In his blog today, Greg Mankiw argues that charging senders a small price to correct this externality is an appealing "Pigovian" solution. For instance, a reader of his blog writes:
[A] penny tax (say) on email would probably generate large amounts of revenue, mitigate an important negative externality, and have minimal inefficient disincentives. Since email servers are necessarily centralized and networked and all email senders are ipso facto connected to an ISP who is charging them for access the transactions costs and evasion problems seem low.
I agree that charging senders is the ideal way to solve the email congestion problem, especially spam. It is superior to technical solutions like filters because it helps align the sender and recipients' incentives and thereby harnesses the knowledge that the sender has about the content.
Unfortunately, the issues are a bit more complicated than he describes in his brief posting and thus this solution requires some new standards and infrastructure. Along with Marshall van Alstyne and Jon Koomey, I wrote an Op-Ed in the WSJ explaining some of these issues and proposing a way forward.
See "You've Got Spam" (WSJ, Sept 6, 2007).
Update: [Here's an un-gated version]