Author: Jack French -- Used under Creative Commons Permission 2.0

Back in May 2009, I showed that Noe Valley was not immune from the slump in prices affecting the rest of the city, despite suggestions to the contrary from real estate agents, mavens and media.

Have things gotten any better?  Well, no.  And maybe.

Here’s a chart showing percentage change in single family home prices for the last 14 months, relative to their all-time highs  (click to enlarge).  (All figures are 3 month moving averages.)

After reaching an all time high in March 2008, prices plummeted.  Just a year later, in the midst of fears of a global Depression, home prices were down 30%.  Did things get better?  No, they got substantially worse.  Despite an impressive  city-wide recovery in 2009, with prices going from 30% down to around 18% down for single family homes at  year’s end (see more detail here) , Noe Valley home prices continued to retreat.  In October and November 2009, prices were down 35%.  At year’s end, they’d barely clawed back two percentage points. Not surprisingly, days on market (DOM) remained stubbornly high for all of 2009.

Still, with cherry blossoms busting loose all over Noe Valley’s quiet streets, there certainly seems to be a change in the air.  There are many more listings coming onto the market and there’s even the occasional feeding frenzy over a clean, well-priced home.  These go in a matter of days, not weeks. Maybe that upturn in prices for January and February suggests a continued warming trend.

In the next few posts, I’ll look at Noe Valley in more detail, including how condos have fared.

Case-Shiller published its closely watched indices yesterday.  Hooray! The broadest CS index shows that the rate of decline in the nation’s largest housing markets has reversed in recent months.  Now we’re only going down 16% year over year instead of 20%.

They also point out that we are now back to 2003 values, which also holds true of San Francisco.  Here’s my chart from an April blog:

Before you go out and celebrate, Case-Shiller has “San Francisco” down a whopping 26.1% year over year.  Why the quotes?  Because it’s really the “San Francisco-Oakland-Fremont, CA Metropolitan Statistical Area” and it includes ALL of Alameda, Contra Costa, Marin, San Mateo, and … San Francisco County. That’s 5 counties folks, a factoid often omitted even by such august publications as the New York Times (see today’s front page article).

Now here’s the “good” news.  My data says that the San Francisco we live in was down “just” 5.7% in June 09 year over year for homes.  Take a look under the Market Trends tab for annual and monthly data for the City and specific MLS Districts.  (By contrast, condos are down 15% year over year.  That also happens to be how much they’re down from their all-time highs, which occurred right about a year ago.  See my previous post.

Total listings is defined as new listings plus anything that’s under contract but still  “contingent” in the parlance of realtors.  Total sales is exactly that.  The chart reflects the raw monthly numbers with no averaging.  This really highlights the seasonal fluctations:  ie. the very evident drop-off in activity at the end/beginning of each year.

Other than the seasonal dips, maybe you can conclude that both listings and sales are trending downward, but I sure don’t see any evidence of a major change of direction in either.

A couple of closing thoughts.  My absorption rate conclusions were based on an analysis of single family homes only.  It’s possible that the conclusion would be different if I’d included condos and TIC’s as well.  ie.  Looking at the broader market might change the results.

On the other hand, it’s possible that correlations between absorption and price would appear if we looked at finer segments of the market.  For example, we might find that absorption rates are longer at the high end of the market and that in fact prices have come down as we’d expect for that portion of the market.

Alas, the MLS database that’s the repository for sales information for brokers/realtors simply doesn’t allow you to do this sort of data-mining easily, so we’ll never know.

I stand by my previous conclusions:  First, San Francisco just isn’t that overbuilt a market. Second, if you take out the seasonal fluctuations, the absorption rate doesn’t seem to have moved that much anyway.  Finally, and perhaps most importantly, absorption rate doesn’t tell you how much activity (offers)  each available listing is generating — in the end, just one property gets sold.

The question is whether there are other metrics that do a better job of tracking whether the market’s “hot.”  Stay tuned.

One can quibble about whether a 30 day lag on listings to sales is sufficient, but the average days on market for 2006, 2007, and 2008 were 29, 41, and 51 respectively.

Here’s the revised chart.  The median price line and the Absorption Rates line up beautifully, don’t they?  Until you remember that there’s supposed to be an inverse correlation between them.  True, there appears to be an inverse correlation over the last few months of 2008, but that could simply be due to the lack of forward data.  It certainly doesn’t negate the lack of correlation over the previous nearly 3-year period.

Stay tuned for a chart showing median price plotted against Days on Market

“Absorption” is basically the number of weeks it would take to sell all the homes available on the market based on the number of homes that are selling at that time.  (I’ve tweaked the formula to diminish the spikes caused by the huge seasonal dropoff in new listings each December/January.)   There are many ways to calculate absorption, but the basic idea is simply to capture how quickly demand is eating supply.  Less time to absorb the supply should reflect a “hotter” market where sellers can demand top dollar. A higher absorption rate, on the other hand, means that there’s relatively more listings on the market than demand for them.  That would tend to suggest a buyer’s market and softer prices.

In the chart above, we’d expect to see  median prices rising when the Absorption Rate line falls and median prices falling as the Absorption Rate line rises.  ie. an inverse correlation.

Well, I’ve looked at this chart long and hard and I just don’t see those lines moving that way at all.  In fact I’ve looked at similar data as far back as 2002 and the only thing that’s clear is that people forget about buying or selling a home at the end of the year. Look at 2006:  the market got tighter but prices stayed pretty flat. In 2007, stuff was being absorbed more slowly (the red line goes up), but prices went up anyway.  In 2008, you’d think that with only two weeks of supply available, home prices would be skyrocketing.  Obviously that aint happenin’.

I’m not an economist or a statistician, but I did get my dear wife, who eats statistics for breakfast, to check my methodology.  I think these results are quite counter-intuitive.  Here are the explanations I can think of.  Please chime in with your own:

• If you cut off the peaks and troughs, the Absorption Rate  mostly stays within a band of around 4 to 7 weeks.   That suggests that supply/demand in San Francisco is actually pretty stable.  And that in turn suggests that something else must be driving prices.  Note, for example, that my chart  doesn’t reflect the number of offers that are made on any particular house.  There might be 10 offers on a house, but at the end of the day just one house gets sold.  Anyone who was playing during the frenzies of 2005 – 2007 doesn’t need to be told how multiple offers affect price, but that sort of demand isn’t reflected in an absoprtion rate.
• Relatively speaking, San Francisco is not a stressed market.  Supply/demand is not hugely out of whack.  Foreclosures are not piling up (yet).  Under these circumstances, prices are “sticky.”  They don’t react quickly to changes in demand.  If people don’t get the price they want, what do they do?  They don’t sell unless they really have to.  And SF home-owners tend to be people who don’t have to sell.  More on this in another post.
• The price increases of the last few years and their recent tumble may are probably most directly attributable to one thing, pure and simple:  easy money.  That aint gonna show up on this graph either.

Conclusion:  “Absorption”  isn’t a good measure of supply/demand.

So is there any other metric that correlates more closely with changes in price?  How about the famous “DOM” — Days on Market.  This is how long a property takes to go from being listed to being sold.  The theory goes that when properties sell quickly the market is “hot.”  Why do properties sell quickly?  Probably because … there’s more demand than supply.  ie.  More people making offers, more people getting the loans they need, more people willing to waive inspection contingencies and buy “as is” just to get the deal done.  So maybe DOM actually does capture those muliple offers where the Absorption Rate just doesn’t.

So will DOM tell us how “hot” the market is and where prices are headed?  Or is DOM dumb.  Stay tuned….

Most of the data covers the state as a whole, and even when it’s broken down by county, Ms.  A-Y stressed that there can be huge differences when you get more “granular” with the details.  (I made the same point in my 10/27 post discussing how misleading the much-quoted Case Shiller Index can be.)

Still, the 2008/9 forecast is worth looking at and, surprise surprise, it ain’t pretty.

Look at that 31.7% drop in median price from 2007 to 2008, with another 6% drop forecast for 2009.  Keep in mind that these numbers are California-wide and San Francisco is holding up relatively well.  Here’s another CAR chart showing the movement of median prices for existing detached homes by county.  Note there are some HUGE discrepancies between some of the data presented here and the data collected by Dataquick (see the chart in my 10/27 post).  Remember that the DQ data includes ALL sales (condos, town-houses, etc.) and that they reflect “non-market” transactions like repossessions, sales to family members, etc.  Still, the difference in figures for some counties –e.g. Contra Costa — are astonishing.

And finally, one more chart for a long view through history.  How bad is this housing downturn compared to others?  Well, so far prices have dropped about midway between the catastrophic fall of 1978-82 (think OPEC oil embargo, Iranian hostage crisis) and the more moderate drop in prices that occurred from 1988-1992.

Are we at the bottom?  Probably not, based on everything I heard from the big brains, but the consensus seems to be that a full-blown depression is unlikely and that the economy will start seeing some daylight towards the end of 2009.

Future posts will return to earth and look specifically at San Francisco.  If there’s a bright spot in any of this, it’s that for once San Francisco’s not as cold as some other places in California.

You want scary?  Median prices are down 45% year over year in Contra Costa County.

Actshooly (actually), the reason’s simple.  As is often the case when widely quoted indexes talk about a city, what they’re really referring to is a Metropolitan Statistical Area (MSA), a much larger geographical area. So “San Francisco” doesn’t mean our little piece of heaven, no.  It means…  all of Alameda, Contra Costa, Marin, and San Mateo COUNTIES as well as San Francisco county.

According to the latest release from Dataquick, median home prices in Contra Costa County were down a whopping 45.6% YOY for September 2008. No great surprise considering that it includes ground-zero overbuilt subdivision train-wreck areas like Pittsburgh and Antioch. Take a look:

With Alameda County down 30%  and even Santa Clara County down 27.4%, it’s clear how the other counties in the San Francisco MSA are dragging the “San Francisco” index down.  Feel better now?

And a PS.  Dataquick shows San Francisco County as down 12.7% for September versus my 11.3%.  I get my  numbers from a service that pulls them directly from the MLS, then I use their spreadsheets to dig deeper when necessary.  I’m looking into why there’s a discrepancy between my numbers and Dataquick’s.

So I thought I’d check how September 08’s median home prices (condos will come later) compared to their all-time highs and to the median prices of a year ago, both by MLS District and for all of San Francisco.  I didn’t include District 8 (North-east) because it doesn’t have enough data to be useful, and I also didn’t include the southern-most districts of SF (3 and 10) because to be honest I don’t follow them closely. Here’s the result:

So clearly prices are down from their all-time highs across the board.  (Most districts were still hitting highs or near-highs well into 2007, by the way, and District 5, which includes Noe Valley had its top 3 highs in 2008!) )  But where the drops are really big (Districts 6 and 7 for example), that could simply be due to the fact that the all-time high was aberrational.

The percentage change from a year ago are interesting because you can see how some districts seem to be doing quite well.  Half up, half down.  Once again, though, with sales volumes down across the board, there are less data points and that can skew the numbers.  But it certainly seems like the tonier districts (1, 5 and 7) are holding up better than the others.  (Take a look at my graph from a coupla days ago to see how the districts compare over time.)

Bottom line(s)?

San Francisco single family homes are down over 11% from a year ago.

The more expensive neighborhoods seem to be doing ok.

The moving average “flattens” fluctuations, but the raw numbers show the March median and average condo prices at all time highs of $765,575 and $914,187, respectively. And this is while volume in terms of units sold is down. Why?

My theory, which is mine and belongs to me — Monty Python fans, you need to go back a long way to pick up that reference — is that it’s precisely because single family home prices have remained high AND buyers are finding it harder to get loans that condos are looking particularly attractive as a relatively –ahem — affordable option. Kind of counter-intuitive, but I think it makes a crazy kind of sense. Let me know what you think!