Core Concepts
Improving rental assistance allocation using ML to prevent homelessness.
Abstract
The content discusses the use of machine learning to prioritize individuals facing eviction for rental assistance based on their risk of future homelessness. It highlights the shortcomings of current reactive allocation processes and the benefits of a proactive approach. The study shows that ML models outperform baselines by at least 20% in identifying those at risk, ensuring fairness and equity. The research aims to inform evidence-based decision support tools in similar contexts.
Abstract:
- Rental assistance programs aim to prevent homelessness.
- Current funding distribution lacks consideration for future homelessness risk.
- ML system identifies vulnerable individuals with higher accuracy.
- Proactive approach reduces administrative burden and prevents evictions.
Introduction:
- Homelessness is a significant issue in the United States.
- Rising eviction rates contribute to the problem.
- Rental assistance programs are effective in reducing homelessness.
- Current programs lack efficiency, targeting, and equity.
Data Extraction:
- "Our ML system that uses state and county administrative data to accurately identify individuals in need of support outperforms simpler prioritization approaches by at least 20% while being fair and equitable across race and gender."
Stats
"Our ML system that uses state and county administrative data to accurately identify individuals in need of support outperforms simpler prioritization approaches by at least 20% while being fair and equitable across race and gender."
Quotes
"We partnered with Allegheny County, PA to explore a proactive allocation approach that prioritizes individuals facing eviction based on their risk of future homelessness."
"Using quasi-random variation in funding availability of a rental assistance program in Chicago, it was shown that individuals who called when funding was available were 76% less likely to become homeless."
"Our models identify 28% of people who are overlooked by the current process and end up homeless."