In very brief, we need to clarify that "public use" in the Fifth Amendment does not mean merely "public benefit." Kelo, adopting that broad view, essentially upheld the forced transfer of property from one private party to another. (I say "essentially" because the Court makes some largely ineffectual noises about how the takings must advance a "carefully considered" development plan and not evince an intent simply to benefit one private interest at the expense of another.) That violates the original meaning of the Fifth Amendment, the plain meaning of the Fifth Amendment, and sound public policy.
It will not suffice to simply add an Amendment XXVIII saying, "In the Fifth Amendment, 'public use' really does means 'public use.' And we mean it, this time!" It's too late for that, given the long and sad train of precedents that has slowly eroded the proper meaning of "public use."
Nor will it suffice to say that the public must have access to any property taken under the Fifth Amendment. Takings for prisons or military installations surely qualify as constitutional no less than takings for highways or parks. And we cannot demand public improvements unless we want to rule out takings for the preservation of natural resources.
Jim Lingren's insightful Volokh Conspiracy post, pointing out that the law does a quite able job of distinguishing between taxable private enterprises and nontaxable charitable ones, suggests that we might import into Takings doctrine distinctions already used in the tax code. We might, in other words, adopt an amendment that says something like, "Any 'public use' in the Fifth Amendment must qualify as tax-exempt." The relative instability of the tax code renders it ill-suited as a foundation for constitutional law, however. Better that we should simply take Jim Lingren's point for what it is worth: A very convincing proof that the Kelo Court erred in claiming, "There is . . . no principled way of distinguishing economic development from the other public purposes that we have recognized." Tax law shows that the Court could have made that vital distinction.
Co-blogger Glen Whitman and I have independently toyed with the idea of an amendment imposing a fixed term of years before any property taken for public use can be transferred to a private party. Glen suggested ten years; I was thinking of twenty. That approach offers a bright-line rule that, if effective, would largely obviate the public choice problems created by Kelo. Still, though, I worry even that would not suffice to prevent abuses.
Consider this legal hack: Government agents could take private property from its rightful owner, hold fee simple title in the public's name, and offer a 99-year lease to a hotel developer. The public would own the property and have the right to use it. True, that use would come at a price. But is Yosemite National Park non-public because it charges an entry fee? No. It remains public because the government owns title and because it must give access to anyone willing to pay the requisite fee—just like the hotel in the legal hack I've described.
When I first approached this problem, some months ago, I put it this way: "[T]he Fifth Amendment's 'public use' requirement means that government officials must actually turn taken property into a public park, a public highway, a government office, or similar facility dedicated to public purposes." Putting that idea into solid legal language proves tricky, though. To close loopholes such as the one I described above, I think we need to resort to some legalese. I thus offer a Kelo Amendment invoking the common law definition of an interest in property, a notion sufficiently broad to cover ownership in fee simple and a leasehold estate (among other forms of ownership):
Amendment XXVIII: No taking of private property shall qualify as "for public use" under Amendment V if another private party obtains an interest in the property within twenty years of its taking.
This amendment would plainly rule out the sort of shenanigans upheld in Kelo. It would also rule out legal hacks of the sort I worried about earlier, wherein the government takes title but offers a private party a long-term lease. Note that this Amendment deliberately says nothing to affect the theory of regulatory takings, pro or con. Nor does it specifically limit itself to real property; takings of private chattel or intellectual property would win the same protections. The reference to "another private party" allows the party who originally suffered the taking to regain it, while still enjoying the "just compensation" guaranteed under Amendment V. I settled on twenty years because it seems easily long enough to avoid public choice problems and because that same span sometimes measures the term necessary to establish adverse possession in the law of real property.