From REO to Ruin: Post-Foreclosure Pathways and the Production of Decline in Detroit, Michigan

ABSTRACT

Although much has been written about the localized impacts of foreclosed properties, few studies have examined the role of the main actors handling mortgage-reverted properties, particularly the parties responsible for their disposition. Fewer still have examined these trends in historically stable but hard-hit neighborhoods where owner practices are implicated in current conditions. This study examines the likelihood of the U.S. Department of Housing and Urban Development and the government-sponsored enterprises selling real estate owned homes in the historically stable neighborhoods of Detroit, Michigan, to investors, as well as the likelihood of tax foreclosure following sales to homebuyers and investors. Whereas federal entities were less likely to sell homes to investors, all parties sold a high percentage of homes to investors. Once sold to an investor, the probability of tax foreclosure is extremely high. These results suggest federal and non-federal entities alike are associated with destabilizing and dispossessory outcomes that irreversibly altered these neighborhoods.

KEYWORDS:

• foreclosure
• neighborhood change
• federal housing agencies

Acknowledgments

I thank Lan Deng and Margaret Dewar for comments on earlier drafts of this article. I also thank the three anonymous reviewers for their constructive comments. All errors remain my own.

Disclosure Statement

No potential conflict of interest was reported by the author.

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.

Notes

1. I drew on CoreLogic for 2005–2008, the Wayne County Register of Deeds (WCROD) for 2008–2013, and CoreLogic again for 2014. I possessed data for imperfectly overlapping years for both data sets, and I drew on the full extent of available WCROD data as they are the source material for commercial data vendors.

2. I drew on WCROD data for for 2005–2015; from records of properties offered at the annual tax auction 2014–2016 from Loveland; and, for 2017–2018, records of properties classified by the Wayne County Treasurer as likely to be foreclosed a few months prior to auction, also obtained from the Loveland Technologies website. Although not all of the properties classified as likely to be foreclosed were ultimately repossessed, they exhibit three consecutive years of tax delinquency without enrollment in a repayment plan, indicating disinvestment.

3. Although I was able to capture tax foreclosures on REOs prior to an observed sale, in this article I only consider tax foreclosures subsequent to the first REO sale, to examine the interaction between seller and buyer in tax foreclosure outcomes.

4. I also searched for sales to state and local government entities and nonprofits. These acquisitions were small in number (less than 1% of sales in strong neighborhoods), so I excluded them from the analysis to focus on the contrast between homebuyers and investors.

5. For each tract I calculated REO concentration by dividing REOs by the number of mortgageable properties, drawing on the 2006–2010 ACS estimates. Mortgageable properties is the sum of attached and detached one-unit structures, one half the number of units in two-unit structures, two-sevenths the number of units in three- to four-unit structures, and the number of owner-occupied units in structures with five or more units (Ellen et al., 2014).

6. To test for spatial dependence, I first estimated nonhierarchical probit models with census tract fixed effects. Following Amaral, Anselin, and Arribas-Bel (2013), I then conducted postestimation tests of spatial dependence among residuals using a trio of tests adapted from the Lagrange multiplier and Moran’s I tests of spatial autocorrelation. To conduct these tests, I constructed a spatial weights matrix selecting REO sales occurring within 500 feet of a given REO sale and located within the same tract as neighbors. I assigned a weight to each neighbor based on the inverse of its distance from the REO sale of origin. None of these test statistics was significant at the 90% confidence level in any model, so I failed to reject the null hypothesis of independence among error terms.

7. I excluded suburban Wayne County from the analysis to retain a degree of homogeneity within each suburban sample, although the results are not substantially altered with the inclusion of REO sales in this location.

8. REO starts are calculated relative to the number of housing units in one-unit detached structures in 2000. Values are aggregated from the tracts included in the regression analysis (included in the Appendix).

9. Many large buyers acquired properties using various LLC names, although in many cases one or two names account for the majority of acquisitions.

10. In 2010, HUD introduced the Distressed Asset Note Program (DASP) to auction pools of seriously delinquent home loans in danger of foreclosure (Edelman, Zonta, & Rawal, 2016; Goldstein, 2016). Freddie Mac created a similar auction program in 2014, followed by Fannie Mae in 2015. Through early 2015, HUD alone had sold roughly 100,000 loans through the DASP program (Perlberg, Gittelsohn, & Benson, 2015). In June 2017, Fannie Mae completed its tenth nonperforming loan sale.

Additional information

Funding

This work was supported by the Robert Kahn Fellowship for the Scientific Study of Social Issues, awarded by the Institute for Social Research at the University of Michigan.

Notes on contributors

Eric Seymour

Eric Seymour is an assistant professor in the program in urban planning and policy development at the Edward J. Bloustein School of Planning and Public Policy at Rutgers University. His research focuses on transformations in postcrisis housing markets and their implications for the housing security and health of low-income and minority populations.