San Francisco Housing Bubble Has Been Deflating Since 2005

Computer prices have been falling for decades, but that's not because demand is falling. Computer manufacturing technology has gotten better year after year, enabling computer companies to produce cheaper (and better) computers, even in the face of rapidly increasing demand.

What does this have to do with real estate? San Francisco housing prices increased rapidly between 2001 and 2005. Most people attribute these price increases to rising demand. As the example above makes clear, however, rising demand alone is not sufficient to drive up prices. If you want to sell your home at a higher price, you need other sellers to cooperate by limiting the supply of comparable homes. In fact, with enough cooperation, you can get a higher price for your home, even in the face of falling demand. This seems to be what happened during the final years of the bubble.

Take a look at the chart, below, which shows the recent history of sales volumes and median prices for San Francisco housing.

The blue bars (left-hand axis) represent annual sales volumes, and the purple line (right-hand axis) represents median sale prices. The entries for 2009 are year-to-date figures, through the end of February.

Sales volumes peaked in 2004 and have been falling ever since. In contrast, median prices continued to rise until 2005, and then remained stable for the next three years, even as sales declined by over 35%. It's hard to imagine how this sequence of events could have occurred without a sustained contraction in demand beginning in 2005. The following series of charts should illustrate the point.

The blue line in the chart (above) represents the supply of housing. At any given price, there is a certain number of owners who are willing to sell. As the price rises, so does the number of willing sellers, hence the upward-sloping line. The purple line represents the demand for housing. As the price rises, the number of willing (and able) buyers drops off, so the demand curve slopes down. Equilibrium is attained at the intersection of the two lines, where supply and demand are equal.

Lending standards were relaxed steadily during the housing boom. This led (for a time) to rising demand for housing. As a result of the relaxed lending standards, there were more buyers who were willing to pay any given price. This is represented in the chart by the dashed demand curve, which sits to the right of the original demand curve. With stable supply and increased demand, prices and sales volumes both increased. This is indicated in the chart by the fact that the new equilibrium is above and to the right of the original equilibrium point.

This roughly describes what happened between 2001 and 2004 (i.e., 'Phase 1' in this story). As prices continued to rise, however, sellers caught on and started limiting the supply of homes for sale.

In the second chart (above), the relative position of the dashed blue line indicates that sellers are now demanding higher prices for their homes. In economics parlance, housing supply has gotten tighter. (That characterization will be easier to understand if you notice that the new supply curve is not only above, but to the left of the original supply curve, indicating that there are fewer willing sellers at any given price.)

Combined with stable demand, the net effect of tighter supply is to move the intersection of the supply and demand curves up and to the left of the original equilibrium. In other words, prices rose while sales declined. That's roughly what happened between 2004 and 2005 (i.e., Phase 2 in this story).

The penultimate act of this little cartoon is represented in the chart, below, which shows the effect of shrinking demand combined with tighter supply. The relative position of the dashed purple line indicates that there are fewer buyers who are willing to buy at any given price. Similarly, the relative position of the dashed blue line indicates that there are fewer owners who are willing to sell at any given price. The net effect, in this case, is stable prices and reduced sales volumes. That's more or less what happened in San Francisco between 2005 and 2008, during the final years of the bubble.

There may be other ways to explain what happened to the San Francsico housing market over the last decade. But this story is simple, and it fits the facts nicely. It strongly suggests that the San Francisco housing bubble began deflating in 2005, long before prices showed any signs of softening. We didn't need the benefit of hindsight to figure this out.

California Job Losses Accelerated in January

California's unemployment rate rose to 10.1% in January. That's the highest jobless reading since the 1982 recession. Perhaps more importantly, it's 1.4 percentage points higher than December's reading of 8.7%. At that rate of increase, unemployment would reach 25% by the end of the year. Nobody is expecting job losses on that scale, but as long as jobs are being lost at the rates we've seen lately, the real estate market is likely to remain under serious pressure.

Bad Start for San Francisco Real Estate Market in 2009

2008 was a rough year for the San Francisco real estate market. Sales of condos and single family homes fell by 18% relative to 2007. 2009 is looking even worse, thus far. Through the first two months of the year, there were 40% fewer sales than during the first two months of 2008.

Two months don't make a year, but over the last 15 years, sales volumes in January and February have provided a reliable leading indicator of full-year sales volumes. Take a look at the chart, below, which shows the recent history of sales in San Francisco.

The blue bars (left-hand scale) indicate sales volumes for the first two months of each year. The purple bars (right-hand scale) indicate sales volumes for the corresponding full-year period. If the historical correlation between these two series holds up, 2009 sales will probably come in around 2,800 units. The lowest previous full-year total was 3,664 units, which occurred in 1995.

Prices haven't suffered as much as sales volumes, but 2009 is nonetheless off to a bad start. The median sale price for the first two months of the year was $660,000. That's 16% lower than the median price for the first two months of 2008 (i.e., $785,000).

The 16% year-on-year price decline is particularly discouraging because it was concentrated into the six-month period since September, when Lehmann Brothers failed and the financial crisis began in earnest. (The median sale price for August was $780,000, which was only slightly lower than the beginning-of-year figure of $785,000.)

Recent price declines don't necessarily indicate continuing price declines. But sales volumes seem likely to come in at a 16-year low. All things considered, 2009 is shaping up to be a bad year for San Francisco real estate.

Update (3/3/09): The National Association of Realtors just released its Pending Home Sales Index numbers for January. They show a 7.7% decline in the number of purchase contracts signed, relative to December. In the West, the PHSI actually rose 2.4%. Remember, however, that much of that increase was driven by sales of bank-owned properties. Unlike private owners, banks are not price-sensitive sellers.

The High End of the Market Isn't Immune After All

Zillow recently published its latest Homeowner Confidence Survey. According to the results, 70% of western region homeowners believe their homes lost value during the last year. Evidently, perception is catching up to reality, but there is still a sizable gap. Zillow estimates that, in fact, 90% of western region homes lost value. And judging from the Case Shiller indexes, the loss of value was substantial. Only one western market (Denver) fell by less than 10%, and five of the eight markets fell by 25% or more. The Zillow survey suggests, in other words, that at least 20% of western region homeowners are in a state of denial.

Some of those mistaken homeowners may live in San Francisco's high-end neighborhoods. Once a week, I hear someone say, "Sure, prices are falling, but the high-end of the market is holding up." I'm not sure why people are so determined to believe that, but it turns out that it's not true. Take a look at the chart, below, which shows an index of median prices for single-family homes in two of San Francisco's real estate districts.

The blue bars represent District 7, which comprises some of the City's most expensive neighborhoods, including Pacific Heights, Cow Hollow, and the Marina. The purple bars represent District 10, which comprises some of the City's least expensive neighborhoods, including Bay View, Excelsior, and the Outer Mission. Both data series are indexed to a value of 100 in 2000.

Contrary to popular belief, high-end prices have in fact fallen, by about 8% between 2007 and 2008. What's more interesting, however, is that on a cumulative basis, the high-end neighborhoods haven't done any better since 2000 than the low-end neighborhoods. The low-end neighborhoods have given up more of their gains recently, but those gains were much larger to begin with.

Another way to show the similarity between high- and low-end neighborhoods (at least as far as market performance) is by considering sales volumes. Take a look at the chart, below, which shows an index of sales volumes for the same two districts as before.

Once again, the blue bars represent District 7 (high-end) and the purple bars represent District 10 (low-end). In this case, however, each series is indexed so that its average value over the 15-year period is 100.

Except for a couple of years in the mid-1990's, the two indexes have tracked each other closely. In particular, both indexes peaked in 2004 and have declined by 30%-35% since then. (It's noteworthy that low-end sales volumes actually rose slightly in 2008.)

It seems, then, that the high-end neighborhoods are not a class unto themselves, but are, in fact subject to the same economic factors that affect the rest of San Francisco. Nonetheless, many homeowners in these neighborhoods remain convinced that their homes are special. Zillow's survey results contain a striking (and funny) illustration of this aspect of human nature. 48% of homeowners think their local markets will decline in the next year, but only 30% believe the same will happen to their own homes.

Note: The supposed immunity of high-end neighborhoods looks like a reincarnation of an earlier piece of wishful thinking, namely, the suggestion that a weakened US dollar would induce wealthy foreigners to prop up the San Francisco housing market. No doubt, the proponents of that theory were thinking about the high-end of the market.

Architecture Billings Index Foretells Big Drop in Construction Spending

This is the kind of data series that I like to monitor. The American Institute of Architects publishes an index of industry activity, the so-called Architecture Billings Index. They announced yesterday that it hit an historic low of 33.3 in January. To put that into context, any score below 50 indicates a reduction in billing activity. Until the current crisis, the index had never fallen below 40.

The billings index is a pretty good leading indicator of non-residential construction activity. According to a 2005 study by AIA economists Kermit Baker and Diego Saltes, the billings index leads construction spending by nine to twelve months. I did a quick search for historical data so I could do my own analysis, and ran across this posting on Calculated Risk. I've reproduced one of the key charts below.

The red line (right-hand axis) is the billings index. The blue line (left-hand axis) is the percentage change in private, non-residential construction spending over the trailing twelve-month period. It's clear that the two series track each other fairly closely.

The billings index has fallen by 30% in the last twelve months. (See the updated chart, above.) Calculated Risk points out that a 30% fall in construction spending would be equivalent to $128 billion. That's small compared to the overall economy, but it is significant when compared to any number that's relevant to the real estate industry. (You can read the Calculated Risk posting for more insight.)

Bay Area Unemployment Still Rising Sharply

Unemployment in the nine-county Bay Area hit 7.3% in December. That matches the high point of the 2001 recession (which was actually reached in June of 2003). It's also the highest rate in the last 19 years, going back to 1990 (i.e., the beginning of my data series).

The national situation isn't any better. In its remarks yesterday, the Federal Reserve said that national unemployment will probably increase to 8.5%-8.8% by year end, and probably won't return to the current level (7.6% in January) until 2011.

From the looks of the chart above, I'd say we'll be doing well if we can keep unemployment from rising by more than a percentage point over the next year. Fed Chairman Ben Bernanke seems to agree with that sentiment, saying that unemployment will climb to 8.0% "for sure."

Fed Says Recession Will Be "Unusually Prolonged"

My last posting referred to a paper by Carmen Reinhart and Kenneth Rogoff, in which they said that recessions stemming from financial crises are typically twice as long as 'ordinary' recessions. The Federal Reserve weighed in on the issue today, saying that recovery from the current crisis will be "unusually gradual and prolonged." My guess is that the Fed is reading the same tea leaves as Reinhart and Rogoff.