Non-Residential ABS: No Problem... Why?

by: Accrued Interest

An e-mailer asked me about non-residential ABS. This paper has been hit pretty hard, and I agree with the e-mailer who characterized the move as "odd and irrational." Anyway, here is a chart.

The blue line is auto-loans, the green is student loans, and the red is credit cards. All are AAA-rated, floating rate, and 2-years to maturity. The quote is in spread to LIBOR. The source is Merrill Lynch.

We can see that this paper has been trading right around LIBOR flat for years. The next graph goes back to 2000.

Notice no major move in this stuff in response to the 2001-2002 credit crunch. Or 9/11. Or even the 2005 downgrade of everything auto.

So why has this paper moved wider in this credit cycle but not the much worse 2001-2002 cycle?

The obvious theory is that the current period has more to do with consumer credit than 2001-2002. Back then corporate bond spreads moved dramatically wider as investors questioned the accuracy of financial reports. This time it has less to do with corporate health and more to do with poor credit standards in consumer lending.

And I think its logical to conclude that if consumer lending practices were getting sloppy in home loans, why not other loans? Put another way, if CDO demand was so strong as to encourage mortgage lenders to throw caution to the wind, the same effect should have encouraged other lenders to act the same way.

Interestingly, however, it appears this isn't what's happened. I did a little unscientific research into the composition of ABS CDOs in 2007 by looking back at all of the marketing materials I was sent in 2007. I found 10 deals (remember that basically all ABS CDO issuance dried up in June), none of which had a significant exposure to non-residential paper. A few had a bucket for "ABS CDOs" but didn't see any where that bucket was more than 10%.

Econometrics professors reading this blog, please forgive the sloppy statistics here, but I figure if out of 10 deals picked out of a hat, none of them had any non-resi ABS, then its likely that ABS CDO collateral overall was dominated by residential paper. Anyone who's seen hard stats on this, let me know.

Why would CDO collateral be primarily resi ABS? Wouldn't adding other types of consumer loans improve the correlation? Well I'd argue that auto loan, student loan, and credit card paper was too tight to interest CDO managers. Remember that a CDO needs wide spreads. He doesn't need to own a bucket of higher credit quality paper, because s/he figures credit losses can be managed within the structure. And as I've said before, both the managers and the ratings agencies grossly miss-estimated the correlation of residential loans.

A fair concern is that the default rate on all consumer loans is likely to rise. When HPA was very strong, consumers could get out of debt by doing a cash-out refi. We all saw the "consolidate your debts" commercials. Now that HPA's fire has gone out of the universe, the "debt consolidation" option is all but extinct.

So while my broad view is that most ABS has plenty of subordination to withstand a modestly higher default rate, there is a non-zero probability that we've all underestimated how much consumers were using their homes to bail themselves out. I'm personally involved in enough other sectors which have widened similarly which don't suffer from the possibility, however likely, that consumers will be even weaker than I currently think.

One option is FFELP student loan paper, which is U.S. Government guaranteed. I think the rate is about 5-10bps tighter than that student loan paper shown in the graph, but 5-10bps to eliminate credit risk ain't bad.