Exploring the relationship between housing choice vouchers and neighborhood housing dynamics in Louisville, Kentucky

ABSTRACT

Since the inception of the housing choice voucher (HCV) program, extensive research has focused on large metropolitan areas serving as demonstration sites for federal programs and concerning the impacts of dispersal on the residents and the types of neighborhoods where residents choose to locate. Little attention has been given to medium and small cities. Our analysis focuses on a medium-sized city, Louisville, Kentucky. We focus on the 170 census tracts of Louisville to examine the variations in neighborhood housing dynamics (i.e., median assessed housing values (MAVs), number of housing code violations, number of foreclosures, and crime rates) and the concentration of HCVs. Our findings indicate a significant negative relationship between census tracts with a greater concentration of HCV units and MAVs as well as a positive relationship with foreclosures. We find that HCVs have no impact on housing code violations or crime rates. A split regression analysis, however, shows that the presence of HCVs does not have a negative impact on the top half of Louisville’s neighborhoods with high property values. Finally, we show that HCVs, which are intended to deconcentrate poverty, concentrate poverty more in neighborhoods that already have a majority of low-income residents.

KEYWORDS:

• Housing choice vouchers
• housing values
• housing code violations
• foreclosures
• crime

Acknowledgement

Our thanks to Darlene Young Grove who did editorial for us and Paul Dries who replicated this research to confirm it for his Ph.D. dissertation and worked on the split regression analysis with us.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. The term “Section 8” is still commonly used to refer to the HCV program (Patterson & Yoo, 2012; Wang & Varady, 2005).

2. The findings presented here are consistent with other studies of Louisville, KY (see Ambrosius et al., 2010).

3. This variable was omitted from the models which contained the test variable percentage of HCV housing units in a neighborhood.

4. This control was omitted from the model when crime per 100,000 residents was used the dependent variable.

5. Due to space limitations, these tables are available upon request.

Additional information

Funding

This work was supported by the University of Louisville [Dept of Urban and Public Affairs and University of Louisville SUN http://sun.louisville.edu].

Notes on contributors

Wesley L. Meares

Wesley L. Meares is an assistant professor of poli–tical science and public administration in the Department of Social Sciences at Augusta University. His research interests include housing, economic development, sustainabil–ity, and urban politics and policy.

John I.Gilderbloom

John I.Gilderbloom is Professor of Planning, Public Administration, Urban and Public Affairs, and School of Public Health.  He was named in an international survey by Planetizen “as one of the world's top urban thinkers.”  He has seven books and 56 peer reviewed articles.