New Mexico's Preventive Analytics Strategy Keeps Unemployment Insurance Fraud at Bay

Since the state's Workforce Solutions Department launched a strategy that marries predictive analytics with behavioral science techniques, it has seen a 40 percent reduction in the likelihood to commit fraud.

by / May 3, 2016
New Mexico Department of Workforce Solutions Google Street View

In nearly every industry, effective communication is critical. Without it, confusion ensues, processes slow down, quality falls short and costs are more likely to shoot through the roof. That’s exactly why the New Mexico Department of Workforce Solutions turned to predictive analytics, which significantly reduced unemployment insurance fraud and improper payments. 

When the department sat down to strategize how to improve overwhelming fraud and overpayment rates, officials realized that about 95 percent of payment issues were traced back to smaller, procedural issues related to customer communications regarding work search, earnings and separation issues. And predictive analytics was a solution that met many of their needs, said Deputy Cabinet Secretary Joy Forehand.

“We went back to look at what we can do that won’t cause harm to the eligibility process. We also couldn’t take on anything that would require additional staff or resources,” she said. “We ran the predictive analytics, and are now able to assess at a certain point of time if an individual may have a higher risk of an improper payment down the road.”

While seeing those “red flags” was helpful, it also posed another question: What kinds of techniques could be used to prevent improper payments from happening in the first place?

The answer came in a form not traditionally seen in the public sector, which required marrying predictive analytics with behavioral science techniques. The department implemented various forms of messaging, including certification boxes and pop-ups, to remind claimants to submit and review their information for accuracy and completeness. 

And since claimants are required to check in with the department weekly to certify unemployment insurance eligibility, the team also implemented an escalation messaging stream to target claimants marked with elevated risk of fraud or inaccuracies. In this way, the tone of messaging escalates to correlate with the risk of fraud or inaccurate information. 

So why are reminders and pop-up messages so effective in motivating claimants? Most likely because they address a bigger issue beyond blatant fraud: human error. “There is a lot of bureaucracy that is attached to the insurance program and it’s a struggle to communicate the requirements,” Forehand said. “Being able to clarify some of that means communication through messaging is critical to preventing over payments.” 

Since launching the strategy in May 2015, the department has seen significant results linked to specific messaging. For example, pop-up messages resulted in a 40 percent reduction in the likelihood to commit fraud. What’s more, claimants who viewed a message that nine out of 10 people report their earnings correctly were twice as likely to report their earnings. 

While the strategy has resulted in positive benefits, it’s also provided the department with key insights into what messaging is most effective in reducing errors and preventing inaccuracies. Case in point: Messages that indicated a consequence for not doing something correctly had the reverse effect — it deterred claimants from providing information or providing it correctly, Forehand said. 

Looking ahead, the department plans to apply this blend of predictive analytics and behavioral science to communications with employers as well. While they will use some of the information they’ve gleaned so far, she said the team also expects to be faced with new challenges, as messaging needs may be significantly different for employers.

“Employers and third party administrators report to us quarterly so we don’t have as much contact with them,” Forehand explained. “Pop-up messaging may work, or it may not in some areas. Traditional correspondence may be the way to go. We’re having to have an intense discussion to see what would be the most effective. This is a different type of customer so the strategies may differ.”

Forehand said she's optimistic that the strategy will continue to enhance operations and outcomes within the New Mexico Workforce Solutions Department, and she foresees its success in other cities as well.

“I think ideally we would love to share our experience with other insurance programs across the country and beyond," she said. "You’re really saving the customer from negative consequences and we’re excited to see if we can share that. It is a nice balance because we’re not impacting benefits; it doesn’t suspend the claim and there’s no harm to the program beneficiary, but it aligns with the integrity effort. It’s also at a lower cost than a system that requires additional staffing.”

And at the end of the day, Forehand said, the department’s strategy is there to better serve claimants and improve efficiency.

“In a broad sense, the biggest service is to the customer. Being able to help them provide accurate and correct information prevents an overpayment,” she said. “So being able to avoid that is a great service in preventing them from having any issues with the program that might result in negative consequences. It’s truly a preventative program.”

Julia McCandless Contributing Writer

Julia McCandless is a journalist passionate about finding the story and telling it well. She currently works as a freelance journalist and communications expert in Northern California, where she lives with her husband and son.