Home Foreclosures as Early Warning Indicator of Neighborhood Decline

Problem, research strategy, and findings: Ideally, planners would intervene in neighborhood processes before substantial forces of decline have gained momentum. Unfortunately, currently there is no guidance about which neighborhood indicators forecast future neighborhood changes. This study seeks a neighborhood early warning indicator that is readily available, frequently updated, and that predicts with substantial foresight and accuracy future changes in key aspects of neighborhood market processes and quality of life. We search over a range of neighborhood indicators for one or more that clearly lead others in a temporal sense, instead of lagging behind them. We employ Granger causality tests with indicators related to public safety, housing market, and economic activity that are based on data from Chicago neighborhoods between 1998 and 2009. Our key finding is that only one indicator analyzed, completed home foreclosures, systematically precedes other neighborhood indicators but does not systematically follow after them. All other indicators are enmeshed in complicated, often mutually causal temporal patterns, which render them inappropriate for forecasting. We conclude that completed home foreclosures is an appropriate early warning indicator of neighborhood decline in several dimensions. Takeaway for practice: Planners interested in identifying imminently emerging problems in their constituents' neighborhoods should regularly acquire and map data on home foreclosures as soon as they become available, and then undertake compensatory actions as quickly as feasible. Keywords: foreclosuresneighborhood indicatorsGranger causalityneighborhood decline Acknowledgments The opinions expressed in this paper are the authors and do not -necessarily represent those of MDRC, the Foundation, or their -respective boards of trustees. We thank Nathanial Roth for his excellent research assistance and the anonymous reviewers for their many constructive suggestions on earlier drafts. Sonya Williams (sonya.williams@mdrc.org) is a research associate in the low-wage workers and communities policy area at MDRC. George Galster (george.galster@wayne.edu) is Distinguished Professor of Urban Studies and Planning at Wayne State University. Nandita Verma (nandita.verma@mdrc.org) is a senior associate in the low-wage workers and communities policy area at MDRC. Notes 1. Several cities tout the existence of websites that permit citizens, community groups, and other analysts to download up-to-date, local information providing richly textured portraits of their neighborhoods over time (see, e.g., CityNews Chicago; Neighborhood Planning for Community Revitalization-Minneapolis; Philadelphia Neighborhood Information System; Neighborhood Knowledge Los Angeles; Providence Urban Land reform Initiative). Although some of these websites proclaim that they provide an early warning system for predicting imminent neighborhood changes, in fact, they leave it to the user to draw inferences from the data about the future or offer arbitrary and unproven weighting schemes for aggregating indicators into some notion of forthcoming changes. Unfortunately, such neighborhood early warning systems have typically not been tested in a statistically rigorous way for their predictive power and thus we do not discuss them here (Dewar, Basmajian, Alter, & Law, 2005; Myott, 2000 Myott, E. 2000. Developing and analyzing: The neighborhood early warning system in Hamline-Midway Minneapolis, MN NPCR. -Retrieved from http://www.cura.umn.edu/sites/cura.advantagelabs.com/files/-publications/NPCR-1143.pdf [Google Scholar]). 2. Hipp (2010 Hipp, J. 2010. A dynamic view of neighborhoods: The reciprocal relationship between crime and neighborhood structural characteristics. Social Problems, 57(2): 205–230. [Crossref], [Web of Science ®] , [Google Scholar]) also found, however, that more crime in 1990 not only predicted more concentrated disadvantage in 2000 but the reverse as well, thereby supporting theoretical expectations about mutually causal relationships. 3. Early investigations probing annual changes in a variety of neighborhood indicators were undertaken by Galster, Cutsinger, and Lim (2007) and Lim and Galster (2009), but they did not attempt to estimate predictive models or identify early warning indicators. 4. These works have recently been challenged by Kirk and Hyra (2012 Kirk, D. and Hyra, D. 2012. Home foreclosures and community crime: Causal or spurious association?. Social Science Quarterly, 93(3): 648–670. [Crossref], [Web of Science ®] , [Google Scholar]), however, who argue that the foreclosures-crime relationship is spurious and disappears when other neighborhood characteristics are controlled in the statistical models. 5. A 10-year, $47 million effort still under way at this writing, NCP is a comprehensive effort to engage community-based groups to attack multiple problems simultaneously in education, workforce development, housing, social services, and public policy. Managed by the Local Initiatives Support Corporation of Chicago (LISC/Chicago), NCP focuses its efforts on 14 neighborhood areas in Chicago with varying challenges. 6. With multiple incidents, the report is classified in the Uniform Crime Report category of the most serious crime (generally, the crime with the highest potential penalty). Note that these are police reports and do not reflect later adjudication of the incident (e.g., an assault recorded on the initial report as a criminal act later adjudicated as justifiable self-defense is still included). 7. The Metro Chicago Information Center accomplished the data collection, aggregation, and standardization tasks for this research, under contract to MDRC. We recognize that these crime data have shortcomings. Unreported crimes are not included and the underreporting rate is likely neither random nor constant across neighborhoods. Nevertheless, these are the sorts of data that have been used in prior scholarship (see the review in Raleigh & Galster, in press Raleigh, E. and Galster, G. in press. Neighborhood disinvestment, abandonment and crime dynamics. Journal of Urban Affairs, [Google Scholar]) and are likely to be available to planners. 8. The first indicator can be thought of as a proxy for the total mortgage capital flow associated with in-moving, home-buying residents; increases in this flow represent a higher valuation and desirability for the neighborhood. The second indicator is a proxy for both changes in capital gains by homeowners and incentives to maintain and improve properties. 9. HMDA data are used instead of sales price data due to the relatively longer time series available; HMDA data was available starting in 1992, whereas the sales price data was only available as two, incongruent time series that covered a more limited period. The HMDA and sales price data exhibited the same trends for the times in which we had overlapping coverage. Galster et al. (2005 Galster, G., Hayes, C. and Johnson, J. 2005. Identifying robust, parsimonious neighborhood indicators. Journal of Planning Education and Research, 24(3): 265–280. [Crossref], [Web of Science ®] , [Google Scholar]) demonstrated the value of HMDA data as a source for constructing neighborhood indicators. 10. Foreclosure filings that are resolved in other ways are not classified as completed. Both filed and completed foreclosure data exclude ownership transfers that occur as the result of financial distress (short sales or deed-in-lieu-of-foreclosure transactions). 11. Unfortunately, neither filed or completed foreclosure indicators give a precise estimate of when a foreclosed property becomes vacant and then (possibly) is occupied again. This is potentially problematic because it is primarily during the period of vacancy when a foreclosed home presents the greatest risk to neighborhood safety and physical upkeep. But, because our two indicators likely bracket the desired figure, we conducted our analyses in parallel using both. 12. Community Reinvestment Act data are less complete and comprehensive compared with the HMDA data, which do limit their utility in analyses of community change. In particular, there were several changes in coverage and scope of the data (i.e., banks required to report on commercial lending) that might mask or accentuate temporal relationship among indicators. 13. The use of "causal" in the name of the test is a reference to the particular statistical relationship, not to be mistaken for conceptual causality or causal mechanisms. This test has rarely been applied to neighborhood analysis; for two applications, see Freeman, Galster, and Malega (2006) and Batabyal (2011 Batabyal, S. 2011. Temporal causality and the dynamics of crime and delinquency. Atlantic Economic Journal, 39(4): 421–441. [Crossref] , [Google Scholar]). 14. Strong associations between foreclosures and crime also have been found by Baumer, Wolff, and Arnio (2012); Cui (2010 Cui, L. 2010. Foreclosure, vacancy and crime, Pittsburgh, PA: University of Pittsburgh. (Department of Economics working paper)[Crossref] , [Google Scholar]); Ellen et al. (2012 Ellen, I., Lacoe, J. and Sharygin, C. 2012. Do foreclosures cause crime?. Journal of Urban Economics, 74: 59–70. [Crossref], [Web of Science ®] , [Google Scholar]); Goodstein and Lee (2010 Goodstein, R. and Lee, Y. 2010. Do foreclosures increase crime?, Washington, DC: Federal Deposit Insurance Corp. (Center for Financial Research Working Paper 2010-05)[Crossref] , [Google Scholar]); Immergluck and Smith (2006b Immergluck, D. and Smith, G. 2006b. The impact of single-family mortgage foreclosures on neighbourhood crime. Housing Studies, 21(6): 851–866. [Taylor & Francis Online], [Web of Science ®] , [Google Scholar]); Katz et al. (2013 Katz, C. M., Wallace, D. and Hedberg, E. C. 2013. A longitudinal assessment of the impact of foreclosure on neighborhood crime. Journal of Research on Crime and Delinquency, 50: 359–389. doi:10.117/022427811431155[Crossref], [Web of Science ®] , [Google Scholar]); Stucky, Ottensmann, and Payton (2012); and Williams, Galster, and Verma (2013). The important role of foreclosures in leading decadal changes in neighborhood racial composition and other characteristics was forwarded by the seminal empirical work of Baxter and Lauria (2000 Baxter, V. and Lauria, M. 2000. Residential mortgage foreclosure and neighborhood change. Housing Policy Debate, 11(3): 675–699. [Taylor & Francis Online], [Web of Science ®] , [Google Scholar]) and Lauria and Baxter (1999 Lauria, M. and Baxter, V. 1999. Residential mortgage foreclosure and racial transition in New Orleans. Urban Affairs Review, 34(6): 757–786. [Crossref], [Web of Science ®] , [Google Scholar]). Finally, Anenberg and Kung (2012); Harding, Rosenblatt, and Yao (2009); and Immergluck and Smith (2006a Immergluck, D. and Smith, G. 2006a. The external costs of foreclosure: The impact of single-family mortgage foreclosures on property values. Housing Policy Debate, 17(6): 57–79. [Taylor & Francis Online], [Web of Science ®] , [Google Scholar]) have identified substantial home sales price discounts as a result of proximity to foreclosed homes. 15. Filed, rather than completed, foreclosures are conceptually a better candidate to serve as a leading indicator, particularly in places such as Illinois where the foreclosure process is fairly lengthy. While filed foreclosures would provide a longer lead time in relation to changes in other indicators of neighborhood quality of life, the analysis indicated that the filed foreclosures indicator had both lead and lag (i.e., reciprocal) relationships with the other indicators, which means that the model does not support using the filed foreclosure indicator as a leading indicator.