Friday, August 21, 2009


A couple commentors mentioned upcoming option ARM resets as a reason to be more bearish on housing. I wanted to make some quick comments about that.

First let me say that option ARM resets are a tricky thing. Most ARMs were underwritten in 2005 to early 2007. During 2005, 12-month LIBOR averaged 4%, during 2006 5.33% and in 1H 2007 it averaged 5.33% again. So the reset itself isn't a worry at all.

Its the recast that matters. The switch from the "Option" period to the full amortization period. The tricky part there is that most option ARMs have a 5-year "Option" period. That would mean that 2010 would be a big year for recasts, and 2011 even bigger. We wouldn't "burn out" on these things until mid 2012.

However, most option ARMs also have a provision where if the negative amortization gets to 15% (i.e., you owe 115% of the outstanding balance), it automatically reverts to the fully amortizing amount. Certainly if a borrower has managed to fall that far behind within the first couple years, the odds that that loan winds up in foreclosure is pretty high. Anyway, it throws off the theory that there is suddenly going to be a bunch of recasts in 2010 and 2011. Many of those loans are already recasting because of the neg-am element.

Here is how I see it. I track a large group of whole loan "prime" mortgage securitizations. My group is from 2006 and about half are Option ARMs, the rest are full amortization ARMs. Most are also limited documentation. So while the credit score said prime, everything else about the loan said "questionable." I created this grouping back in 2007 to track how seemingly good borrowers who took out bad loans performed. Its a pretty good gauge of how Option ARMs are doing.

The 90+ delinquency (which includes foreclosures) is currently 11.5% of the original balance for the whole group. The figure continues to climb month-by-month but the pace has slowed considerable. The last 5 months its increased by about 0.3% per month, vs. over 1%/mo. during most of 2008. So that's point 1, that the pace is clearly declining. Worth noting that the pools with mostly option ARMs are about 2.5 times the delinquency level of the full amortization loans.

Second, I track a similar group of sub-prime loans. That series has completely burned out, with the 90+ figure sitting at 17.5% for the last six-months. So basically what's going to happen in sub-prime has already happened.

Put these two together and you could conclude that the continuing rise in price foreclosures is just making up for the lack of rising sub-prime foreclosures. I don't know if the math of that exactly works, and I am sure that the sub-prime and prime loans were not typically in the same neighborhoods, but point is that a lack of new sub-prime foreclosures is at least something of an off-set.

If you drill down into some specific Option ARM deals, you find some very interesting info. Here are some stats on one of the deals in my list.

Pool Factor: 63.2% (meaning 36.8% has paid off)
# of Loans: 496
WALTV: 83.5% (so few loans are being forcibly recast)
90+ Delinq: 39.5%

Now here is what's interesting to me. 37% of this loan has paid off. Another 40% is not paying. That means the potential new problems are only 23% of the remaining principal. I pull several other deals with the same kind of circumstances. What we're seeing here is that the loans that never should have been made are already turning bad (the 40%). The loans that were made to actual good borrowers are paying off rapidly. What remains in between isn't a very large number.

That's the facts as I see them. I'd love to hear the other side. Just post a comment!


Anonymous said...

when people start talking about how great is the market and credit cards start throwing money with no interest, it is time for the down move. I am selling calls(way out of money calls!), with free money from credit cards.

Anonymous said...

most (as of 07) negam loans were minimum pays from the start. the option was really a misnomer. the whole point was to have the smallest monthly payment possible. with minpay, loans recast in 2-3 years depending on the negam ceiling. they did not "fall behind" since minpay was the overwhemlming payment choice for OA borrowers. the idea was ALWAYS to have a low payment that rolled out into another loan. that is what you are seeing in your portfolio. the good got out and the people left are the ones who couldnt escape.

it will depend on who originated the product, but chances are the borrowers were not qualified to the fully am payment so its not that shocking.

Anonymous said...

Examples of option-ARM payment shocks, even with low interest rates:

In Debt We Trust said...

All this talk of option ARMs got me thinking about student loans. Like option ARMs, borrowers can choose to pay the interest only. But w/job prospects for recent grads declining, this does not bode well for consumer spending.

SD Scientist said...

What we're seeing here is that the loans that never should have been made are already turning bad (the 40%). The loans that were made to actual good borrowers are paying off rapidly. What remains in between isn't a very large number.

I fail to see why you'd identify paid-off 37% as "good borrowers". A better description is, 37% of the original pool has been foreclosed on or short-sold, 40% aren't paying and will join the 37% eventually, and 23% are either good borrowers or too stupid to walk away.

In Debt We Trust said...

I found this interesting comment on zerohedge:

So, in a nutshell, Bernanke has been slowly withdrawing dollars from foreign central banks - forcing the other central banks to sell dollars and buy back their own currency. Selling dollars forces up commodities (since all commodities are traded in dollars). And the Open Market Operations monetizing debt has the same effect (long debt = short dollars).

So the effect of these two activities at the same time has been very bearish for the dollar and bullish for commodities and stocks.

The question is, where are we now with respect to Liquidity Swap levels and OMO.

Accrued Interest said...

37% of the principal has been returned, so that's a real honest to God pay off. Anyway, my point isn't to really say they are "good" borrowers so much as to say they are out of the Option ARM loan, so whether it was a short sale or what, it still can't create marginal problems.

Accrued Interest said...

Anon: Barry and I are talking about the same thing. The interest rate portion of ARM resets aren't a problem. Its the recast into full amort. I just think more of the problem has occured already than what people think, actually for the same reasons Barry mentions. As soon as the loan gets to 115% LTV, it recasts immediately. If a borrower has been making minimum payments all along, s/he'd hit that 115% number quickly. Long before the official "recast." If the borrower were making the full payment anyway (supposedly a small number, but whatever) then that borrower isn't a problem anyway.

Anonymous said...

So let's see if I got this straight.. the 37% who are 90+ days is 37% of the amount that's left, not the original 100% with ARM's. Did the 40% who paid off get a new deal in the mushroom cloud that was more of the same house of cards or not? Does 90+ at 37% mean that even 23% of the original deal is going to default in part or whole (closer to whole cos the borrower couldn't afford in the first place, let alone the ballooning interest on reset). Aren't the guys who got out already in another ARM because, with house prices already well through the 15% reset point on a market to market basis? Weren't the original pools 90% prime and 10% aa to equity? So even at 23% of original loan, isn't the AAA holder impaired by 13% already? Are these going into PPIP too? Do you think this is really representative of the entire market? Is the deal to assume that the principal when you buy the pool will only repay 93 cents in the dollar? What is the correct price of this based on the risk of contamination from the 40% who migrated to better deals on other ARMS? (assuming they did). Why on earth would someone even want an ARM if the could afford a repayment mortgage? Sounds like "please kick me in the nuts, cos I like it...later"!

Wriiight said...

Weren't there a lot of 2/28s with a teaser rate, so there will be a rate spike for many borrowers, even though the underlying index has moved in their favor?

SD Scientist said...

37% of the principal has been returned, so that's a real honest to God pay off.

What about losses/write-offs suffered because of foreclosures? If the original loan was foreclosed and the property was sold for 50% of the original loan value (not uncommon in some areas), how is this reflected in your numbers?

How can anyone possibly pay off an option ARM? For every person who pulls off an upside-down refi or brings a few hundred grand to the table, there are probably ten or twenty who walk away.

Anonymous said...

It is not just the 90+ delinquency rate. Take a hard look at this report on the defaults that are going into foreclosure. Prime Cure rates are 6% and Alt A cure rates are 4%.

This bodes ominously for all.

What we are witnessing is a complete change in social and economic behavior with individuals recognizing that they can stay in their home for 12-24 months without making a payment or being foreclosed.

The only reason the consumer has some spendable income is because he quit making his house payment and is living rent free.

195 loans in your delinquent pool out of the remaining 311 loans (496-182) is 62.7% in arrears.

Accrued Interest said...


What I'm saying is if you owned $100,000 of this bond, $37,000 in principal has been repaid. Cash in your pocket. So let's say that the whole $37,000 was short sales for 50% of the original loan amount. Then really 74% of the deal would have been "resolved" with 37% in principal returned and another 37% written off.

I don't want to sound too rosy. I just want people to look beyond the first glance. I think Anon from 8/21 6:19PM has it right. The Option ARMs were a shitty product, we know the losses are going to be huge. The question is are the problems "coming" or are they "here." Because if its the later, the marginal increase in foreclosures will be small, and thus the impact on home prices will be small.

BR said...

There was a big discussion on this over at Matt Padilla's blog -- the summary seemed to be that neg-am has halted with the low interest rates of today (see comments by a mortgage broker "Dealtracker" and the counter-arguments by Liem@May22}. Sort of similar to what AI is saying -- that the bad problems are already out there.

BR said...

More here:

Anonymous said...


For the benefit of your ignorant readers, can you please explain the significance of LIBOR rates during the 2005-2007that you referenced at the beginning of your post?


SD Scientist said...

Is there a way to tell how many loans are really not there any more (as opposed to knowing the percentage of principal returned)? Is the delinquency rate the percentage of all loans there were to begin with, or the percentage of all loans that remain?

It seems that your numbers can be interpreted in the following way. 60% of all loans were foreclosed/short sold with the average pay-back of 62% of original loan value (60% * 62% = 37%). All remaining loans are 90+ days delinquent by now. In other words, there are NO paying loans left in the whole package!

Anonymous said...

Its a very useful discussion. Such discussions allow us to see through the fog and mirror more clearly.

However, here are some questions:
1. What about 30+ defaults? What is the pace of the growth/decline of these loans?
2. What's the percentage of 60+ defaults? What the pace of the growth/decline of these loans?
3. How representative is your sample?
4. Do you discuss your numbers with your colleagues working on different pools? Do their numbers corroborate your conclusions?
5. What are the "famous" charts of the reset/recast based on? Are they using the nominal figures when they were originated without deleting the already paid off and/or recast loans? Doesn't Credit Suisse (and others) who published the "famous" charts track their own internal pools?
6. Would a proper phrase to define the "coming tsunami of recasts" less as what it will do to these particular mortgage holders and more to with the effect of even 2-3 foreclosures affecting the equity of the neighborhood, thus lowering the paper wealth effect of the area residents or bringing more negative equity for the neighbors?


Anonymous said...

Very interesting post.

Agreeing that Recasts and not resets are the issue now (though far enough out if Bernanke starts withdrawing liquidity resets may become an issue again - those loans will still be underwater 5 years from now I'd bet), and taking your numbers straight, Those 40% may still be a current problem.

How? If banks know the loan is no good, and thats marked, but they haven't dumped the OREO inventory yet as they are praying for a bounce.

I don't have the data to know if that's the case. But I could see a situation where your numbers were correct, and still in keeping with a flood of depressed inventory into the market in the near future.

Be nice to see good stats on the various forms of shadow inventory out there.

Anonymous said...

электронная почта без регистрации