It's In The Math: More Tax Audits May Reduce Some Workers' Incentive

GAINESVILLE, Fla. — Recent news that the Internal Revenue Service plans to randomly audit 50,000 extra returns this tax season may discourage cheaters, but it could also have some unintended consequences, say University of Florida economists.

Raising the probability of being audited may spur some taxpayers to work less and earn less, says Ira Horowitz, a UF professor emeritus of decision and information sciences. That’s because these taxpayers may feel that if they earned more and tried to hide the extra dollars from the IRS, they’re likely to get caught and have to fork it over anyway.

“The extreme example is someone who works overtime and does not intend to report any income if there is no chance of being audited,” Horowitz said. “This person might give up the overtime and fully report all regular income when he is certain to be audited.”

In research published in Public Finance Review and other journals, Horowitz and his wife, Ann Horowitz, a UF professor emeritus of economics, probe taxpayer behavior through the very academic lens of economic theory.

Like tax-paying itself, it’s an approach driven by formulas. Rather than polling taxpayers or poring over demographic data, the researchers transform assumptions about what people can reasonably be expected to do when faced with certain choices into mathematical equations. Then they run the equations and see what shakes out.

The increased potential for being audited would only encourage some taxpayers to work less, so it probably wouldn’t have an effect on the economy, Horowitz said. Still, he and Ann Horowitz have concluded that the IRS’ all-around most effective strategy to combat cheating may be to jointly determine audits and penalties.

If the chance of audit is low, the penalty should be high, but if the chance of audit is high, the penalty can be lowered, the researchers found. The researchers even worked out a formula for the right combination. For example, to adequately deter cheating, if the probability of an audit is 10 percent, the cheating penalty ought to be nine times the taxpayer’s unreported income.

“Economists recognize that no taxpayers sit down and work all this out in a methodical way,” Horowitz said. “What we try to do is make reasonable assumptions as to how people will behave and then see what the equations tell us.”