One of the first things you learn in Means-Tested Welfare Economics 101 is that means-tested welfare programs produce enormous marginal tax rates. Say you have a $5,000 benefit that's available only to people making less than $10,000. If you have market income of $9,000, this means you have total income of $14,000. But if you have market income of $11,000 you have total income of $11,000. You're essentially paying a marginal tax rate of over 100%.
Of course, you could spread this out. Instead of just taking away the entire benefit when you cross the $10,000 threshold, you could take away, say, $500 for every $1,000 in additional income. This means that for every additional $1,000 you make, your actual income only goes up $500. That's better, but it's still a 50% marginal tax rate, which is higher than the tax rate paid by any other income class on any kind of income.
This is all basic stuff, recognized by everyone since the dawn of time. Megan McArdle comments on it today:
Note two things: first, that in this case, at least, the supply siders seem to be completely right. Everyone I've spoken to about the problem seems to agree that the poor respond to these high marginal tax rates by either taking lower-paying jobs than they could, or working less — not in every individual case, but in aggregate.
And second, that this is not a problem that supply siders seem to be applying much brain power or political capital to fixing.
Nope. In fact, they mostly want to make it worse by applying means testing to programs like Social Security and Medicare. Unfortunately, despite plenty of high-wattage brainpower being applied to this problem, nobody's ever figured out how to avoid this basic problem of means-tested benefits. If the means testing is strict (i.e., benefits phase out quickly) it creates a big incentive to simply save less or not work as hard. If the means testing is loose (i.e., benefits phase out slowly), the perverse incentives get smaller, but you end up barely saving any money. It's an especially big problem for Social Security, which is a pure cash benefit.
This is why I'm pretty skeptical of means testing either of these programs more than we already do. Any version of this that avoids big negative incentives would phase out so slowly that it just wouldn't save much money. So why bother?