Detecting Fraudulent Unemployment Claims with Intel:ID

As the pandemic unfurled last spring, millions found themselves out of work — and turning to their states’ unemployment systems.

But in many cases, the states were unprepared to handle the sheer volume of new claims. In Texas, for example, claims surged from an average of 13,000 per week in February 2020 to more than 400,000 in a single week in April. Nationwide, 16.8 million unemployment claims were filed between March 15 and April 4 — 13% of the entire U.S. workforce.

Meanwhile, scammers pounced on the confusion, filing millions of fraudulent claims. LexisNexis Risk Solutions reports that at least 70% of fraudulent claims originated overseas, and organized crime rings sell step-by-step instructions on how to exploit each state’s unemployment systems. In many cases, these scammers are using ID information harvested during years-ago data breaches or phishing emails.

By February 2021, more than $63 billion in fraudulent claims had been paid out, according to a watchdog for the U.S. Department of Labor. In some states, 35 to 40% of all new applications are fraudulent.

Even then, thousands of filers are still waiting for their legitimate claims to be paid. In fact, only 11 states meet the federal target of paying 87% of approved claims within 21 days.

So what’s a state to do? How can states — often coping with decades-old legacy systems — combat fraudulent claims?

In many cases, it comes down to accurately identifying claimants. In December, Congress mandated that states must verify claimant identities before dispersing funds. Some states have turned to facial recognition to help solve the problem — but now scammers are 3-D printing masks of identity theft victims.

So where is the silver bullet? And how do you arm a program integrity department with tools that will reduce improper payments without delaying applications or payments?

Entity resolution and relationship detection solutions like IMT’s Intel:ID can analyze data from disparate data sources, linking and resolving all the records pertaining to a single person. This real-time analysis can fill in the gaps to verify if the information in a person’s unemployment claim matches the data in their DMV or state tax records. The sophisticated algorithms can overcome complex multi-cultural names and other identity variations — such as nicknames, juniors/seniors, or previous addresses — to flag cases where additional review is needed.

By retaining historical data, the Intel:ID solution is purpose built to target synthetic identity fraud where perpetrators combine fictitious and real data to create new claimant identities or modify the identity of a legitimate claimant by gradually introducing bogus demographic data to the record.

Intel:ID has been redesigned with a streamlined interface that delivers need-to-know information exactly when it’s needed. And with our IBM i2 Analysts Notebook plugin, program integrity analysts can quickly visualize claimant entities and other non-obvious relationships. Having data reconciled from different sources for Individuals applying for unemployment benefits or Small Businesses applying for economic relief through stimulus checks mitigates the risk of fraud.

To learn more about how IMT Intel:ID can help protect the integrity of employment claims, please contact Peter Huber at