Zillow’s Look at Third Quarter Real Estate Performance

Today, Zillow released our third quarter real estate market reports for five metro regions (Seattle, San Francisco, Los Angeles, Phoenix and Miami) and an additional report covering the top 36 metro areas studied this quarter. Click here for all six reports.

San Francisco and Phoenix area real estate experienced a drop in values in the third quarter, relative to the prior quarter, while LA real estate values were essentially flat, quarter-over-quarter.  Seattle and South Florida, while slowing slightly, continue to show robust appreciation, significantly outpacing the majority of the nation.  The U.S. report gives you insight into the top 36 metro areas, showing that median home values in the top markets are still on the rise, measuring average appreciation of 4.8 percent year-on-year. While this is still considered healthy appreciation by many standards, it’s a dramatic slowdown from the double-digit appreciation many homeowners have come to expect in recent years.

For more information on our methodology, continue reading…

As with last quarter, we are using the Zindex as our indicator of general home price levels in an area.  Most analyses use the median sale price of actual transactions as the key indicator of home price levels.  While popular and certainly feasible for Zillow, our analyses led us to believe that median sale prices can be distorted by the changing mix of homes that were sold at any given time. For example, the high-market in an area might slow while the low-end market continues to see robust sales.  In this case, one will be left computing the median from amongst the predominantly lower priced homes and will arrive at an estimate of the general price level for a community that is lower than the “true” price level.

The Zindex, in contrast, looks not just at the limited number of sales that take place in an area but at valuations made on all homes in the area (the Zestimates). Having produced a valuation on every home, it’s relatively straightforward then to capture the “true” price level for the community of homes by taking the median of these Zestimates. 

But wait…isn’t there error in the Zestimate and wouldn’t this distort the Zindex? In our testing, we’ve found that roughly the same number of errors are above the actual sale price as are below. When all the Zestimates in an area are aggregated into a Zindex, these estimation errors tend to cancel each other out to produce a very stable general price index for a set of homes. In short, our choice of the Zindex as the basis for our evaluation of real estate markets is about more than just another “Z” branded product.  It’s about us trying new ways of looking at real estate in order to unlock value for consumers.   

Anyway, check out the reports and as always, let us know what you think.