Abstract
In the early 2000s, many policymakers and researchers became concerned about suburban decline. The recent national subprime, foreclosure, and economic crises have intensified these concerns. In this study, I analyze the 2010 Neighborhood Stabilization Program 3, the 2005/2009 American Community Survey, and other databases with descriptive statistics and weighted least squares regression models. Differentiating among tracts in central cities, mature suburbs, and developing suburbs in the 100 largest metropolitan statistical areas, I examine what factors determine the Census tract foreclosure risk rate and what differentiates these factors. Results show that mature suburbs have foreclosure rates similar to central cities and that similar factors determine the neighborhood foreclosure risk rates among central cities and mature and developing suburbs to a different degree. These results demonstrate the need for place-based interventions.
Disclosure statement
No potential conflict of interest was reported by the author.
Acknowledgements
The author thanks the participants of the Urban Affairs Association (UAA) in April 2013 in San Francisco, California, for constructive comments. The author also thanks the two anonymous reviewers and the editor for helpful suggestions.
Notes
1. Mature suburbs are also often called first-ring suburbs (Rokakis & Katz, Citation2001); inner-ring suburbs (First Suburbs Consortium Housing Initiative, Citation2002); inner-ring cities (Advisory Commission on Intergovernmental Relations, Citation1984); inner suburbs (Sutker, Citation1974); first suburbs (Puentes & Orfield, Citation2002; Puentes & Warren, Citation2006); first-tier suburbs (Hudnut, Citation2003); older suburbs (Kotkin, Citation2001; Lucy & Phillips, Citation2001a; among others); older hubs (Listokin and Beaton, Citation1983); or mature suburbs (Listokin & Patrick, Citation1983), among other labels. These labels are used interchangeably in the literature (Hudnut, Citation2003), indicating that there is not a generally accepted definition.
2. “A lis pendens is a notice filed on public record to provide warning that the title to a particular property is in litigation. Lis pendens notices on mortgaged properties are filed by the lender holding the note after the borrower has defaulted, but before a court judgment of foreclosure has been rendered. These notices are commonly known as ‘pre-foreclosures’ […]” (Newman & Wyly, Citation2004, p. 59).
3. The author thanks one of the anonymous reviewers for this suggestion.
4. The dependent variable (SDQ_RATE) is calculated through a regression analysis based on the following model: SDQ_RATE = –2.211 – (0.131 * percent change in MSA OFHEO current price relative to the maximum in the past eight years) + (0.152 * proportion of total loans made between 2004 and 2006 that are high-cost) + (0.392 * unemployment rate in respective county in June 2008). Note the differences among (1) the dependent variable based on variables utilized by HUD, as enumerated below, and (2) the independent variables utilized by the author, enumerated above. HUD used a July 2010 extract of the rate of seriously delinquent mortgages at the county level provided by McDash Analytics to develop a predictive model using publicly available data for every Census tract in the United States. The model, based on a weighted number of mortgages in each county, predicted most of the variance between counties with regard to their rate of seriously delinquent mortgages with an R2 of 0.821. HUD’s model at the county level was as follows: 0.523 (intercept) + 0.476 unemployment change 03/2005 to 03/2010 (based on BLS LAUS) – 0.176 rate of low-cost high-leverage loans 2004 to 2007 (based on HMDA) + 0.521 rate of high-cost high-leverage loans 2004 to 2007 (based on HMDA) +0.090 rate of high-cost low-leverage loans 2004 to 2007 (based on HMDA) – 0.188 decrease in home values since peak (FHFA metro and nonmetro area). HUD then applied this model at the Census tract level to calculate the rate of mortgages that were seriously delinquent. See http://www.huduser.org/portal/datasets/nsp_foreclosure_data.html
5. While the US Bureau of the Census provides a breakdown for Asians by subgroup, regression models that include Asian subgroups lead to inconsistent results.
6. The author is thankful to reviewer #1 for making this point.