Bryan Caplan's essay exploding the "myth" of the rational voter points out a series of facts I agree with and comes to a series of conclusions I'm skeptical of. Obviously I, like most coastal-bred elitists, don't think voters make terribly good decisions. But I also don't think economic actors are particularly rational. Bryan does. And that's where we part: I look at the savings rate, the economic decisions of those with the most acute financial vulnerability, the findings of behavioral economics, and conclude that folks aren't particularly rational, and a bit of asymmetric paternalism goes a long way.
As for his proposals to insulate fiscal policymakers from public opprobrium, creating a sort of economic Supreme Court, that appears an obvious non-starter. Setting aside questions of controversial economic policy (economist actually disagree on a lot of things), it's obviously unfeasible. At some point, appointments will have to be made, and as we've seen in the Supreme Court, the resentment built up during in the interregnum hyperpoliticizes the process during the few moments of electoral input.
Rather than trying to ineffectually chalk out corners where democracy can't go, if you feel voters are working off incomplete or distorted information, it would seem the place to concentrate is on improving or enlarging their knowledge base, not ignoring their preferences. Elites, after all, already ignore their preferences, and rarely lose a fight to popular opinion. But if the disconnect concerns you, there are surely ways to better educate, or at least persuade, the populace. Bringing the contempt of technocrats and their desire to de-democratize into sharper focus will only further estrange them from the public, possibly with disastrous results.
And one last thing. Bryan singles out religion as a place "where irrationality seems especially pronounced." This gets said occasionally, and it's poppycock. It might be factually wrong to believe in God, but it's certainly not irrational (using rational here in the economic sense, as an action that maximizes your utility). Studies universally find that religious belief and participation offer positive returns for individual health, happiness, finances, personal satisfaction, etc. Obviously, the causality is unclear, and probably has a lot to do with social involvement and capital, but since few have suggested more direct routes to network building, becoming religious and taking advantage of belief's institutions and side benefits is perfectly, totally rational. Indeed, not doing so is the stranger course of action.