July 2008 Quarterly Forecast
> Housing
By Ryan Ratcliff
By most measurements, the Bay Area was the last of California’s housing markets to peak. Mostly, this is because it was the last market to get going: in the aftermath of the tech bust, the Bay Area actually saw a small dose of home price depreciation in late 2001. After some ups and downs in 2002, it was only in late 2003 that the Bay Area experienced the dizzying acceleration of price growth that was sweeping the rest of the state. Examining the path of Bay Area home sales and home price appreciation (Figures 2 and 3) suggest that the housing market decline in the Bay Area has proceeded in three phases.
Figure 1: Year-over-Year Change in Median Sales Price of All Homes by Region
Source: Dataquick, UCLA Anderson Forecast
Figure 2: Year-over-Year Change in Median Sales Price of All Homes by County
Source: Dataquick, UCLA Anderson Forecast
Figure 3: Total Homes Sales by County (SA)
Source: Dataquick, UCLA Anderson Forecast
Phase 1: A Typical Housing Correction?
In the first phase, from the peak in summer of 2005 through 2006Q3, the year-over-year growth rates in median sales prices of all homes in the Bay Area dropped from over 20% to close to 0% by 2006Q3, with home prices essentially treading water around this 0% appreciation through the summer of 2007.
This is exactly what we expected to see in the Bay Area, for two reasons. First, while the economy slowed a bit in the latter half of this first phase, there was not major across-the-board job loss. The infrequent declines in resale home prices that we’ve seen in the past have always occurred in the wake of major job loss, like the early 1990s in Southern California, or the brief spate of depreciation in the Bay Area in 2001. In the absence of being forced to sell by job loss, owner-sellers preferred to just wait the slump in prices out: resale sales volumes collapse, but prices just go flat.
Of course, Figure 1 shows that the Bay Area’s experience was hardly universal: over this same Phase 1 period, median sales prices in the Central Valley were declining at between -5% to -10% a year, while Southern California’s median home prices were still growing at 5% a year thanks to LA County. Figure 4 offers some insight into these differences: in this first phase, home prices were weaker in markets where new homes were a bigger share of total sales. As we’ve argued before, a builder does not have the luxury of waiting out a weak market the way an owner can: they have to move inventory, and in a weak market that means cutting price. So in markets like Sacramento and Yolo Counties, where new homes represented 26% and 40% of total sales, the aggressive price cutting on new homes provides a double whammy to average house prices. There’s the direct effect of having 30% of the sales prices in the sample fall, and then there’s the secondary effect that requires owners who do want to sell existing homes to price to compete against these cheaper new homes.
There has been virtually no growth in non-farm employment in the past three months. The principal culprits creating the drag in payrolls are construction and finance, and the housing sector related occupations, losing almost 120,000 jobs in the first quarter. With the exception of Retail; Trade, Transportation and Warehousing employment growth has slowed, but remained positive. Retail employment is down due in part to weak auto and department store sales. Manufacturing is still losing jobs, but at a rate no greater than in the boom days of 2004/2005. If we focus only on those sectors losing payroll employment, over 75% of the job loss is in our culprit sectors.
Figure 4: New Home Sales vs Median Sales Price in Phase 1
Source: Dataquick, UCLA Anderson Forecast
This is the second reason why we expected to see flat prices in the Bay Area: new homes accounted for 10% or less of sales in most Bay Area counties in 2005 – among the lowest levels in the state. However, Contra Costa County is the exception that proves the rule: new homes represented 20% of 2005’s total sales in Contra Costa, and Figures 2 and 4 demonstrate that Contra Costa had the weakest home prices of any county in the Bay Area during this first phase of the Bay Area housing slump.
At the end of Phase 1 in late 2006 and early 2007, sales volumes were starting to look like they had found a bottom, and home prices looked like they were settling in for a long period of stagnation – a scenario we’ve often seen in the historical record. But then the bottom fell out: sales volumes hit new record lows, and median prices in the existing home market started plummeting again. Even markets like San Mateo and San Francisco Counties, which we thought would be relatively insulated from price declines, are now down over 15% from their respective peak median sales prices as of June 2008. This second precipitous drop in home prices marks the beginning of Phase 2, which began in the early summer of 2007 and lasted through 2008Q1.
Phase 2: The Foreclosure Storm
The unique feature of Phase 2 is the emergence of substantial declines in resale home prices, leading to a second round of overall home price weakness. While the California economy has avoided recession caliber job losses so far, we have seen an explosion of foreclosure activity in the state that dwarfs the 1990s. This surge of foreclosures/distress sales in the absence of substantial job loss is historically unprecedented: now, widespread mortgage defaults are occurring not because of layoff-driven income losses, but because the slowdown in price appreciation shut off the only escape route from mortgages that many homeowners simply had no hope of ever servicing. We always knew that the cocktail of lax underwriting, low down payments, and exotic high-leverage mortgages that did not require even full interest payments would inevitably lead to some increase in foreclosures when the market turned. But not even the most rabid housing bears anticipated that these problems were so widespread that foreclosure rates would surpass the records set during some of the most severe recessions in California history.
Figure 5: Trustee Deeds of Sale per 10,000 Households
Source: Dataquick, US Census, UCLA Anderson Forecast
Figure 5 shows the number of Trustee Deeds of Sale filed per 10,000 households in 2008Q1. This number measures the intensity of the local foreclosure problem: Trustee Deeds of Sales mark the actual repossession of a home by the bank, and the scaling by household controls for population instead of letting LA County dwarf the rest of the state. While almost every county in California saw a massive increase in foreclosure filings in 2007, the worst of the problem once again seems concentrated in the inland areas.
Figure 6: Notices of Default Filed per 10,000 Households
Source: Dataquick, US Census, UCLA Anderson Forecast
Where does the Bay Area fit within the context of these larger trends? For the most part, the rise in Bay Area foreclosure activity has been among the mildest in the state. The intensity of mortgage defaults (Figure 6: Notices of Default per 10,000 households) in San Francisco County is actually still below the record set in the 1990s. In San Mateo, Santa Clara, and Alameda Counties, the increase in defaults has been more severe, but is still fairly mild when compared to the rest of the state. The exception is Contra Costa County, where the surge in defaults and foreclosures is more akin to the hardest hit inland areas. Of course, even within the counties, the foreclosure experience is far from uniform. Figure 7 presents Zillow.com’s estimates of the share of buyers in the Bay Area who purchased homes between 2005 and 2008 in a particular ZIP code and now have negative equity (i.e. they owe more on the house than its current estimated value). These concentrations of negative equity are highly but not perfectly correlated with foreclosure activity. Much of the negative equity is concentrated in more affordable areas within these regions: east Contra Costa County, Oakland, Hayward, South San Francisco, etc. Nevertheless, Contra Costa once again stands out relative to the rest of the Bay Area.
Figure 5: Trustee Deeds of Sale per 10,000 Households
Source: Zillow 2008Q1 Home Value Report
Why is Contra Costa County having so much more trouble with foreclosures than the rest of the Bay Area? As we’ve argued before, the common denominator among the foreclosure hotspots seems to be the widespread use of adjustable rate mortgages to purchase relatively modestly priced homes. High usage of ARMs alone was not a risk factor: in 2005, more than 80% of purchase mortgages in the Bay Area were some sort of ARM, as even relatively wealthy families stretched to afford some of the most expensive homes in the US. And not just any ARMs: interest only and negative amortization loans (which allowed borrowers to pay less than the full amount of interest due every month) were especially popular in the Bay Area. According to Loan Performance, 40% of purchase originations in 2006 in the Bay Area were some sort of interest only or negative amortization loan, compared to the national average of 23%. But only Contra Costa combined the highest usage of ARMs in the Bay Area (85%) with some of the most affordable homes in the nine-county region. Across California, this combination of using the riskiest loans to stretch to buy the cheapest homes has proven to be recipe for the worst foreclosure problems.
Figure 8: 2006 Share of Interest Only and Neg. Amortization Purchase Mortgages
Source: Credit Suisse
Unfortunately, the hardest hit areas are trapped in a feedback loop: high levels of foreclosures depress resale prices, which push more homeowners into a negative equity position where foreclosure becomes the only way out, creating even more distress sales. According to estimates by Dataquick, resales of foreclosed properties accounted for 33% of all resale activity in California last quarter, and up to 70% in some of the hardest hit counties in the Central Valley. Figure 9 is a primitive attempt to gauge the impact of foreclosures on California county resale prices more directly. The horizontal axis is our approximation of this Dataquick metric: we compare the number of trustee deeds filed in a county in 2008Q1, divided by the number of existing single-family homes sold in the same period. This is an imperfect indicator of the importance of foreclosure sales in the resale market: since the trustee deed represents the transfer of the home to the lien holder, it’s an inconsistent leading indicator of future distress sales, since it takes time for the repossessed property to be sold in the resale market. Nonetheless, this measure is consistent with our previous story about foreclosures: high foreclosure markets like San Joaquin County show that trustee deed filings were 200% higher than resale volume in 2008Q1, while a low foreclosure area like San Francisco County comes in at only 23%. The vertical axis of Figure 9 is the percentage change in resale single-family home prices from 2007Q1 to 2008Q1: as we’d expect, the markets where foreclosure activity is high relative to sales are the markets where resale prices are weakest. Contra Costa County is highlighted in the Figure 9: with foreclosures in 2008Q1 representing 135% of total resales in 2008Q1, an above average level of foreclosures sales are a major factor in Contra Costa’s above average price weakness.
Phase 3: A Flicker of Light in a Dark Tunnel (Summer 2008)
The feedback loop between foreclosures and price declines has made Phase 2 of the real estate bust feel a lot like free fall, with no end in sight. However, the past couple of months have actually seen a small dose of good news: sales volumes in Southern California and the Bay Area are actually rising again. Several factors have combined to produce this small dose of upward momentum: steep price declines, lower interest rates, and an easing of the credit crunch may now be bringing bargain hunting buyers back into the market. Ironically, the areas with the steepest price declines are the areas with the biggest jump in sales: for instance, Contra Costa County and Riverside County have seen sales volumes in 2008Q2 higher than in 2007Q2.
It may be a little early to be celebrating the beginnings of Phase 3 – perhaps we’re better off thinking of this as Phase 2.5. Obviously, the driving force in California housing markets in the near term will continue to be the shadow cast by a glut of foreclosed properties on the market. So the first question we have to answer is how much longer the foreclosure problem will last. In part, the answer depends on the rest of the economy, both nationally and locally. If our forecast that California will avoid major job loss in 2008 holds up, foreclosures will primarily be an issue of working through the unviable low down payment loans made in the last days of the boom. Two issues are relevant to this question: when did the spigot of bad loans get turned off, and how long on average does it take for an unviable loan to go bad?
The Fed’s Senior Loan Officer Survey gives some insight into the first issue. The relevant question from the survey is “Over the past three months, how have your bank's credit standards for approving applications from individuals for mortgage loans to purchase homes changed?” Figure 10 shows the net percentage of respondents who replied that they have tightened lending standards (the break in the series at the end reflects a change to the questions in 2007Q2). 2007Q1 is where we see this percentage start to rise towards its current highs, giving us a reasonable estimate of the end of the funny money party.
Figure 10: Fed Sr. Loan Officer Survey: Net Share of Respondents Tightening Mortgage Lending Standards
Note: In 2007Q2, the survey was changed to disaggregate loans of different credit quality.
As for the second issue, Dataquick reports that most of the loans that defaulted in 2008Q1 in California were originated between August 2005 and October 2006, with a median age of just under two years – suspiciously close to the reset date of many of the 2/28 subprime ARMs so popular during the boom. Figure 11 shows Credit Suisse’s estimates of the volume of ARM’s that will be resetting nationally, starting in January of 2007. Since we’re now at 19 month mark, we’re still near the peak of resets – but on the back side. Of course, resets are only part of the foreclosure problem. But combine the Fed survey, the Dataquick estimates, and this reset pipeline, and a dim light starts to emerge at the end of the tunnel. If the end of 2006 was approximately the last bulge of bad loans, and these loans take an average of two years to go bad, we may see declining foreclosure rates in early 2009.
Figure 11: Volume of Adjustable Rate Mortgages Resetting per Month, Starting Jan. 2007
Source: Credit Suisse
Even so, another 9-12 months of distress sales will continue to wreak havoc on home valuations, as the combination of foreclosure discounts and the skewed mix of homes selling in this market render even the most carefully crafted measures of current home prices suspect. 2009 will likely be devoted to picking up the pieces, and discovering what houses are really worth in a market that isn’t dominated by short sales and foreclosures. We still have a long dark road ahead, but at least we can see a flicker of the light at the end of the tunnel. While it’s hard to find much good news about Phase 2, there is one tarnished silver lining in this dark cloud: the unprecedented speed of the price adjustment means that instead of several years of slow bleeding (like the 1990s), we have compressed the necessary adjustment into two years of intense housing pain. Mom always said it’s better to just rip the Band-Aid off…
| 
