Association for Public Policy Analysis & Management

Housing cost burden among renter households has typically been evaluated in the United States as the share of household income spent on rent and utilities; if this amounts to over 30%, the household is said to be experiencing housing cost burden. Due to federal government policy and custom, this has become entrenched as the primary measure by which one determines if “too much” of one’s income is expended on housing.

The ease of this standard calculation provides a rough sense of who is paying more than they can reasonably afford. However, it has been argued in the literature – primarily by Kutty (2005) and Stone (2006) – that this is a flawed measure. This is because not all household expenses are infinitely scalable to income; there is a minimum amount of money that a household must spend to ensure that all of its members remain healthy and productive, no matter how little income the household earns. It is, therefore, imperative to determine whether one’s housing costs compromise their ability to meet these other needs. This is known in the extant literature as “shelter poverty,” or a “residual measure” of housing cost burden.

The Self-Sufficiency Standard, computed by the University of Washington, generates estimates of this minimum level of spending to sustain a household without government assistance for most states. This is done at the county level for hundreds of household configurations. This information is merged with 2012-2016 American Community Survey (ACS) microdata to determine whether a household’s housing costs exceed their household income less necessary non-housing expenditures. Using this approach, it is possible to compute rates of shelter poverty, as well as average and aggregate amounts of additional income that would be required for households in shelter poverty to no longer be cost burdened.

Once computed for all Public Use Microdata Areas (PUMAs) in the state of Ohio, this information is utilized in two ways. First, it is compared to the established ratio measures of housing cost burden, as published in Census Bureau ACS tables. It is hypothesized that there will be more cost burdened renters with larger affordability gaps using the shelter poverty measure than the 30% ratio measure, and that this disparity will be larger in PUMAs with larger shares of various disadvantaged populations. Second, these cost burden measures are tested for correlation with individual and PUMA-level demographic, economic, governmental, and housing stock characteristics.

Ultimately, an understanding of how housing costs affect households with fixed essential expenditures is developed, as well as how these vary with demographic and geographic characteristics. The normative question of which method of evaluating housing cost burden is preferable to policymakers is explored, with the potential implications of shifting legal definitions toward a shelter poverty approach receiving a preliminary discussion, as the impacts on U.S. affordable housing programs of such would be substantial.