High Tax, Low Tax

There has been a lot of talk among economics bloggers about Ezra Klein's informal poll regarding the Laffer Curve. The key is that how high we set our highest income tax rates (called "top marginal rates" by those who like using economic-y words like "marginal") matters a lot for how much money the government actually brings in, especially from the wealthiest citizens. Klein summarizes this favorite tool of conservative political discourse:

The idea, popularized by economist Arthur Laffer and writer Jude Wanninski in the 1970s and '80s, is simple. Tax rates of zero percent produce no revenue, for obvious reasons. Rates of 100 percent should produce no revenue either, as no one would bother making the money that falls into that bracket knowing it would all be taken away. Thus, presumably, there is some rate in between the two that maximizes revenue. Go above it and revenue would fall because people would avoid taxes or stop working; go below it and revenue would fall because less money would be taxed.

That is, as we raise taxes we increase government revenue, but at some point this has to stop being the case. The reasons put forward for this are many, among them the following:
(A) if taxes are too high people have a very strong incentive to lie about their income levels;
(B) if taxes are too high people have very little incentive to work more / harder (they won't see most of the benefits anyway);
(C) if taxes are too high the highest income earners have a strong incentive to take their business somewhere else more wealth-friendly, like Singapore or Hong Kong;
(D) if taxes are too high government will be too big a part of the economy, and since government is not as good at promoting long run growth as the private market, long run growth and overall income will suffer, and tax revenues will suffer along with them.

Some of these are much stronger arguments than others. (A) is a big one on a practical, year-in-year-out level. (C) probably plays an important role in the long run, on the order of several decades, and becomes more of an issue the more globalized the world becomes and the more high income earners get paid for their skill with information (which means its easy to do what they do from anywhere). (B) and (D) are tempting philosophically, but there's just not much empirical evidence to support them.

Still, one thing is certain: we definitely don't want to be on the wrong side of that peak, so knowing where it is could be very important. Thus, Klein decided to ask some experts where the peak is.

There were a lot of bad answers to his question. To be sure, figuring out where precisely government revenues can't go up anymore is hard, but a lot of the people Klein asked gave numbers based almost purely on ideology---as opposed to, I don't know, data, or even theory---but acted like it was based on research. Others gave a single number as if they'd seen the number written on stone tablets somewhere. But there were some pretty good answers among the bizarre ones.

Probably the best numerical answers were given by Emmanuel Saez and Bruce Bartlett, both of whom offer a way of getting a range, and both of whom put the number somewhere around 60% to 80%; in other words, the US was probably close to the edge at the beginning of Reagan's presidency (top rates were around 70%), but we're nowhere close right now.

Greg Mankiw pointed out that in a longer term situation people will probably respond more to the tax rates they face, although he does seem to put a lot of stock in the economic growth argument that, as I mentioned, isn't great empirically. Bruce Bartlett adds a pretty important insight:

I think 50 percent is an important threshold and I would be very reluctant to go higher even if it raised net revenue.

This is key, in my view. Yes, we might be able to raise revenue up to a certain point without crossing the threshold, but there are good reasons to not just stay on the left side, but stay significantly on the left side. If we reach the peak, we've already gone too far. For this reason, my favorite response is Marty Feldstein:

Why look for the rate that maximizes revenue? As the tax rate rises, the "deadweight loss" (real loss to the economy) rises so as the rate gets close to maximizing revenue the loss to the economy exceeds the gain in revenue.

If only my micro principles students could think like this.  Yes, it might bring in revenue. It might not even be hurting long term growth. There are still important economic reasons---distortions to the economy---not to go that far.

There's a moral dimension here, too, but that is, I suppose, a post for another time.