Examining housing discrimination across race, gender and felony history

Abstract

Those who have been convicted of crimes are subjected to a stigma that affects many aspects of their social lives. The “felon” label brings collateral consequences that make it difficult to obtain basic human needs, including housing. This study uses the audit method to examine the effects of race, gender, and criminal history on housing outcomes. Testers, exhibiting characteristics suggestive of race and gender and disclosing one of three offenses, placed phone calls to rental property owners across the Midwest to inquire about renting a property. We found powerful negative effects for those with a criminal record seeking apartments, regardless of whether the offense was sexual or drug-related. However, we found no differences between minority and non-minority testers. We explain these findings in the context of housing as an essential resource for formerly incarcerated individuals.

Keywords:

• Felony
• conviction
• race
• gender
• housing access

Acknowledgments

We would like to thank Jasmin Whitfield, Roy Ruiz, Mabel Mendez, Anthony LaRosa, Richard Pena, Cheradyn Pettit, Baboucarr Bah, Michael Madden, Kimberly Jones and Chanel Garcia for their help with this project. Their work was indispensable in completing the study. This work could not have been completed without their efforts.

Notes

1. The difference between the two is that landlords own the properties and rent them directly to tenants, whereas property managers operate and handle the business end of real estate property that someone else owns.

2. Many audit studies use paired testers – one tester displays or conveys the treatment/experimental manipulation and the other tester acts as a case control. Pairs are rotated with each case. For the current study we opted against paired testers because our initial research, in which we paired testers, and a follow up study in which we did not, both indicated near unanimous landlord and agent consideration in the control (no offence) condition (Evans, 2016; Evans & Porter, 2015).