Home Values Are Up. Repeat: Up

Posted by: Peter Coy on November 27, 2006

Would you believe that home values are up?

According to Zillow.com, home values rose 4.8% in the third quarter in the 36 major metropolitan areas covered by its Zindex. That may come as a surprise if you’ve been reading about falling sales prices, but there’s a good explanation for the discrepancy. I think that while Zillow’s number may make things look a little better than they really are, the median sales price numbers from the Census Bureau and the National Assn. of Realtors probably make things look a little worse than they really are.

Obviously one difference is that the Zindex covers just 36 metro areas, not the whole country. But the more interesting difference is that Zillow’s index includes updated estimates of the values of all homes in an area, not just the ones that were sold. That’s important because the sales price indexes can be skewed by changes in the mix of homes sold in a period. If for some reason there are relatively few high-priced homes sold during a quarter, that will skew a sales price median downward.

This ain’t just some technical debate. Richard Berner of Morgan Stanley, who argues that housing will “rust not bust,” points out a big skew in the Census Bureau’s widely quoted figure of a 9.7% decline in median new-home prices from September 2005 to September 2006. As Berner points out, the sales of new homes priced between $300,000 and $400,000 fell 44% over the past year, while sales of those priced under $300,000 fell only zero to 20%, depending on the price bucket. In other words, the sales mix shifted to cheaper homes, and that pulled down the median.

Zillow’s Zindex isn’t the only measure out there that corrects for this mix problem. The Office of Federal Housing Enterprise Oversight index and the S&P Case-Shiller index do, too. But it’s a good one to pay attention to.

(One other thing. This blog has gotten lots of complaints from homeowners who say that the Zestimates for their own houses are wrong. Whenever a house is sold, Zillow checks how close its value estimate was to the actual sales price. It says that in 62% of cases, the estimate was within 10% of the actual sales price. Not bad.)

TrackBack URL for this entry: http://blogs.businessweek.com/mt/mt-tb.cgi/

Reader Comments

jeff

November 27, 2006 11:28 PM

Zillow will always....repeat always....inflate the value of houses in areas where the executives and employees of Zillow own property.

Repeat always. Why? Absolute power corrupts absolutely.

Cal

November 28, 2006 03:02 AM

How is the Zindex not affected by the mix of homes in a particular neighborhood? If they apply a sale to all similiar homes, and their are a bunch of those type of homes in that area, it will skew the zindex one way or another.

Also, zillow takes a very long time to get new homes added in, in my particular area (so cal) the lag time from construction completion to zillow addition is at least 18 months.

Personally, from what I've seen from zillow, I think it has quite a lag from current market conditions to local market prices. I dont think it is a "better" market indicator, just a different proprietary way of looking at the same data.

Sally

November 28, 2006 05:25 AM

I took an interesting tour of downtown Chicago condos last week. One condo is subsidized - yes! subsidized by the City of Chicago Housing, about 1/3 of their units. Plus the condo conversion was throwing incentives like 2 yrs free parking and one yr assessment paid. Some condos are even paying taxes for one year. This tells me that there are efforts to prop up the housing market. if it's happening in downtown Chicago, then it must be happening in all the major cities and surrounding suburbs.

John Beule

November 28, 2006 08:21 AM

It's also important to remember that Zillow tracks property values based on local government appraisals to determine your property tax. So while it might seem that home prices are going up...it's most likely going up because your local government has raised the appraisal value to collect more taxes.

Jan B

November 28, 2006 01:02 PM

Ummmm....HELLLOOOOOOO??? Have you read the headlines today?? Home prices have fallen sharply-(by the most in years), and every Economist worth their salt that (dosen't work for the Real Estate Industry)--is predicting that any tiny increase reported at this juncture is merely a "dead cat bounce"...

When you look at the big picture--at all the factors that are about to collide,
(Trillions of dollars of ARMs set to adjust in the next 2-3 years, foreclosure rates up by 23% in the 3rd quarter, horribly loose lending practices over the last 3 years, and the majority of American's Debt to Income ratios skyrocketing--you will realize that we have the makings of the perfect storm and are no where NEAR a bottom in this thing.

There are logical reasons for these little blips on the screen, but it would be nice to see you guys doing a little more research so that you can report more accurately about the big picture of the housing market!!

Tom

November 28, 2006 01:04 PM

Zillow always overestimates ALOT here in California. Houses are listed below the Zesitimate and sit for months and don't sell. Even realtors and real appraisors laugh about zillow. Houses on my street is listed at the Zestimate for 2 months now and has only see like 2 people come by.

lizziebeth

November 28, 2006 01:55 PM

John,

that's exactly why we aren't buying in Bradenton, florida yet. We found a great deal on a home, while checking into insurance and taxes, the county appraisers office told us that even if we pay xxxxx for a home, they can still charge us more for taxes. If the house was appraiesed during the overinflated real estate market of 2004, 2005, we're screwed even if we get the home at 2001 prices!

mark Jones

November 28, 2006 01:57 PM

explanation:
Zillow is highly unaccurate...takes it data from things like comps just like real estate agents
go visit a home on Zillow site and you'l see home improvements, views and other upgrades not taken into account by Zillow..poor obsolete, and unsubstantiated data like Zillow is VERY misleading!

ForceOfReason

November 28, 2006 03:25 PM

Neither Zillow nor the notoriously biased numbers from NAR reflect the fact that developers and many houseowners are offering significant sweeteners to move inventory. Car leases, flat screen TV's, a few months worth of mortgage payments have all been offered as essentially stupid gimmicks to lure stupid buyers. These giveaways are NOT factored into the median sales price, and therefore blur the TRUE situation with housing prices. Don't be confused by what the media reports. House prices are sinking much faster than being reported.

Rob

November 28, 2006 04:40 PM

Is Zillow's record of coming within 10% of the sales price 62 per cent of the time really "not bad"? Seems pretty bad to me. If I sell my house for $500k, that means that Zillow valued it anywhere between $450k and $550k (plus or minus 10%), but only 62% of the time. The other 38% it would have valued it below $450k or above $550k. Is that range of valuations useful?

jf.sellsius

November 28, 2006 05:02 PM

John Beule makes a great point. Also, public sales data fed to zillow lags up to 6 months. Thus, the zestimate values are lagging 6 months.

The Zindex Housing Report is seriously flawed BECAUSE it includes "unsold" homes of unknown error rates (error rates are determined only after a sale is made). Combine this with the fact that
zillow does not disclose the percentage of sold vs. unsold zestimated homes included in the figures. In areas with low sales the Housing Report is heavily weighted with zestimated homes & the report becomes simply a Zestimate Report. Unreliable, in other words.

Also consider their use of "median" error rate of 7.5% That means 50% of the zestimates (in the Zindex & therefore in the Housing Report) are above 7.5%. How much above 7/5%? ---No one knows because zillow doesn't say. It can be 10%, 100%, 1,000%. Have Zillow tell us what the "average" error rate is in this half of the median and we will be amazed at the error percentage, IMO.
And while you're at it, ask them of the 40% nationally which are 10% off the selling price--what is that average error rate.
To be fair to consumers, zillow should disclose the average error rates in the higher median for all areas it reports. "Median" error rate alone is not enough since there is, by definition, a 50% chance my home may be in the group whose error rate is above the median.

In those metro markets (NY,NJ,PA,Fl) & for the majority of states having admitted high error rates, the average error rate in the upper median will be even higher. Any report that is too heavily weighted with these zestimates is suspect.

Remember, "a" starting point is not a "good" or "best" starting point.


D Bush

November 29, 2006 12:12 PM

Zillo has been WAY off on values in the Phoenix metro area as some of my clients bring this crap to me before they list. Also, 10% on a $300K house is $30K. I wouldn't call that very accurate of a comparison basis to price your house. You might as well throw zillow, the comps and most everything else out the window. Do you want to know what a house is worth? Put it up for sale and lower the price each month until you get an offer. It's worth what someone is going to pay for it. Don't belive the hype!

David G from Zillow.com

December 1, 2006 03:31 PM

Hi, it's David G from Zillow.com

ALL - Thanks for the feedback -- it helps us prioritize improvements to our website. Please remember that Zestimates are a starting point for researching what houses are worth -- they are not intended to replace the advice of an experienced specialist with a deep knowledge of your local real estate market.

Jeff - There is no such bias in Zillow's algorithms. It's technically impossible to manually manipulate Zestimates on individual properties. I respectfully request that you correct your opinion.

Cal - The Zindex does not look at the same data that median sales reports do. The Zindex considers the values of ALL homes, not just those that have sold. Because all homes are considered every time we report these metrics, the "mix" is constant or rather, there is no "mix" in the Zindex that could skew the median result as what happens when the rate of sales in sub sectors of the market change.

John Beule - While tax assessments are an input to Zestimates, they're just one attribute we consider in calculating a valuation. Our analysis of sales transactions should offset any weirdness in tax assessments.

Mark Jones - views (and their types) are actually included in the public records used to calculate Zestimates. Click on "show all facts" for the complete list of details that are taken into consideration by Zillow. That said, the public records we use as our primary input to Zestimates are not always complete and can be inaccurate. Because of that, we now allow homeowners to publish updated facts along side the public record. I'd encourage you to do so if there are major features about your home that we're not aware of.

jf - We measure the time it takes us to receive and import public sales data. It's around 3 weeks, not 6 months. Further, since we use median value in the Zindex, inaccurate Zestimates are statistically excluded from these trend measurements. You seem to be confusing the median calculation with the mean or average -- which would be thrown off by Zestimate inaccuracies.

jf.sellsius

December 6, 2006 11:29 PM

Hello David.

The time for Zillow to receive & import is NOT what I mean by lag. The sale documents do not get to the public database for many months, depending on the county---in NY county, where I practice real estate law, it takes anywhere from 6-8 months for the data from my sale to make it to the database. THIS is the lag. NO ONE gets the public data until it's filed in the county records. The fact that you can retrieve it then & run it through your algorithm in 3 weeks or 1 day is beside the point and meaningless. It means that your data is STILL 6 months and 3 weeks stale. It would help if you addressed my points accurately.


I completely understand the difference between median and average --they wouldn't let me get my degree in Finance without it. The reason Zillow uses median is because it looks better than average, for public consumption. Median is simply a middle value---half the errors are below it and half are above it. I'm asking you about the half above it. Do you know the average of the error rates in this upper 50%---of course you do. Will you ever reveal it? Of course not. Why? We all know why. Read Rob's comment again.

Interesting that you failed to address my point on the weighting of the report. If both sold and unsold houses are in the Report it would be relevant to know the proportion of solds v. unsolds in the Report so consumers could determine how heavily weighted they are with zestimates of unknown accuracy. An extreme example to focus you on the point: If an area had zero sales, the Zillow Housing Report becomes a Zestimate Report. True? And if that Report is for an area your zestimates are highly inaccurate, that report is highly inaccurate. An analogy perhaps to the point --if pure drinking water is actual sales and impurities are unsold zestimates, I'd like to know the mix in the glass of water before I took a swig. How about you?

As always, interesting to read your rebuttals, DAvid. I bet you're a good dancer.

dan

December 11, 2006 11:15 AM

Over the last five years, how have commercial property values change in the south of Chicago?

Jan B

December 11, 2006 09:22 PM

This NY Times article is a very interesting read on WHY we are nowhere near a bottom in the current market....

http://www.nytimes.com/2006/12/06/business/06leonhardt.html

The happenings over the last 3 years have plainly been unprecidented, and we WILL have consequences of the same magnitude.
Fasten your seatbelts folks!!

Keith

December 14, 2006 07:53 PM

I sold my home on Long Island on May 22nd 2006, for $440,000, Zillow's Market assessed value is $290,000. How long does it take to update there database, over 7 months and counting! (Real Dynamic system they have!) Hey but at least they got the data right from 1994, and the price I paid for the house back then.

Thank you for your interest. This blog is no longer active.

 

About

BusinessWeek editors Chris Palmeri, Prashant Gopal and Peter Coy chronicle the highs and lows of the housing and mortgage markets on their Hot Property blog. In print and online, the Hot Property team first wrote about the potential downside of lenders pushing riskier, "option ARM" mortgages and the rise in mortgage fraud back in 2005—well ahead of many other media outlets. In 2008, Hot Property bloggers finished #1 in a ranking of the world's top 100 "most powerful property people" by the British real estate website Global edge. Hot Property was named among the 25 most influential real estate blogs of 2007 by Inman News.

BW Mall - Sponsored Links

Buy a link now!