|
|
Pre-award vetting using data analytics can protect taxpayer dollars
We estimated 1.4 million questionable Social Security numbers (SSNs) were used in applications for pandemic relief payments. Here's what we found.
Using the PRAC’s , our data scientists randomly sampled 662,000 identity records from 67.5 million applications that received funding from major pandemic relief programs: the Small Business Administration’s (SBA) COVID-19 Economic Injury Disaster Loan program, the SBA’s Paycheck Protection Program, and the Department of Labor’s pandemic unemployment insurance programs.
We then provided the Social Security Administration (SSA) with the SSN, name, and date of birth from the randomly selected records, and asked SSA to verify the following:
- Is the SSN valid?
- If valid, does the name associated with the SSN on the application match SSA records?
- If valid, does the date of birth associated with the SSN on the application match SSA records?
- Is the SSN on the application associated with a deceased individual?
SSA’s verification process flagged nearly 24,000 of the 662,000 sampled records as potentially fraudulent—either the SSN was never issued or didn’t match the applicant’s name or date of birth, indicating that they were either stolen or being used without authorization. Based on those results, our data scientists used statistical modeling to estimate how widespread these issues were in the full set of pandemic relief applications. The analysis suggests that around 1.4 million applications used potentially stolen or invalid SSNs, leading to over $79 billion in potentially fraudulent payments.
Why it matters
The PRAC’s analysis can be used to verify identities and flag potential anomalies in applications before money is disbursed through federal programs.
The PRAC’s shared analytics center houses innovative tools, subject matter experts, and data scientists focused on strengthening program integrity and ensuring funds are paid to legitimate applicants and not fraudsters.
|
|
|
|
