Standard Affordability Measure Misrepresents Households Living in ‘Shelter Poverty,’ Unable to Meet Basic Needs

A paper in Housing Policy Debate,Shelter Poverty in Ohio: An Alternative Analysis of Rental Housing Affordability,” finds that the standard housing costs-to-income ratio measure of housing affordability, which is 30%, can underestimate the problem of affordability in low-income areas and overestimate it in high-income areas.

The study tests two measures of housing affordability by applying them to renter households in Ohio. The ratio-based measurement identifies households as housing cost-burdened when they spend more than 30% of their income on housing and utility costs. The residual-income approach identifies households as being in “shelter poverty” when they cannot meet basic non-housing needs after paying for housing. A household could be in shelter poverty even if it spends less than 30% of income on housing if it cannot afford fixed costs of food, transportation, and other basic needs. For this particular study, the author used the “Self-Sufficiency Standard for Ohio 2015” to determine minimum expenditure levels for non-housing needs (e.g., food, child care, transportation, health care).

The study uses 2012–2016 American Community Survey (ACS) microdata for Ohio. The author assembled data on 67,706 renter households that included rent, income, household size, age of household members, and location. These data were merged with data drawn from the Self-Sufficiency Standard about minimum expenditure levels. Rates of shelter poverty and ratio-defined cost-burdened status were calculated for all households. Affordability gaps—how much money would be required to bring households out of shelter poverty or cost-burdened status—were calculated for each household.

The analysis found a strong positive correlation between housing cost-burden and shelter-poverty measures. Most housing cost-burdened households were experiencing shelter poverty. In addition, however, 8,878 households (13.1%) were identified as experiencing shelter poverty but were not considered housing cost-burdened by the housing cost-to-income ratio, and 3,889 households (5.7%) were identified as cost-burdened but were not in shelter poverty.

The results varied depending on an area’s median income. In urban core areas of Cleveland, OH, where median household incomes are lower, shelter-poverty rates were 11.2% to 14.4% higher than housing cost-burden rates, indicating that the traditional 30% cost-to-income ratio underestimates unaffordability (overestimates affordability). The author suggests that the ratio method underestimates unaffordability because it does not account for the extent to which household spending on basic non-housing needs is fixed, regardless of income. In contrast, in the outer-ring suburbs of Cleveland, where median household incomes are two or three times larger, the cost-burden rates are 0.5% to 4.8% higher than shelter-poverty rates, indicating that the standard 30% ratio overestimates unaffordability (underestimates affordability).

The two measures also produce different estimates of the depth of the affordability gap. The shelter-poverty measure estimates that Ohio renter households would need to earn $14.9 billion more each year to close the rent gap, whereas the ratio method estimate is only $3.2 billion. The author takes the results to imply that prior analysis has underestimated renters’ financial challenges. The implication is that some households that spend less than 30% of income on housing are still unable to afford health care, child care, transportation, adequate food, or other non-housing needs. He calls for a complete national data set of Self-Sufficiency Standard estimates, so that the shelter poverty measure could be used nationwide.

“Shelter Poverty in Ohio: An Alternative Analysis of Rental Housing Affordability” is available at: