The problem here is bias to one’s own biases, I think. After all, we’re all stupid some of the time, and realising this is surely a core component of the Overcoming Bias project. Robin Hanson may not think he’d ever be stupid enough to walk into the Banned Shop, but we all tend to assume we’re the rational one.
You also need to consider the real-world conditions of your policy. Yes, this might be a good idea in its Platonic ideal form, but in practice, that actually doesn’t tell us very much. As an argument against “regulation”, I think, with a confidence value of 80, that it’s worse than useless.
Why? In practice, you’re not going to have “Banned Shops” with big signs on them. If enough people want to buy the banned products, and we know they do want them because their manufacturers are profitable, the rest of the retail trade will instantly start lobbying for the right to sell them, maybe on a Banned Shelf next to the eggs. That’s an unrealistic example, but then it’s an unrealistic proposal.
What’s more likely is a case of Pareto inefficiency—if you relax, say, medicines control on the grounds that it’s a step towards the ideal, the growth in ineffective, dangerous, or resistance-causing quackery is probably going to be a significant disbenefit.
I, for one, imagine that I could easily walk into the Banned Shop, given the right circumstances. All it takes is one slip up—fatigue, drunkness, or woozy medication would be sufficient—to lead to permanent death.
With that in mind, I don’t think we should be planting more minefields than this reality currently has, on purpose. I like the idea of making things idiot-proof, not because I think idiots are the best thing ever, but because we’re all idiots at least some of the time.
Certain types of content labeling might work a lot like Hanson’s Banned Shop, minus the trivial inconvenience of going to a different shop: the more obvious and dire the label, the closer the approximation. Cigarettes are probably the most advanced example I can think of.
Now, cigarettes have also been extensively regulated in other ways, so we can’t infer from this too well, but I think we can tentatively describe the results as mixed: it’s widely understood that cigarettes stand a good chance of killing you, and smoking rates have indeed gone down since labeling laws went into effect, but it’s still common. Whether or not we count this as a win probably depends on whether, and how much, we believe smokers’ reasons for smoking—or dismiss them as the dribble of a hijacked habit-formation system.
The problem here is bias to one’s own biases, I think. After all, we’re all stupid some of the time, and realising this is surely a core component of the Overcoming Bias project. Robin Hanson may not think he’d ever be stupid enough to walk into the Banned Shop, but we all tend to assume we’re the rational one.
You also need to consider the real-world conditions of your policy. Yes, this might be a good idea in its Platonic ideal form, but in practice, that actually doesn’t tell us very much. As an argument against “regulation”, I think, with a confidence value of 80, that it’s worse than useless.
Why? In practice, you’re not going to have “Banned Shops” with big signs on them. If enough people want to buy the banned products, and we know they do want them because their manufacturers are profitable, the rest of the retail trade will instantly start lobbying for the right to sell them, maybe on a Banned Shelf next to the eggs. That’s an unrealistic example, but then it’s an unrealistic proposal.
What’s more likely is a case of Pareto inefficiency—if you relax, say, medicines control on the grounds that it’s a step towards the ideal, the growth in ineffective, dangerous, or resistance-causing quackery is probably going to be a significant disbenefit.
I, for one, imagine that I could easily walk into the Banned Shop, given the right circumstances. All it takes is one slip up—fatigue, drunkness, or woozy medication would be sufficient—to lead to permanent death.
With that in mind, I don’t think we should be planting more minefields than this reality currently has, on purpose. I like the idea of making things idiot-proof, not because I think idiots are the best thing ever, but because we’re all idiots at least some of the time.
Certain types of content labeling might work a lot like Hanson’s Banned Shop, minus the trivial inconvenience of going to a different shop: the more obvious and dire the label, the closer the approximation. Cigarettes are probably the most advanced example I can think of.
Now, cigarettes have also been extensively regulated in other ways, so we can’t infer from this too well, but I think we can tentatively describe the results as mixed: it’s widely understood that cigarettes stand a good chance of killing you, and smoking rates have indeed gone down since labeling laws went into effect, but it’s still common. Whether or not we count this as a win probably depends on whether, and how much, we believe smokers’ reasons for smoking—or dismiss them as the dribble of a hijacked habit-formation system.