Copyright updates

Paul Heald's excellent graph has brought commentary from all over the web; it's been fun watching where it's popped up since he gave the talk here at Canterbury. If I had to guess, a talk for 15 economists at Canterbury hit an audience, for that graph, in the hundreds of thousands.*

TechDirt usefully notes that things are even worse than pictured. While copyright protection extends back to the 1920s, that's only for works where the rights-holder has renewed his copyright. Rights to most works aren't renewed. But it isn't always easy to figure out whether or not the rights-holder has renewed the rights, and getting it wrong can be costly. Masnick writes:

This is something most copyright supporters ignore: entering the public domain can actually renew the value of art, and can (and does) stimulate the economy by allowing others to exploit additional commercial value from a work beyond what was possible under copyright. The commercial usefulness of a monopoly on a book has a shorter shelf-life than the monopoly actually granted by copyright law. Based on Patry's findings, that shelf life is somewhere under 28 years, otherwise more people would have renewed their registration—but copyright lasts much longer than 28 years. Thus you get the giant gulf on Heald's chart: in between the pre-1923 public domain books and the books that are new enough to still be actively sold, there are several decades of titles that are no longer worth anything to their rightsholders, but can't be offered by anyone else because they are still effectively under copyright.

Yes, just effectively—not actually. As you may have noticed, there seems to be a contradiction here: if the majority of copyright registrations went un-renewed, then the majority of books published between 1923 and 1963 have lapsed into the public domain alongside the books from 1922 and earlier, so the drop-off in Heald's chart should be much, much smaller. This is not a conflict in the data, it's a symptom another massive and entirely separate problem with copyright law which I discussed in a recent post: the difficulty of determining a work's status.

Copyright is a good thing. But not at its current duration or its current scope.

*An incomplete summary: Marginal Revolution, Matthew Yglesias, Kevin Drum, Rebecca Rosen [made the most popular on The Atlantic's front page on the Saturday after posting], Brian Doherty, Kevin Kelly, and FAIR. It's hit Memeorandum, Reddit, Hacker News, The Glittering Eye, the CEI's Open Economy Blog, Politikon, and LISNews. Topsy tracks the tweets; here are the +Ripples. And TechDirt and Information Liberation and Right to Read and Habr. At 15000 post views and counting, if the click-through rate to Offsetting from the other sources is maybe 5%, then multiply my views by 20 to get a ballpark audience of 300k.

Copyright stagnation

Paul Heald demonstrated the effect of the stagnant US copyright wall in seminar at Canterbury last week.

Recall that books published through 1922 are in the public domain in the US; those published since then are covered by copyright.

Heald dug through some Amazon stats to see what happens to books as they come out of copyright. Here's the rather stunning graph.

So any arguments about underexploitation of unprotected works seem untenable.

If this were a moving wall, maybe it wouldn't be so bad: eventually, books would come out of copyright and be released in new editions. But Disney does keep going back and insisting that nothing can ever be returned to the Commons from which they so liberally drew, and Congress loves Disney; we might reasonably expect another copyright term extension act to keep the wall fairly rigid.

So while I can get Pride and Prejudice in remix with either vampires or zombies,* I'm not betting on being able to read a version of Good-bye, Mr. Chips in which he protects his students from the werewolf menace as well as offering them solace through the Great War. The werewolf version practically writes itself - the Germans infect some injured British soldiers with lycanthropy just after a full moon, knowing they'll be back in Britain by the next full moon....

Here's Paul's SSRN page. The chart above isn't in any of his released papers, but is an update to some of the matters he covered here. His talk for the department is embedded below; the audio isn't fantastic, but all the slides are there.

* Pride and Prejudice is unreadable except in remix.


Update: Paul Heald clarifies the chart source data:

Hi, I just wanted to note that Amazon does not know when a book it sells was first published. It only knows the date of publication of the volume that it is selling, e.g. Treasure Island could have a date of 2002, if that’s the edition Amazon is selling. I had to check each of the 2500 books at the Library of Congress to determine the actual initial publication date. This is why stats taken from an Amazon “year of publication” stats don’t match up. Cheers, Paul Heald

See also discussion at Marginal Revolution, Matthew Yglesias, Kevin Drum, Rebecca Rosen [made 8th most popular on The Atlantic's front page], Brian Doherty, Kevin Kelly, and FAIR. It's hit Memeorandum, Reddit, Hacker News, The Glittering Eye, the CEI's Open Economy Blog, Politikon, and LISNews. Topsy tracks the tweets; here are the +Ripples. And TechDirt.