Here's the least controversial principle in all statecraft: Never pass up an opportunity to make everyone happier at the same time. Admittedly, opportunities like that are vanishingly rare--almost anything you propose is likely to make someone unhappy--so the principle is as useless in practice as it is compelling in theory. But if you want to talk about the difference between good and bad policy, it helps to have an uncontroversial starting point.
This particular starting point has proved useful enough to earn a proper name. Economists call it the Pareto Principle, in honor of the 19th-century economist Vilfredo Pareto. The party line among economists is that although we disagree about a myriad of issues, at least we’re all good Paretians: If you can think of a way to make everyone happier, we’re behind you all the way.
How disturbing, then, to realize that the Pareto Principle conflicts with another value that some of us hold at least equally dear--the right of individuals to make free choices. Here's a stylized example: Suppose some people (call them the "prudes") cherish their freedom of religion, but not half so much as they would cherish a general ban on pornography. Others (call them the "lewds") cherish their right to read Lady Chatterley's Lover but not half so much as they would cherish a general ban on religion. Then if you outlawed both pornography and religion, you'd make everyone happier, while simultaneously making everyone less free.
The Pareto Principle demands that we embrace such a law while the fundamental precepts of liberalism demand that we recoil from it. As Nobel laureate Amartya Sen pointed out 30 years ago, there's no such thing as a Paretian liberal. When some people get pleasure from controlling other people, the Paretians and the liberals must part ways.
That's no problem in practice, because in practice there will always be liberals who are offended by book-banning--and the very existence of those liberals leads the Paretian to withdraw his endorsement of the ban. (Remember, the Pareto Principle endorses only changes that make everyone happier, including the liberals.) But in practice, nobody is guided solely by the Pareto Principle anyway. Instead, we're guided by more flexible criteria such as the cost-benefit principle: A policy is good when its benefits exceed its costs, with benefits (or costs) measured by what the proponents (or opponents) would be willing to pay to see the policy enacted (or defeated).
Under quite general circumstances, it can be proved (though not in the space of a single magazine column) that free markets yield exactly those outcomes the cost-benefit principle recommends. That makes it easy to reconcile a taste for economic welfare with a taste for individual liberty: Usually, markets deliver both at once.
Usually, but not always. With just a slight modification, the paradox of the Paretian liberal reasserts itself in the cost-benefit context. Suppose I'm willing to pay $20 to read the subversive works of Paul Krugman and you're willing to pay $40 to stop me. A strict cost-benefit analysis suggests that Krugman's writing should be banned.
(There's no need for a ban if you can locate me and offer me, say, $30 to change my reading habits--and if you can be assured that I won't just take the money and run. But let's suppose the impracticality of such arrangements leaves book-banning as the only realistic alternative.)
That's a conclusion we liberals find repugnant, and it would be nice to avoid it. One way out is simply to declare that "psychic costs don't count." If you don't like getting your nose punched, your aversion goes into the cost-benefit calculus and inspires us to write laws that discourage nose-punching. But if you don't like knowing I read Krugman, that's your own problem.
Appealing as that position might sound, it's also suspiciously incoherent. If my habit of reading Krugman and my habit of punching your nose are equally painful to you, why should public policy discourage one and not the other?