Eric Hagermeyer, one of my students, had a very clever suggestion after class. It's so interesting that I have to write it down while I still have it clear in my head.
The purpose of intellectual property law is to encourage the production of ideas. Patents encourage invention. Copyright encourages art, literature, etc. Trademark allows firms to establish a reliable identity that is useful to consumers for assessing quality. The way they do this is by giving the holders of the intellectual property right an effective monopoly over the product, allowing them to earn profits from the work. So, for example, Rush's copyright of "Tom Sawyer" allows them to collect royalties whenever it is played. The classical argument for copyright is that a longer period of copyright protection increases the incentive to create stuff like "Tom Sawyer", because bands like Rush can earn those monopoly profits over a longer period of time. Copyright protection currently lasts for the life of the author, plus 70 years (that is, his heirs or assigns receive the copyright as an asset, and it can be enforced 70 years past the creator's death).
Eric pointed out, however, that it may not work like this for long copyright periods. Instead, lengthening the copyright may encourage some creators to have one big hit, and then rely on it for income for a long time. That is, really long copyright protection might create "one-hit wonders". These creators are, in a sense, substituting leisure for labor. The copyright allows them to earn so much from their hit, for so long, that they need not create anything else. This would tend to make copyright look like less of a good idea; we want creative people to keep creating, after all, rather than resting on their laurels.
But wait, there's more! On the other hand, this long period of profit may increase the incentive for people to try to create these one-hit wonders. So a really long copyright period may reduce the creations from each individual, but increase the number of people creating things. If a diverse array of one-time hits from many people is better than many creative works from a few people, then this could actually be an improvement. It depends on how much people value the diversity, and how much diversity we get.
This is a difficult theory to test, both because it's hard to measure the value of diverse published works, and because there aren't good natural experiments to test the effect of changes in copyright law on creative behavior. I still think, however, that it's quite clever. Good work, Eric. I do wonder if anyone has investigated this before; I haven't bothered to search Econlit for related papers. Still, this could be a good idea for a senior project, if a way can be found to test it...