Note: This article was updated and expanded in December 2012. You can view the new version here.
One of the biggest problems with today's copyright system (and patent system, for that matter) is that there is no economic pressure pushing works toward the public domain. Instead, the incentives are all structured toward copyrighting for as long as possible — far past the point of really benefitting even the copyright holder, let alone the public. Below is a proposal for a truly balanced policy, one that that would give copyright holders a motivation to release works sooner, and give the public a way to influence how long any particular work stays under copyright.
This article describes a new copyright system, one designed to reflect the public's interest in moving works into the public domain, while at the same time preserving the most useful market characteristics of the current system. It works through a tempered monopoly: instead of offering today's automatic, fixed-term, and essentially zero-cost copyright grant, it offers a monopoly whose value and duration are based on a combination of the public's and the creator's interests, which are not necessarily at odds.
Best of all, it is simple and low-overhead. It does not require any complex negotiations, third-party assessments, or intricate pricing formulas. Here's how it works:
First, new creations would still get an initial automatic copyright term, but fairly short: on the order of three to five years. Today's copyright terms automatically last many decades — holders are given years of monopoly power regardless of whether they actually want or need it. Thus the first step is to arrange things so that, if the owner does nothing, works will enter the public domain much sooner than they would today.
Later we'll look at the question of initial term lengths in more detail. For now, let's focus on the moment when a work is about to default into the public domain under this new system.
Assuming the copyright owner takes no further action, the work just enters the public domain .
But there's an alternative: instead of letting the monopoly lapse completely, the copyright owner can choose to register the work each year for continuation of copyright, with a registration fee proportional to the self-declared value of the work. That is, the copyright owner picks a number of dollars that she claims the work is worth. It can be any number at all, but the yearly registration fee will be a percentage of it — for discussion's sake, say 2%. (The exact proportions don't matter here: it could be 1% or 5% instead of 2%, registration could be semi-yearly instead of yearly, etc. The idea is the same, regardless of the details.)
Now comes the key: since that declared value is now a matter of public record, anyone can pay it to the copyright owner to liberate the work into the public domain. This would be a mandatory transaction: the copyright holder has declared what the work is worth, and has an incentive not to declare too high, because she'll have to pay a percentage of it to renew every year. Anyone willing to pay the total amount can cause the work to go into the public domain.
Because the value of a work may change over time, the registrant may adjust the declared value up or down each year when renewing the registration . This is also one of the reasons behind that initial brief (but free) monopoly term: it gives the copyright holder a chance to see what the work is worth, and she can use that information to decide how much to register the work for when the initial term ends.
Let's run through an example. Suppose I write my great novel, "The Helprin Affair". For a brief period of time, I have a regular copyright, just like today. After that, I decide to register to keep my copyright. Calculating (based on sales so far) that my novel is worth US $100,000, I therefore pay $2000 to the copyright office, which I can afford, because the novel has been selling strongly and I'm sure I'll make that back in royalties this year too. The next year, I make the same decision... But eventually, there will come a time when it's no longer worth it to me. At that point, I can either reduce the declared value, so as to pay a smaller fee, or release the work entirely. (There should be an upper limit on renewals, however; no work should be kept out of the public domain permanently. Monopolies are still monopolies, after all.)
At any time after my initial registration, some third party can look up the declared value of "The Helprin Affair" in the copyright office's records, and pay me $100,000 to liberate it on the spot, or in later years perhaps less or more, if I've changed the declared value in subsequent registrations.
Note that this is not a purchase of the copyright itself, but rather a liberation of the work from copyright entirely. People would still be free to lease or sell their copyrights as before, for whatever price they can get (which, interestingly, may be higher or lower than the registered value — the market dynamics behind that decision are just as rich as those involved in determining a work's value under today's copyright system). But whoever the owner is, whether the author or someone else, they're responsible for keeping up the registration. And while the work is still under registration, anyone can come along and pay the registered owner the declared value to liberate it. The justification for making this a mandatory transaction is that since the copyright owner chose the price in the first place, it is by definition fair, and the public's ultimate interest is in having the work be available to everyone, without restriction.
The initial term, for which no registration is required, has some purposes beyond just giving the creator a chance to assess the market value of the work.
First, requiring explicit registration from the very beginning would penalize small-scale artists disproportionately: they are likely to be short of resources and to either forget to register, or not know how. They would thus be unfairly disadvantaged in comparison to those working with publishers, who have a legal staff to remind them of registration requirements. But after an artist has a chance to see a work succeed, they're not likely to forget to register it (though a few may still forget — no system is perfect).
Another reason to grant a free initial period of copyright is conservatism. Although I'm not convinced that copyright is truly necessary at all in the Internet age, there is at least an argument for giving new works a "clear landing space" for a few years after their creation: a default exclusive copyright, much like today's but lasting a shorter time, can help give the work time to establish its own identity. Again, I'm not sure this is really necessary — but it wouldn't be excessively harmful, and it might make a lot of people more comfortable with reform.
There is a balancing act going on here. It may be useful for the original work to have time to become known on its own terms, before derivatives appear that capitalize on the momentum of the original work's initial appearance but that might also eclipse it before it has a chance to make an impression. On the other hand, we also want derivatives to become possible during the "reaction cycle" of the original work, that is, during the time when the original work is still relevant, so that derivatives have a chance to be meaningful responses. (This is very similar to the right to fork in open source software; see The Wind Done Gone for an example of why this right is important in art).
So what's an appropriate amount of time for the original work to make its splash, before having to contend with derivatives? I won't pretend to know for sure, but for the purposes of discussion, let's say three years. Instinctively, anything over five years feels too long to me, and less than a year feels too short. But it doesn't matter here what the exact number is; three will do for now. That's a period of time within which most royalty-generating works make the majority of their revenue anyway. Registrations could then continue the monopoly for another ten even fifteen years (though I think the latter number still too high, because it's so punishing to independent derivative works).
While this proposal is a compromise, it's at least a compromise tilted toward the public interest. By analogy, think of a homeowner who cuts a driveway opening onto a public street, in order to gain access to a private garage. If I take a streetside parking space away from the public, I expect to pay the city (that is, the public) a fee, and usually annually, too, not just a one-time fee. Similarly, a copyright owner who wants to keep a work out of the public domain should pay for that privilege. But unlike a garage, this privilege need not be permanent, because losing monopoly control over a work after roughly a decade is not as serious as losing one's indoor parking space.
This system would go a long way toward alleviating the orphan works problem, by ensuring that the copyright owner of a work could be found (someone must be paying the fee over at the registry), and toward alleviating the ghost works problem (in which derivative works are suppressed), by setting a maximum amount of money that, for the vast majority of works, would probably still be affordable for a motivated party wanting to see that work in the public domain.
The copyright lobby frequently talks of finding an appropriate "balance" between the needs of creators and the needs of the public. Like many appeals to balance, it is usually a smokescreen for something else: in this case, for efforts to increase copyright terms and restrictions beyond their already grotesque lengths. The "balance" they're talking about neatly presupposes that creators and the public are somehow on opposite sides, while publishers are, curiously, absent from the picture altogether. (Their portrayal is also historically suspect, as copyright was designed to subsidize distribution more than creation — that is, it's been about publishers, not artists, all along.)
Nevertheless, why not take them at their word? If balance is such a good thing, let's have it: a true economic balance between the desire of copyright holders to retain exclusive control, and the desire of the public to let creative works roam freely. If that kind of balance is what we want, then the Declared Value system comes much closer to it than do today's one-sided copyright laws.
 Entering the "public domain" does not, of course, mean that accurate attribution is no longer protected. Crediting and copying are entirely separate things, and while the public's interest is served when works can be copied and derived from, it is not served if people take credit for what they did not do. Thus it may be wise to legally protect attribution separately from copyright, although note that the absence of dedicated attribution laws laws has not resulted in a plague of plagiarism on the Internet. This is probably because widespread distribution and search tools are a far better protector of attribution than any law: there is no plagiarism problem in the world of open source software, for example. We've written elsewhere on the protection of attribution and how plagiarism is not related to copyright, in this article about Google Book Search, this article about how digital analysis resulted in the swift detection of a major instance of plagiarism, this article on distinguishing between copyright violation and plagiarism in the classroom, this article on how New York University apparently fails to make that distinction, this reader's comment on that article, and this article about the history of copyright.
 Alternatively, the owner could be allowed to adjust the declared value at any time (perhaps even as a reaction to liberation offers), with the provision that any upward adjustment would require immediate payment of the difference between the old and new registration fees. However, the public domain would be better served by simply allowing adjustment only at fixed intervals: if the owner of a work can't figure out its market value and set the fee accordingly, that is no reason to favor the owner over the public when the work is being liberated at a price the owner once clearly thought sufficient.