Individual Submission Summary
Share...

Direct link:

Who Is Captured by Eviction Data? Rethinking Housing Instability Measurement in Maryland

Friday, November 6, 3:30 to 5:00pm, Property: Boston Marriott Copley Place, Floor: 3rd Floor, Room: Boston University

Abstract

Eviction filings have become a widely used proxy for housing instability in both research and policy contexts. Administrative court data are frequently relied upon to estimate the scope of housing instability, allocate resources, and inform intervention strategies. However, the extent to which eviction filings accurately capture the broader and multidimensional nature of housing instability remains insufficiently examined. This study addresses this gap by critically assessing the validity of eviction filings as a population-level indicator of housing vulnerability, with a focus on Maryland, a state characterized by unusually high eviction filing rates and pronounced racial and spatial inequality. Drawing on principles of construct validity, this study evaluates whether eviction filings adequately represent the broader concept of housing instability or instead reflect a more limited and socially patterned subset of housing experiences. Using annual time-series data from 2000 to 2018, the analysis examines the relationship between eviction filings and a key measure of housing vulnerability: households threatened with eviction. The study employs a parsimonious modeling strategy, including lagged correlation, baseline regression, and models accounting for temporal persistence and shared trends. Additional diagnostics, including trend analysis and stationarity testing, are used to assess the robustness of observed relationships.Findings indicate that while eviction filings and housing instability measures exhibit a positive association in level-based models, this relationship weakens substantially once temporal dependence and shared trends are accounted for. In models incorporating a lagged dependent variable and a linear time trend, eviction filings are no longer statistically significant predictors of subsequent housing instability. Furthermore, analyses using differenced data show no meaningful relationship between year-to-year changes in eviction filings and changes in housing vulnerability. These results suggest that observed associations may reflect shared temporal dynamics rather than a stable or predictive relationship. Substantively, the findings raise important questions about what eviction data capture. In Maryland, high filing rates are driven in part by institutional practices such as serial filings, which may inflate administrative records without corresponding increases in displacement. Combined with well-documented racial disparities in eviction exposure, these dynamics suggest that eviction filings may reflect the interaction of structural inequality, demographic composition, and legal processes rather than serving as a comprehensive measure of housing instability.This study contributes to ongoing debates about the use of administrative data in social policy research, highlighting the need for more nuanced and multidimensional approaches to measuring housing vulnerability.

Author