Here’s the fundamental weakness in arguments justifying government power to prevent we adults from taking medications of our own choosing: as flawed as Mr. Doe’s judgment might be, and as uninformed or even misinformed as he might be, there is no plausible case to be made that a stranger can make decisions for Mr. Doe that are generally better for Mr. Doe than those that Mr. Doe would make himself for himself.
Let’s grant the validity of the findings of behavioral economics  – biases in perception, cognition, and judgment that cause many human choices and actions to differ from those predicted by standard neoclassical economics. Let’s also grant that these biases cause people to make decisions that they will later genuinely regret – that is, decisions that they would not have made had their perception, cognition, and judgment been bias-free. (In more formal language, let’s grant that many decisions are ‘bias-induced’ rather than ‘preference-induced.’)
What’s the relevance of these facts (assuming, again, that they are indeed facts) for policy? It’s tempting to conclude that because each of us is not as ‘rational’ as we are in economic-textbook models, we would be better served by government regulation and restrictions.
But if I’m biased in making decisions for myself, might I not be at least as biased in making decisions for someone else? In fact, might I be even more biased in making decisions for others – for people whom I don’t know, who reach my purview only as data or as abstractions?
There are four ways in which you can spend money. You can spend your own money on yourself. When you do that, why then you really watch out what you’re doing, and you try to get the most for your money.
Then you can spend your own money on somebody else. For example, I buy a birthday present for someone. Well, then I’m not so careful about the content of the present, but I’m very careful about the cost.
Then, I can spend somebody else’s money on myself. And if I spend somebody else’s money on myself, then I’m sure going to have a good lunch!
Finally, I can spend somebody else’s money on somebody else. And if I spend somebody else’s money on somebody else, I’m not concerned about how much it is, and I’m not concerned about what I get. And that’s government. And that’s close to 40% of our national income.
Although focused on the act of spending money, Friedman’s point is more general. Each of us has better ability and stronger incentives to make good decisions for ourselves than we have to make good decisions for others. I know myself better than you know me, and I care about myself more than you care about me. Likewise, you know yourself better than I know you, and you care about yourself more than I care about you.
So uncover all the decision-making biases you can. If they distort decisions Mr. A makes for Mr. A, surely they distort even more any decisions Mr. A makes for Mr. B.
It’s true that Mr. A can often be more emotionally detached than is Mr. B from Mr. B and Mr. B’s problems, thereby sometimes giving Mr. A some advantage over Mr. B in making decisions for Mr. B. But this fact doesn’t salvage the case for generally giving Mr. A decision-making power over Mr. B’s life.
First, emotions do not always distort decisions; emotions exist for a reason – often to inspire us to choose actions that are best over the long-run. (Robert Frank’s Passions Within Reason  remains my favorite reference for this important point.)
Second, when Mr. A is charged with making decisions for Mr. B, Mr. A’s judgment might be distorted by his own emotional biases – biases that differ from those that distort B’s judgment, but biases nevertheless.
Third, each of us has incentives to try to overcome our own harmful emotional biases when making decisions for ourselves; when others make decisions for us, the incentives for these others to overcome whatever harmful emotional biases that affect their decisions regarding us are unlikely to be as powerful as are our own incentives to overcome our own biases.
Fourth, consistent application of this justification for government intervention would intrude into people’s private lives in ways that almost no one wants. For example, romantic passion is charged with high emotions, emotions famous for distorting sound judgment – and emotions that not infrequently result in unwanted serious consequences, such as pregnancy or killer diseases. Should we police against romance and lust? Do we give Mr. A the authority to decide whether or not Mr. B may sleep with Ms. C or have sex with Mr. D? After all, the consequences of an emotionally distorted, unwise decision can be very serious, perhaps lethal, for Mr. B.
Eric Crampton and I say more about this issue in this paper .
(Thanks to Jack Wenders for pointing out Asman’s interview of Friedman.)