Prof. Byrne & Colleagues Examine Relationships Between Homeless Rates & Community Factors
Although national analysis of homelessness has been the norm, these broad overviews of the homeless crisis do little to guide communities in proactive prevention policies. In a recent study by BU School of Social Work Professor Thomas Byrne, Senior Managing Economist at Zillow Chris Glynn, and University of Pennsylvania Professor Dennis P. Culhane, communities with similar factors impacting their homeless rates were grouped into peer clusters in order to develop and evaluate policy interventions. This approach aims to provide more effective opportunities to reduce and decrease the risk of homelessness. “Quantifying the association between homeless rates and covariates of a community is practically useful along two dimensions,” the authors explain. “First, it sharpens public focus on the social forces related to homelessness—leading to improved monitoring and intervention opportunities to help the most vulnerable citizens. Second, it provides a set of measurable objectives to guide public policy.”
“Putting in place effective and lasting policy responses to homelessness requires understanding what structural factors are driving the problem,” says Prof. Byrne. “This study is important because we show, on the one hand, that housing affordability is a key driver of homelessness and, on the other hand, the structural drivers of homelessness can differ depending on the characteristics of a community. These findings help underscore the need for both broad-based strategies to address widespread problems of housing affordability and more locally oriented responses as well.”
While the results of the study are an important first step in a more thorough understanding of homelessness causality, the authors stress that it is “not a substitute for the invaluable local knowledge of continuums of care coordinators and service organizations in addressing the needs of homeless populations in individual communities.” In addition, more studies using these methods must be done to better determine social policy to prevent poverty and homelessness.
Community Clusters Offer Unprecedented Balance of Healthy Data Size & Local Homelessness Factors
Previous attempts to quantify causes of homelessness have either been too broad or too specific to provide accurate guidance in local social policy. At one extreme, analyses use one parameter (i.e. the relationship between homelessness and housing costs), assuming that this one variable is responsible for the homeless crisis nationwide: “Assuming a single global parameter is rigid, and it ignores the possibility that local social structures mitigate (or exacerbate) the role that housing costs play in housing vulnerability.” Others studies forego similarities between communities altogether, and instead select unique parameters for each community. “As there is scarce data on the size of the homeless population in each community, it leads to imprecise estimates of model parameters,” the authors explain.
To find a reliable middle ground between these two extremes, this study groups similar communities who share the same parameters to increase the amount of data available without neglecting the local effects on poverty and homelessness. By identifying structural changes in the relationship between a cluster’s parameters and homelessness, communities can more easily evaluate the effectiveness of policy interventions.
The Three Dominant Community Clusters
In this study, the authors saw patterns emerge connecting homeless rate, housing cost, and poverty rate, which resulted in three dominate community clusters in the U.S.:
- The midwest, mid-Atlantic and southeast. Cluster one tends to have the lowest homeless rate, most affordable housing, and lowest extreme poverty rate.
- Most of New England, Florida, the mountain west and central United States. Cluster two has intermediate homeless rates and housing costs on par with the national average.
- Much of the west coast and large metropolitan areas on the east coast. Cluster three communities have, on average, the highest homeless rate, the least affordable housing, and the most poverty.
“Identification of distinct clusters of communities suggests there is potential value in implementing strategically differentiated policy interventions based on community characteristics which would be a departure from current approaches,” the authors explain. New approaches could result in higher prevention of poverty and homelessness in similar communities. “Prior research in this vein operated under the implicit assumption that pulling the same levers with the same strength and in the same direction will have an identical effect regardless of the community in question.” In fact, these practices could be detrimental to community efforts trying to reduce poverty and homelessness. The research “points to the potential need for multiple policy responses that target the needs and structural determinants of homelessness in individual communities.”
Next Steps
The implications of a data-supported approach that directly addresses a community’s unique needs could make way for more effective, compassionate, and realistic social policy: “Our results serve as an important first step in developing improved early warning systems for monitoring housing market conditions and forecasting their impact on the demand for homeless services.”
However, the authors stress that more research must be done to determine next steps. “Causal inference on the relationship between community level characteristics and homelessness is an important direction for future research,” say the authors. “While our results may not be directly used to determine policy, the inflection point analysis is important to improve homeless population monitoring and forecasting systems.” Encouragingly, the success of this study suggests that other community-based research could benefit from a cluster approach such as “crime, real estate transactions, and onset of rare disease [that] occur infrequently, and data is scarce at the community level.”