Computer scientists' tax code algorithm could make it easier for IRS to catch partnership tax cheats
Computers and taxes are inseparable.
Most of us use our computers to file our taxes every year.
The Internal Revenue Service relies on its computers, which it keeps telling Congress it needs money to upgrade, to process those millions of electronically completed and filed returns.
Now some computer scientists are working on an algorithm that could help catch some tax cheats.
Partnership tax shelter troubles: Specifically, researchers from the Massachusetts Institute of Technology and the nonprofit technology research and development organization Mitre Corporation, are focusing on a certain type of tax shelter used by partnerships.
"We see the tax code as a calculator," Jacob Rosen, a researcher at MIT, told the New York Times. "There are lots of extraordinarily smart people who take individual parts of the tax code and recombine them in complex transactions to construct something not intended by the law."
Bad basis, bogus deduction: Rosen and his fellow computer experts are concentrating on an installment sales shelter that artificially inflates the basis value of an asset on a tax return to wipe out taxable gains when the asset is sold.
To nab tax cheats who use these and other types of tax evasion techniques, the IRS now uses data mining. Basically, the agency collects data from filed tax returns and analyzes it for patterns.
The data mining approach, however, depends on the tax investigators already having some kind of evidence of possible illegal tax activity, such as a suspicious deduction on a return.
AI as preemptive tax filter: The artificial intelligence approach, however, doesn't need the preexisting evidence. Instead, it focuses on rule mining.
In rule mining, individual tax code regulations are lined up against one another to determine if they can be used collectively to create a sophisticated tax dodge.
"It's incredibly difficult to have a computer algorithm that duplicates the enormous creativity of taxpayers, but it's very promising," Robert A. Green, a tax professor at Cornell Law School who read the MIT/Mitre paper, said.
You also might find these items of interest:
- Would you trust your taxes to a robot?
- IRS getting sneakier in tracking tax cheats
- IRS awards tax whistleblower $11.6 million
Find more tax news and tips at the Don't Mess With Taxes home page.