Tax amnesties: Revenue drivers or duds?

Tax amnesties, which have been offered in 35 states and the District of Columbia since the 1980s, are enjoying a wave of popularity. Data from the Federation of Tax Administrators show that since 2000, states have offered 35 tax amnesty programs. Under these programs, errant taxpayers face no penalty, no prosecution and, sometimes, no interest on delinquent taxes they pay. The state, on the other hand, gets its money and, possibly, some new taxpayers on the rolls. This may sound like a winning proposition for all, but some researchers at the W. P. Carey School of Business question the value versus the cost. Tax amnesty programs don't come without a price tag, they say mdash; and the downside of these deals actually may outweigh the minimal new revenue that amnesties usually bring in.

When state revenues slump and coffers dwindle, lawmakers often will latch onto what sounds like a perfectly plausible fix: an income tax amnesty. Under these programs, people who cheated on their taxes or those who didn't file — but should have — gain the chance to 'fess up with impunity. Generally, errant taxpayers face no penalty, no prosecution and, sometimes, no interest on delinquent taxes they pay.

The state, on the other hand, gets its money and, possibly, some new taxpayers on the rolls. This may sound like a winning proposition for all, but some researchers question the value versus the cost. Tax amnesty programs don't come without a price tag, they say — and the downside of these deals actually may outweigh the minimal new revenue that amnesties usually bring in.

Amnesty: a sign of the times

According to Charles W. Christian and Sanjay Gupta, two accountancy professors at the W. P. Carey School of Business, 35 states and the District of Columbia have conducted tax amnesties since the 1980s. In fact, such programs are currently enjoying a wave of popularity. Data from the Federation of Tax Administrators show that during the 1980s, 31 amnesties took place.

The number dropped to just 14 amnesties in the prosperous 1990s but, since 2000, 35 of the pardon programs have been offered. Arizona, Massachusetts and Missouri each have had three amnesties since the start of this half-over decade. "In times of fiscal stress, states look everywhere for money. That's when amnesties pop up," Christian says. He explains that lawmaker justifications for the programs almost always include raising new money in two ways.

First, legislators hope to get people who understated income or overstated deductions to come clean. In addition, legislators hope to hook people who didn't file a return at all — nonfilers — into the system. The thinking, Christian says, is that "once people start filing their taxes, they continue to file, bringing the state money for years to come. "It's like an annuity," he adds.

To find out if the annuity analogy reflects the reality of amnesty programs, Christian and Gupta studied an amnesty program in Michigan and published their research on subsequent-year filing behavior of participants. Their paper, "Evidence on Subsequent Filing from the State of Michigan's Income Tax Amnesty," reports that more than two-thirds — 68.6 percent — of the non-filers did indeed return to the tax table and file again three years later. Among those amnesty participants who filed amended returns, 88.5 percent filed again three years later.

Both groups topped the number of repeat filers in a group of people tracked through Internal Revenue Service data. In that group, only 82.1 percent of the taxpayers who filed a return during the amnesty tax year of 1983 stayed in the game for tax year 1986. However, Christian says that number reflects natural attrition rates due to moves, deaths and other changes in circumstance.

He also maintains that ongoing tax compliance among those Michigan amnesty participants was pretty good. Does that mean the program was a revenue boon? Not necessarily. "Amnesties don't really raise much new money," Christian says. "They accelerate the collection of revenue because people who owe money do come forward. But many of them would have paid anyway."

Gupta adds that the boost to revenues is a "shortsighted" goal. "It doesn't take into consideration what will happen in the future," he says. "What kind of signal are you sending out to the marketplace?" It might not be a message of magnanimity, but one of weakness that will invite further non-compliance down the road.

Show me the money

Revenue collection is the goal of an amnesty, but most researchers, including Gupta and Christian, have found that collections aren't always particularly significant. In the case of that 1983 Michigan amnesty the W. P. Carey School researchers studied, revenue increased only 0.1 percent.

"It's remarkable that these amnesties are publicized so much and that so many states have gone through these programs," Gupta says, "yet when you look at the revenue yield, it is astonishingly low." He adds that this is a pattern well established nationwide and recorded in accounting literature.

Michigan's 1983 amnesty brought in mostly new filers — 776 newbies vs. 227 who amended returns. About half of the new filers had already paid some of their state taxes, and their prepaid liability averaged about 60 percent of their total liability. Most of the amending filers — 86 percent — had prepaid most of their liability. Their prepaid taxes averaged 82 percent of the total liability.

What's more, none of the participants studied had huge outstanding tax bills to pay. New filers averaged $367 for their amnesty payment; amending filers averaged $250, indicating that the amnesty net didn't exactly catch a lot of big fish in the tax cheaters' pond. They were "small fries," Christian says. Worse, most of the taxpayers were already in the state revenue system in some way. They were known by the state as being non-compliant.

"If they're already known to the system, what is preventing the system from going through normal enforcement mechanisms and pursuing payment from them?" Gupta asks. Although Christian and Gupta don't focus on this in their Michigan paper, both note that for the small increase in revenue — one-tenth of 1 percent — Michigan and any other state considering an amnesty also should consider the costs of the program and any ripple effects amnesty might have.

Does it pay to play nice?

Among hard costs an amnesty program could incur, Gupta counts additional staff to manage the out-of-the-ordinary processing that might need to be done and "the whole publicity machine." And that publicity machine could spark negative reaction.

According to Gupta, state amnesty programs can prompt at least two undesirable responses from the noncompliant crowd. Some cheaters, he notes, will say, "Well, if they announce an amnesty today, they'll surely have an amnesty again. I could participate then." Meanwhile, some honest taxpayers may resent the break given to the crooks and wind up wondering, "What's the incentive to be honest?"

Anybody who has ever filled out a 1040 knows that incentives to be honest include the penalties, interest and possible punishment that thumbing your nose at tax laws could bring. But as Christian points out, if the threat of those noncompliance consequences doesn't make people fill out their forms correctly and send them in, why would scofflaws come forward when the repercussions are so much smaller? He notes that many experts believe something must have changed for the culpable taxpayers: more guilt, more fear of detection, more morality — something happened to make taxpayers respond.

To give amnesties more power, Gupta says that many states up the ante for post-amnesty noncompliance. Then they make sure that all the publicity clearly warns offenders to pay up now or face bigger fines in the future. But the fact is that many states don't have strong collection mechanisms to back such threats. As Christian explains, "State income taxes are not enforced at nearly the level of federal income taxes."

That's disconcerting when you consider that even the Feds have trouble getting taxpayers to do the right thing. According to the latest statistics available — figures released Feb. 14, 2006 — the 2001 tax gap is now $345 billion. The tax gap represents the difference between what the IRS thinks people should have paid and what they actually paid on a timely basis. Because the IRS expects to collect $55 billion of that $345 liability, the net 2001 tax gap is actually $290. IRS data show that these uncollected revenues come from the 16 percent of people who were non-compliant in paying their federal tax liability.

"If you have a taxpayer who is federally noncompliant, that person is unlikely to come forward in a state amnesty for fear of revealing their federal noncompliance, for which they're sure to get caught and punished," Christian says. "The state can't give you amnesty from federal prosecution."

Given the lack of revenue-collection punch of state amnesties, and the potential for inspiring future scofflaws, is there a better way to get people to pay up? Take the money you'll plow into a state amnesty program and apply it to the state's own enforcement mechanisms instead, Gupta says. "It seems to me that states would get a similar bang for their buck" with that approach, he concludes.