[ganeumann]

The part about "democratic choice" should be handled by our "democratically elected leaders." But I don't think I really understand your point at all. What does selling the IP v. giving it away have to do with democratic choice?

My point was that selling the IP instead of giving it away is revenue neutral to society (and, in fact, a positive to the taxpayers who are not buying the IP) and so selling it does not decrease welfare and may well improve it if it provides some data to the agency on what the users of their research value. This is entirely unlike what happens with NSF-funded research which is given away to journals who charge a fortune to read it. In this latter case (unless you think that peer-review is incredibly valuable, which no one besides the publishers seem to believe) the overall effect is welfare reducing.