The 1710 UK Statute of Anne established a term of protection of 14 years for an original period of protection, followed by a 14 year renewal term. This is the same durational structure adopted in the U.S. in the 1790 Act. The length of copyright has been on an upward arc ever sense, with the current regime in the U.S. at the obscene level of life of the author plus 70 years, a term utterly divorced from the reasonable need for incentives to create.
During the briefing leading up to the U.S. Supreme Court's oral argument in the Eldred case, 537 U.S. 186 (2003), which upheld as constitutional the 1998 term extension, a fine brief by economists such as Ronald Coase and Milton Friedman was submitted. The brief concluded that the additional twenty years provided was a windfall that would lead at most to a statistically trivial incentive to create, little if any different from a perpetual term. At the same time, the economists asserted that there were significant negative social costs associated with term extension, principally in the form of increased costs for existing works whose term was extended, thus leading to an overall reduced level of innovation (e.g. prohibiting the creation of derviative works building on existing works).
The brief did not attempt, though, to set out an optimal term of protection. Last week, Rufus Pollock,who is pursuing a Ph.D in Innovation and IP at Cambrdige University and who recently accepted a three-year fellowship at Emmanuel College (also Cambridge), has made available (here, at his website) a copy of the paper doing just that. The paper is called, ""Forever Minus a Day: Some Theory and Empirics of Optimal Copyright." The reference to "forever minus a day," is to a remark attributed to the late Jack Valenti (may his Rock keep and save him). Mr. Pollock concludes that the optimal term of protection to be around 14 years, half of that possible under the two early acts mentioned above. One interesting point about his analysis is that it is based on the dramatic (and he believes permanent) fall in the costs of production as a result of the savings from digital means of production.
He does of course also discuss the difference between low costs of production and the other side of the equation, the low costs of reproduction, and this is precisely what I find interesting: in in the policy debates about digital technology, going back in my experience in Congress to 1992, digital technology was regarded as a monolithic monster poised to forever destroy copyright unless some powerful weapons (read DMCA) were put in place. Mr. Pollock's paper is unique for looking at the larger picture, i.e., that is production costs and how those costs impact on the extent of incentives required. The paper is chock-full of statistics which is likely to turn many away, but there are other papers on his site (link provided above) that may prove easier going.