PHILADELPHIA -- Families are more likely to be driven into homeless shelters by increased unemployment and by hikes in rental-housing costs than by welfare reform or by the occurrence of substance abuse or disabilities in heads of households.

Those are the findings of a Philadelphia study by Dennis Culhane, professor of social welfare policy at the University of Pennsylvania.  He and his team found a significant association between a rise in unemployment and the rate of families use of homeless shelters.  Increases in rental housing costs also had a direct relationship to the rate of shelter admissions, he said.

The study, published in Cityscape: A Journal of Policy and Development Research by the U.S. Department of Housing and Urban Development, also looked at other factors, including welfare reform.

Culhane's team used data collected by the Office of Emergency Shelter and Services, Philadelphia's central management agency for homeless emergency services.  To see if shelter utilization changed significantly after March 1997, when major welfare reform was implemented in Pennsylvania, they looked at the number of families admitted by family size, race, age of household head, income, household-head disability and average length of stay.  The researchers also checked these factors after three, six, nine and 12 months of welfare-reform implementation.

"There was an expectation that the number of families seeking shelter would rise after welfare reform," said Culhane, who examined families' shelter-utilization patterns before and after the laws were changed.  "We found no change."

The Penn researchers looked at whether an increase in shelter admittance could be tied to an expected decline in the state welfare caseloads as the new Temporary Assistance for Needy Families guidelines took effect.  Under TANF, lifetime benefits are capped at five years, and those receiving benefits have two years to find employment before being cut off.

"We found the state caseloads decreased without any corresponding increase in shelter admissions," Culhane said. "The results from this analysis fail to support the existence of any significant relationship between TANF caseload dynamics and admissions to and exits from the shelter system."

The researchers also found a small but negative relationship between self-reported substance abuse in the household heads and shelter admission, as well as a small but positive relationship between household heads with a disability and shelter use.