Special Studies March 1, 2018
By Carmel Ford
NAHB Economics and Housing Policy Group
Report available to the public as a courtesy of HousingEconomics.com
Homeownership has been and continues to be an important driver of wealth creation for US households, with owner-occupied housing accounting for approximately a quarter of total household assets in 2016. However, the economic and financial benefits of homeownership have been uneven across demographic groups. For example, homeownership rates are known to vary by race, ethnicity, educational achievement, and other demographic characteristics, and may partly explain why Hispanic or Latino and black family average wealth lagged behind at just 21 percent and 15 percent, respectively, of white family wealth in 2016. 
To better understand the divergence in homeownership rates by race and ethnicity, this special study analyzes the geographic variation in homeownership rates of black and Hispanic or Latino households using Tableau interactive maps (each figure in the report links to a Tableau map showing geographic differences). The report identifies regional clusters where black and Hispanic or Latino homeownership rates are higher (or lower) than their national averages and takes a look at the demographic and socio-economic variables affecting those patterns.
Using these same variables, this report also explores homeownership rate gaps between white and black households, as well as between white and Hispanic or Latino households by employing a regression analysis. This analysis shows that, even when controlling for the selected demographic and socioeconomic variables affecting homeownership rates at the county level, black and Hispanic or Latino households still have lower rates of homeownership, suggesting that reasons beyond demographic and socioeconomic factors are preventing these groups from attaining homeownership at rates similar to that of white households.
Where is Black Homeownership the Highest (Lowest)?
According to the Census Bureau’s American Community Survey (2012-2016), the national black household homeownership rate was 41.9 percent, about 29 percentage points lower than the white household homeownership rate (71.0 percent) during the same time period. The first tab of Figure 1 shows black homeownership rates at the county level.
Click on image to link to Tableau map
Black homeownership rates are highest in the South and lower in the West, Midwest and the Northeast regions. Among the southern states, it is clear that counties in Virginia, North Carolina, South Carolina, Georgia, Alabama, and Mississippi have the highest black homeownership rates. When switching the map view to focus only on counties outside of metropolitan statistical areas (MSAs) (filter buttons to the right side of each map), it is apparent that counties outside of MSAs in the South have higher black homeownership rates compared to counties in MSAs. However, only about 15 percent of black households in the South live in counties outside of MSAs.
The White and Black Homeownership Rate Gap
Figure 1, on the second tab, shows the gap in homeownership rates between white and black households (white homeownership rate minus the black homeownership rate expressed in percentage points) from the 2012-2016 American Community Survey. White homeownership rates exceed black homeownership rates in almost all of the counties in the U.S., with few exceptions. In general, the smallest gaps between white and black homeownership rates are in counties in the South, while the largest gaps are in the Northeast region.
Of the few counties in which the black homeownership rate exceeds the white homeownership rate, almost all of them are in the South and most do not have a sizable number of black households, many with less than 250 in total.
Mapping Demographic Variables Affecting Black Homeownership
To help explain the geographic variation in black homeownership rates, the following variables are employed: price-to-income ratios (PTI), age, and marriage rates. These variables are used because of a previously completed regression showing that home values, income, marriage, and age are strongly associated with homeownership rates at the county level.
To construct a PTI ratio, the median home value in a county is divided by the county median household income. In this case, the median household income of black households is used. According to a Freddie Mac analysis, a ‘normal’ PTI ratio is 3.5, with higher values indicating lower affordability. Price-to-income ratios among black households are negatively correlated with black homeownership rates, or in other words, as price-to-income ratios increase, black homeownership rates decrease.
Figure 2, on the first tab, displays price-to-income ratios by county among black households, with blue gradients in the map signifying PTI ratios higher than 3.5. PTI ratios among black households are wide-ranging, going from a low of 1.08 in Webb County, Texas to a high of 30.0 in San Francisco County, California (removed from map because of extreme value; map only highlights PTI values up to 10.0). In general, counties with the lowest PTI ratios are in the South, where black homeownership rates are the highest, while counties in the Pacific West and the Northeast have the highest PTI ratios. When further teasing out geographic variation, it is clear that the counties with the lowest black household PTI ratios in the South are mostly outside of MSAs.
Click on image to link to Tableau map
The median age of the black population is positively correlated with black homeownership rates. When looking at the second tab in Figure 2 (gradients of blue indicate a median age older than 35), there are older black populations in southern states, especially in counties in Virginia and North Carolina, two states with relatively high black homeownership rates. There are also counties stretching into upper Appalachia (Kentucky, West Virginia, and Ohio) with relatively older black populations.
In the previously constructed model, higher marriage rates are strongly associated with higher rates of homeownership on the county level. The third tab of Figure 2 shows marriage rates among black households. Across the board, marriage rates among black households are low (national rate of 27.4 percent) compared to the rate among all households (48.2 percent). Even in the South, which has a relatively older black population, marriages rates are still very low.
Hispanic or Latino Homeownership
Where is Hispanic or Latino Homeownership the Highest (Lowest)?
The national homeownership rate among Hispanic or Latino households is 45.8 percent (American Community Survey 2012-2016), about 26 percentage points lower than the white homeownership rate. According to Figure 3, which shows Hispanic or Latino homeownership rates at the county level, rates are highest in the Southwest region of the US, especially in Texas, New Mexico, and Arizona. This comes with little surprise given that a large share of the total Hispanic or Latino population live in these three states (23.2 percent). The percentage living in these states is about the same share as in California (22.9), where homeownership rates among this group are generally much lower (signified by gradients of red). Besides the Southwest region of the US, a cluster of counties in the upper Midwest region as well as Florida have relatively high homeownership rates among Hispanic or Latino households.
Click on image to link to Tableau map
The White and Hispanic or Latino Homeownership Rate Gap
The second tab of Figure 3 displays the homeownership rate gap between white households and Hispanic or Latino households by county (white homeownership rate minus the Hispanic/Latino homeownership rate). When looking at the map, the “red counties”, or counties where the Hispanic or Latino homeownership rate exceeds the white homeownership rate, are scattered throughout the US. However, many of them are located in Texas and New Mexico, the same states where Hispanic or Latino homeownership rates are relatively high. The regional cluster with the largest gap between white and Hispanic or Latino homeownership rates is in the New England region, especially in Connecticut and Massachusetts.
Mapping Demographic Variables Affecting Hispanic or Latino Homeownership
To help explain homeownership rates among Hispanic or Latino households, the same set of demographic and socio-economic variables are employed: price-to-income ratios, age, and marriage rates. Figure 4, on the first tab, displays PTI ratios for Hispanic or Latino households by county. When using the threshold of 3.5, PTI ratios for Hispanic or Latino households are generally low in Texas (gradients of blue), a state in which homeownership rates are fairly high for this group. Although Hispanic or Latino homeownership rates are also relatively high in counties throughout New Mexico and Arizona, many counties in these states have higher PTI ratios among Hispanic or Latino households. For example, Taos County, New Mexico, has a Hispanic or Latino homeownership rate of 76.5 percent, but a PTI ratio of 7.0.
Click on image to link to Tableau map
Also displayed in Figure 4, on the second tab, is the median age of the Hispanic or Latino population by county, which is positively correlated with homeownership rates. The map illustrates that the Hispanic or Latino population is relatively young across the board, with most counties highlighted in red, signifying a median age of less than 35 years. However, several counties throughout New Mexico, Texas, and Colorado have relatively older Hispanic or Latino populations, the same counties where homeownership rates among this group are relatively high. The third tab of Figure 4 shows Hispanic or Latino household marriage rates, with gradients of red signifying a rate higher than 50 percent. The household marriage rates vary from region to region among this group. Counties with lower marriage rates are seen in New Mexico, Arizona and Colorado, states where homeownership rates are relatively high among Hispanic or Latino population. Marriage rates for this group are also low in the Northeast region. Marriage rates are relatively high for this group in California, Texas, and counties scattered along the east coast.
The previous sections provided a geographic analysis of homeownership rates among black and Hispanic or Latino households as well as of demographic and socio- economic variables spatially associated with the rates. However, a natural question arises from this analysis – do these variables wholly explain the homeownership rate gaps between white and black households, as well as between white and Hispanic or Latino households?
To answer this question, a county-level fixed-effect method regression model is employed, with dummy variables for black and Hispanic or Latino households used to show if homeownership rate differences still exist after controlling for price-to- income ratios, median age, and share of married households. Figure 5 displays the results. All of the variables included in the model are statistically significant; the model is also jointly statistically significant with the F-Statistic = 3.11.
The main finding from the model is that, even when controlling for the variables measuring income and home values (PTI ratios), marriage rates, and age, black and Hispanic or Latino household homeownership rates are still lower than homeownership rates for white households. More specifically, when compared to white households, black household homeownership rates are 22 percent lower, and Hispanic household homeownership rates are about 11 percent lower, holding all other variables constant.
This analysis shows that homeownership rates among black and Hispanic or Latino households vary from region to region. Homeownership rates are the highest among black households in southern non-metropolitan areas, while rates are highest among Hispanic or Latino households in the Southwest part of the US. Some of the variables shown to help explain these spatial patterns include income levels, home values (PTI ratios), marriage rates, and age. However, even when controlling for these demographic and socio-economic factors, statistically significant differences still remain in homeownership rates between white and black households, as well as white and Hispanic or Latino households.
4] Continuous variables in the model are logged – explanatory variables: Price-to-Income ratios, Share of Married Households, Median Age; and the dependent variable, homeownership rates