Greg Mankiw (among others) points to new NBER working papers by Casey Mulligan that point out that marginal tax rates go up under Obamacare. I have not read the papers, but I assume that he counts as an increase in the marginal tax rate the fact that you lose out on subsidies as your incomes goes up. That is legitimate economic analysis, but try to do satisfy the following:
1. Use “means testing” in order to provide a significant benefit that is aimed at the poor.
2. Keep the marginal tax rate low.
3. Keep the budget cost low.
Those of us on the right tend to argue separately for all three. But collectively, they are not so easy to satisfy. (My undergraduate economics professor, Bernie Saffran, pointed this out, and I have not forgotten it.)
If you want to offer a means-tested benefit at low cost, then you have to scale-back the benefit rapidly as income rises, meaning a high marginal tax rate.
If you want to keep the marginal tax rate low and and the budget cost low, then you cannot offer a sizable benefit to the poor. So you can’t do much in terms of means-testing.
If you want to provide a significant benefit to the poor with a low marginal tax rate, then you have to phase the benefit out very slowly as incomes rise. So the budget cost is high.
If we want to, we can play “gotcha!” with any proposal that is aimed at helping people who are poor. It is bound to fail (1), (2), or (3). But how can we be constructive?
My solution was offered in the essay Bleeding-Heart Libertarianism. The idea was to offer a significant benefit with a low marginal tax rate. To hold down the budget cost, I shift away from in-kind benefits (such as food stamps or Medicaid) toward a cash benefit.
That essay is worth re-reading.