Wednesday, April 20, 2011

Uh oh

Matthew Yglesias writes about a topic about which I know something -- taxes and growth.

He concedes that it isn't a topic which will be resolved by a blog post, then argues that taxes affect the level but not the growth rate of GDP. His claim is supported by economic theorizing using words not equations. The absence of equations is good, as far as it goes, but it doesn't go far.

It's not a topic which will be resolved by a blog post. Also I think it is clearly not a topic which can be resolved by theory. Here (as once before related to New Jersey) I am upset by what appears to be the illusion that economic theory can tell us something about the real world.

I agree with the conclusion, but I agree because of empirical evidence. Basically no one has managed to demonstrate a bad effect of taxes on growth after considering possible conditional convergence (that's something which is very definitely in the data).

The idea that we can tell a priori about what affects levels and growth rates is a mathematical error. Economists can convert anything which affects a level to something which affects a growth rate using the tricks of endogenous growth theory. Just the most nearly reasonable example is one in which growth is fundamentally based on technology, and technology developes due to costly research and development. Taxing away the profits of the firms which have developed new products (or less costly ways of making old ones) reduces the returns to R&D and, typically in the model, reduces growth.

Look if the problem is "take this level effect and make a growth rate effect" I'm pretty sure I can do it when woken up in the middle of the night, when standing on one leg, and when totally drunk). It takes a bit of effort to make a model in which taxing the profits of firms which have invested in R&D causes faster growth, but it isn't really hard (I can send you a *.pdf with the model and proofs, or look up a related published model

Alessandra Pelloni and Robert Waldmann (2000) "Can Waste Improve Welfare ?"
The Journal of Public Economics. vol. 77 pp 45-79.

Also and less eccentrically standard growth models have effects of taxes on growth. In particular, in standard endogenous growth models taxes on capital income cause lower growth rates. This is a well known and very simple mathematical result. In fact (as noted in the paper cited above) the result depends on the assumption that labor supply is exogenous, an assumption which is basically always made in endogenous growth models and which was assumed to be harmless. But I mean I am talking about developments in the theory of taxes and growth which were new in 1986.

Now one might argue that, after a good look a the data, it is reasonable to conclude that all this endogenous growth theory doesn't fit the data as well as the model which was new in 1955

but your post is written on the assumption that endogenous growth theory doesn't exist and can't exist (not just that your claims about what we can tell about the world using economic theory are contradicted by published economic theory but also that there is no model in hyperurania for which your claims are false).

No comments: