A Flaw in the Democratic Coase Theorem?

So, as I explained in the previous post, Cooter's democratic Coase Theorem is that in the absence of transaction costs, legislators will reach an efficient outcome. The basic idea is that any proposal that "makes the pie bigger" can fully compensate anyone harmed by the legislation, with some pie left over (I used the word surplus earlier, but that's confusing because I don't mean a fiscal surplus).

I thought I had identified an objection to this theory, but after more thought I'm not so sure. Imagine this scenario. One state elects a minority senator. Every other state is racist against that minority. Now, their distaste for a minority in the Senate might be strong enough that it's worth it for them to pay the state to elect a white person. Meanwhile, the state's desire to be represented by this particular senator might not be very strong (there might be a white candidate who is almost as good, as far as they're concerned).

There's room for a bargain here, so absent transaction costs, the state will elect a white senator in exchange for a payment from the other states. [If you're troubled by the idea that it might be "efficient" to satisfy racist preferences, just switch the example and make the senator himself a bigot, with most of the country bothered by the presence of a bigot in the Senate]. The reason this isn't satisfactory to me is that everyone recognizes that there are prohibitive transaction costs for individual voters. The question is whether transaction costs for legislators can be reduced to the point that some fruitful bargains are made. I think that this is possible, which is why the democratic Coase Theorem might be useful. I don't think the theory is at all applicable, though, to the voters in a general election (I should note that it still might be useful conceptually, but I find this particulay application inapt).

Now, say the theory is only applicable to the legislature. Now how does the country get rid of the minority senator? Again, strictly speaking, the theory has an answer. The senator should resign in return for some kind of payment to his constituents (in the form of federal spending or whatever). This counterintuitive result follows from the unrealistic assumption that legislators merely seek to represent faithfully the interests of their constituents.

Given this assumption, my objection doesn't stand. The "efficient" outcome is reached through a bargaining process. Instead of identifying a technical problem with the theory, I've proposed a situation in which application of the theory leads to implausible results. That's less of a strike against the theory than you might think, because the Coase Theorem (and Cooter's derivative theory) don't claim to describe the world accurately. Rather, they set out an ideal and then examine the ways in which we fall short of it, and the consequences of those shortcomings. I'll come back to this theme again, no doubt, but for now I'll just note that such theories can be misused rather badly when they are mistaken for accurate descriptions of the world.