Advanced Analytics for Child Support Programs
From reactive enforcement to proactive prevention — Part I
It is no exaggeration to say that predictive modeling and the mining of large databases using advanced analytics has become a major cultural trend that is changing the world in which we live. In areas as disparate as marketing, insurance, professional sports, entertainment, education and medicine, powerful statistical equations are routinely developed to anticipate various types of events and human behaviors. As consumers, such equations touch us almost daily: they can generate book and movie recommendations, insurance quotes, the catalogues and bank offers that come in the mail and the coupons we receive in the grocery store checkout line. And as taxpayers, we increasingly see predictive models used to make local state, and federal government agencies run more smoothly and economically. Today, many government agencies use predictive models to more effectively allocate resources, help identify opportunities and target decisions in such realms as tax collection, workers compensation insurance, criminal recidivism prevention, Medicare, Medicaid, unemployment insurance and Child Support.
The rise of analytics
Our world is rapidly changing because of technology advancements, enhanced data capture, increased advanced analytics ability and a shift in focus from the use of hindsight indicators to the use of predictive indicators (i.e., looking out of the front of the car versus looking out the rear view mirror). Advanced analytics loosely defined as the use of statistics, data and computing power to anticipate outcomes and provide new insights, is really nothing new. After all, banks have been using credit scores to sell loans and determine interest rates for decades; and actuaries have been plying their advanced analytical trades even longer. Still, analytics have entered the mainstream consciousness only recently. A number of factors are driving the buzz.
- Analytical Tools and Techniques
- Media Blitz
- informatics Evolution
Data as the foundation
Building a powerful predictive model can be a lot like building a dream house: it starts with a strong foundation. Data is the foundation of a strategic predictive modeling initiative: both the agency’s internal case data (the soil), as well as publicly available externally third-party data (the crushed gravel that helps strengthen the soil). Child support agencies across the country possess treasure troves of historical data on the cases they manage. States gather such extensive case-level information as monthly support obligation, employer information, asset information, arrears information, income, prior enforcement actions taken and so on. Though highly valuable, this data typically lies fallow rather than being used to improve efficiency and drive more productive decision-making. This presents an opportunity for forward-thinking agencies. The large amounts of high quality data available for analysis can be leveraged to effectively guide the panoply of enforcement measures that may be available to child support professionals, such as national and state new hire reporting, tax offset programs, financial institution data match, credit bureau reporting, license suspension and liens on real estate. In short, the data is rich, broad, available — and valuable. The challenge — and opportunity — of predictive modeling can be to build valuable decision support tools on this foundation.