Wednesday, June 23, 2010

Decline in our house values: worst in the City?

It seemed like a good idea at the time. A few years ago I signed up on (the house values site) as the owner of my Madison Park abode. One of the side benefits of claiming home ownership on the site is that Zillow sends you periodic updates on the supposed value of your property (Zillow calls these Zestimates). So, for many a month after signing up I received email notices stating how much my house had risen in value since the last report.

This was all well and good—actually rather encouraging—until the inevitable happened. The local real estate market started its inexorable downward turn somewhere in mid-2007, and the valuation news started to get grim. The emails, unfortunately, have the Zestimates in the subject line, so there is no opportunity to treat them the same way you might have dealt with that 401(k) statement you got in the mail when the stock market was tanking (throwing the envelope unopened into a drawer or straight into the garbage, I mean). No, you don’t even have to open these Zillow missives in order to find yourself on the receiving end of the bad news.

So it was all right there in all its gut-wrenching immediacy when I noticed a Zillowgram in by inbox last week: “Zestimate for [my address] decreased 1.1%.” Since I’d received an email less than a month earlier along similar lines (down .7%), it was beginning to look like a trend. I had been hoping that the market had bottomed out—and several times earlier this year I actually received new Zestimates stating that my house had increased in value. So what happened? Is Seattle’s market still declining while many other cities have begun moving up? Is Madison Park out of line with the rest of the City? Or is it just me?

I decided to do some research, but the news was not good. In Zillow’s estimation, Madison Park’s houses have on average declined in value more during the past year than any of the 85 or so other Seattle neighborhoods the site tracks: down 19.9% year over year as of April 2010. Our neighborhood was followed by Hawthorne Hills (down 13.6%), View Ridge (down 11.6%), and Madrona (down 10.6%). This is in comparison to the Zillow Home Value Index (ZHVI) for Seattle as a whole, which was down only 3.3% year over year.

Stan Humphries, Zillow’s Chief Economist, notes that Seattle’s upper-tier market has not been behaving like the lower two tiers, with house values declining more in neighborhoods such as Madison Park than in lower-priced neighborhoods such as Greenwood (which is actually up .6% year over year, according to Zillow). Seattle’s market, he says, is also not acting like the markets in many other US cities, where the upper tier has been outperforming the lower market in recent quarters. Indeed, a recent article in the Wall Street Journal (“Luxury Sales Bounce Back” 5/28) reports that “in some markets, sales of high-end homes [have returned] to levels not seen since the boom.” San Francisco is cited as having a robust upper-tier market, for example.

With regard to Seattle’s high-end market, it’s hard to say what’s going on. But with regard to Madison Park in particular, Humphries offers some solace, suggesting that we take a longer-term view: “In terms of year-over-year decline, Madison Park has the highest among Seattle neighborhoods, but this is because it had less depreciation for the first year or two after the peak than many other areas (i.e., it was initially better, now worse—but the cumulative fall from peak is similar).”

He notes that from the market peak in September 2007, Madison Park’s houses have declined 28.6%--the worst performance of any Seattle neighborhood alright, but well in line with many other Seattle neighborhoods, including Windermere, Magnolia, Laurelhurst and Leschi. And several eastside neighborhoods performed significantly worse than Madison Park in Zillow’s analysis: Hidden Valley (Bellevue) was down 37%, North Bellevue was down 36.3% and Downtown Bellevue was down 33.5%, perhaps due in part to a very soft condo market. Medina, Bridal Trails, Meydenabuer, and Overlake each declined at levels in line with Madison Park’s. So, apparently we can take consolation from the fact that we are just part of a general upper-market downturn for the Seattle area.

Nevertheless, Madison Park's position looks a little bit frightening on the chart Zillow provided, which shows the fall of various neighborhoods since the peak (click to enlarge). For those who aren't into reading graphs, we can simply note that Madison Park is towards the bottom--and the bottom is not a good place to be.

If you don't like the implications of this analysis, you may want to take a pot shot at Zillow and its ZHVI process. Many realtors certainly do (one recently exclaimed “Who cares about Zillow?” when I mentioned the big downturn for the neighborhood that Zillow had reported). Realtors, in fact, often find themselves having to react to Zillow’s home value information since both buyers and sellers are increasingly using the site as a valuation tool. Although Zillow itself cautions about reading too much into any single-home Zestimate, that doesn’t stop buyers from discounting a selling broker’s higher opinion of a house’s value if Zillow is lower. Nor does it cause sellers to necessarily lower their prices to what a broker may argue is a number more in line with the market than Zillow’s.

I asked Humphries about this often-negative realtor attitude. He responded that “the accuracy of Zestimates in Madison Park is pretty good. About 50% of sales are within 10% of the Zestimate and about 80% are within 20% of Zestimate.” To back this up, Humphries provided me with a chart showing that for 68 recent Madison Park sales there was a “very high” correlation between the Zestimate and the actual sales price. The correlation was much higher for houses sold at under $2 million that at the higher level.

At least one Seattle realtor is dismissive of Zillow’s accuracy claim: “Saying that you’re within 20% of being right 80% of the time. How is that okay?” he asks. In a market like Madison Park’s, being 20% under or over the actual value might mean the difference of hundreds of thousands of dollars, he notes. And 20% of the time, by Zillow’s own admission, the Zestimate is off by more than 20%. Doesn’t this lack of precision throw into question Zillow’s data for changes in the neighborhood values?

Not according to Humphries. “The Zillow Home Value Index,” he notes “is an aggregate based on the individual Zestimates. In the aggregation process, the individual errors of Zestimates tend to cancel each other out since, statistically, we ensure that there are likely to be just as many valuations that are above the actual sale price as there are valuations that are below the actual sale price. This makes the ZHVI a very robust indicator of local home values.”

Robust, maybe--but definitely gloomy. And the bottom may not yet be in sight if you believe a report last month from Goldman Sachs stating that Seattle housing prices will fall another 22% on average by the end of 2012. Humphries discounts the Goldman prediction, believing that the methodology they employed is flawed. We can only hope he’s right.

Fortunately, there is also a slight bit of good news to report. Redfin recently concluded that Madison Park during May had the highest median selling price of any neighborhood in Seattle and the second highest for King County as a whole. Feeling better?

[Note Zillow’s Home Value Index for Madison Park does not include Broadmoor, which was also not tracked as a separate neighborhood by Zillow. Broadmoor is included in Redfin’s analysis of Madison Park.]


  1. zestimates are not very good, most people in the mortgage industry use cyberhomes as a quick and dirty estimate, not zillow. That being said, no doubt any larger decline for Madison Park has to do with a trend you touched on in a previous post - the tendency to build 4000 sq ft homes on 4000 sq ft lots. These types of homes are typical of a bubble, and fall further and faster during a downturn. Check out the sale history of 3805 E McGilvra St - old house torn down in 2005, a 3050 sq ft house put up on a 3000 sq ft lot in 2006, sold for 1.5M, recently sold for 1.05M, a 30% decline in 3 years.

  2. So true - who would want a mansion without a garden?!

    That said, this is the typical post-bubble ditch. Belts get tighter everywhere, the smart money seeks out bargains instead of baubles and starts pricing in future decreases in value (as opposed to the bubble times "it's going to appreciate at least 20% every year".

    Prices will get worse and then, eventually, better. Ultimately, the only real losers are those who can't wait to sell for a decade. So - everyone, don't worry, you'll be fine.


Note: Only a member of this blog may post a comment.