Estimating a population at-risk of homelessness: The roles of income and rent in four metropolitan areas.

Abstract

The extent of homelessness appears to vary widely across metropolitan areas. This study hypothesizes that the increase in the population of homeless persons is related to changes in both the supply of rental housing which is affordable to low income renters, and to an increase in the dem and by these renters for low-cost rental housing. These factors will vary across metropolitan areas. The extent of homelessness is proposed to be inversely related to the proportion of the metropolitan area population who are homeowners, and positively related to the proportion who are minorities. Changing demographic composition of the low income renter population is expected to contribute to the disproportionate representation of blacks and Hispanics, and increasing proportion of women and children, among those who become homeless. The data are the machine readable St and ard Metropolitan Statistical Area files of the Annual Housing Survey. Changes in the stock of low-cost rental housing and in need by low-income households for affordable housing within four metropolitan areas across three time periods are examined. These areas, Baltimore, Chicago, Houston and Seattle, represent a range in the extent of homelessness from 3 per 10,000 in Baltimore to 28 per 10,000 in Chicago as estimated by the 1984 HUD survey. Income and rent are all adjusted to 1986 dollars using the CPI-XI. Three levels of vulnerability are assessed among households with incomes below 125 percent of the poverty level. The vulnerable are those who pay in excess of 45 percent of income to rent; the very vulnerable pay in excess of 60 percent of income to rent. The severely vulnerable are those households who have less than $50 dollars per person per month remaining in residual income after paying rent. The population at risk is compared with the number and composition of homeless in individual metropolitan areas, as estimated by HUD and by other surveys, to determine how closely the vulnerable population parallels the literally homeless. Findings indicate that rent has risen beyond the rate of inflation in all four metropolitan areas. The result of regressing gross rent on positive and negative indicators of quality show that, in the aggregate, subst and ard quality does not reduce the price of housing to the extent that it had in the past. The population of vulnerable renters has risen in all four metropolitan areas, but at a rate which is consistent with the extent of homelessness in the four areas. Women and minorities are disproportionately represented among the vulnerable. Logit regressions of vulnerability indicate that over time, households with children have become significantly more vulnerable.