Early Intervention to Prevent Persistent Homelessness: Predictive Models for Identifying Unemployed Workers and Young Adults who become Persistently Homeless

112 Pages Posted: 9 May 2019

See all articles by Halil Toros

Halil Toros

Economic Roundtable

Daniel Flaming

Economic Roundtable

Patrick Burns

Economic Roundtable

Date Written: March 20, 2019

Abstract

Two new predictive screening tools that are based on analyzing records of over one-million people who experienced homelessness have been placed in the public domain by the Economic Roundtable. The two groups targeted by these tools are low-wage workers who have just lost their jobs and youth entering adulthood who will become persistently homelessness within the next three years. While most people in both groups don’t become homeless at all or are able to get out quickly, eight percent of each group become persistently homeless.

The tools are highly accurate. For the top 1% of people with the highest probability scores, the models can predict with 81% accuracy which newly unemployed workers will become persistently homeless, and with 72% accuracy which young adults will become persistently homeless, nine times more accurate than random selection.

Predictive tools can solve the difficult problem of telling apart people who will be homeless only a short time and those who will be homeless a long time. These two groups look much alike at the onset of homelessness. The common face of homelessness is someone who has lived on the sidewalk for a long time, but there was a first day of homelessness for that person, when he or she was more hopeful, less damaged, and looked much like more fortunate counterparts who found early exits. These predictive screening tools can accurately tell these two types of individuals apart.

This is a front-end approach that jumpstarts the current model of progressive engagement where progressively more help is given to individuals as they remain homeless longer. If individuals become chronically homeless, they are offered permanent supportive housing, if any of these scarce units are available. The new tools prevent chronic homelessness by targeting people who are most likely to stay homeless and helping them early on when there is far less economic, social, medical, and legal wreckage in their lives, and exiting homelessness costs far less.

Keywords: Aid Recipients, Algorithm, Behavioral Health, Cost Avoidance, Data Integration, Disabilities, Early Intervention, Employment, Health, Homelessness, Hospitals, Jail, Job, Justice System, Los Angeles County, Mental Illness, Opportunities, Probability, Predictive Analytic, Predictive Model, Prevention

JEL Classification: C15, C22, C32, C33, C51, C53, C55, D63, H11, H51, H53, I18, I31, I32, I38, R21, J23, J64, J71, J78,

Suggested Citation

Toros, Halil and Flaming, Daniel and Burns, Patrick, Early Intervention to Prevent Persistent Homelessness: Predictive Models for Identifying Unemployed Workers and Young Adults who become Persistently Homeless (March 20, 2019). Available at SSRN: https://ssrn.com/abstract=3370634 or http://dx.doi.org/10.2139/ssrn.3370634

Halil Toros

Economic Roundtable ( email )

315 W. 9th Street, Suite 502
Los Angeles, CA California 90015
United States

Daniel Flaming (Contact Author)

Economic Roundtable ( email )

244 S. San Pedro St., Ste. 506
Los Angeles, CA 90012
United States
2138928104 (Phone)

HOME PAGE: http://economicrt.org/

Patrick Burns

Economic Roundtable ( email )

244 S San Pedro, Suite 506
Los Angeles, CA 90012
United States

HOME PAGE: http://economicrt.org

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
189
Abstract Views
851
rank
232,532
PlumX Metrics