In Defense of Supply Siders

Hoo boy, now there is a headline I never thought I would write. I am not a supply side economist by any stretch of the imagination, in that I don't believe that tax cuts will pay for themselves, certainly not in the short term. That said, I think James Surowiecki, a writer I like, did a bad job explaining the true trade offs between taxes, growth, and government spending. Surowiecki's key facts are correct

To begin with, the absurd idea that tax cuts pay for themselves is based on an idea that is not at all absurd, which is that tax rates can have an impact on people’s behavior. Increase taxes too much, and people may work less (since they get to keep less of the income they earn) and invest less (since their gains will be taxed more heavily), and so the economy will grow more slowly. The opposite can happen if you cut taxes. (How much of an impact tax rates have—and how high taxes have to get before they have an impact—is a subject of much debate in economics, but it’s inarguable that they do matter.) What supply-siders have done is start with that reasonable idea and extrapolate it to unreasonable lengths.

What are those unreasonable lengths? Let's continue:

They’re aided in that extrapolation by the simple fact that the American economy grows over time. As a result, even if you cut taxes the federal government will eventually take in more tax revenue than it once did. And that allows supply-siders to fashion a spurious syllogism: taxes were cut in 2001, government revenues are higher in 2007 than they were in 2001, therefore the tax cuts increased revenue. The comparison that really matters in analyzing the impact of the tax cuts, of course, is not between government revenue in 2001 and government revenue in 2007. It’s the comparison between actual tax revenue in 2007 and what tax revenue would have been in 2007 had there been no tax cuts in 2001.

I think the time element that Surowiecki raises is critical.

It's easy to build a simple model economy and see what impact taxes and growth rates have on it over time. You can set a low tax rate (30%) and assign that a particular annual growth rate (say, 3%). You can set a high tax rate (40%) and assign that a lower annual growth rate (say, 2%). As the article mentions, there is a great deal of debate on exactly what impact different tax rates have on growth, and this does not capture the myriad of regulations that impact actual business processes, and so also impact growth, but are not captured in a tax rate. Nonetheless, this model captures the essential elements in the argument.

In year 1, there is no question that a high tax rate economy will capture more revenue from the government than a low tax rate country. But the growth rates make a difference -- in 30 years (one generation) the low tax, high growth economy is 50% larger than the high tax economy, and the government is collecting the same amount of taxes in both. From the 31st year onwards, the low tax economy in every year collects more in taxes than the high tax economy. Does this vindicate the Supply Siders?

Not quite, because we need to determine how much we care about the next generation (year 30) compared to the current generation (year 1). If one tries to guess from current US spending patterns, and the large, hidden liability in medicare and social security, the implicit discount rate is very high -- ie. the government cares very little about the next generation compared to the last one. With a high discount rate, the net present value of the high tax model is higher than the low tax model -- all that extra growth and wealth does not count for anything because the benefits fall to the next generation.

If we use a much lower discount rate, say zero (which says we care about future generations as much as this one) then the low tax regime is better. Incidently, a zero discount rate is required for any environmental intervention to be worthwhile (ignoring opportunity cost) so it's not a crazy option.

So we're left in a difficult position. If you believe that there is a trade-off between tax rates and growth, then whether or not tax cuts pay for themselves is entirely a matter of what your time horizon is, and how much you care about today vs tomorrow. If you don't care about the environment (high discount rate) then high taxes are a good idea. If you care about the environment (low discount rate) then low taxes are a good idea. These combinations make strange bedfellows.