Even though insurers have invested heavily in fraud-fighting technology, the percentage of referrals from automated systems accepted by special investigation units has declined to 15 percent from 22 percent in the past two years, the Coalition Against Insurance Fraud reported last week.
The Coalition’s biennial benchmarking report on insurer SIUs notes that investigators’ caseloads also declined in 2019 — to 10.7 per month from 12 for field investigators and from 24 to 17 for desk investigators. At the same time, insurers increased the number of desk investigators, who grew to 16 percent of SIU staff from 9 percent in 2017.
“One possible explanation is that with technology and training, desk and field investigators may be handling fewer but larger and more complex cases,” the report says. “Such cases require considerable time and resources. With the increase of insurer major-case units, the decline in caseload would be one result.”
The coalition said 48 percent of carriers used major case units that specialize in complicated cases, up slightly since the last survey but a significant increase from 2015.
Bang for Buck
The focus on major cases will likely give insurers more bang for their buck, which they will need because carriers are also spending less money on their SIUs. The report says SIU budgets dropped by 20 percent from 2017 to 2019, measured as a percentage of premium.
“This doesn’t necessarily mean SIUs faced budget cuts, but it appears that SIU budgets at many companies may not be keeping pace with premium growth,” the study says. “Insurers reported that areas where budgets expanded the most was in technology and investigative staff.”
The Coalition said 40 carriers participated in the study.
Dennis Jay, the retired Coalition executive director who wrote the report, said the drop in the percentage of accepted claims referred by automated systems is surprising. He expected that percentage to increase, as it did from 2015 to 2017.
“With the number of insurers that are comfortable using technology, that number should be going up, especially now that we are involved in predictive analytics and artificial intelligence that would be finding fraud that before this went undetected.”
Jay said the numbers may be skewed somewhat because insurers have been stepping up their investment in fraud-detection technology, and insurers that first start using automated systems won’t get the same percentage of accepted claim referrals from the technology as more experienced insurers.
Jay said it takes some experience — and training — to learn the difference between a quality fraud referral and a false positive.
“One of the biggest concerns that SIUs have is they are going to have more cases than they can handle,” he said. “So it’s kind of a chicken or the egg thing.”
Brad Barron, a fraud and financial crimes consultant for SAS, said an increased focus on major fraud rings may reduce the number of accepted referrals.
“Complex schemes of this nature typically involve multiple claims, adding up to tens or hundreds of thousands of dollars in potential losses – sometimes millions,” Barron said in an email. “They often start as a single referral, but the investigation can unravel dozens, if not hundreds, of affiliated cases.”
While complex cases require more time and resources to investigate, they account for a significant amount of the fraud committed. “So it’s not just the quantity of referrals, but also the quality that matters,” he said.
Barron said the increasing number of desk investigators shows the increased focus on technology. He said that focus also explains the reduced share of premium dollars going to SIU budgets. Insurers expect technology to increase efficiency.
“Fraudsters love digital, yet as they increasingly use online and digital channels to defraud insurers, investigators gain a significant advantage,” Barron said. “Leading-edge analytic technologies, including machine learning and AI, enable investigators to analyze internal data along with a growing array of external data, such as street view maps, links to public records, and social media. With these capabilities, SIUs can operate more efficiently and effectively while delivering more return on investment for the insurer.”
Dan Gumpright, vice president of global products for Friss, said in a telephone interview that the percentage of automated referrals accepted by SIUs may be dropping because investigators are better at discerning referrals that are likely to bear fruit.
“As people deploy these, you start accepting a lot more of the referrals and then you learn what you need to accept over time,” he said.
Gumpright said fraud detection systems also prevent fraud by catching potential scams at the underwriting level, sometimes even before a quote is given. The nature of fraud-detection systems has also changed in the last several years, he said. A decade ago, most systems were custom built and highly specialized, which made them effective at first but difficult to upgrade as fraudsters changed strategies. In the past few years, more insurers are turning to software-as-a-service detection systems that are less specialized but updated more often.
Gumpright, like Barron, also said the move toward major crime units may decrease the number of referrals that are accepted. He said major crime units like to receive a large number of referrals because they have the resources to sift through the data, but that doesn’t mean they necessarily accept those referrals and launch investigations.
“I think in general they are going to have a slightly lower acceptance rate but the fraud they are finding will be significant,” he said.
Some other highlights of the Coalition’s report:
- Salaries for field investigators averaged $85,557 in 2019; $73,101 for desk investigators. The average salary for all SIU employees climbed to $84,715 in 2019 from $74,383 in 2015 — a 14 percent gain.
- Insurers outsourced 21 percent of investigations in 2019, up from16 percent in 2017. Outsourced legal work increased to 40 percent from 30 percent during that period.
- Suspected fraud referrals made up 1.8 percent of claims for personal auto, 1.9 percent for property, 2.3 percent for commercial auto and 2.3 percent for commercial property.Those percentages have remained steady in the past three studies.