I'm also a little leery of anything coming out of the Cato Institute. If this was a published study, that would be different. Given that the author is a professor, I'm willing to give him a little slack, but I'd rather see the opinion peer-reviewed. I don't follow economics well enough to know of any contrarian points of view.
The author's main point seems to be that just because some people can vote, doesn't mean that they should vote. What he doesn't address is why the noise from the uninformed doesn't cancel out. Or more accurately, why wouldn't more noise be better noise? He states that the errors compound, but he never really give a good reason why (at least in my skimming).
I still maintain that the current distribution of uninformed voters isn't uniform. There was a study a while ago (but I can't remember the citation) that said that an increate in even uninformed opinions strengthened predictive power. So in this case more idiots were better idiots. I wish I remember what field this was in though...