Ok, finally got through Bryan Caplan's The  Myth of the Rational Voter. Here's my one-sentence summary of  his conclusions:

That might sound unfair (unfair to Caplan, that is)  but I don't think it is. Caplan never comes flat out and says that people who don't know economics should be disenfranchised, but he comes as close as you can and maintain some professional standing. For example, after mooting the question of  why there should be a universal franchise, he points out that we license car drivers -- why shouldn't we license voters? (Ignore the question of who would do the licensing, since the problem he's trying to solve is bad government policy). Backing off from that, Caplan suggests instead that people with university degrees or who run businesses might be given additional votes.

All of  this is to counteract the supposedly pernicious influence of voter bias,which systematically demands policies that are anti-market, and which cynical politicians are  usually happy to give.

I dunno. I'm actually willing go grant Caplan the empirical claims: that voters have anti-market biases, that voting these biases serves quasi-religious ideological needs in many voters, and that democracy fails in part because politicians give voters what they ask for. But I  don't see how any of Caplan's normative recommendations follow.

These come in two main types: First, there are the voting reform ideas I mentioned above. At the very least, Caplan would like to see an end to the "get out the vote"  campaigns, which simply bring  more ignoramuses into the polling station.

Second, Caplan suggests that we should have less government and more market. Why? Because unlike shopping, voting is a commons. A vote for a bad policy is like throwing garbage into a reservoir -- it might unburden the thrower, but everyone suffers.

My two quick replies to these recommendations:

1. Caplan doesn't get why we have universal enfranchisement. It isn't  so that everyone can have a say in policy; it is so that everyone can concede to being governed. Giving everyone the  vote is not a way of figuring out what should be done, it is a method for ensuring the legitimacy of our system of government.

2. Caplan ignores the way various cognitive biases affect consumer choice, which in turn requires more government. One simple example: Our poor evaluations of risk lead us to make poor investment/savings decisions, which necessitates various forms of public insurance. The point is, our cognitive biases have externalities in the consumer realm as well as the voting realm. The upshot: Caplan's brand of libertarianism would paradoxically lead to more government, not less.