Let me state right up front that Jason is awesome, a guy I’ve met several times, and in no way someone I’m trying to pick a fight with. Primarily because I like him a lot, secondarily because I have no doubt he’d kick my ass from here to Tucumcari. This is just a friendly rebuttal of something he said earlier that didn’t seem to make sense to me. We’re all good, I hope. May it be received in the spirit intended.
Anyway, he made two assertions:
- The US corporate tax rate is too high.
- Lowering corporate tax rates creates jobs.
I did a little digging earlier and found that America’s effective corporate tax rate is the sixth highest in the world. Whether that’s too high or not, I dunno, but it’s clearly high by relative measure to our peers. I learned something today.
The second point is one I wanted to find out a little bit more about. Maybe he’s right, after all. In this environment where we have an entire political party dedicated to seeing who can propose the lower, flatter tax, it’s probably smart to rethink what I assume to be true. Especially since there’s a good chance one of those yokels is going to be the next president.
So, a test of the hypothesis is in order: do lower tax rates lead to lower unemployment rates? I started with the tax rate data that I found earlier today from a PWC study. Next, I went to Wikipedia and found a list of unemployment rates by country. I then crunched it all up into a google spreadsheet and did a simple Pearson analysis to see how the variables correlate. (A scatterplot is featured above.)
For the uninitiated, the Pearson coefficient, usually called r, is a measure of correlation. It’s scaled from -1 to +1, where values near -1 and +1 mean the two things have very strong correlations, and values near 0 mean there is no correlation: they are unrelated. The way it’s set up here, Jason would want to see an r of near 1: higher tax rates correlate strongly to higher unemployment. If r were near -1, that would mean higher tax rates correlate strongly to lower unemployment, exactly the opposite of what he’d want to see. And if r is near 0, that means the two variables don’t have a lot to do with each other at all.
So, I did the math (and you can check it if you want), and I came up with r = 0.0557. This is extremely close to 0 as these analyses go, and indicates that the two variables don’t appear to have anything to do with each other. You can see that in the scatterplot above; there’s no clear line you can draw that fits those points. It’s a blob. To me, really, this isn’t too surprising: the unemployment problem is incredibly complex, influenced by dozens of variables. There’s unlikely to be a silver bullet, or someone would have used it.
Now, some disclaimers:
- I haven’t had a statistics class since 1986. This may be the wrong analysis, or it may be done wrong.
- I don’t know shit about economics or taxes.
- Who knows what the veracity of the PWC or Wikipedia data is.
Basically, this whole analysis may be horseshit, and I am not presenting myself as an authority on any of the above. I’m just a guy who tumblrs pictures of his kids and makes smartass comments, who also happens to be going through a sort of crisis of confidence regarding domestic politics, and I thought for my own benefit I’d test a core belief of mine at a time when it’s under assault from many vectors. Based on what I know, and even accounting for what I don’t, I feel OK with my position on this.
(edit: I really wanted to put a “Read More ->” thing on this, but I can’t since it’s a photo. Thanks, tumblr.)