For any of you who haven’t gone there, the Calculated Risk Blog is a great resource for info related to housing and mortgages.

At a recent visit, I came across this great post which I will quote in it’s entirety.

Tanta: Credit Suisse "Not drinking the Kool Aid"
CR Note: To put the title into context, last December during the Toll Brothers conference call, Ivy Zelman asked CEO Bob Toll: "Which Kool-aid are you drinking?"From Tanta:

Ivy Zelman and several colleagues at Credit Suisse have put together a very big and highly detailed equity research report called “Mortgage Liquidity du Jour: Underestimated No More,” which unfortunately is not available on the free internet. Although the title may suggest otherwise, the report is not really about mortgage credit issues; it is an attempt to look at the rapidly deteriorating conditions in the mortgage market for their impact on homebuilders, which are the equities Zelman covers. (Sure, CR basically has already done exactly this analysis, but Zelman has the advantage of access to databases and CSFB’s mortgage industry analysts.) The upshot is that Zelman is not drinking the kool aid that the problems are limited to subprime only, or entry-level housing only, or are going to be over any time soon.

First, the report looks at overall recent trends in the large categories of general credit quality (prime, Alt-A, and subprime):The overall share of prime conventional loans has declined from an estimated 66% of total purchase dollar originations in 2002 to 45% last year. The GSEs’ share loss has been largely attributed to the proliferation of “exotic” mortgage products such as high CLTV loans, low/no documentation mortgages and interest-only/negative amortization loans, which the GSEs have typically chosen to limit their exposure to given the high risk profiles of these products.

[T]he Alt-A market has expanded from just 5% of total originations in 2002 to approximately 20% in 2006. Although the credit profile of Alt-A borrowers is stronger than that of the subprime market (717 average FICO score for Alt-A borrowers versus 646 for subprime), we believe that there is considerable risk associated with the lax underwriting standards and exotic mortgage products utilized in this segment of the market in recent years, both in the form of continued credit deterioration and reduced incremental demand resulting from tightening lending standards.

• The combined loan to value on Alt-A purchase originations was 88% in 2006, with 55% of homebuyers taking out simultaneous seconds (piggybacks) at the time of purchase.

• Low/no documentation loans (stated income loans) represented a staggering 81% of total Alt-A purchase originations in 2006, up significantly from 64% just two years earlier. . . .

• Interest only and option ARM loans represented approximately 62% of Alt-A purchase originations in 2006.

• . . . 1-year hybrid ARMs represented approximately 28% of Alt-A purchase originations in 2006, setting the stage for considerable reset risk.

• Investors and second home buyers represented 22% of Alt-A purchase originations last year, which is the largest non-owner occupied share among the various segments of the mortgage market.

In the past five years, subprime purchase originations have more than doubled in share to approximately 20% of the total in 2006. Over this time period, subprime lenders eased underwriting standards in an effort to gain market share. . . . . In the third quarter of 2006, the Mortgage Bankers’ Association reported that 12.6% of subprime loans were delinquent.

• 2006 subprime purchase originations posted an alarming 94% combined loan-to-value, on an average loan price of nearly $200,000.

• Roughly 50% of all subprime borrowers in the past two years have provided limited documentation regarding their incomes.

• In 2006, 2/28 ARMs represented roughly 78% of all subprime purchase originations according to data from Loan Performance. According to our contacts, homebuyers were primarily qualified at the introductory teaser rate rather than the fully amortizing rate, which for many buyers was the main reason they were even qualified in the first place.

The analysis continues by drilling down into the question of “exotic” mortgage products within and across those general credit categories:Based on data from SMR Research, approximately 40% of home purchase mortgages in 2006 involved piggyback loans (through the third quarter), compared to 20% in 2001. . . . Based on a survey of our builder contacts, their average combined loan-to value ratio on home sales in 2006 was 91%, with 49% of homebuyers taking out a simultaneous second mortgage at the time of purchase. . . .

A misconception that we commonly hear is that the growth in piggybacks has generally been isolated to the subprime arena. While more than half of all subprime mortgages had a simultaneous second mortgage associated with them, Alt-A and jumbo loans have seen similar growth in piggyback prevalence in recent years. In fact, 55% of securitized Alt-A mortgages in 2006 had simultaneous seconds attached to them. . . .

An estimated 23% of total purchase originations in 2006 were interest only or negative amortization mortgages. Similarly, according to our private builder survey, interest only and option ARMs represented 24% of new home sales in 2006.

Low/no documentation loans increased from just 18% of total purchase originations in 2001 to 49% in 2006 according to Loan Performance. Based on a survey of our private homebuilders, the percentage of buyers providing limited-to-no documentation was

similar on the new construction side of the business to the overall market, at 46% in 2006. . . .

The report then summarizes the current situation of guideline tightening and regulatory changes that are contracting the mortgage market (a familiar list of items to CR readers). It concludes that

[T]ightening liquidity puts current builder backlogs at considerable risk for fallout, which should lead to another surge in cancellations and additional spec inventory on the market. We are already hearing anecdotes from builders in California, Florida, Nevada and Texas of buyers in backlog being unable to obtain financing because their loan program is no longer being offered by the lender (or the lending requirements have changed), which could lead to the next tranche down in pricing.Moving beyond changes in the origination market, the report looks at issues of performance of the current mortgage book for its further impact on new home sales:We estimate that there are approximately 565,000 homes in the foreclosure process around the country that have the potential to be added to inventory within the next two to six months in the form of an REO, and another 135,000 that are already listed or on the verge of being listed as “must-sells.”

To put this into perspective, the National Association of Realtors reported existing inventory of 3.55 million units in January, implying that total inventory may be 20% understated when taking foreclosures into account. . . .

Rising subprime and Alt-A delinquency rates will likely keep foreclosure levels elevated for the foreseeable future. Subprime 60+ day delinquencies and foreclosure rates for 2006 vintages are running more than 3 times the levels form 2004 vintages given the sharp downturn in home prices and underwriting standards that continued to ease through much of the year. Delinquencies of 90 days or more, foreclosure and REO rates on 2006 vintage Alt-A ARMs are running 3 to 4 times above the levels from 2003 and 2004 vintages.

Roughly $300 billion of securitized subprime mortgages (36% of outstanding subprime MBS) are set to reset in 2007 alone, with $500 billion in total mortgage debt (6% of outstanding) scheduled to reset during the year.

Finally, the report turns consideration of this data into a set of projections of the effect on new and existing home sales:

In our base case, we assume that 50% of the subprime market is at risk, taking originations back to 2003 levels, which would impact total purchase volume by 10%. Similarly, we estimate that 25% of Alt-A and 10% of prime loans would not be approved under tighter restrictions for various combinations of investor purchases, piggybacks, low down payments and low documentation, and the impending ripple effect down the entire housing market food chain. In aggregate, the total fallout of incremental originations would be 21% over the next one-to-two years.

Related to speculation, investors’ share of the market climbed to roughly 18% in 2005 and 2006 from an average of 7% from 1998-2001, implying that a return to the mean would remove 11% of housing demand.

Combining the two yields a 25-35% reduction in peak housing production. This would likely be exacerbated by declining consumer confidence, investor demand falling below historical norms, the risk of a softening economy and supply pressures weighing on demand (all of which seem present today), suggesting at least a further 10% drop. Aggregating the various impacts would result in a 35-45% drop-off in new starts from the peak of 2.1 million homes to roughly 1.2-1.4 million, as compared to the 16% decrease thus far on a trailing twelve month basis. For comparison, starts during the last three downturns ending in 1991 (down 34%), 1982 (down 32%) and 1980 (down 37%) fell by an average of 34%.

Expressed differently, if we assume that the full impact of mortgage lending tightening will be felt in 2007, all else equal, we would expect new home sales to fall roughly 20% from December’s seasonally adjusted rate of 1.123 million to an annual rate of 887,000 homes (236,000 reduction from tightening lending standards). . . .Click on chart for larger image.

This chart shows the derivation of the 21% drop in new purchase originations.

I have been meaning to comment on this post since it was posted. I think CR is great and the fact that Tanta posts about the inside of the MBS area is the best and love her for it. Getting info from the underbelly in an industry which tries so hard to keep that info away from the public makes her posts priceless. So I will add one minor thing that is not that bad but I will add a lot more for the ugly.
First thing the 28% 1 year hybrid ARM isn’t that scary. They have been priced accordingly to the risk involved for about two years. In other words they have higher interest rates compared to other mortgages. Since they are based on the six month libor and the start rates were higher than margin plus libor if anything they would see a rate reduction rather than increase upon reset.
Now for the ugly. Homebuilders in socal had a much higher rate than 40% for piggy back loans. I am not exaggerating in any way what so ever but it is more like 90%. We had about seven loan officers when I worked for the builder and if any of us got a deal in which someone had 20% down or more we made an announcement and everyone knew about it. Yes it was so uncommon that we had to tell people we had somebody with 20% down. We had buyers with the 20% down but why would they do that when they could "qualify" for more of a mortgage and have that money for savings. I would say 1 in 10 move up buyers with a down would still do 100% financing. It didn’t matter if these were financially savy people with master’s degrees or blue collar workers with a high school diplomas. The percentage of stated income borrowers were about the same as well since you could get them "qualified" without having to provide income. Why would you bother to burden the borrower with income documentation. National numbers do not matter on a macro level and this will be evident in the next few years. Being an order taker is not fun and is one of the many reasons why I got out.