Inspired by the United States Census Bureau’s graphic and animation showing the movement of the mean center of the U.S. population, we used Zillow data to calculate the mean center of U.S. home values.
First, we should clarify what we mean by “mean center.” The Census Bureau does a great job describing the mean center of the U.S. population, so we’ll use that as a starting point: “The center of population is the point at which an imaginary, weightless, rigid and flat (no elevation effects) surface representation of the 50 states…and the District of Columbia[1] would balance if weights of identical size were placed on it so that each weight represented the location on one person.[2]”
The mean center of U.S. home values can be described in a similar way. However, instead of weights of identical size with each representing the location of one person, we would have weights of size relative to home value with each representing the location of one home. Put more intuitively, we can represent each home’s value as a stack of physical dollar bills, place each stack in the location of the home it represents and let the weight of the stacks determine where the surface would be balanced.
The top panel of the interactive visual below shows the movement of the home value center[3] between 2000 and 2013.
The bottom panel of the interactive visual shows changes in the total market value of homes[4] by state. We see that California’s total market value drives the center’s movement along the east-west axis—specifically, the movement west from 2003 to 2006, east from 2006 to 2011 and back west from 2011 to 2013. To a lesser degree, Florida and California seem to influence movement north-south, namely the movement south from 2004 to 2006 and north from 2006 to 2008. (It’s difficult to disentangle the effects of the two states without deeper analysis.)
California’s strong influence is due to four major factors. First, it has a lot of homes: about 10 percent of the U.S. total.[5] Second, its homes are worth a lot: California’s total market value of homes is 22 percent of the U.S. total, which is greater than the rest of the West and almost as much as the entire Northeast.[6] Third, its large and pricey housing stock was subject to especially volatile price swings during the boom, bust and recovery of the housing crisis. The drop in California’s median home value from peak to trough was 42 percent, which is significant when compared to the U.S. drop of just 23 percent.[7] And fourth, it has high leverage due to its location west of most other states. Using the analogy from earlier, each weight (or stack of dollar bills) representing the value of a California home tips the imaginary surface more than it would if it were located closer to the center of the country.
In contrast, the Midwest doesn’t have much influence on the center’s movements. Midwestern home prices are generally lower and less volatile than elsewhere, and the region’s central location doesn’t allow for much geographical leverage.
Methodology
We first combined Zillow’s county-level home value averages[8] with the Census Bureau’s county-level housing stock estimates. [9] Multiplying these two metrics gave the total market value of homes by county.
To calculate the location of the center, we used a spherical model of the Earth.[10] Specifically, we converted the county-level centroids, expressed in the latitude-longitude coordinate system, into three-dimensional Cartesian space. We then calculated the total market value-weighted average separately for each of the three dimensions, giving us the three Cartesian coordinates for the U.S.-level total market value-weighted centroid (which represents the home value center). We converted the coordinates back to the latitude and longitude system by projecting the point onto the sphere representing the Earth.
To calculate state-level total market value, we simply aggregated county-level total market value to the state level.
[1] Alaska and Hawaii are excluded from the calculations for the mean center of the U.S. population made before 1960.
[2] Source: United States Census Bureau (link).
[3] For convenience, we’ll refer to the mean center of U.S. home values as the “home value center” and the mean center of the U.S. population as the “population center” for the rest of this post.
[4] The total market value of homes in a region is the sum of the value of all homes in that region.
[5] Source: The United States Census Bureau, 2013 estimate.
[6] Source: Data generated for this post, 2013 estimate. See Methodology section for more details.
[7] Source: Zillow. A public version of the Zillow Home Value Index (ZHVI) data can be found at zillow.com/research/data.
[8] This included almost 2200 counties representing 94% of all homes (according to Census estimates).
[9] July was used as the basis for the country-level home value averages to be consistent the time scale used for the Census Bureau housing stock estimates.
[10] Taking a weighted average of the latitude and longitude coordinates would lead to inaccuracies due to the curvature of the Earth. This isn’t a problem at small scales, but the United States is large enough to make these inaccuracies non-negligible.
