Market Health Index Methodology

Posted by: Skylar Olsen    Tags:      Posted date:  December 13, 2013  

The Market Health Index, an ordinal measure ranging from 0 to 10, illustrates the current health of a region’s housing market relative to other markets across the country. For example, if a metro area has a value of 8 on the Market Health Index, the metro is healthier than 80% of all metro areas covered by Zillow in the US. By construction, the index says nothing about the degree of health; it cannot be claimed that a region with an index value of 8 is twice as healthy as a region with an index value of 4.

The Market Health Index is formed from ten different metrics. To capture the recent and continued rebounds from housing value troughs we include the month-over-month and year-over-year change in ZHVI, the percentage of homes selling for a gain, as well as the percentage change in the Zillow Home Value Index (ZHVI) forecasted for the coming year.

Areas where home values are skyrocketing are often areas where inventory is restricted and homes are selling in record time. Listings languishing on the street are also not indicative of a healthy market. For a given region, we begin by taking the homes sold in the last 90 days and estimating the median number of days between listing and contract signing. We then find the historical days on market: the median in a region over the entire history of home sales since January 2010. The difference between these two values tells us if homes are now selling faster or slower than in the past.
However, when we discuss the difference between the current days on market and the historical days on market, we want to know how unusual that difference is for that area. The expected days on market could fluctuate greatly in some regions while remaining steady in others due to the structural characteristics of the markets. Metros with cold winters and hot summers, for instance, would exhibit greater seasonality and so larger fluctuations in days on market. Metros with heavy investor participation in the real estate market and a history of frequent local bubbles would also exhibit more volatility in the time it takes to sell a home.

To control for these structural features, we need some measure of the usual deviation from the historical median in each region, the median absolute deviation (MAD). The days on market MAD is the median absolute difference between the days on market from individual sales and the historical days on market in that region. Now that we have an idea what is a usual fluctuation from the average, we can take the difference between current days on market and historical days on market, divide this difference by the MAD, and we’ll get a standardized version of the current days on market relative to the historical average to be compared across metros. Once we take the absolute value of this standardized days on Zillow, we can interpret the largest values to correspond to those metros that are selling either much, much slower than usual, or much, much faster than usual, neither of which is very healthy. Values near zero correspond to metros whose transactions are taking just as long as they always have, steady as a rock.

To incorporate the most painful component of the housing bubble burst – negative equity leading to delinquency and the spread of foreclosures – we include the percentage of mortgage holders with negative equity; the percentage of mortgage holders delinquent on their loans; the number of foreclosures out of 10,000 homes; the percentage of sales composed of previously foreclosed homes,;and our estimate for the number of foreclosures out of 10,000 still held by banks, i.e. unsold REOs.

The Market Health Index is created separately for metros, counties, cities, ZIP codes and neighborhoods.

For each of the ten metrics used in the Market Health Index, we assign each region a score along a continuous scale from 0 to 10, where 10 corresponds to the healthiest value and 0 to the least healthy among all regions in the US with available data. To ensure sufficient data availability, we only retain regions with at least 5 scores available. We take the average of these scores and then rescale the average to range from 0 to 10. This becomes the Market Health Index.


About the author
Skylar Olsen
Skylar is a Senior Economist at Zillow. To learn more about Skylar click here.