Over at Underbelly, Buce, in commenting on Bruce Bartlett’s new book, makes a point about what he calls “supply side lite” that I’m inclined to agree with:
First, supply side lite. How much extra income will we garner if we tax all incomes over $200,000 at 100 percent? The answer is none: no one will work for nothing. Taxes are incentives, and incentives change behavior. As Bartlett correctly argues, this is really nothing new. Taxes on heroin, shotguns, human trafficking, are designed not for revenue: they are designed to discourage disfavored behavior.
But the devil is in the details. Suppose we have a gross domestic product of 100 and a tax rate of 25 percent; we collect 25 in taxes.* Suppose we cut the rate to 20 percent. How much revenue will we lose? You can’t say in the abstract, but the answer is almost certainly not five. Some people, incentivized by the new, lower taxes, will work harder and make more money. So in the end, we may lose only three or four. It’s an empirical question and it is not easy to answer precisely in the abstract, but the general principle is inarguable.
It works the same the other way around. Suppose we have GDP of 100 and a tax of 20 percent, so we collect 20. How much extra income will we get if we raise the rate to 25? Again, the answer is unknowable, but it is almost certainly not five. At the margin, some people will lighten up on work, and we may get only three or four.
By now this ought to be axiomatic although it is amazing to recall that the government just didn’t do this kind of accounting until about 30 years ago. It’s tricky: it is hard to get right and easy to get wrong (and, therefore, vulnerable to corruption). But the principle is beyond dispute.
Like I said, I agree. But consider Buce’s example. In the first case, the higher rate is ‘normal’, and the tax cut is not as effective as one would predict. In the second case, the lower rate is ‘normal’, and the tax increase is less effective. I realize that this is a ‘toy’ example, but why is one rate optimized, but the new rate suffers from a psychological effect? To clarify, a nominal 25% rate doesn’t have to actually yield 25%, it could yield less, but the point is that cuts and increases are never as effective at least initially as one might think. So at what point, if any, does the psychology adjust, and the amount of taxes captured change?
Somebody might want to study that.