Monday, January 11, 2010

Ooops! Talf gaffe at the Fed

The NY Fed press release:

As part of the process for reviewing requests for TALF loans to be collateralized by legacy commercial mortgage-backed securities (CMBS), the New York Fed conducts a risk assessment of the proposed collateral. The assessment considers whether the estimated value of the CMBS would fall below the loan amount should economic conditions turn out to be much worse than expected. The New York Fed currently obtains these “stress value” estimates from two separate vendors.

The New York Fed continuously reviews the stress value estimates and recently identified and corrected a methodological error. The New York Fed has determined that as a result of this error, one legacy CMBS — CUSIP 059497AX5 — was accepted as collateral that would not have been accepted using the current methodology. However, the New York Fed continues to expect no losses on the loans backed by this CMBS because the stress value is based on extremely unlikely economic circumstances, and because the market value of this CMBS is well above the TALF loan amounts.

The New York Fed will not accept CMBS CUSIP 059497AX5 as collateral for new TALF loans at or around its current market price. The New York Fed continues to reserve the right to reject any legacy CMBS in the future, whether or not the legacy CMBS was previously accepted.

Accepted and Rejected Legacy CMBS Archive »

And some commentary posted on FT Alphaville by Tracy Alloway:

A major mea culpa from the Federal Reserve on the legacy CMBS portion of its Talf programme...

CUSIP 059497AX5 corresponds with BACM 2007-1 (Class A4). The Fed previously accepted some A4 tranches of the BACM 2007-1 deal in September and August, then mysteriously rejected the A3 tranche in October, prompting analysts to ask for transparency on the Fed’s selection process.

On the plus side, and at the very least, we do know that the Fed is applying some sort of stress-testing methodology when it chooses CMBS bonds to accept for the Talf. The downside is that we still don’t know what that methodology is, and it, err, might be prone to errors.

No comments: