In reaction to the latest Case-Shiller home price graphs, a few readers have asked how a property worth not much more than $400,000 can be considered a member of the "high-priced tier."
The answer is that there is no considering about it. Each month, the Case-Shiller price tiers are calculated by separating all sold homes into thirds by price. The high-priced tier represents not someone’s subjective idea of what comprises a high-priced San Diego home, but rather the most expensive one-third of homes sold during the measurement period.
For January’s Case-Shiller index, the cutoff between the top one-third and the middle one-third was $419,143. The cutoff between the middle one-third and lowest-priced one-third of homes sold was about $284,375.
The tier cutoffs, and especially the one between the high- and mid-priced tiers, used to be a lot higher.
Sounds about right. So 66% of
Sounds about right. So 66% of all the homes being sold are now below ~$420K. 33% are below $285K. It’s where the action is now.
Rich,
Thanks for the
Rich,
Thanks for the explanation. But if the price range used by Case-Shiller for the thirds is determined by the prices of homes sold in the measurement period, how do they calculate an index for each third that makes any sense over a long time period?
Surely they’d want to use a static definition of each third that tries to ensure that any one home always shows up in the same third, regardless of the measurement period. If they used a dynamic distribution of sales for each measurement period, the index for any third in one period could represent a very different group of homes than the same ‘third’ index in a prior measurement period, leading to terrible bias in a time series of the index, reminiscent of all the issues we had with raw median prices and averages back in 2007 / 2008.
Ideas? I have probably just misunderstood your explanation, because I know Case-Shiller is constructed carefully.
You can find the answer to
You can find the answer to your question(and a whole lot more)…
http://www2.standardandpoors.com/spf/pdf/index/SP_CS_Home_Price_Indices_Methodology_Web.pdf?vregion=us&vlang=en
The index works off of the price paid for the two most recent sales of the same home or “Sales Pair”. The tier a home falls into is based on the first of those 2 paired sales. Over time homes can move between tiers, but not during the indexing period.
Hi PR –
I’m not certain if I
Hi PR –
I’m not certain if I understand what you are describing (though i think i do). Let me copy the relevant from the methodology doc that SpringSession linked to:
They do it based on the first sale price, which makes sense because you want to see how much something appreciated based on what tier it would have been in at purchase time. They must do some adjustment to account for the fact that different first-sales happen at different times, if I am understanding this all correctly.
I think that houses could jump tiers depending on when they sold… eg, it’s easier to get into the high tier now than it was a year ago, because so many cheap houses are selling. As such, the tiers only provide a rough guide (but they are nonetheless useful… if only to demonstrate the huge price gains in the low tier during the boom).
Does this clarify things at all? Any thoughts?
Rich
Rich Toscano wrote:….
Does
[quote=Rich Toscano]….
Does this clarify things at all? Any thoughts?
Rich[/quote]
Thanks, Rich and SpringSessionM. This clears it up. I have no excuse for failing to find the source myself, except rank laziness.
As I think about the method, I can see how massive shifts in the market could create at least a temporary bias, but it’s still a lot better than most other indexes.
Right now, the number of home sales used to compute the top third must be very small, and even the middle third must be sparse. Few homes being sold now would have fallen into the top 1/3 or 2/3 when they were purchased – which was probably somewhere in the peak years, given that many are foreclosures. So I’d be very cautious about drawing conclusions from the higher tier indexes right now.
I think another factor from
I think another factor from many sales being from banks is that there’s few if any “move-up” sellers to buy in the next higher catagory once they sold their lower cost home.
A few months back Rich had some fellow do quite an extensive write-up on the market. His insight was very keen. Perhaps we could get an update from him sometime soon?
Rich also I am under the
Rich also I am under the impression, but not completely sure, that as you implied the total number of sales is simply bundled and for the most part cut into thirds.
Are the CS statistics seperate for attached verses detached housing?
SDR, Case Shiller doesn’t
SDR, Case Shiller doesn’t deal with attached housing, it’s single family only.
patientrenter, first of all, after your stories of working 100 hour weeks I don’t think laziness is your problem. 🙂
Anyway, I don’t think this bit in your post is correct: “the number of home sales used to compute the top third must be very small”
… because it conflicts with this bit from the CS methodology: “price breakpoints between low-tier
and middle-tier properties and price breakpoints between middle-tier and upper-tier
properties are computed using all sales for each period, so that there are the same number
of sales, after accounting for exclusions, in each of the three tiers.”
So, it seems there are the same number of sale pairs in each tier — but that the tiers are separated based on the initial price in the sale pairs. (IE each initial sale is compared to the other initial sales among all the sale pairs, not to the other sales that happened the same month as the initial sale). It’s not clear to me exactly how they perform that comparison, given that the initial sales have all happened at different times, but it does sound like they are doing it somehow.
Rich
I think you are seeing an
I think you are seeing an artifact of the methodology on the high end as I just dont see anything close to 35% declines on the hi end as a whole as of yet. Declines up therea re still pretty slow and orderly. IMO, what is happening with these C-S numbers is the sales distribution is so skewed to the low end that you have what most would consider to be low end homes falling in the highest tier.
FWIW, esmith has a 23% decline on his C-S equivalent which seems alot more accurate to me based upon what I’ve seen.
While I have not done a
While I have not done a detailed analysis, the areas that I follow (Del Mar, La Jolla and parts of Carmel Valley) seem to have had a 10-20% decline in just the past few months. It seems to have started in November, just after the stock market crashed (and hit the wallets of the wealthy) and accelerated in January (when quarterly statements arrived in the mail from the brokerage houses, 401ks, etc). I would love to see an analysis on homes in the $1 million to $1.5 million range. Does Case-Schiller provide such data?
dfreshdee wrote:While I have
[quote=dfreshdee]While I have not done a detailed analysis, the areas that I follow (Del Mar, La Jolla and parts of Carmel Valley) seem to have had a 10-20% decline in just the past few months. It seems to have started in November, just after the stock market crashed (and hit the wallets of the wealthy) and accelerated in January (when quarterly statements arrived in the mail from the brokerage houses, 401ks, etc). I would love to see an analysis on homes in the $1 million to $1.5 million range. Does Case-Schiller provide such data?[/quote]
Nope, just the three tiers.
[quote=sdrealtor]I think you are seeing an artifact of the methodology on the high end as I just dont see anything close to 35% declines on the hi end as a whole as of yet. Declines up therea re still pretty slow and orderly. IMO, what is happening with these C-S numbers is the sales distribution is so skewed to the low end that you have what most would consider to be low end homes falling in the highest tier.
[/quote]
Yep, this is pretty much what I said in the article.
Rich
My mind is still on Island
My mind is still on Island time. I have to chuckle when I think of low end homes registering in the highest tier. C-S is the best we got (outside of our reports from the field) but the flaws in its methodology are more glaring than ever…..
sdrealtor wrote:I have to
[quote=sdrealtor]I have to chuckle when I think of low end homes registering in the highest tier. C-S is the best we got (outside of our reports from the field) but the flaws in its methodology are more glaring than ever…..[/quote]
The Case Shiller methodology is not flawed, not at all. It is not the methodology’s fault that volume in the high end has decreased. If the volume of, say, $1M houses drops to zero, what is the value of the houses? Like a stock that becomes thinly traded, it becomes difficult to say what individual houses are worth.
Even if one goes through individual sale pairs, to get “reports from the field”, buyers over- and under-pay for houses. The indices only become reliable over large transaction volumes; as the data becomes thin it becomes noisy. Case Shiller has even eliminated tiered reporting for entire metro areas because the data became too noisy.