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.

That’s useful if you’re looking at an average-priced TIC and you’re curious about how much of a premium you’d have to pay for an average-priced condo.  But how about reducing that to a per square foot premium?

For those who just want the bottom line, here’s the answer, but it’s worth reading on for the caveats.

$37 a foot doesn’t sound like much of a condo premium to me, that’s for sure.  And as my astute readers will note, the drop in price on a per square foot (from around $225 per sf) is obviously much more than the drop in median sales prices shown in the previous chart.

What’s going on?  It’s really simple:  there’s a lot less information on sales price per square foot for TICs.

All my data comes from the MLS (Multiple Listing Service) that real estate brokers use to find and market properties.  When a sale’s completed, they are required to enter the sales price.  If there’s information on the square footage of the property — provided by the owner or more frequently from the property records — the database calculates a per square foot price.  Roughly 80% of condo sales have a recorded price per square foot in the MLS.  Only 45% of TIC sales have a recorded price per square foot. How bad is that? In September 09, there were just 27 TIC sales.  Only 9 of them had a recorded price per square foot.  For all of 2009 through September, there were 275 TIC sales.  Only 113 – 41% – show a per square foot price.

There are lots of people — mostly on other blogs   — who love to trash statistics and say they’re meaningless.  Medians don’t reflect home values, etc etc.  I disagree.  Provided you have enough data  and you understand what you’re measuring, statistics help make sense out of what is otherwise undifferentiated data.  But I am afraid that in the case of measuring the condo premium on a per square foot basis, we are in dangerous low on data.

One final reminder:  For this series of posts, my TIC data includes the handful of stock cooperative sales that occur in this market.

And thanks for sticking with me on this long series of posts….

Look at that!  Excluding those wonderfully regular dips every Xmas, condo sales are generally at around 200 units per month.  TICs rarely break 40.

Here’s how TIC and condo median prices stack up against each other on a monthly basis.

Dueling spaghetti you say?  That was my reaction, too.  The huge variability in prices from month to month on the TIC line is a direct result of the paucity of sales.  And this chart certainly doesn’t help get at the key question, which is this:

Given that TICs are riskier and less flexible than condos, what’s the premium that you pay for buying a condo vs.  a TIC?

In fact many TIC buyers do so with the hope of being able to realize this “premium” by converting their TICs into condos down the road.  Fat chance unless you’re buying a TIC in a two unit building which — for now at least — remain exempt from San Francisco’s byzantine annual lottery system.

Luckily, I have a bona fide statistician mathematical genius phd for a wife, and she always lends a hand on methodology when I need it.  She suggested that where one set of data (condos) is so much larger than another, using averages provides a more reliable “apples to apples” comparison than medians.   Also, with so few monthly TIC sales, I decided to look at annual rather than monthly trends.

Here’s attempt number two.

Much more useful!  (By the way, the fact that TICs were more expensive than condos in 2003 and 2004 can be explained by a few massively (in excess of $8 million) expensive TIC sales in those years.  This is a great example of how using medians or averages can really affect the results.)

So, can we drill down further and come up with a condo premium per square foot? Stay tuned….

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

As I mentioned in my previous post, I’ve had several questions about per square foot prices recently.  There’s no doubt that it’s a very useful metric, for the obvious reason that it allows you to get closer to an “apples to apples” comparison of the value of two different properties that are different in size.  Of course, that leaves all sorts of other variables — location, amenities, etc.  But if, say, you’re looking to make an offer on a property, certainly you’d want to start by looking at what other properties in the same area have been selling for on a per square foot basis, and then use that to see if the property you’re interested in is in the ballpark.

Since I live in Noe Valley, I’ll readily admit that I tend to follow my neighborhood more closely than other areas.   No great surprise there.  So here, without further ado, is a chart showing the price per square foot for single family homes in “core” Noe Valley (click to enlarge).

Rather than run the chart as Percentage Change From All-Time High, as I usually do, this simply shows price per square foot as a 3 month moving average.  I’ve added the  “number of sales” per month, plotted on the right-hand axis as well.   Note that low monthly sales volumes ( no surprise, given the small geographic area) will make the data less statistically reliable.  In most months, there are less than 15 sales.

While we’re on the subject of sales volume, I recently read an advertisement in the local rag, The Noe Valley Voice, from a local real estate company touting how sales volume in Noe Valley is up, compared to San Francisco as a whole.  So what?  As I’ve stated before in the context of the luxury home market, I really haven’t found any correlation between volume and price.  Though it may be a little difficult to tell from this chart, I don’t see it here either.  For example, sales volumes were down and falling during the autumn months of 2007, but that’s when prices started climbing towards their all-time high in early 2008.  Likewise, sales volumes were increasing through the first 6 months of 2008, even as prices were sliding.

The bottom line is that Noe Valley homes are still fetching north of $700 a square foot, and that’s after a protracted slide.  Sure, the price per square foot is  down substantially from the near- $900 a foot that they hit back in January 2008, but it’s a pretty well-heeled foot nonetheless.

At the same time, I noted that there was only about $100,000 difference in median value between condos and homes.  That seemed like a small delta and I was interested to see whether it was, historically speaking.  Turns out that it is.

Since, until recently (ahem!), home prices along with everything else have tended to go up, I decided not to look simply at the difference in price between condos and homes.  Instead, I converted the price difference to a percentage of the median value of condos sales for the given period.  This represents the “premium” for owning a home rather than a condo.  Here’s the result.

Sure enough, you’d normally expect to pay around 20% more for a home than for a condo.  But starting in 2008, the home “premium” started dropping significantly.  I believe that drop was a direct reflection of the housing market decline that began with homes and only subsequently spread to condos.  As I postulated in my blog, condo values possibly held up for longer as people got squeezed out of the single family home market by tightening credit standards.

But what about 2009?  The chart above shows the premium based on all sales for the year to date.  The picture looks a little different if you look at values on a monthly basis.

Again this is consistent with my previous blog.  It suggests while that condo values may have held up longer, they too have fallen so that the premium paid for a home is now heading back to its historic norm.  Of course, the other possibility is that home prices are beginning to recover.  It may well be that both explanations are true.

I looked at percentage change from all time highs for condos and single family homes (sfd’s) since January 2003 and here are the results for the city as a whole.

Until June 2008, condo and home prices were in lock-step in terms of price appreciation and decline.   Thereafter, homes fell first and further. (Do I hear a lithp?) In March 2009, the delta between condos and home prices was a whopping 13%.  Since then, however, home prices have recovered smartly:  as of June, homes are about 4.5% further off their all-time highs than condos.

What does this all mean?  First of all, I wouldn’t take too much consolation just yet in the upward spike in both condo and home prices since the beginning of the year.  If you take a look at the chart, this happens every Jan/Feb when people start buying out of the winter doldrums.  I wouldn’t predict a bottom until we see what happens this winter.

Secondly, given the woeful condition of the economy and the credit markets, together with the fact that San Francisco is not a badly overbuilt housing market, it sort of makes sense that condos are holding their value relatively well as people are finding themselves priced out of more expensive single family homes.

Still, the current delta of only $100,000 between median condo and median home prices seems rather small.  If people are just begging to know what the historical average is, let me know and I’ll find out.

Say what you want about the two valleys, you’re going to pay more if you live in Cole. About $200,000 more, to be precise. Before you conclude that the premium all but vanished in 2009, remember that there have only been five Cole Valley home sales so far this year. Way too few to conclude anything other than that nothing’s selling or that homeowners are so addicted to the breads and pastries at La Boulange that they can’t leave.

After doing a guest post on Noe Valley price trends at theFrontSteps a few weeks ago, Alex, tFS’s friendly editor, suggested that I do a side-by-side comparison of sales trends in Cole Valley and Noe Valley.

Great idea, I thought! Trouble is, Cole Valley sits within a tiny subdistrict of the MLS  (see the pink area below?) and as a result, there very few transactions from month to month.

That makes data crunching hard.  Maybe even meaningless. Check out the white bars in this chart (click). They represent the number of single-family home sales per month back to January 2003.  (Number of sales is tracked on the right side of the chart; percentage change from “high” is tracked on the left side.)

You can see that there are many months where only one or two houses sold. There are some months where there were no sales at all. It’s tough to extrapolate monthly sales trends under those circumstances and dangerous to assume that an “all-time high” is meaningful when it’s based on only one or two data points.

So instead of running percentage changes off of median monthly values, as I had done for Noe Valley, I ran the percentage changes off the “95th Percentile” value of all sales occurring between January 2003 and April 2009. The 95th Percentile value represents a “high”, while excluding the potentially aberrational top 5% of sales.  Aren’t you glad you asked? (Special thanks to my wife, Nina, who looks over my shoulder at a lot of my statistical analyses — she’s the one with the one with the PhD in data-crunching.)

After looking at this chart, I sort of threw up my hands.  With only 179 sales in over 6 years, it’s not sensible in my view to draw conclusions about monthly trends in Cole Valley, let alone to compare them to Noe Valley, where the “core” area alone — Subdistrict 5C — had over 900 sales during the same period.

So I re-ran the numbers and calculated medians based on annual sales.  The second chart (click) shows the results.

I think this is much easier to understand.  Again, with so few sales, one should be careful about drawing any conclusions, and with only 5 sales in 2009 so far, I think it’s too early to conclude that the apparent drop in median prices for 2009 will continue to be accurate.  Rather, I’d say that Cole Valley seems to have been holding up pretty well.

Stay tuned.  I can’t help myself.  Coming up, Cole Valley and Noe Valley go head to head.

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

Noe Valley’s been my home since 1991 so I’ll admit that I track it with more attention than other neighborhoods.  For quite a few years now it’s also had a reputation for maintaining home values even as the rest of the city stumbles.  The question is whether that’s still true.
Back when I bought my two-unit home at a probate sale with one other bidder in attendance, the “hood” was closer to its working class roots.  Lots of single-story cottages in the 1200 square foot range and lots of duplexes.  It had a thriving lesbian community and was popular with singles who couldn’t afford to rent in the Marina.  It was San Francisco’s stealth neighborhood, just out of yuppy radar range.
By the mid-90’s it was growing increasingly popular with young families who were drawn to its slightly sleepy streets, decent weather, cute parks.  The shops along 24th Street had everything you could possibly need, but without the attitude of Fillmore Street or the grunge of the Haight.  All this, and decent Muni service downtown to boot.
When the dot-commers discovered that it also had great access to 280 and points south, prices took off.  Two-unit buildings started converting to single family homes (that’s what we did).  Little cottages would suddenly grow an extra floor, or a built-out basement, or both. Suddenly, we started seeing 2,500 foot homes and even the occasional trip-K (3,000 sf).
Prices inflated along with the floor-plans.  The million dollar mark was crossed some time around 2000.  The dot-com bust barely caused a blip.  By 2005 we were seeing feeding frenzies over houses approaching $1,000 per foot.   Even after the market started cooling off in 2007, Noe Valley prices seemed to defy gravity.
There’s nothing more local than real estate.  Could it be that Noe Valley was situated in some sort of socio-economic sweet-spot?  Many thought so, including me.  And until recently it was true.  But no longer.  Take a look (click to enlarge):

The data reflects sales of single family homes for MLS subdistrict 5C only.  Though the subdistrict includes areas as far up as Diamond Heights Boulevard and as far out as Guerrero Street, it captures the neighborhood pretty precisely. Here’s the MLS district map.

For the data-geeks among you – and I say that with the deepest affection – I should point out that the volatility of the line for Noe Valley prices is related to the much smaller volume of monthly transactions relative to the All-Districts data, and this is particularly evident around January of each year, when transaction volume really falls off.  Also, the three-month moving average is not weighted to reflect the number of transactions in any given month – it’s a simple average, as is the moving average for All-Districts.
These caveats aside, the trend seems pretty clear.  Noe Valley home prices continued to climb for nearly a year after the city’s as a whole had reached their peak.  But, starting in March 2008, they’ve dropped like a stone.  As of March, they’re actually doing slightly worse, on a percentage basis, than the city as a whole.
I’m not exactly thrilled about the thought that my house is worth about 30% less than it was a year or so ago (and, being both an optimist and a home-owner, I don’t really believe that’s true ☺).  Nor would it surprise me if Noe Valley recovered more quickly than some other neighborhoods.  But, at least for now, its gravity-defying days seem to be over.