On Selling Votes
I said No. It is true that both parties in the transaction must be better off if they agreed to the deal. Nonetheless, the standard argument for unfettered voluntary exchange does not apply because there are externalities. That is, when one person sells his vote to another, that transaction may affect unrelated third parties through the electoral process.
Here is an example of what I had in mind. Suppose three voters are deciding whether to provide a public good that costs $9, which would be financed by a $3 tax on each voter. Andy values the public good at $8, while Ben and Carl do not value it at all.
Under majority voting, Ben and Carl vote against, and the public good does not get provided, which is the efficient outcome.
Suppose, however, that Andy could buy Ben's vote for $4. He could then ensure the project gets passed. Andy is better off by $1 (the $8 benefit minus the $3 tax and the $4 price of the vote), Ben is better off by $1 (the $4 price of the vote minus the $3 tax), and Carl is worse off by $3 (the $3 tax). The Andy-Ben vote deal has negative externalities on Carl.
One can concoct examples in which selling votes enhances efficiency. (Increase Andy's valuation from $8 to $10). So I do not mean to suggest that vote trading is always inefficient. The purpose of this example is only to show that externalities abound. Standard conclusions about the benefits of voluntary exchange do not readily apply to the vastly complicated situation of democratic voting.
Update: Michael Sandel emails me some commentary on this post:
Professor Mankiw argues that we should not allow the buying and selling of votes, despite the fact that both buyer and seller would be better off, because such exchanges create externalities. I agree that we should not allow a market in votes. But the reason cannot be that vote-selling creates externalities. Persuading people to change their minds about which candidate or policy to support also creates externalities. But I doubt Professor Mankiw would say that we should therefore prohibit political persuasion.
Consider his example: Three voters are deciding whether to vote for a public good that costs $9, to be financed by a $3 tax on each voter. Andy values the public good at $8, while Ben and Carl do not value it at all. Suppose the policy in question is a tax override to support the public schools. Andy, who has two children in the public schools, strongly favors the override. Neither Ben nor Carl has children in the public schools, which explains their initial reluctance to assign any value to the lower class size, improved school library, and renovated science labs the override would finance.
Now suppose Andy persuades Ben that everyone in the community has a stake in the quality of the public schools, regardless of whether they have children who benefit directly. Perhaps Ben comes to believe that a well-educated citizenry will make for a more prosperous local economy, or a more healthy democracy. Or perhaps he is simply persuaded that strong public schools will bolster the local real estate market, and enhance his property's value. He now comes to value the public good associated with the override at $4. Since this is more than the $3 tax he will have to pay, he joins Andy in voting for the override.
Carl, who remains unpersuaded of the override's merits, votes against it and loses. His taxes go up by $3, for a public good he does not value. Andy's persuasion of Ben has negative externalities for Carl.
This example shows that Professor Mankiw's reason for opposing vote selling--that it imposes externalities on third parties--can't be right. Whether Andy buy's Ben's vote or wins it through persuasion, the effect on Carl (the "negative externality") is the same. In both cases, Carl loses a vote he would otherwise have won, and is saddled with the $3 tax. So unless Professor Mankiw would prohibit political persuasion, the negative externality on Carl does not explain why vote-selling should be prohibited.