Here's a dilemma. Recent research by Douglas Bernheim of Princeton University and John Scholz of the University of Wisconsin finds that college-educated Americans are far more responsive than high school graduates to individual retirement accounts and other savings-targeted tax incentives. Yet college grads also tend to enjoy considerably higher incomes and are thus far more likely to garner tax deductions for money they would have saved anyhow. Partly for this reason, current law allows no tax deductions at all for contributions to IRAs made by relatively affluent taxpayers, such as couples with adjusted gross income exceeding $50,000.
Is there a way to capitalize on the responsiveness of the college-educated to IRAs while limiting their ability simply to shift current savings into such accounts? In a recent National Bureau of Economic Research study, Bernheim and Scholz offer a proposal to do just that.
Under their scheme, couples with moderate incomes up to $34,000 could still put as much as $4,000 into IRAs, as they do now. But higher-income households would have to save a minimum amount on a nondeductible basis before they became eligible for the tax benefit. If a couple with an income of $80,000 made a "floor" contribution of $8,000 to an IRA, for example, they could obtain a tax deduction on the next $4,000 saved in the account. More affluent couples would have to save even more before their tax deductions kicked in.
Because it is designed to enhance savings behavior at the margin, such a system, claim the two economists, would limit the government's loss of tax revenues. At the same time, it would vastly increase savings by adding over a million savings-prone households to the group currently eligible for IRAs.