Fraud and Data Analytics

How Predictive Analytics are Discovering Fraud, and Saving the Government Money

Public Benefits programs are often a target for criminals who have stolen or false identities. They illegally obtain, trade and/or sell government benefits. This results in a great deal of wasted funds. According to the Governing Institute, the federal government estimates that $17.5 billion is wasted annually on improper payments.

Although the use of online services is very helpful and handy for most of us, unfortunately, identity fraud has also increased due to this.

Luckily, this is finally being addressed. The government is now using predictive analytics and integrating external datasets to better identify individual’s identities, as well as determine who is at high risk of fraud. For example, upon initial verification steps, if an identity is at “low risk” then it merely has to pass through low-risk indicators; otherwise it will be routed for further authentication.

By introducing external data, the government has implemented a multi-factor authentication process. This allows for the focus to be on fraud preventions rather than fraud investigation. Furthermore, predictive analytics helps investigators sort through fraud referrals, so they don’t waste their time on “fruitless investigations.”

By using external data, the government also places less reliance on self-reported information (which can of course be reported inaccurately). It helps them to discover things like hidden bank accounts, undisclosed earners, changes in family and household composition and even relationships to known criminals.

Teaming Up

One way they are accessing external data is by teaming up. Several states, including Mississippi, Arkansas, Florida, Georgia and Louisiana all compare their data within the NAC repository. This way they are able to determine if an identity, or an individual is receiving benefits in more than one state.

Predictive analytics help the government determine what factors they need to keep an eye for, that may suggest that fraud is more likely. They have found that ownership of luxury vehicles and of more than one property, are indicators of fraud.