Abstract
During the past decade, housing markets across the United States experienced dramatic upheaval. Housing prices rose rapidly throughout much of the country from 2000 until the start of 2007 and then fell sharply during the next 2 years. Many households lost substantial amounts of equity during this downturn; in aggregate, U.S. homeowners lost $7 trillion in equity from 2006 to 2009. Aggregate home equity holdings had fallen back to 2000 levels by early 2009. Whereas this intense volatility has been well documented, there remain unanswered questions about the variation in experiences across racial groups, particularly among those who purchased their homes before the boom and kept them through the collapse of the market. Did this housing market upheaval widen the already large racial and ethnic gaps in housing wealth? Using the American Housing Survey, we analyze differences in the changes in home equity experienced by homeowners of different races and ethnicities between 2003 and 2009. We focus on homeowners who remained in their homes over this period, and find that blacks and Hispanics gained less home equity than whites and were more likely to end the period underwater. Black–white gaps were driven in part by racial disparities in income and education and differences in types of homes purchased. Latino–white disparities were most dramatic during the market’s bust.
Acknowledgments
The authors would like to thank Samuel Dastrup, Tina Park, and Claudia Ayanna Sharygin for their valuable contributions to this project. We are also grateful for the feedback given by Jaclene Begley, J. Michael Collins, Christopher Herbert, and Signe-Mary McKernan. Finally, we thank the Furman Center for Real Estate and Urban Policy for ongoing support in this and other work.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1. Flippen (Citation2004) argues that appreciation is lower over the long run in neighborhoods with larger minority concentrations and higher poverty rates.
2. Rugh and Massey (Citation2010) rely on this association in using racial gaps in subprime lending in a metropolitan area as an instrument for segregation.
3. There is a variable in the AHS that indicates whether the people living in the unit were also living in that unit in the previous survey wave, but we found that the variable had substantial error.
4. The U.S. Census Bureau top-codes home values using different methodologies each year.
5. Information about the mortgage terms for mortgages beyond the first two is insufficient to estimate remaining debt.
6. Although AHS collects information about the cost of replacement and additions, it does not identify housing units that have undergone major remodeling that might impact the value of the home. We define major remodels as replacement/additions made to the unit of a cost greater than 2% of the self-reported value of the home. We also estimated models including the 2003 value of the home as a covariate, which further controlled for unit quality, but decided to use baseline equity instead.
7. Racial differences in the percentage change of home value follow a similar pattern. Average value rose more dramatically in percentage terms for minorities than for whites during the boom, and fell more sharply during the bust. Across the whole period, average percentage changes in home value were largest for nonwhite homeowners in our sample.
8. Findings were insensitive to how the starting equity categories were constructed. Among white households in the sample, 151 were underwater in 2003, 618 had starting equity between $0 and $50,000, 399 had 50,001 to $100,000, and 446 had over $100,000. Among nonwhite households, 130 were underwater in 2003, 305 had equity between $0 and $50,000, 140 had 50,001 to $100,000, and 135 had over $100,000.
9. Housing units built in the 1940s gained significantly more equity, on average, than units built in other decades. No other coefficients for dummy variables indicating the decade in which the unit was built were significant. To conserve space, we do not show these coefficients in the table.
10. Baseline (i.e., 2003) equity is well correlated with 2003 value (0.60) and years between purchase and 2005 (0.30). However, the variance inflation factors in Model 6 were low: 2.32 for value, 1.92 for starting equity, and 1.25 for number of years since purchase of the home, so we are not concerned with multicollinearity biasing our estimates.
11. In results not shown, we estimated racial differences in home equity trajectories within samples stratified by starting equity position. Although these smaller samples resulted in less-precise estimates, they were consistent with Figure .
12. Average LTV among white homeowners, who were the least leveraged in 2003, fell from 0.62 to 0.54. Hispanic and Asian LTV declined from 0.73 and 0.75, respectively, to 0.57—almost on par with white households. Blacks had the highest LTV in 2003 (0.79) and 2007 (0.62).