They claim the presentation calls for 45% losses in 2010 in the CRE space, "worse than the 1990s real estate recession" -- yeah, that is nearly 6 times worse than the 8.1% 10-year loss during the 1990s, and 30-40 times worse than any one year from the early 90s! Something must have been misunderstood here.
Also, people keep talking about IO loans being the problem. Partial IOs are one thing, due to payment resets, but just straight IOs make a for a hard argument that they are going to lead to significantly greater losses. Let's look at an example.
Let's compare an IO loan from 2006 to an Amortizing Loan in 2006. 2006 is more or less the start of the bad underwriting, and we'll imagine we have a loan at 6% per year with a 30-year amortization schedule and a 10-year balloon payment due.
- At Year 3, 2009, the principal paid down through amortization is just 3.92% of the loan balance.
- Year 5, principal returned amounts to 6.95%
- Year 10, but before the balloon, is 16%
At Year 3
Posts
3 comments:
Dark Space, your post disappoints me. First, most insurance and lend to own banks were underwriting with 25 year amortization periods, not 30... But even using your metrics a 7% principal pay down by year 5 is significant when you consider the average severity of loss on a foreclosed property is only in the 20% range. I agree it doesn’t matter on transitional property deals that have 80 to 90 percent downside values, but that makes up a tiny fraction of the universe.
In addition many borrowers would buy a $50 million dollar property by borrowing $35 million from CMBS then another $10 million in mezz financing. That insignificant +7% of amortization translates to a doubling of the equity they have put in.
In some ways we won’t know until post 2020 when the toxic 2007 CRE mortgages have matured, extended, and realized losses… but I’ll bet 2 weeks pay that 2007 underwritten amortizing loans will have fared significantly better.
Sorry to disappoint. I like your stuff - you should post more, I'd be an interested reader.
I realize that 25 years is a typical amortization period on CRE loans, but most partial IOs seem to be longer. Just looking at the 28k loans in Conduit deals issued from 2006 - 2008, the median amortization terms, regardless of amortization type, is 27.5 years. If you break it up by amortization type within that dataset, and look at the average amortization period, the various amortization terms range from an average (by amortization type) of 23.4 Years to 30.6 Years (for Partial IOs).
I would guess, but have no hard data to support, that CMBS has a much higher concentration of Partial IOs when compared to portfolio lenders.
I guess the point I'm trying to make is that Full IOs shouldn't be the "biggest" concern - there are a whole list of concerns that should be higher on the list.
We probably agree more than disagree... but a property that is defaulting while I/O (especially with low rates from 06/07 origination) is going to take a massive loss. An amortizing loan that can't keep up with payments may have an option to restructure, refi, or if foreclosed will on average not be as big of a loss.
It still makes me sick that Fitch, Moody's, and S&P rate a 1.3x dscr I/O loan higher than a 1.2999x 10/10 fully amortizing loan (though the 1.2999x 10/10 deal would likely get upgraded due to the LTV proxies in stressed cap rate tables). Giving amortizing resi and CRE loans a 1 or 2 tick upgrade could have greatly subdued or even prevented the entire capital market crisis we are in.
Post a Comment