Because the transactions in question are voluntary and among adults, it isn't anyone else's business what transpires between A and C. Would Arnold applaud a DHS ban on the sale of Big Macs, for it's quite easy to construct an argument, very much like Arnold's above, for government to prevent unsophisticated people from buying food that those of us in the know know a sophisticated diner would not buy. I understand well and sincerely that one might draw out differences between the buying of Big Macs and the buying of credit-card payment-protection plans; the question would be just how compelling those differences turn out to be upon careful reflection. Why, for example, would government action on the Big Mac front be objectionable nanny-statism while government action on the credit-card front be applause-worthy government intervention?
Consider two actions.
1. B sells C a product that A does not like (say, cigarettes or jet-skis).
2. B sells C a product on the basis of fraud (say, snake oil).
One approach is to say that in both cases a third party should interfere via regulation. Another approach is to say that in neither case should there be third-party regulation--it's buyer beware all the way. But one also can try to differentiate between the two.
In the case of (1), my thought is that A has no right to impose his values on B and C.
In the case of (2), my thought is that if B goes unpunished, then we are headed for trouble. With no punishment, the upside for B is profit, and the downside for B is little or no loss. Since I want resources going into productive innovation rather than fraud, I think I have an interest in punishing B, even if I was not the victim.
The products in question fall somewhere in between (1) and (2), but in my view they are closer to (2). If I sold someone cigarettes or jet skis, I could easily say to myself, "I do not like what I am selling. But if I had the same preferences as my customers, I would buy them." With the Capital One credit card add-ons, I cannot say that.
So I think I have a test that would allow for some consumer protection without heading down a slippery slope toward paternalistic regulation. That test is, "Can you imagine a set of preferences that would make you want to buy this product?" If the answer is "no," then regulation is justified. However, if you can imagine a set of consumer preferences that would make you want the product, then regulation embodies paternalism and should be subject to libertarian opposition.
Incidentally, the CFPB's mortgage regulations, which I discussed in an update to my original post, do not meet this test. They do cross over into paternalism.