Time To Catch Up On Predictive Analytics
A special report from the Social Security Administration Office Inspector General about detecting and deterring disability fraud has what some may consider a surprising recommendation: Invest in predictive analytics tools to help identify likely fraudulent claims on the Social Security Disability Insurance (SSDI) books.
That’s surprising because very few people think of Social Security as being an early adopter. Predictive analytics is not exactly a term or concept in common knowledge after all.
Yet the recommendation, and its playing out, will likely be of interest to life insurance and annuity professionals. That includes those who are not involved in the disability insurance business.
First, a few details about the proposal. According to the report, Social Security has already begun testing the analytics waters. In May, it completed a 90-day “proof of concept” initiative to use data analytics to prove known fraud using disability claims.
The result is illuminating: The study produced a “match rate” of 81 percent, 91 percent and 86 percent in the three states examined. That means the analytics successfully identified the large majority of fraudulent claims in each scheme studied.
Those were known disability frauds. Now the agency is engaged another analytics study. This one aims to see how well predictive analytics performs in uncovering unknown fraud. This new 180-day project is still under way, the report indicated. But the outcome — and the follow-up — should help enlighten a lot of life and annuity professionals who want to know if predictive analytics “really works.”
In the past few years, LIMRA has provided insurance company executives with education and data about using predictive analytics in business decision-making. But although some execs tell me they are now using analytics, others say they’re still on the fence about it or only now exploring the possibilities. Some believe their property-casualty insurance colleagues are way ahead on this score.
That’s anecdotal information, not scientific, but it seems to be the status on the company front right now. It’s still a frontier.
As for insurance and financial advisors, many with whom I speak have little insight about whether analytics can help them at the agency or practitioner level. Some say they even don’t know what predictive analytics is. They’re curious though. Very curious.
The fact that Social Security is already probing predictive analytics is meaningful to insurance for several reasons. For one thing, if the analytics do help nab heretofore unknown fraudsters, that could provide powerful incentive for insurance professionals to give it a whirl. They could, perhaps, use it to help nab their own fraudsters or, as analytics experts put it, use it for a myriad of company-level operations.
For another, the fact that a federal government agency is looking into this could be incentive enough to take a look. After all, government agencies are not known for being technology leaders. If Social Security is probing the possibilities, maybe industry firms should too.
Finally, the initiative already has an insurance stamp on it. The authors of the Social Security Administration’s Ability to Prevent and Deter Disability Fraud report cited the analytics experience of “private insurance firms, such as the U.S-based Unum Group” as one of the factors in support of Social Security moving in this direction. If it’s useful for Unum, might it also be useful for other insurance firms?
Most non-techies are late to the party on analytics. Time to catch up. Even if it’s imperfect, even if it’s hard to learn, it’s already here and some wings of the life and annuity business are already using it. Just knowing about it would seem to be better than not knowing.