How Housing Finance Actually Works

I’ve gotten a bunch of emails this week regarding my piece on the credit outlook for the bonds that make up the 2006 ABX indices, expressing wonderment that nearly 50% the bonds have already been paid down.

“How do almost half of the sub-prime loans get ‘repaid',” asks a typical correspondent. “Are these ‘refinanced'?”

In a word, yes—one way or another. In the real world, the typical 30-year mortgage doesn’t last anything like 30 years. Rather, people sell their houses and move--and pay off their loans with the proceeds from the sales, then take out new loans to finance their new houses. Or, if they’re subprime, they might raise their FICO scores and refinance the existing property on better terms.

Regardless, the way housing finance works, the vast majority of loans are repaid with the proceeds from new loans. In all, the typical 30-year mortgage loan lasts something like seven years or less. (A lot less, depending on the rate environment.) You shouldn’t be surprised that subprime mortgages tend to be repaid even faster.

Repaid = Undefaultable

Anyway, one of my points was that 2006-vinatge subprime loans that have been paid down--$282 billion of the total $600 billion originated--cannot default. It is, as they used to say on the Saturday morning political chat shows, a metaphysical impossibility. Which is one reason I believe eventual cumulative losses on 2006 subprime originations will be materially below what the conventional wisdom seems to believe.

Ah, but skeptical readers come back, all this loan-refinancing mumbo-jumbo means is that the problem isn’t solved, but just being pushed into the 2007 vintage. That’s where things will really get ugly.

Maybe. But that argument has a few problems. First, it implicitly agrees that, well, yes, my analysis is correct and eventual losses from the 2006 vintage won’t be as high as expected. I take that as quite a concession. Thank you.

2007 Originations Much Lower

But there are other problems, too. For starters, there just isn’t as much stuff to blow up among the class of 2007. Subprime originations came to just under $220 billion last year, according to Inside B&C Lending, or only a third of originations in 2006. Granted, the bonds aren’t performing as well as 2006’s (My own analysis, using the same bond-by-bond methodology I used to look at the 2006 vintage, shows an expected 21.2% cumulative loss on 2007 originations.) But even on that higher number, the sheer dollar amount of 2007 losses figures to be manageable assuming that 2006-vintage losses come in as I expect. For the cataclysm the bears expect to occur, both years have to blow up badly.

And, of course, no one expects the problem to be refinanced into 2008.

What’s more, the latest data (which we pull off monthly servicer reports, via the Bloomberg, by the way) seems to show that the 2007 ABX vintage might actually be improving. Repayment rates are rising, for starters. As noted, those bonds won’t default. Also, delinquency-bucket roll rates are falling which means more delinquent loans have cured.

Will the 2007 vintage be worse than 2006? Of course. But for true subprime Armageddon to occur, both years have to implode. On the available evidence, they won’t.

Tom Brown is head of BankStocks.com.

Comments

[ZenInvestor]

This is a great point, and one on which most bankers would agree. I see this as the reason so many are seeing 'buying' opportunities in MER, LEH, ETFC, and others who have taken large write downs and reserves on mortgage paper in general, and should show solid 'upside' surprise potential as their loan portfolios are reduced by early payouts that are natural and not forced by foreclosure.

[rdial54]

Really need the market to rebound in real estate..

[Chris1206]

Great work! I have not seen anyone else give that type of an analysis to the situation.

Isn't it likely though that the 2007 defaults maybe even higher because of the fact that there are fewer sub prime options for that borrower today to refinance into?

[ArnoldCountry]

On the surface this makes sense from a historical perspective. However, I think it has many simplified flaws. The issue is more complext. Traditionally the average life of a loan is somewhere around 4.3 years, could be different for a sub-prime loan but overall that is the number thrown around for years. What is different is, depending on the market, how much value has been lost. Take some parts of California, values have declined by 30% and still declining. That sub-prime or prime mortgage that was done at 80% or more, is done. The lender would have to take a short sale for the consumer to sell it. I would say that your analysis is good on values that orginated in 2006 and 2007 on sub-prime loans that were less than 70% LTV in those areas (similar analysis is similar areas at higher LTV). Adding another wrinkle is the decline in values does not appear to be done. So loss mitigation efforts to give some relief to buy time could blow up again if consumers see that they have lost even more value a year from now. I guess, I find your analysis risky. No one can look to the future at the moment and say that things will be ok. 9 months ago many people did and they were wrong. The historical persepective of markets has a different animal before it and this downturn will profile many future lessons. I guess I am just not as optimistic as this article presents. Consumers/Sellers who move or need to sell that normally do statistically are just unable to sell without huge losses is way too many cases.

[cfish]

This article demonstrates the mentality that got us in this mess.

[Rhett]

What are you, Tom, some kind of realtor or something?

[pochovilla]

FACT OR FICTION?

To get a better Understanding of what Tom Brown is “pitching” here put all three of these articles in your memory bank and than make a more informed conclusion:

Subprime Mortgage Losses: Not Destined to Pitch World into Abyss, After All
Thomas Brown April 22, 2208

www.bankstocks.com/art...


S&P Subprime analysis 2000-2006: Mar 2007 Article
Given these observations, combined with our expectation that the U.S. economy will not experience a recession in the next two years, we believe 7.75% should be at the high end of our subprime loss assumption for the 2006 vintage. Our loss expectation also considers the relatively weak housing market. We also assume stagnant home price appreciation for 2007 and that home prices will return to a more traditional growth rate of 3%-4% after 2007.
If losses for the 2006 subprime vintage reach 7.75%, this vintage will become the worst-performing vintage in recent history, realizing 50% more losses than any other. To gain some perspective about the impact of an economic slowdown on the 2006 vintage, we can look at Detroit's recent economic slide. Over the past five years, Detroit's unemployment rate has risen above 8% and house price appreciation has hovered in the 2%-4% range. In this environment, subprime loans have experienced cumulative losses of approximately 6% (source: Jan. 9, 2007, CSFB Market Tabs). Comparatively, the high end of our 2006 subprime loss expectations for all of the U.S. is higher than the worst-performing microeconomy over the past five years.

www2.standardandpoors....


Subprime Time Bomb Feb 2007 Article

“Despite the headwinds, lenders like Countrywide, which only has 10% of its business in subprime production, could weather the housing slump better than nearly pure subprime companies, like New Century and Novastar”.

www.businessweek.com/b...

THE BIG QUESTION: FACT OR FICTION?

[AvlGuy]

There seems to be a psychology-play afoot here. I''ve heard other blogs cite Mr. Brown's track record ala PonchoVilla's post above.
The history Mr.Brown cites may be correct, but the $million-question is: is it still relevant to the credit situation in 2008? I do not think historic mortgage lifespans or those of the 2006 vintage will apply to the balance of vintage 2006 or to those beyond. And at this point, we know the issue no longer is exclusively subprimes, it's now Libor -pegged ARMs , Alt-A's, and ARM recasts.

[tcornelison]

Don't look for disasters this year from LIBOR ARM resets. Based on the current index most would reset in the 5.5% range. Many of them will go down. Don't forget that historically those with adjustable rate mortgages pay less than those with fixed rate loans because of the 4.3 to 7 year average loan life listed above. I personally haven't had a fixed rate loan in the last 14 years and have always paid below market rates.

Prime and Alt-A products do not have the punitive margins which cause subprime defaults at reset. A larger issue is the Alt-A loans which were issued as I/O loans with I/O periods which match their first reset. A lot of 5 year interest only ARMs go to full amortization at the same time they can see their first increase in rate. As long as the Fed keeps rates low the indices will stay low and the impact from resets is eliminated except for subprime loans with 6-8% margins. The problems for these loans come when we return to the real world and the fed starts adjusting rates upward to control inflation.

I believe that Tom's analysis holds a lot of truth but home price declines are a big concern. In reality large home price declines are confined to a few markets and the majority of the nation will see a 0-10% adjustment. Investors who speculated in Florida, California, Arizona & a some other areas are being and will continue to be devastated by the fall back to reality. Homeowners who bought and live in those areas are the real victims in this situation. The downturn in the economy which has been agravated by this mess will continue to hurt other markets like Cleveland and Detroit but the rest of us will come through with most of our equity intact.

By the way, the reason Countrywide did not follow New Century and Novastar into the abyss is stated above but the 10% figure may be a little low.

[Malkiel]

Tom's analysis of loan lifespans is important because there's way too much being written (especially invective against sellers) without a perspective on how real estate transactions really work--but ArnoldCountry above makes an important reference to the problem of the upside-down seller in bubble markets. I've commented elsewhere that sellers who are too upside-down on their mortgages (meaning they lack the cash to pay off their old mortgage; in typical housing transactions the new lender makes payoff of the old loan a precondition and expects it to happen at the closing table simultaneously with the activation of the new loan) have no viable strategy other than to list their house at a noncompetitive price and hope to outlive the trend. Short sales are coming into the mix but I'm not sure many lenders are giving in to those. We end up with a body of "hidden" troubled loans whose outcome is suspended in time; they don't resolve until the holders give in to foreclosure or outlast the trend. In bubble markets, where some homes will never recover their inflated value, there may be a hump of foreclosures 2 or 3 years beyond the expected reset dates for the known problem loans...

[billddrummer]

Arnoldcountry makes an excellent point about upside-down homeowners in bubble markets. As home values drop further, many of those homeowners (or mortgage renters, as I like to call them) will be unable to refinance or sell because their homes will have fallen in value beyond the ability of a new lender to finance them. The homeowners at greatest risk for this purchased their homes at the peak of the bubble, during 2005-2006. (Not surprisingly, the period when lending standards were loosest.) With the more exotic mortgage products gone, the only altenative for these people is to stay in their homes and continue to make their payments. When prices finally hit bottom in the bubble areas, inventory comes down to a balanced level, and appreciation begins to resume, it may be 5 or more years from now before those homeowners see real equity in their properties. It may take even longer if they purchased with little or no money down. Then, look for them to stay permanently, with their first home and their climax home being the same address.

[Pent up demand]

Tom's analysis falls apart when home prices decline by 20-30%. What happens to your refinancing when your formerly 80% LTV loan is now a 95% LTV? Two years ago that was no problem mon. Not so much anymore. And I don't care what your FICO score is.

The loans that have paid off are the people who made it to the exit in time. It will be interesting to see what becomes of the members of the "class of 2006" who didn't/couldn't refinance.

[paularush]

Once again a moron speaks! I just wish financial types with their heads in the sand, would leave them there and shut up. Misleading investors into thinking everything is fine, so they will continue to give money to idiots to mismanage, is outrageous. What isn't mentioned here is the fact that many non subprime loans are defaulting, and no mention of the NEUTRON BOMB loans, the pay option arms, negatively amortizing crap. These loans feature a payment of 1%-2% of the actual loan payment, with the rest negatively amortizing into the balance. So thanks to GAAP the people holding this crap can claim they are collecting full payments when they are not. In the end the "phantom profits" will disappear as loans recast to 110% LTV (Countrywide vintage) or 125% LTV ( American Home Mortgage vintage) at a time when home values are plummeting. American Home did about 65% of their business in these loans in many of the worst hit markets, CA, FL, etc.
Also, the mortgage insurance, all these junk loans are counting on to bail the investors out, is about to implode as insurers and credit default swap partners are fighting back. In the case of American Home Mortgage, TRIAD is in the middle of a battle not to pay claims, and Bank of American refused to pay out on a credit default swap on AHM Broadhollow and Melville LLC loans. You can read all about it in the legal filings on EPIQ. Have you looked at the Mortgage insurers stocks. TRIAD, RADIAN, PMI, MGIC, lately? Can they weather the avalanche of claims coming? YSo the 25% to 30% insurances which lenders locked in to insulate the portfolio, may be disappearing, the home values are certainly plummeting, and the loan balances are rising. A recipe for disasters yet to come. When will all this implode? Well vintage 2007 is some of the worst loans ever, and pay option arms take three years to recast in some cases, and payments will then quadruple on homes not worth what is owed on them. What do you think will happen then? The worst is over. I don't think so! Don't let the industry fool you.

[P2P-Loans.com]

Subprime is only part of the problem. We (America) have a huge asset class (trillions of dollars) that is incredibly overleveraged. And, that category is currently in a nuclear winter (sales plumetting, values following suit, >25% declines in a number of very large markets - SO FAR). Thus, if you leverage an asset 9x - 10x to 1 (sometimes more) and it declines in value in any material way, the equity is wiped out and the loans take a bath. This is bad for consumers and really bad for lenders who stand to lose 40% or more when they foreclose. Also, it's not just about subprimes. While this article from Money.com is largely anecdotal, there is a lot of truth to it:

money.cnn.com/2008/04/...

I simply don't agree with Tom's analysis here - there are more shoes to drop in my estimation.