Renewed attention to the revenues that could be raised by reducing tax expenditures makes this proposal -- Raise Taxes, but Not Tax Rates -- from Martin Feldstein very timely. (The article draws on an NBER working paper by Feldstein, Daniel Feenberg, and Maya MacGuineas.) The idea is very simple: cap the aggregate tax savings granted to any taxpayer through tax expenditures to 2 per cent of that taxpayer's income.

This is relatively easy to understand and very easy to administer. Another advantage, until comprehensive tax reform comes along, is that it requires no tinkering with the existing structure of deductions: they can all stay in place, which lightens the political burden of getting it done. Another is that capped taxpayers would have no reason to alter their spending for tax-saving purposes: no incentive, for instance, to take out a bigger mortgage, or to delay paying their existing one down. Also, according to Feldstein's simulations, the policy would lead more than 35m taxpayers to switch from itemizing to taking the standard deduction, which would undermine support for itemized deductions and prepare the ground for a more thoroughgoing reform.

The policy would increase revenues by $300 billion a year, and broadly maintain the present degree of progressivity. Revenue as a percentage of adjusted gross income rises by about 3 per cent across the income scale. The richest taxpayers would get a slightly smaller hit, but this could be fixed at no great cost in extra complexity, by making the cap 2 per cent or $10,000, whichever is less. This brings in another $20 billion in revenues too.

It's a good idea, and I can see it catching on.