Existing home sales yesterday, new home sales today:
New-home sales in the U.S. in June were higher than forecast and the number of properties on the market dropped by the most in four decades, indicating builders are making some headway in clearing out inventories.
Purchases decreased 0.6 percent to a 530,000 pace, from an upwardly revised 533,000 in May, the Commerce Department said today in Washington….Builders are offering more incentives and lower prices to attract buyers and help reduce a glut of unsold properties. Still, stricter lending rules and rising mortgage rates may prevent sales from rising much more in coming months.
And a paragraph that shouldn’t catch any plugged-in person by surprise:
The median sales prices last month decreased 2 percent from June 2007 to $230,900. These figures can be influenced by changes in the mix of sales at the regional level. For that reason, economists prefer price measures that track the same home over time.
SocketSite too. Apple Jacks for breakfast anyone?
∙ New-Home Sales in the U.S. Fell 0.6% to 530,000 Pace in June [Bloomberg]
∙ U.S. Home Resale Rate Falls To Ten Year Low (And 33% Below 2005) [SocketSite]
“These figures can be influenced by changes in the mix of sales at the regional level. For that reason, economists prefer price measures that track the same home over time.”
Ha! I have been one of the more vocal critics of the use of the mix argument on this site, as it’s only used to explain higher prices. I haven’t seen one posting where someone says prices would have been higher if it weren’t for mix. We get the math and understand how mix works, it just that it works both ways. For example, mix is arguably keeping prices down in many regions. Is there any hard evidence that this is the case here or not as well?
OK, back to my Apple Jacks 😉
[Editor’s Note: Mix can work both ways, but up until recently we haven’t seen any evidence that it has in San Francisco (and on those numbers we’re still checking).]
Until recently, meaning for the past year or so?
Total SFRs for sale in San Francisco — 611.
Total SFRS for sale in San Francisco minus District 10 — 382.
[Editor’s Note: Nope, as in possibly over the past month. And listings? Are you kidding? An awfully ironic comment coming from a reader who is always quick to point out the relative strength of the upper end of the market versus the lower (i.e., a shift in the mix of what’s selling.]
No. I’ve been talking District 10 as relatively isolated for a very long time, and you know it.
Listings? I can show sales too. Shoot, plug in the recent median price of 727K and look at where the sales come from. It’s not unclear.
372 SFRs sold for the median of 727K or under so far this year. 155 are from District 10. Mix?
“372 SFRs sold for the median of 727K or under so far this year. 155 are from District 10. Mix?”
That is, 155 are NOT from District 10. Meaning 217 of the 372 homes sold for median or under are from District 10.
1-9 — $psqft for SFR properties @ or under median this year, $577 per foot.
10 — $psqft for SFR properties @ or under median this year, $ 455 per foot.
[Editor’s Note: Now compare that to last year and don’t forget to include condos (which are predominantly outside of District 10 and account for half the market).]
Come on, fluj, you know better than that. Providing the number of D10 houses that sold at or below the median in 2008 so far tells us exactly nothing about how the mix may, or may not, be affecting the median sales figures.
Tell you what, why don’t you just provide all the relevant data on every house sold — district, price, sf, etc. — and we can all run the numbers. Let me guess, the SF MLS won’t let you do that precisely because they don’t want anyone to be able to run the numbers. Not your fault, of course.
@ socketsite,
I made my point. I can’t believe you said I haven’t been hollering District 10 this and District 10 that tho!
Trip,
I actually tried to do something similar. The problem for me wasn’t what you said there. The problem is that exceeds 1000 sales and the database won’t display data over 999.
Let’s not forget that condos have other factors going on, such as market saturation. Condos and SFRs are related markets but not precisely the same market. And yeah, District 10 doesn’t have a whole lot of condos. It does account for over 1/6 of the SFRs sold in the past month, 34 out of 190. We’re talking 441 a foot for those 34 versus 663 a foot for the remaining 156. That’s gonna affect mix for SFRs, clearly.
[Editor’s Note: Now compare that to last year and don’t forget to include condos (which are predominantly outside of District 10 and account for half the market).]
Last year District 10 saw 264 solds 1/1/ through 7/25. And the average $psqft was $529. That’s 20 % higher than this year.
Fluj, what exactly are you reporting with respect to D10? Up above you noted the 2008 D10 sales at median or under at $455/sf. Then you note the 2007 D10 sales at $529/sf, which you say is “20% higher than this year.”
First, it is unclear what you are comparing to what. 529/sf (2007) to 455/sf (2008) is only a 14% drop — really not that far off from the overall 11.9% drop in SF median for SFRs (remember, a 20% rise followed by a 20% decline does not leave you right were you started — different denominators). From this very limited bit of data, it sounds like the city as a whole is moving down right in line with D10.
But from what you wrote, it sounds like you are only presenting the 2008 D10 $/sf data for those homes that sold below the citywide median, not all D10 homes that were sold. That would skew the 2008 numbers lower. But you do not say you’ve done the same thing with the 2007 numbers.
Just giving D10 data really doesn’t tell us anything about how changes in the mix might be affecting the median citywide — we need much more. But the numbers you’re giving are very misleading and appear to contradict the point you’re trying to make in any event.
Yeah Trip you always disagree with everything I say.
I presented one point, plugging in the median.
Then I showed the flat sales in another post.
“really not that far off from the overall 11.9% drop in SF median for SFRs” — HUH?
Absent District 10 — 728 sales in 2008 so far at 684 a fot.
For 2007 it was 802 sales and 690 a foot.
That’s like a 1 point drop in price.
Then, plug in District 10 with its sizeable drop.
Get it? Mix.
Please someone make fluj’s life easier and give District 10 its independance. Let’s call it “Not Really San Francisco”.
“Please someone make fluj’s life easier and give District 10 its independance. Let’s call it “Not Really San Francisco”.”
Foolio? I think you’re being summoned!
Dude, come on. The point was to show that mix works downwardly too. OK. That’s it from me.
Fluj, I don’t disagree with everything you say. I actually agree with much of what you say. But I do call foul when I see it.
You are comparing apples and oranges in your latest posts. There was an overall drop in median sales price YOY of 11.9% citywide. Then you point to the smaller drop in $/sf excluding D10. And you say “That’s like a 1 point drop in price.” But $/sf is not the same thing as median sales price — they measure two different things. Your hypothesis about a change in mix being responsible may be correct, but you are presenting irrelevant, selective data to test that hypothesis.
Well, if we want to “adjust” the numbers and get rid of oddball situations, we could remove a big chunk of Noe and PH that are profiting from fresh cash injected by successful companies around the bay.
Just like macro-economics is making District 10 plummet, micro-economics is also affecting the 2M+ SFHs.
But the more we crop out, the more we’ll need to crop out. Because you’ve got tear downs in good nabes, great houses in bad nabes. There’s no end to this.
Actually, another issue with a mix is the “recency” of the construction/renovation.
In the same nabe, a fixer at 800/sf is not the same as a fresh redo at 800/sf. With the redo, you’re sure not to have upfromt expenses. They are expected to be more espensive.
If much of the latest sales are redos, then the prices are slanted to the upside.
A 3000sf complete rebuild of a SFH that sells for 2.8M does not mean that YOUR 3000sf fixer will get as much as that.
“There was an overall drop in median sales price YOY of 11.9% citywide. ”
Look, I can’t get median figures for districts or I would. I can get averages. Averages are affected by mix, big time.
You say that $psqft and median measure different things. I’d say that one measures something and the other, nothing. At least around here. What does median measure in a city of extremes? where the median number itself buys such a small percentage of the properties available?
And once more I fail to see how such a precipitious drop in 10 does not affect median. Look at what’s on the market right now. Those are the “solds” of tomorrow. Ten is dominant.
Let’s look at the definition of ‘median’. I’m not sure how a median house price is determined but this is how I remember it from stats class.
If we assume that every house sells for a different price then the ‘median’ house, out of 611 sales, is house number 306 (counting upwards beginning with the cheapest house and ending with the most expensive house).
According to fluj, there were 229 sales in D10 and 382 outside of D10. If we assume D10 composes the entire bottom of the list then the ‘median’ priced house, number 306 is not in D10 (which ends at house # 229).
If this is true D10 would have huge impact on the mean but no impact on the median; unless it grew to consist of over 50% of the market.
I’m happy to be corrected by those who know how real estate medians are actually calculated.
Medians in District 10 CAN affect the overall median
Take a simple sample of 5 sales:
200K 300K 800K 1600K 1800K
Current median is 800
Take out 200 and 300, and median goes to 1600
uh oh, I think I agree with fluj
as a professed neutral to bear, I generally find fluj’s cheerleading (and writing style) a bit of a pain.
but fluj’s statement:
“You say that $psqft and median measure different things. I’d say that one measures something and the other, nothing.”
is dead on.
median is always influenced by mix, up down sideways, everywhere. it goes without saying that in a market of homes, in a city like San Francisco, that ‘mix’ will play a very significant role in determining median price.
A commodity good, such as ‘commodity corn’ has a price that rises and falls due to changes in supply of the good, and demand for the good. one kernel of commodity corn is just the same as another (they are all apples to apples in the SS sense). mix can never give a false median price for commodity corn. there is no mix. all commodity corn kernels sell for the same price because they are indistinguishable (a commodity if you will).
houses, OTOH, are unique (due to location, property size, layout, finishes, age, view, etc). one house is very different from another house. they are not commodities (though in the bubble years they began to display commodity-like characteristics, and in their fall from grace they are/will doing/do the same). mix always determines the median price of housing. that is because there is no such thing (especially in SF) as homogenous homes.
Did housing medians rise in the 2002-2005/6 years? yes? are they falling now? yes what does that tell us? in the absence of other information – nothing.
when fluj gives out $/sqft data, as s/he says, it is something. and correctly stated, because it is far from everything. $/sqft data from a particular district is even more of something. with each variable included, you are controlling for more variables. apples to apples is best. this is because then you are presumably controlling for all variables (though this is virtually impossible due to changes in the area surounding the house – oh such as the house next door getting fixed up as was highlighted in a recent topic).
the D10 debate will rage on forever. I do think it is relevant, if onyl because I know of very few folks who are looking to buy in SF who would rather buy in D10 in SF compared to a suburb in the East or South Bay. That is just my opinion. It doesn’t change the fact that prices are falling in D10, or that these falling prices are likely having a slightly disproportionate effect on the median SFH price in SF, or that $/sqft is a measure of something but of course does not tell the whole story at all.
the truth is out there. it just seems impossible to find.
dodd- assumptions are everything, so this will also make some assumptions – but your analysis is incorrect
let’s say your starting assumptions are correct, yes house #306 is not in D10, and yes it is more expensive than any D10 house. but the median house is lower on the list b/c of 229 ‘bottom dweller’ houses in D10.
i’m basically saying what sanfronzi said, but worse 🙂
to put it another way, say D10 wasn’t part of “the real? (jk) San Fran. then you would have #191 in D1-9 being the median house. Assuming a general spread of house prices, this median would be way higher than the price of the D1-10 median.
again, median is a bad measure any way you slice it.
Do we really need another thread where everyone steps over themselves saying how bad a measure median is?
It’s bad. Stop using it.
And people said I was repetitive. Yeesh.
“median is pointless” that was my point from last week and it’s still true.
This thread is actually about new homes, which doesn’t really exist is SF anyway. And you are almost always tearing something down to build new if you do. So, I think any numbers stated above are even more meaningless if that’s possible.
The median new construction home in SF, what are there 5 a year. Next years numbers will be greatly swayed by the 7 new West Portal houses being built now.
I thought median sale prices by district were available?
I am sure I have seen comparisons YOY and MOM ofsales volumes and median by district?
Certainly on the front steps.
[Editor’s Note: They are. And as you might recall, here’s how we responded when you asked the same question three months ago:
“Neighborhood mix is but one part of the overall “mix” equation. That being said, according to the San Francisco Association of Realtors (SFAR) sales volume dropped 16% on a year-over-year basis in February. If we rank order listed sales by district based on median sales price in 2007 (one each for condos and single-family homes), in 2008 sales volume dropped 7% in those districts above the median in 2007, and dropped 25% for those districts below the median in 2007.
Seriously folks, we don’t simply make this stuff up (or just report what we “want” to see).”
We do the calculations every month, but unfortunately SFAR data lags behind DataQuick which can hinder our ability to respond to certain inquiries as quickly as we’d like.]