Our new book of conservative policy ideas, Room to Grow, has generated a lot of coverage, some friendly amendments from other conservatives, and some criticism from points left. The most compelling criticism so far—although we’re only a few days in—is Josh Barro’s. He says that the math of the book’s tax-reform proposal doesn’t add up.
Robert Stein wrote the proposal, which centers on an expanded tax credit for children. Stein notes in the book that Senator Mike Lee’s version of the plan has been estimated to widen the deficit quite a bit over the next ten years. Stein then suggests several ways to make up the gap. Barro dismisses them in turn.
Stein says that the level at which the Lee plan’s top income-tax rate kicks in could be lowered. Barro notes that if you go far enough with that idea, the marginal tax rate on a lot of middle-income people, combining the income tax and the payroll tax, would be roughly 50 (where it is now roughly 40). This higher marginal rate would create both a work disincentive and a political problem. Stein says Lee’s plan could be modified to scale back the mortgage-interest deduction further, or to end the tax preference for municipal bonds. Doing those things, Barro points out, won’t raise enough money.
All fair points. But there are other ways to narrow the gap. Lee’s plan gets rid of the alternative minimum tax. He could instead leave part or most of it in place. That would make up a lot of lost revenue. Or he could do a little of everything—lower thresholds, scale back the mortgage deduction more, end the municipal-bond preference, and leave more of the AMT in place.
I don’t see why you couldn’t do those things in a way that lowered both the top marginal tax rate and the average marginal tax rate. Some people, it is true, would face a higher marginal tax rate under almost any version of the Lee plan. Even many of them, however, would not pay a higher tax bill, because they would be benefiting from the expanded child credit.
Getting the math to work, in budget and political terms, is important, but it’s always a back-and-forth process with the modelers who score such proposals and it should be feasible.