The Corner

Taxes and Entrepreneurship and Causality

Andrew Sullivan has a post up that quotes two different authors attacking the argument that I made in City Journal that higher tax rates will tend to produce, all else equal, less entrepreneurship.

The first of these at least attempts to make the argument from evidence. The author asserts that tax rates were higher under President Clinton than under President Bush, and there was a greater increase in net new businesses under President Clinton than under President Bush, so therefore my conclusion “just isn’t true.”

There are some problems with the analysis, but more fundamentally, with the logic.

First, the analytical problems. As the author notes, he is looking at “net increase in number of firms”, which is new formations less eliminations, so it does not actually capture new starts. I think that a much more direct analysis of this kind in regard to my assertion would be to compare government spending as a percentage of GDP (I referred in the article to current tax laws plus the rational expectation of future tax increases) to total dollars of new venture capital commitments (I described a venture-backed entrepreneur, and quoted the number of jobs created by venture-backed start-ups as evidence for the importance of the issue). I did this only as a response to some of these criticisms. You’ll find that there is a negative correlation with R-Squared of 0.28 between these two variables — that is, as government spending becomes a larger part of the economy, venture capital funding of businesses tends to drop.

Did I just prove my thesis to be correct? Of course not. This is very slightly less naïve than the alternative analysis, but shares the same fundamental logical problem. There are many, many factors that drive new company formation, and the expected future tax bite is only one them. This is why argument that “Hey, we had higher tax rates and lots of new companies under Clinton, so therefore tax rates don’t affect new company formation” is so empty.

Consider an analogous case where Andrew and I disagree about taxes. Andrew has frequently argued that we should increase gas taxes in order to reduce climate change, dependence on unstable oil suppliers, and so on. He believes that higher gas taxes will, all else equal, discourage use. But the federal tax on a gallon of gas was stable at about 9 cents per gallon from 1983 to 1990. It was roughly doubled in steps from 1990 to 1997 to 18.4 cents per gallon. But, look, gasoline consumption in the U.S. kept rising steadily over this period. So can we conclude that proposed future increases in taxes on gas won’t lead to a reduction in use? Andrew doesn’t think so, and neither do I. The price of gasoline and its relative economic benefits are impacted by many other factors. The same is true for entrepreneurship.

The open question is what these rates of company formation (or gas usage) would have been but for the tax changes.

The second item that Andrew cites (a quote from Matt Miller in an article by Kathleen Parker) makes the same essential argument, but without a real attempt at analysis:

“We know from the Clinton boom of the 1990s that marginal tax rates of 39.6 percent put no brakes on entrepreneurship or growth. And the modest limits Obama is proposing on the value of itemized deductions for mortgage interest and charitable donations puts their value exactly where they were under Ronald Reagan, which no one would say was a ‘socialist’ interlude for the U.S. economy. So everyone jumping up and down about how supposedly ‘radical’ Obama’s plan is should calm down and look at the facts.” [Bold added]

Once again, this is a claim for causality with no evidence — what would these have been under a different tax regime?

This paragraph also elides between a directional and a quantitative argument, when it argues that these are “modest.” First, in my article, I was clear that I was referencing current increases plus the rational expectation of future increases. Second, the argument that “Well, this is so small, how could it change anybody’s decision in practice” is a fallacy of pricing. Do you think you would not have bought your last car if it cost literally one dollar more? Why doesn’t every car company in the world raise the price of every car $1? That would be a lot of incremental free cash flow. Once at that price level, why don’t they apply the same logic and raise price $1 again? This road leads to infinitely expensive cars.

We don’t always consciously understand our decisions, and you only have to deter the marginal car buyer (or entrepreneur), not the average one, to experience price elasticity different that zero.

I was carefully circumscribed about my claim:

Higher tax burdens raise the price of entrepreneurship. When you raise the price of something, then, all else held equal, you usually get less of it.

That is, I didn’t claim that we would have fewer new company formations in the next decade than we did in the last. I didn’t even claim that the net effect of the overall Obama political program would be positive or negative for entrepreneurs. The only reason I wrote the article was to provide an estimate for the quantity of “price increase” that would be experienced by a typical potential entrepreneur, because I thought this quantity might be surprisingly high to lots of people. I’m quite surprised at the specific point of controversy that my article has engendered. I thought that non-zero elasticity of economic behavior to price was pretty uncontroversial.

Jim Manzi is CEO of Applied Predictive Technologies (APT), an applied artificial intelligence software company.