States Use Big Data to Nab Tax Fraudsters

Technology has made it easier for people to commit tax fraud and for governments to catch it.

It’s tax season again. For most of us, that means undergoing the laborious and thankless task of assembling financial records and calculating taxes for state and federal returns. But for a small group of us, tax season is profit season. It’s the time of year when fraudsters busy themselves with stealing identities and electronically submitting fraudulent tax returns for refunds.

Nobody knows for sure just how much tax return fraud is committed, but the amount is rising fast. According to the U.S. Treasury, the number of identified fraudulent federal returns has increased by 40 percent from 2011 to 2012, an increase of more than $4 billion. Ten years ago, New York state stopped refunds on 50,000 fraudulently filed tax returns. Last year, the number of stopped refunds was 250,000, according to Nonie Manion, executive deputy commissioner for the state’s Department of Taxation and Finance.

What’s behind this surge in fraud? Experts credit two reasons: automation and the role of tax departments. First, technology has taken some of the pain out of filing tax returns, thanks to sophisticated tax preparation software programs. But it has also made it easier for fraudsters to submit fake returns at the push of a button. Gone are the days when government workers would fan through paper returns, looking for similar handwriting styles. Today, tax thieves can create one fraudulent return, then change a few data elements and resubmit over and over, making it much more difficult to detect fraud in a computer-generated return.

Second, the role of state tax departments has shifted from collecting money to paying our refunds, thanks to the growth in tax credits. Taxpayers want those refunds as quickly as possible, and states have worked to process returns and submit refunds within 30 days. But processing lots of returns so quickly raises the risk that fraudulent returns will go undetected. “Once a state lets a refund go out, the chances of recovery are slim,” says John Feldmann, compliance manager at the Federation of Tax Administrators.

To combat the problem, state revenue and tax agencies are using software programs to sift through mounds of data and detect patterns that would indicate when a return is not valid. Just about every state with a tax fraud detection program already compares tax return data with information from other state agencies and private firms to spot incorrect mailing addresses and stolen identities. Because so many returns are filed electronically, fraud spotting systems look for suspicious Internet protocol (IP) addresses. For example, tax auditors in New York noticed that similar IP addresses in Fort Lauderdale, Fla., were submitting a series of returns for refunds. When the state couldn’t match the returns with any employer data, they were flagged for further scrutiny and  ultimately found to be fraudulent.

High-tech analytics is one way states keep up with the war on fraud. The other is accurate data. The third component is well trained staff. But it takes time and money to put together the technology and the expertise to combat the growing sophistication of fraudsters. “I would like to say that we are at least in step with the fraudsters, if not ahead of them,” says Manion. “But we have to keep investing in technology in order to be both fast and accurate.”

Of course, not every state has been equally successful in fighting tax return fraud. “A lot depends on the technology available, the personnel available and leadership,” says Feldmann.

In New York’s case, state officials were able to increase detection after the state legislature changed the time that employers had to report tax withholding information, giving the Department of Taxation and Finance a bigger window of time to analyze returns, which can arrive at the rate of 1 million per day at the peak of tax season. Manion says the extra time means more accuracy. But it also means that crooks are more likely to be caught: “We have heard,” she says, “that the fraudsters are told not to bother filing in New York because the state will stop them.”

Tod is the managing editor of Governing and the contributing editor of our sister publication, Government Technology. He was previously the editor of Public CIO, e.Republic’s award-winning publication for IT executives in the public sector, and is the author of several books on information management.