Last-minute taxpayers beware: States are deploying data analytics in the fight to collect delinquent taxes.

New York state is among them. In 2010, the state’s Department of Taxation and Finance implemented an IBM analytics tool to help recover $83 million in delinquent taxes — an 8 percent increase from 2009 and double the annual increase from previous years, according to an announcement from the company. IBM and New York state have been working together for more than a decade on analytics technology.

Each year, the state’s finance department is responsible for collecting $3 billion in tax revenue.

The Department of Taxation and Finance chooses different methods to alert a debtor when his or her taxes are delinquent — whether by phone calls, letters or confronting the debtor in person. Shaun Barry, IBM’s global leader for fraud management, said the software inserts an algorithm into the department’s debt case management system. The software determines on a case-by-case basis the best course of action for collecting a delinquent tax given the department’s limited resources, while maximizing the amount of revenue collected.

The department then develops an action plan for each case — delinquent or fraudulent taxes — based on the analytics data, said Brad Maione, the department’s director of public information.

“The collaboration has improved the quality of our enforcement actions and helped us fight fraud in terms of fraudulent tax returns,” he said.

The department doesn’t necessarily need to target people and businesses that owe an exorbitant amount of money, Barry said. For example, the analytics could show that it might be less efficient to try to collect money from a business owing $100,000 compared to another business owing $10,000. Some companies go out of business and have no financial means, so going after those businesses wouldn’t produce any collected taxes.

“A lot of it is about what kind of ‘pulse’ does the taxpayer have, and does it make sense to go after him or her,” Barry explained.

The predictive modeling tool used in the IBM Tax Collections Optimizer is like what private-sector companies use for gathering predictive analytics. But the tool’s distinguishing feature is that it factors in budget and resource limitations in its decision-making, Barry said.

New York isn’t the only state using analytics for tax collection. For example, last year Hawaii officials announced they had collected more than $100 million within a three-year period through a partnership between the Department of Taxation and CGI Technologies and Solutions.

Sarah Rich, Staff Writer Sarah Rich  |  Staff Writer

In 2008, Sarah Rich graduated from California State University, Chico, where she majored in news-editorial journalism and minored in sociology. Since 2010, Sarah has written for Government Technology magazine and covers a spectrum of public-sector IT topics, including cloud computing, transparency, broadband, and other innovative projects and trends. She currently lives in Sacramento, Calif.