The most recent issue of The Accounting Review (March 2012) contains a study by Barth, Ormazabal, and Taylor that examines the association between two measures of credit risk and asset securitizations.  I found the study to be interesting and relevant to standard setters.

Accounting for asset securitizations is controversial, and critics have charged that GAAP allows companies to artificially deflate leverage ratios and thereby, effectively mask their inherent risk.  However, GAAP permits securitized assets to be treated as sold only when the following summarized conditions are met: 1) the transferred assets must be isolated from the transferor— put presumptively beyond the reach of the transferor and its creditors, even in bankruptcy or other receivership, 2) the transferee must have the right to pledge or exchange the assets without conditions or constraints, and 3) the transferor does not maintain effective control over the transferred assets.  Meeting all of these conditions would appear to transfer all of the risks and rewards to the transferee and therefore, accounting for the transfer as a sale would seem appropriate.

In their study, Barth et al. find a significant association between a transferor’s S&P credit rating and the retained portion of securitized assets.  The retained portion includes only those assets where the transferor retains a contractual interest and does not remove from their balance sheet.  Of primary interest, however, is the finding with respect to the portion of securitized assets that is removed from the balance sheet – the non-retained portion.  The authors find no significant relation between the non-retained portion and the transferor’s credit rating.  This finding is consistent with the accounting for these assets as sales, but inconsistent with claims by critics of “sale accounting” that the transferor retains the credit risk of transferred assets because of non-contractual but implicit obligations.

The findings of the study become more interesting when the authors use their second measure for credit risk in their tests – the firms’ quarterly bond spread, which is measured as the annualized bond yield less the compounded risk-free rate on one-month T-bills.  In this case, the authors find statistical significance for both the retained and non-retained portions of the securitized assets, which indicates that the transferor retains the risks inherent in the assets that they swept off the balance sheet.

The contrasting findings from using different measures of credit risk suggest that either the credit rating agencies or the bond market (or BOTH) got it wrong in their assessment of credit risk associated with securitized assets.  Both groups can’t be right (although, I suppose, they both could be wrong).  One could argue that credit rating agencies possess private information and expertise that facilitates their developing an accurate rating of the transferor’s credit risk.  However, the popular press has criticized the agencies’ ability to understand the underlying complexity of these transactions and their conflicts of interest, which might have compromised their objectivity.  In the wake of the financial crisis, the rating agencies are facing several lawsuits for substantially inaccurate ratings.  On the other hand, bond spreads might efficiently impound publicly available information about the transferor, but this information might be incomplete.  However, the case for the bond investors being right is supported by prior research that finds that securitized assets are also relevant in explaining market measures of equity risk.

The authors also find that the type of securitization – residential mortgages, consumer loans, and commercial loans – does not affect how the bond market views the credit risk.  But credit ratings are associated with only the retained portion of residential mortgages.

So, what are the implications for standard setting?  The results from this study (combined with research related to equity risk) might suggest that all transfers of financial assets should be accounted for as secured borrowings instead of sales, regardless of the contractual terms.  However, for those who believe there are significant economic differences across transfers (i.e., some transfers are effectively sales while others are secured borrowings), this might be going too far.  It should be noted that the sample period of this study was 2001-2006.  The FASB issued FAS 166 in 2009 with the expected outcome of reducing the number of transfers that are treated as sales and requiring additional disclosures for investors to properly assess the transactions.  Under FAS 166, the transferor must consider its continuing involvement in the transferred financial asset even if certain agreements were not entered into at the time of the transfer.  An interesting research question is whether FAS 166 has changed the way the bond and equity markets view asset securitizations.

Readers might be interested in this related article printed in Accounting Today (here).