credit risk of any sale or conveyance of assets through the securitization process.
To its credit, the bill does attempt to differentiate between various types of assets, while giving regulators broad authority to exempt some classes of issuers. (This almost certainly will include the GSEs, although the current bill does not specifically exempt Fannie Mae or Freddie Mac from the requirements.) By focusing on the securitizers of ABS (and originators that sell assets to them), the bill also exempts sales of non-securitized assets.
The legislation's focus on "skin in the game" is, however, unfortunate. First, the notion that securitizers' lack of credit risk exposure was a primary cause of the financial crisis is patently untrue. Both originators and securitizers had plenty of credit exposure, given their portfolios of loans and their holding of MBS residuals and CDO tranches. The failure of institutions such as Washington Mutual, Countrywide, and Merrill Lynch arguably took place because they had too much exposure to products backed by flawed and nonperforming assets. (As a friend put it,"They already had skin in the game -why do you think none of them are around?")
The capital requirements and costs associated with risk retention will also put smaller lenders at a competitive disadvantage vis-Ã -vis larger originators, forcing them to sell their production in whole-loan form. This will ultimately lead to the development of a multi-tiered, highly consolidated, and consumer-unfriendly mortgage industry.
The requirement of risk retention also creates the probability, albeit small, that in the future the entire framework of asset securitization could be called into question. While the consensus is that the 5% retention requirement is too low to prevent sale treatment for most securitizations under current accounting practices, it's possible that future statements from the FASB could directly or indirectly change the accounting treatment for securitized assets.
Taken together, the risk retention proposals risk creating the illusion of a solution to problems related to securitized finance while offering little additional protection to investors. Moreover, the bill misses an opportunity to truly improve the securitization process by strengthening and codifying the representations and warranties associated with ABS deals. While a section of the bill does address the issue, it is very limited in scope. This provision requires the rating agencies to include a description of the deals' reps and warranties "in any report accompanying a credit rating," as well as requiring securitizers "to disclose fulfilled and unfulfilled repurchase requests across all trusts aggregated by the securitizer."
A robust enhancement of deals' reps and warranties would require a few straightforward provisions. For example, there is currently little standardization with respect to when and how investors can gain access to loan files - which are necessary for a full forensic analysis of any deal's compliance with its governing documents - from the servicers and trustees.
Rep and warranty language should also be standardized. For example, most PSAs contain a clause stating that loans collateralizing a deal must conform to their program guidelines; however, the loan programs are typically outlined in very general terms. A requirement that the loan programs must be outlined in each deal's governing documents, and that all loan data contain ways to identify each loan's program, would be a major improvement.
In my mind, stronger and more enforceable reps and warranties would have a much greater impact on securitization quality than forced risk retention.
Bill Berliner is a mortgage and capital markets consultant based in Southern California. His Web site is www.berlinerconsulting.net
