Varying Effects of Housing Policy: Rural–Urban Differences In the Impact of Hud Program Expansions on Child Maltreatment

Background/Purpose

Prior research suggests that child maltreatment risk—and the effects of social policies on maltreatment—may differ by rurality. However, relatively little is known about whether the effects of U.S. Department of Housing and Urban Development (HUD) programs vary by rurality. Using county-level panel data, this study examined whether associations between state-level HUD program expansions and county-level child maltreatment report (CMR) rates differed by rurality in the United States. Specifically, we assessed four dimensions of HUD policy: participation rates among eligible populations, program coverage beyond households below 50% of area median income (AMI), housing availability (HUD-assisted units per eligible population), and utilization (occupancy rates of HUD-assisted units).

Methods

We assembled a county-by-year panel dataset covering all counties in the 50 states and DC from 2014−2023 (N=10,450). Linear mixed-effects models were used to estimate associations between within-state changes over time in HUD policies and changes in county-level CMR rates. Models adjusted for between-state differences in long-term average HUD policy levels, year fixed effects, rurality, county socioeconomic/demographic characteristics, and county/state-level random intercepts. All HUD policy measures were entered simultaneously in a fully adjusted model to estimate their independent associations. Interaction terms between each HUD policy measure and rurality were further included. All models were weighted by county child population.

Results

Associations between HUD program expansions and CMR rates varied by rurality, with protective associations concentrated in large urban counties or stronger there than in small urban and rural counties. Increases in HUD-assisted housing availability were significantly associated with reductions in CMR rates in large urban counties (β=−3.94, p<.001), but not in small urban (β=−0.52, p=.394) or rural counties (β=0.26, p=.767). Higher utilization of available HUD-assisted units was also significantly associated with lower CMR rates only in large urban counties (β=−0.26, p<.001), with no significant associations in small urban (β=−0.03, p=.740) or rural counties (β=0.11, p=.326). Expansions in program coverage beyond households below 50% of AMI were significantly associated with lower CMR rates across all rurality categories, with somewhat stronger associations in large urban counties (β=−2.28, p<.001) than in small urban (β=−1.03, p<.001) and rural counties (β=−1.70, p<.001). Unexpectedly, increases in HUD participation rates were significantly associated with higher CMR rates in large urban (β=2.61, p<.001), small urban (β=1.51, p<.001), and rural counties (β=2.46, p<.001).

Conclusions/Implications

State-level HUD program expansions were associated with lower CMR rates for some policy dimensions, but these associations varied substantially by rurality and by type of housing support. Increases in housing availability and utilization were linked to lower CMR rates primarily in large urban counties, whereas broader program coverage beyond the lowest-income households was associated with lower CMR rates across all geographic contexts. The positive association between HUD participation rates and CMR rates should be interpreted cautiously, as participation may capture underlying housing need rather than protective program effects. These findings underscore the importance of distinguishing among dimensions of housing assistance and suggest that expanding housing resources alone may not produce uniform benefits across urban and rural communities without attention to local implementation capacity and need.