Why median prices are misleading

We've been down this road so many times together, but let's go down again: median prices present a somewhat misleading picture of the market. In the current market, distressed properties
(think Palmdale and Lancaster) are selling, and homes in established neighborhoods (think Santa Monica, as in the photo) are not.  Median sales prices are thus skewed toward the distress. By how much?

Here's an exercise: In Palmdale, population 140,000, MDA DataQuick counted 319 single-family homes sold in September, or one home for every 438 people. In Lancaster, population 143,000, 366 homes sold -- or one home for every 390 people. But in Santa Monica, population 87,000, only 15 homes sold -- one for every 5,800 people.

There are two headlines here: Palmdale and Lancaster have an outsized influence on Los Angeles County median sales prices. They are weighting the overall county numbers toward lower prices and larger price declines (The median price in the county is now 31% below year-ago levels).  The second headline is worth noting: the residential real estate market in Santa Monica, one of the larger collections of high-priced homes in Los Angeles, was essentially frozen in September. Frozen. I can't imagine October will be much different.

--Peter Viles

Your thoughts? Comments? E-Mail story tips to Peter Viles.

Photo Credit: Michael McCreary

Comments

[MyLessThanPrimeBeef]

Not only are median prices misleading, but the media is misleading as well.

It's funny how no one won a Pulitzer exposing the risks of CDS's, SIV's and CDO's before we were ambushed by them.

And now, we are supposed to rely on the media to report on the election?

[patient_vulture]

Things that also skew the median up unrealistically is fraud and upgraded homes (by flippers, etc.).

I know of a specific case of fraud in which a $780k home was bought then sold for 1.2m 4 months later! Yes, it was fraud...buyer walked (well, 'ran') away. With no $$ down, all parties (appraiser, agent, mortgage broker, buyer/seller) split the $400k proceeds. I'm sure they were doing this on multiple properties and of course, many were doing it. Point is, that 1.2m sale is still registered as a sale and that skewed the median up, even though the bank later sold it for $650k.

That guy on 'Flipping out' bought a home in Los Feliz for something like $700k, put a good $500k into it and sold it for something like 1.5m....another case of medians skewing up.

During the boom there were a lot of remodeling/upgrading going on....and many of those properties were sold. It can be argued that as values are rising, medians are more skewed than when prices are falling....as fraud and upgrades won't be happening with falling values.

[Drew]

In addition to the "median price" explanation that you so clearly explain here, there is also the "median house" factor.

What exactly is the "median house"? In some communities, it's an older 2br/1ba cottage in rundown but fixable condition.

In other communities, it's a 3br/2ba house with twice the square footage and in notably better physical condition with "all (or most) mod cons."

And median prices -- when calculated on a community-by-community basis -- should reflect that difference.

As Peter suggests, when we substantially overweight the Palmdales and Lancasters as we seem to be doing, the county-wide median becomes so badly skewed that it becomes meaningless as a way to describe the market.

It's like a recent presidential poll that purports to show that Obama is 1 point ahead of McCain nationwide. A closer examination of the poll internals reveals that people who describe themselves as "fundamental Christians" were grossly oversampled in comparison to their actual proportion in the population, and that the pollsters did not correct for this oversampling.

[Drew]

And regarding the "frozen" market in Santa Monica (and presumably other higher-end markets) -- no one is selling in this market unless they HAVE to sell.

Or in the case of a friend of mine, is selling at what he knows will be into a distressed market for the express purpose of buying a larger home in the same "distressed" market.

Had he decided to wait for the market to improve to sell, he'd also be buying into a strengthening market.

But too many people are afraid of their own shadows in this market. Not that I blame them, but the term "frozen in fear" seems applicable here.

[Tim K.]

"median is misleading" is misleading, because you don't say over WHAT AREA.

Of course, if you include a very large population of disparate homes, you are not going to get a very clear picture of perhaps the areas *you* are interested in.

But that doesn't mean "the median" is wrong, it just means you need to pay attention to the values which applies to you. For example, Redfin's Neighborhood pages, which you yourself Peter pointed out to us, have excellent graphs of the price SOLD per square foot of not only specific cities, but even specific neighborhoods and zipcodes within the city! It's incredibly specific and very instructive. In some areas of Pasadena I'm watching, it's extremely clear what's happening - prices are falling by about 25% in 6 months. And when I actually examine specific homes that *I* want to buy, I can see the asking prices falling too. How much more accurate do you need to be?

[Rational Renter]

So by frozen, you mean f***ed. Sorry. But no sales = falling prices.

[Rational Renter]

Another thing, I don't understand discounting the median sales prices. Sure, the Palmdales of the city are dominating sales. But why do you think that is? You think there's just so much more demand to buy in Palmdale than there is in Santa Monica? Of course not - that's a ludicrous suggestion. The truth of the matter is, houses are moving in Palmdale b/c they are affordably priced. Until Santa Monica moves out of bubble territory, houses aren't going to sell.

What's so hard to understand about that?

[Sally]

Thank you, Mr. Viles! Of course the market in Santa Monica is frozen. Who out there can afford a $1.2 mil. shotgun shack? There can only be so many hedge fund managers and corporate lawyers in the world. The sad thing is that prices will have to come waaaay down before ordinary people are once again able to buy in the LA area, and no seller in their right mind is going to sell their house for anything approaching affordable.

[x-man]

Before a ball thrown into the air starts to come back down, it comes to almost a complete standstill at the top of its apex.

Witness Westside LA.

What's that Inland Empire? Down 50% in real terms? But that couldn't happen on the westside? NO WAY!

Those 4 million dollar homes in the 'Calles' of the Palisades, the one that sold for 1.2 six years ago, they can't fall back down to 1.6 can they? Not from 4 million to 1.6! They can't, right? RIGHT!

Because if they did, people would be underwater by seven figures! Think of what that would mean! Savings immolated! Or just mailing in the keys...

People in the top 1% of income in the United States being FORECLOSED on...coming soon to a PLATINUM TRIANGLE NEAR YOU

[Wooster]

Doesn't seem to skew the data that much. Both areas are not that big relative to the entire county. Regardless, it's affect was baked into the numbers on the way up as well, so let's not go throwing out the medians for the county. The medians for small areas, however, are definitely subject to the noise. You see that in Santa Monica and Manhattan Beach a lot where slow citywide sales paces and strength in the most expensive streets drives numbers up in the face of actual declines in the market.

[anonymous]

Peter, it would be interesting to see the mean, median and mode here. I think it would better illustrate that there are two markets right now. Icky foreclosures and decent homes. It would be nice to see the statistics broken out by foreclosure versus non-foreclosure.

I don't ask for much, do I?

[Cal]

Median price is just a stupid metric for homes. What was happening in late 2006 and early 2007 was the low end had stopped selling (in relative terms) which pushed the median up. As the prices on the low end kept falling and a few people decided to jump in places like DQ and realtor groups point to that saying "look the low end is selling more, it is skewing the median!". In actuality the median didnt reflect the falling prices that had been ongoing and now it looks like it "caught up" all at once.

As far as values for areas not selling, you don't know values you just know houses arent selling (again, relatively speaking) and prices are too high. Even indexes such as Case/Shiller can't help in situations where the market has no volume. Low volume of sales is telling you some of what you need to know, but it cant tell you everything.

You wont know true values until sales volumes and inventory return to somewhat normal (i.e. with little to zero distressed selling).

Right now the discretionary seller is the minority player in the market.

[Drew]

"But no sales = falling prices."

Actually, no sales = no data.

[Drew]

I think many people assume -- and I believe assume incorrectly -- that if prices fall by x% in one area, they will necessarily fall by that same x% in all areas.

This assumption, if indeed incorrect, leads to erroneous assumptions about the future of some of these markets.

[sfvrealestate]

Patient vulture, perhaps I misunderstood your point about upgrades. Most upgrades DO add value. It's not necessarily equal to the amount spent on the specific upgrade, however.

[AsafaJeez]

yes, this seems to be nonsense about the median. If the Antelope Valley skews these price drops so much, then why is there a 37% drop in the Valley media (as mentioned a few La Land posts ago) vs the 31% in the county. your logic is just so flawed without actual data to support it. next time, break up the numbers and figure it out.

[patient_vulture]

"Patient vulture, perhaps I misunderstood your point about upgrades. Most upgrades DO add value. It's not necessarily equal to the amount spent on the specific upgrade, however. "

Posted by: sfvrealestate
---------------

The point is the value upgrades add skews the medians up. Medians don't take into account upgraded properties and most forget that. Most are looking at medians as the appreciation of (existing) property values as a whole, yet during the boom many of these properties were upgraded and some of that 'appreciation' was due to the upgraded properties (and fraud).

And yet those skewed median values tack on 'additional' appreciation to properties that don't deserve it (non-upgraded, etc.) and the rest of us have to pay the price for that....while others (home owners) got to cash in extra on that if they sold in the bubble.

How much this is a factor? Hell if I know, but the last 8 years certainly must have been the biggest era for home upgrades EVER!

[bkl]

There is nothing misleading about medians, but some people misinterpret them. If the median of 6000 sales last year was $450,000 and the median of 5000 sales this year is $315,000, it means exactly what it means. Of the 6000 sales last year, half were for $450,000 or less; of the 5000 sales this year, half were $315,000 or less. The median fell by 30%, but that doesn't mean the most expensive house fell by 30% and doesn't mean the least expensive house fell by 30%. It also doesn't mean that the median in Santa Monica fell by 30% nor that the median in Palmdale fell by 30%.

Medians are used because the mean (average) is even worse. Ten houses sell for $200,000 and one sells for $2,000,000. The median is $200,000 and the average is $363,636. Which is a better representative measure of the market? Of a typical house that sold?

[Drew]

bkl: "There is nothing misleading about medians, but some people misinterpret them."

Hear, hear! (I'd say the median level of misinterpretation is about 70%.)

Have you ever considered a career as a epidemiologist?

[Laker]

If median sales price in LA county is $340,000 for example. That means that the middle point of houses sold is $340,000.
If that figure was $600,000 in 2007, it means that in 2007, half the houses sold for more, half for less.

The big take from this is that whatever is selling is what people can afford. As someone here suggested, the demand for housing in Lancaster is not greater than the demand in Santa Monica. The opposite is true. However, it is not about want to buy in SM, it is more like can afford, and can get a loan to buy in SM.
At the end of the day, prices in SM and other "nice" areas will have to give way on the price and decline, since on average, sales volume would compensate.

I highly suggest tracking the price per square foot. Property shark has a nice tabulation once you research a house, you can see the neighborhood was selling for median per sf at every year from today going back 10-20 years ago.
You could clearly see the 20 or 30% declines. Remember that this is based on actual sale prices, not refinances or zillow"remove 20% of the zestimate" bogus guesstimates.

[srla]

I'm not really sure what you point is here, Peter. Medians are the most reasonable measure to gage a market. Even a few hundred additional houses selling in Palmdale will likely have a negligible effect on the median, as the median is simply the price of the house in the middle -- if you were to line up all the sales by price. It's likely the houses near the middle range of this "lineup" would be somewhat close together anyhow, and thus by sliding the median lower or higher a few hundred spaces, the effect would also be minimal.

Much as the average is disproportionally influenced by very high priced sales, it is dragged down by a large number of low price sales. But the median holds up well even in the current scenario.

In wealthier areas, residents have the resources to hold out longer, but not indefinitely. And once they realize just how pervasive and profound the current correction is, even those with the resources will lose the incentive to hold out and begin to sell. This will incite a downward spiral, in which anyone needing to move will try to get out before prices fall even lower.

Peter, you seem unwilling to believe that the higher end areas will actually drop. But history as well as economic analysis tell us that the higher end areas will eventually fall too, and when they do, their decline is bound to be just as steep as the outlying areas.

Remember, prices in West LA and Santa Monica were 1/2 to 1/4 their current levels as recently as 7 years ago. Just because we currently have a certain familiarity with these price levels doesn't mean they are either economically justified or sustainable.

[syscom3]

Median price information should only be listed for zip codes, or a group of geographically coherent zips.

[shockg]

Peter, How dare you suggest the high end will not fall! Many here bet their home on it.

[It All Happens on the Margin]

I love shockg.

Hey shock: riddle me this, if Obama's "95% of Americans" are below $250k, who will buy all these $750k stucco boxes?

Yeah, I know. Keep telling yourself LA is full of movie stars and rich Chinese.

[Crash and Burn]

I think that's kind of like saying that your passenger driver's rear mirror is misleading because "objects may be further than they appear". The median is the best data you have, and of coarse it will be a little skewed in terms of things like foreclosures and such, but there will always be factors skewing it (and many that cancel each other out) and it won't be a huge error. Those values are the only real values you have and come from real data, unlike speculation like lack of foreclosures in Santa Monica skews the median.

And if you are so worried about it, just publish the Case-Schiler index data. There will be no bias there.