Science and Mr. Shriver

Adam Shriver discusses the earnings tax over at the 'Hub, and doubts whether its negative effect on cities has been adequately proven:

So if one wanted to honestly study such a phenomenon he or she might begin with some research before performing analysis. How many cities have earnings taxes in the U.S.? answer 14 states and the district of Columbia a list encompassing hundreds of cities. Has there ever been a documented scientific study that demonstrates that employers avoid such cities? In truth I do not know.

The Sun King visits the Academy of
Sciences--but where are the marginal
value model equations?

He doesn't specify what he means by "scientific study," but I suspect he has in mind something like this: an empirical study demonstrating a correlation between cities employers avoid and earnings taxes. In the physical sciences, where work is done by isolating variables and demonstrating causation, such an approach can be quite productive.

But human societies cannot be readily reduced to controlled laboratory conditions -- which is why the constant reminder about statistics is that correlation does not necessarily prove causation. Say Portland enacted an earnings tax and didn't see any avoidance on the part of employers. Hypothesis demolished, right? But hold on -- we don't know what else Portland is doing. Are its other taxes (or wages!) attractively low? Does it ply potential employers with tax breaks? Is it the only game in town for certain industries, which allows it to extract (in an essentially monopolistic way) its earnings tax?

Such messy variables are the reason Austrian economists, for example, reject statistical studies of correlation as a valid basis for economic inquiry. Instead, economics must seek to understand the commonsense principles all human action is based on (what Mises called "praxeology"), and then derive economic conclusions from those principles. Such commonsense principles intuitively include the idea that, if I can do the same business in Clayton I can do in the City of St. Louis, but have less of my earnings taken away from me, then, all other things being equal, that's where I'm going to go!

Now, we can call that analysis "scientific" in the sense that it's pretty obviously true, but my suspicions are that Mr. Shriver wants something with charts and graphs. There are excellent reasons to think, though, that human diversity is more obscured than clarified by attempting to reduce it to colored bars and pie sections.