The Case-Shiller Index for October, 2011 was published today, and the media is swamped with negativity. CR prefers the seasonally-adjusted; here are both for San Diego:
| San Diego CSI | Sept | Oct | MOM % chg | YOY % chg |
| SA | 151.67 | 151.61 | ||
| NSA | 153.72 | 152.86 |
There’s the soundbite – “prices” are still going down, according to the Case-Shiller Index.
We’ve picked apart their methodology before – today let’s examine how many sales are excluded in their rolling three-month counts.
They compare the most recent sales price to the previous sales price of existing single-family homes only (no condos). They then weight the data based on the time interval, and any extreme price changes. Typically 85% to 90% of the sales pairs receive no down-weighting.
But they also exclude sales too.
They state that the excluded ‘non-arms-length’ sales pairs are “usually less than 5%” of the total, and that new-builts and flippers could exclude another 0 to 15% of the total sales too. (See pages 8 and 19 here).
So let’s say that they think 2% to 20% of the actual sales are left out. Or is it more?
Standard & Poors/Case-Shiller does publish their counts of sales pairs, but they don’t add up:
| SD Sales Counts | Case-Shiller | SD MLS |
| Aug | ||
| Sep | ||
| Oct | ||
| Total |
Their published counts can’t be just the one-month total, because they are way too high. If their published number is the 3-month total, then they are excluding more than half of the detached sales, according to the MLS count.
Sure, a survey of half of the sales is worthy. But when the index is only moving 1% to 2% per month, it wouldn’t take many of the excluded sales to drastically influence the outcome in either direction. Yet, that isn’t mentioned anywhere – instead, the media uses the CSI like it is a gold-plated AAA-rated fact about “prices”.
Just like with the NAR data, don’t make decisions solely based on what you think the Case-Shiller index says. The best gauge is the on-the-ground survey done with your own eyes and ears.
Can it be that the CSI count includes sales that are not in SD MLS (e.g. foreclosures, short sales, SBOs)? I’m not an expert. Just want to hear your opinion, Jim.
It’s actually pretty unlikely the numbers would be influenced by much with that large of a result set. If we have 2500 sales and add in 100 sales that all showed price gains of 10% we’d only move the weighted average up 0.4% YoY (so -4.1% versus -4.5%). For 92126 real sales I have -5.2% YoY for Oct, so it’s in the right ballpark. I’m sure there’s local markets that are slightly up YoY, but I think the best optimistic take on the market right now is sideways along the bottom.
But it isn’t insignificant, the actual count is about 100% more than they state. Their October count is 2,397 sales pairs, and the MLS alone shows 5,457 sales for the 3-month rolling total. I don’t like that they exclude flippers either, because they are a vital part of the marketplace.
Can we look forward to your blog in 2012?
Eugene,
Both the CSI and SD MLS include REO listings and short sales sold, as long as there is six months in between sales, plus most FSBOs pay to get on the MLS.
I think the missing link is them finding the previous sales price.
What properties don’t have a previous price?
Long-time owners, and the affluent whose agents think it is their job to omit the sales price from the tax collector – a practice that probably isn’t used as much today as before.
If you go back to the 1980s and 1990s, nobody in La Jolla or RSF had their sales price published – it was a badge of honor.
This reminds me of how the government uses the U3 number for Unemployment when the U6 number is the accurate gauge.
There was a guy who ran his own San Diego home price index, using the same matched pairs method. He got similar results to CS, but not exactly the same numbers, and 3 months earlier and without the smoothing. Unfortunately he hasn’t updated his site in a year:
http://sdhpi.blogspot.com/
The challenge is do you include improvements? And how do you tell, just a fast sale doesn’t tell you. I actually think it’s better to just report the numbers, improvements and flips included. The fancier you try to get in the algorithm, the more likely you are to introduce a bias.
When you say “Swamped with negativity,” what exactly do you mean?
Jakob is 100% correct; complexity in measurement introduces inaccuracy.
House Prices Fall:
http://www.nationalmortgagenews.com/dailybriefing/2010_504/house-prices-fall-1028033-1.html
Nasty Case-Shiller:
http://www.forbes.com/sites/afontevecchia/2011/12/27/nasty-case-shiller-shows-home-prices-barely-off-their-crisis-lows/
Home Prices Fall:
http://latimesblogs.latimes.com/money_co/2011/12/home-prices-fall-in-october-says-case-shiller-report.html
Housing Pain Remains:
http://247wallst.com/2011/12/27/case-shiller-report-housing-pain-remains-kbh-dhi-tol-len-phm/
Home Prices Fall in Most Major US Cities:
http://www.cnbc.com/id/45795202
“In light of the more positive housing numbers we’ve seen in the last week or so, this might be a bit of a disappointment,” said Omer Esiner, chief market strategist at Commonwealth Foreign Exchange in Washington.
The data on Tuesday showed prices declined in October in 19 of the 20 cities.
Want me to keep going? My complaint is that the mainstream media NEVER looks any deeper than rise/fall, high/low, hot/cold etc.
I’m a part-time blogger who in 10 minutes can find reasons to think that the whole CSI is riddled with inaccuracy, but does anyone else?
I wasn’t attempting to attack you about CSI. My point was that in any kind of statical analysis might have a slight flaw, but it tends to average out pretty quick. I don’t think Case-Shiller is out there trying to make the data look a particular way, I think the media puts a spin on it but that’s not the fault of Case Shiller. You can take data and compile it a bunch of different ways but it’s probably going to tell the same story.
Case-Shiller tells me that home prices were down slightly YoY in San Diego and have gotten a bit softer into the slow months of the year. Is that a fair assessment of the current situation? If home prices are flat to down 5% YoY, does it really effect buyers or sellers all that much. Maybe it will be good and sellers will feel like lowering their price.
Omitting half of the sales is a slight flaw?
We’ve covered this publically and privately. I don’t mind that you have opinions, and agree that your base content is usually accurate.
But you always present them in a way that make you out to be the expert, and I’m wrong.
Start your own business blog, and I’ll come there every day and undermine what you say in front of your potential clients and see how you like it.
OK, just two more:
http://www.latimes.com/business/realestate/la-fi-housing-prices-20111228,0,541656.story
Home prices in the nation’s largest cities fell in October for the second straight month, suggesting that prices will head down further next year and dashing hopes that the sluggish housing market was headed for an upturn.
“Add to this the currently high unemployment and underemployment rates, one gets a recipe for further price declines,” Newport said in a research note. “Our view is that foreclosures, excess supply and weak demand will drive prices down another 5% to 10%.”
http://www.signonsandiego.com/news/2011/dec/27/san-diego-home-prices-take-tumble/
Here’s an analysis from Patrick Newport, an economist from IHS Global Insight:
“Home prices are falling after stalling earlier this year … The stall earlier this year probably resulted from a slowing of the foreclosure pipeline after the media reported that lenders had cut corners in processing foreclosures.”
JtR: Pure garbage from idiots pontificating from the ivory tower, and being proliferated by lazy media types who don’t bother to check with reality.
The reality around here is tight supply, excess demand, and easy money. Prices should be going up if all you consider are the hard cold facts.
Fair enough. If we compare Case-Shiller to just using the Average for San Diego county YoY it’s really close. Case-shiller has -4.5%, County average is -4.2% for Oct. It’s probably safe to say that if a North County Case-Shiller index existed it would probably be close to the NCC average which was -1.6%. That proves what you’re saying, NCC is doing better and may go positive soon.
“Tight supply”
That tight supply is artificially created by the banks & govt stalling on foreclosures and short sales. If the flood gates were opened on the homes in limbo, there would definitely not be “Tight supply” and prices would drop further. This means the prices are being artificially kept high by artificially influencing supply/demand curve. In a fair market, home prices would drop until the Fair Market Price (much below today’s artificial price) is reached. Isn’t that right?
Yes you are right that its the widely-believed myth pushed incessantly by the media to scare you into thinking that there is no hope for housing. But it is wrong, at least around here where demand out-strips supply by a wide margin today. If they opened the flood gates it would whip the market into a frenzy with buyers scrambling to get a “bank deal.”