Tweet This
While most of us are worried about staying within the law at tax time, the IRS itself has the opposite problem. Its staff is responsible for identifying those who cheat the system, figuring out what’s owed and collecting.
In 2014 alone, the IRS identified and prevented payment of refunds for more than 2 million confirmed cases of identity theft or fraud in tax returns. That’s about 1 in every 100 returns, and a total of more than $15 billion dollars. An audit also found that millions of dollars of refunds were issued for potentially fraudulent returns.
Who knows how much more fraud may go undetected, cheating the public out of money and public services? The IRS wants to know, and it’s using advanced analytics to seek out hidden fraud.
Sophisticated analytics is nothing new to the agency. It maintains a staff of economists and statisticians, and has done so for decades. SAS founder Jim Goodnight points out that the IRS uses SAS analytics products for fraud detection, as do the Medicaid programs of every state. Nor is the agency a one-product shop; it’s also a long-time user of IBM’s SPSS tools.
In a 2012 presentation, Jeff Butler, then Director, Research Databases, IRS Research, Analysis and Statistics organization, outlined several areas where analytics is useful. Among these were:
- Predicting patterns of filing and payment compliance
- Estimating U.S. tax gap (taxes not paid on time)
- Detect fraud and taxpayer identity theft
- Simulating the impact of legislative changes on taxpayer behavior
- Optimizing case management inventories
So what’s happening in analytics at the IRS today? A new, improved fraud detection system, the Return Review Program (RRP) is under testing now.
The good news is that in the new system’s pilot year, 2014, it identified more than 10,000 identity theft cases and more than 300,000 potentially fraudulent returns that were missed by the old system. The identity theft cases alone totaled more than $43 million. And it can handle roughly 8 million returns per day.
But it’s not time to retire the old system yet. The IRS is conservative about its analytics, and not without reason. Extensive testing is necessary before making any major operational change.
There’s bad news, too. The RRP, classified as a research system, rather than a major system, still requires security improvements. A 2015 audit from Treasury Inspector General for Tax Administration (TIGTA) identified security vulnerabilities that must be addressed.
So, you can look forward to better and faster detection of tax fraud that hurts taxpayers, thanks to analytics. The question is: when?
