What a headline, “A computer model predicts who will become homeless in L.A. Then these workers step in.” For agencies struggling to improve benefit utilization, this Los Angeles Times story illustrates the power of using data and predictive analytics to identify populations that could use government benefit programs — but may not typically apply for them.
TransUnion recently published a report, Improving Government Assistance Program Access and Utilization, also highlighting opportunities to identify communities underutilizing available benefits based on patterns revealed by aggregated credit data. We found areas with higher levels of “credit underserved” individuals — a definition crossing multiple demographics — had 20%–40% lower participation rates in government benefits programs compared to other areas despite having similar eligibility rates. Agencies seeking to increase engagement with traditionally underserved communities would do well to consider the 38 million credit underserved Americans who provide a more inclusive picture of those struggling for equitable participation in opportunities to improve their livelihood.
How to use data and analytics to improve benefit utilization
Typically, the barriers agencies face to improve utilization are identifying eligible constituents; reaching and engaging those who are eligible; and verifying applicants’ identities to minimize fraud. Relevant and actionable data is often an obstacle to overcoming these challenges.
The team at UCLA California Policy Lab (which developed the predictive model used by Los Angeles County to predict those at risk of losing housing) evaluated over 85 million service utilization records on 1.9 million single adults from 7 agencies covering health services, benefits payments, law enforcement, and homeless services. The model also uses hundreds of risk factors to accurately predict someone who may lose housing. The sheer scale of this data would overwhelm most agencies.
Similarly, to identify communities likely underutilizing government benefits, we created a proxy based on millions of anonymized, TransUnion-aggregated credit records to find areas with higher levels of credit underserved individuals. The credit underserved are Americans who have between zero to two credit products and only one type of product (e.g., an auto loan) in their credit file. These are individuals with access to credit but do not use it. Our hypothesis was that areas with low credit utilization would have significantly lower participation rates in government benefits programs — a theory confirmed by our analysis.
The significant overlap of those who don’t use credit and those who don’t use government benefits highlights the ability, using aggregated credit data, to identify counties or ZIP Codes where eligible constituents may be underutilizing available benefits.
Well-timed outreach is especially important to improve benefit utilization
The Los Angeles County Homelessness Prevention unit uses their predictive model to target and prioritize outreach activities by case workers. This direct outreach can have immediate impact on someone facing the risk of losing their home, as evidenced in the LA Times article.
Similarly, aggregated credit data provides unique insights on the financial circumstances facing individuals in each county or ZIP Code. Understanding geography-specific credit trends can also help agencies and their communications partners improve the impact of outreach and education campaigns by targeting communities where individuals are likely to be eligible for (but not enrolled in) programs.
For example, the Chicago Housing Stability Dashboard, a tool developed by the University of Chicago Inclusive Economy Labs, uses aggregate credit data along with other data sources to highlight neighborhoods facing housing instability so agencies can target assistance more quickly.
Getting started to drive sustained benefit utilization through analytics
Timely and relevant data and analytics can help uncover insights into benefits-eligible populations, a critical component for increasing and sustaining program utilization over time.
While many jurisdictions may not have the resources of university labs or L.A. County, organizations like TransUnion offer the ability to leverage existing consumer data and analytical resources to help develop predictive models.
Using aggregated credit data for analysis to uncover areas of benefit-underserved populations for targeted prioritization is a great place to start. Program eligibility can be evaluated more deeply with predictive analytics to help agencies target constituents for outreach. For more information on how government agencies can use data to better engage with underserved communities and increase program participation, download TransUnion’s Improving Government Assistance Program Access and Utilization report.