Housing Assistance and Eviction Prevention During Compound Crises: Evidence from the COVID-19 Pandemic

During economic crises, urban policymakers must assess whether housing assistance helps prevent

evictions. This study explores the link between housing assistance and eviction risk and how

household financial capacity influences this relationship. Using 207,203 observations from the

Understanding America Study during COVID-19 and applying recursive bivariate probit models

with instrumental variables, we find that housing assistance correlates with a 4.3% decrease in the

likelihood of eviction (a 15.6% relative reduction). Notably, among households with credit access,

this effect increases to 6.2%, or 44% more than the overall average. We observe a notable inverted-

U pattern: households with one to two credit sources see reductions over 39%, whereas those with

three or more sources experience much smaller effects, around 4.1%, or a nearly tenfold difference.

This challenges the idea that greater financial access always enhances outcomes. The mechanism

analysis suggests that this pattern reflects financial stability rather than over-leveraging:

households with three or more credit sources tend to be more creditworthy and less dependent on

assistance. The results hold across various estimation methods. Although our instrumental variable

approach has limitations, the consistent heterogeneity patterns across multiple strategies increase

confidence in the finding that housing assistance and moderate credit access work well together.

These insights imply that housing assistance programs could be more effective when combined

with financial inclusion efforts that focus on credit quality over quantity, with important

implications for preventing urban displacement during economic downturns.

Keywords: housing assistance, eviction prevention, financial capacity, instrumental variables, and

urban policy.

JEL Codes: I38, R21, R28, C26, H53.