How long should copyright be? Should we increase or decrease the strength of copyright during periods of rapid technological innovation? These are all questions I address in my paper entitled Forever Minus a Day? Some Theory and Empirics of Optimal Copyright, which I will be presenting at the 2007 SERCI Congress in Berlin this week.
For those who want to know more, the full abstract is below and the latest version of the paper can be downloaded from:
http://www.rufuspollock.org/economics/papers/optimal_copyright.pdf
Update: (2007-07-13) there are a set of summary slides available here:
http://www.rufuspollock.org/economics/papers/optimal_copyright_talk.pdf
Update: (2007-07-16) for those interested in republishing, redacting, or otherwise reusing the paper I should state clearly that it (and this blog) are licensed under a Creative Commons Attribution (by) license v3.0
Update: (2007-08-07) I’ve produced an updated version of the paper. This includes a variety of small corrections (typos etc) and some more substantial reworking. In particular due to the reinclusion of the ’small’ middle term in the statement of Theorem 13 and in subsequent calculations (previously in the proof but omitted from the statement as small) optimal term has increased from 14 years to 15 years (which bears out the original assumption that the term was small but it is nice to explicitly include it). There is also a new figure (Fig 1 in the updated paper) which gives the most accurate representation of the results in the form of the probability distribution of optimal term given the parameter range being used.
Abstract
The optimal level for copyright has been a matter for extensive debate over the last decade. This paper contributes several new results on this issue divided into two parts. In the first, a parsimonious theoretical model is used to prove several novel propositions about the optimal level of protection. Specifically, we demonstrate that (a) optimal copyright falls as the costs of production go down (for example as a result of digitization) and that (b) the optimal level of copyright will, in general, fall over time. The second part of the paper focuses on the specific case of copyright term. Using a simple model we characterise optimal term as a function of a few key parameters. We estimate this function using a combination of new and existing data on recordings and books and find an optimal term of around fifteen years. This is substantially shorter than any current copyright term and implies that existing copyright terms are too long.
July 12th, 2007 at 3:16 pm
[…] You can download a copy of Pollock’s full paper here. […]
July 12th, 2007 at 3:56 pm
[…] Es gibt tatsächlich Menschen die sich wissenschaftlich damit befassen, wie lange das Urheberrecht gelten sollte. Das ist nicht nur erstaunlich, sondern führt auch zu erstaunlichen Ergebnissen, die ein wenig ähnlich dem was man unter The Long Tail versteht, durch Einführung einer zeitlich-ökonomischen Funktion letztendlich sagt: 14 Jahre. Klingt akzeptabel, und je schneller die Technologie avanciert, desto Kürzer sollte es laut Rufus Pollack werden. (Nein, wir haben das unten verlinkte Paper weder komplett durchgelesen noch verstanden s.o., aber wir arbeiten dran). Aber, hey, unsere Regierung besteht nur aus solchen Menschen wie wir. The optimal level for copyright has been a matter for extensive debate over the last decade. This paper contributes several new results on this issue divided into two parts. In the first, a parsimonious theoretical model is used to prove several novel propositions about the optimal level of protection. Specifically, we demonstrate that (a) optimal copyright falls as the costs of production go down (for example as a result of digitization) and that (b) the optimal level of copyright will, in general, fall over time. The second part of the paper focuses on the specific case of copyright term. Using a simple model we characterise optimal term as a function of a few key parameters. We estimate this function using a combination of new and existing data on recordings and books and find an optimal term of around fourteen years. This is substantially shorter than any current copyright term and implies that existing copyright terms are too long. miscellaneous factZ » Blog Archive » Forever Minus a Day? Some Theory and Empirics of Optimal Copyright […]
July 12th, 2007 at 5:57 pm
[…] Optimal copyright! 12Jul07 miscellaneous factZ » Blog Archive » Forever Minus a Day? Some Theory and Empirics of Optimal Copyright Specifically, we demonstrate that (a) optimal copyright falls as the costs of production go down (for example as a result of digitization) and that (b) the optimal level of copyright will, in general, fall over time. […]
July 12th, 2007 at 7:01 pm
[…] July 12th, 2007 A copyright is a monopoly, granted in order to stimulate work in “science and the useful arts.” It is in the interests of users of such work that the monopoly is granted for a limited time. So what should the limit be? Rufus Pollock has researched this question, and posted his answer. Using a simple model we characterise optimal term as a function of a few key parameters. We estimate this function using a combination of new and existing data on recordings and books and find an optimal term of around fourteen years. This is substantially shorter than any current copyright term. […]
July 12th, 2007 at 7:17 pm
rufus, is it possible to make a summary of this paper suitable for popular reading?
I think the reddit.com audience would appreciate it
July 12th, 2007 at 7:31 pm
[…] La durata ottimale del diritto d’autore non deve andare oltre i quattordici anni, secondo Forever Minus a Day? Some Theory and Empirics of Optimal Copyright, uno studio presentato nei giorni scorsi a Berlino. L’analisi, scandita da teoremi, equazioni e concetti econometrici, stata condotta da Rufus Pollock dell’universit di Cambridge e nell’abstract si legge: In the first, a parsimonious theoretical model is used to prove several novel propositions about the optimal level of protection. Specifically, we demonstrate that (a) optimal copyright falls as the costs of production go down (for example as a result of digitization) and that (b) the optimal level of copyright will, in general, fall over time. The second part of the paper focuses on the specific case of copyright term. Using a simple model we characterise optimal term as a function of a few key parameters. We estimate this function using a combination of new and existing data on recordings and books and find an optimal term of around fourteen years. This is substantially shorter than any current copyright term and implies that existing copyright terms are too long. […]
July 12th, 2007 at 9:08 pm
Urheberrecht: Mal was Neues…
Gerade bin ich ausnahmsweise mal im Netz und im Fernsehen unterwegs: Ich habe mich kürzlich schon einmal anlässlich des Internet-Inkompetenz-Outings unserer Bundesminister und über den dringenden legislativen Nachholbedarf im Bereich des…
July 12th, 2007 at 10:37 pm
Your paper makes for interesting reading.
I was wondering if a similar line of reasoning would also work for patents. The effort put into creating, manufacturing and distributing work based on a patent seems remarkably similar to creating, reproducing and distributing a work under copyright.
July 12th, 2007 at 11:06 pm
[…] July 13, 2007 at 04:03 e · Filed under IP, Economics miscellaneous factZ » Blog Archive » Forever Minus a Day? Some Theory and Empirics of Optimal Copyright The optimal level for copyright has been a matter for extensive debate over the last decade. This paper contributes several new results on this issue divided into two parts. In the first, a parsimonious theoretical model is used to prove several novel propositions about the optimal level of protection. Specifically, we demonstrate that a optimal copyright falls as the costs of production go down for example as a result of digitization and that b the optimal level of copyright will, in general, fall over time. The second part of the paper focuses on the specific case of copyright term. Using a simple model we characterise optimal term as a function of a few key parameters. We estimate this function using a combination of new and existing data on recordings and books and find an optimal term of around fourteen years. This is substantially shorter than any current copyright term and implies that existing copyright terms are too long. […]
July 13th, 2007 at 4:23 am
The reply posted on 12-jul-2007 19:07 calls copyright a monopoly, but I think that’s true only in a limited sense. A /patent/ is a monopoly granted by the State; regardless of how one creates something that has been patented by someone else, he cannot sell it without the patent-holder’s permission (licensing) without risking a lawsuit. A copyright that U get on your work, on the other hand, does not prevent me from creating a similar work; rather, it prevents me from /copying/ your work. Your lawyer may sue on grounds that I violated your patented software, but assuming patents are not involved and there’s no proof I copied your work, he will refuse to sue on copyright grounds.
July 13th, 2007 at 6:26 am
[…] miscellaneous factZ » Blog Archive » Forever Minus a Day? Some Theory and Empirics of Optimal Copyright How long should copyright be? Should we increase or decrease the strength of copyright during periods of rapid technological innovation? These are all questions I address in my paper entitled Forever Minus a Day? Some Theory and Empirics of Optimal Copyri (tags: business copyright economics law paper research) […]
July 13th, 2007 at 10:41 am
lofi: I’ve put up some summary slides at here
http://www.rufuspollock.org/economics/papers/optimal_copyright_talk.pdf
Unfortunately it is a pdf file (they come from latex and latex2html does not seem to work well on them) but they do provide a simple overview of the results. All of my papers (and presentations) are under cc-by so please feel free to modify/adapt etc etc.
Arno: I think a similar argument could apply for patents but one would need to be careful since it is not necessarily obvious that costs have gone down (Pharma and Biotech are still pretty expensive …). Another big difference is with patents, the degree of ‘interaction’ between different works, due to cumulative innovation and componentization, .is much greater This is something one can abstract from a lot more in copyright as, crudely, copyright is for ‘expressions’ not ‘ideas’ (unlike patents) so the overlap and dependency between different works is much less than for patents. I have other work which deal with such questions in more detail, in particular my paper on ‘Cumulative Innovation, Sampling and the Hold-Up Problem’ which you can find on this page:
http://www.rufuspollock.org/economics/
July 13th, 2007 at 2:39 pm
When you say the optimum copyright term should go down in times when the technology of copying improves, this
statement flies in the face of “Five Information Ages”
from The Gutenberg Press to the High Speed Steam Press
to the High Speed Electric Press to the Xerox Machine,
and, finally, to the Internet.
Each one of these “Five Information Ages” has been met
by a legal response that simply outlaws competition of
the new technology by reducing available reprints.
In fact, the entire history of reprint houses could be
a launch pad for serious research on this subject.
Please advise,
Thank You!
Michael S. Hart
Founder
Project Gutenberg
July 13th, 2007 at 2:44 pm
great paper and nice framework. Especially the acknowledge of technology which is usually ignored in papers.
July 13th, 2007 at 4:30 pm
Major congrats, Rufus. Quick question: Is that link to the summary slides correct? It doesn’t work for me…
July 13th, 2007 at 5:25 pm
Is it simply coincidence that the original Copyright Act of 1710 (the Statue of Anne) specified a term of 14 years?
July 13th, 2007 at 6:40 pm
There’s a typo on page 17, in footnote 15, which begins “The the”. Remove the redundant word.
July 13th, 2007 at 9:10 pm
[…] As the second day of the conference is winding down, I’m happy that I have this blog to document the great presentations I’ve seen today. Michael Geist is answering questions right now - someone has brought up the Rufus Pollock paper on “optimal copyright duration” and Michael has pointed out how immensely long copyright periods are and how strange that is, for example in regards to software. […]
July 15th, 2007 at 2:08 pm
[…] announcement on his blog […]
July 15th, 2007 at 8:43 pm
[…] As reported in the Chronicle of Higher Education last week, an Oxford graduate student in economics is soon to publish a paper arguing that the “optimal” term of copyright protection is just 14 years. This is vastly shorter than the current term of protection in the US, where the term is life of the author plus 70 years, or in nearly any other nation of the world. Although his conclusion may be too radical to be practical, Rufus Pollok’s calculations add some weight, if any was needed, to the argument that copyright protection has moved very far from its original goal of providing an incentive to authors to create and now nearly exclusively serves the economic interests of large commercial distributors. […]
July 15th, 2007 at 10:50 pm
[…] [Link via Boing Boing] […]
July 15th, 2007 at 11:20 pm
[…] miscellaneous factZ » Blog Archive » Forever Minus a Day? Some Theory and Empirics of Optimal Copyright “Using a simple model we characterise optimal term as a function of a few key parameters. We estimate this function using a combination of new and existing data on recordings and books and find an optimal term of around fourteen years.” (tags: copyright reform theory research data yarp) […]
July 16th, 2007 at 4:55 am
I have an opposing view; I think copyright should last forever and be transferable/inheritable, but it should not be restrictive (in the way that Patents are). On the contrary, technology should facilitate the use of copyright material while ensuring the I.P. owners get paid for usage.
I come from the music industry and also worked for over a decade in design, advertising and ideas creation. I think the sophisticated copyright model built into the music business could and should be extended into other forms of copyright exploitation.
For example; if I create and ‘release’ an original piece of music, I have no power to stop anyone else from either recording their own version of it, broadcasting it, or performing it, despite the fact that I own the copyright. This is because I have assigned those rights to various collection societies and organisations around the world who have sophisticated systems in place to ensure that if any of those things do happen, I will be paid for it (as long as industry standard licenses are obtained). These systems can also monitor the duplication of my original work and collect payments from licensed duplicators.
The fact that I can rely on my intellectual property still being valuable in the future allows me to trade on it and continue to contribute to an ever growing knowledge based economy. I can see no benefit in or any reason why the value of my asset should suddenly be zeroed fourteen years after I create it. If I buy land and build a house, I own it in perpetuity. There should be no economic difference between a tangible and an intangible asset.
Some may say that the music industry model is already a failed one because new technology allows no cost duplication and distribution of the copyright work. However, in my opinion, we are simply in a transitional period and within a few years, the technology that facilitates royalty free copying and distribution will have become sophisticated enough to monitor copyright ownership and also aggregate micro-payments for usage on behalf of the copyright owners (and I know there will always be some usage that slips through the net - there always was).
On a side note, imagine radio stations being able to play any music that is over fourteen years old without having to pay royalties - what would happen to the new artists trying to break through? How would that in any way stimulate the future economy or the creative energies of the artists?
Long copyright periods coupled with statutory licensing and sophisticated royalty gathering technologies are the solution. Short copyright periods undermine the financial stability of the I.P. economy.
July 16th, 2007 at 8:10 am
The problem is: how much should IP owners be paid ?
You cannot set up a market if you cannot restrict use.
Bad example. Houses don’t last forever. They need expensive upkeep, so that after a few decades it’s usually more economical to tear them down and build them new.
Even land (which does last forever) is subject, in some jurisdictions, to expropriation if unused or abandoned.
July 16th, 2007 at 8:44 am
StefanoC Says:
The problem is: how much should IP owners be paid ?
In the music business, there are industry-wide agreed rates for most usage with individually negotiated rates for specific uses (licenses for commercials, featured music in film etc). This system could be extended. Sure, it’s not easy, but that’s no reason to destroy the value of copyrights after such a short period. Better systems build stronger economies.
You cannot set up a market if you cannot restrict use.
Isn’t that an argument for copyright protection? The economy of the future will be very largely a knowledge based, intellectual property market. It will need legal protection in order to flourish.
A house may not be a perfect example, but regardless of whether it costs money to maintain or not (artistic works don’t usually suffer from the weather, but often do need to be re-packaged, re-marketed, re-recorded etc), no householder would accept their home becoming forfeit after 14 years. Let’s not get lost in picking apart analogies - my point is that I.P. is an asset that shouldn’t go from full value to zero value because of government edict. Or perhaps all assets should be nationalised? Didn’t someone try that already?
July 16th, 2007 at 10:41 am
First, thanks to everyone for all the comments. I’ll do my best below to respond to those who had particular questions.
Michael Hart: as I comment in my paper, I think it is very hard to see copyright policy over the last couple of hundred years (at least in relation to term) as being determined by a careful evaluation of the costs and benefits for social welfare. Rather, the main controlling factors are those from political economy (i.e. lobbying power). Thus I’m not sure we can conclude very much from what has actually happened over the last couple of hundred years regarding what should have happened.
Kevin Anderson: Yes it is a coincidence that the point estimate came out at 14 years. I should emphasize, as I do in the paper and presentation, that clearly such an estimate which is itself based on a specific model and set of parameter estimates will be subject to error. In the paper I examine how robust the term is to varying the parameters within the plausible range (Table 1). My basic conclusion is that while there is substantial variation (directly flowing from variations in the estimate of e.g. deadweight loss) we can be fairly confident that current terms are too long.
Scrutin Eyes: thanks for the typo-spotting.
Steve Baker: in general I agree with the main point you are making, namely that some kind of blanket-license-based alternative compensation system would be a better way of paying for and disseminating creative work. Such an approach, by moving to an up-front payment system (i.e. one which is not pay-per-use), could well deliver a better deal both to society and to artists.
The key requirement here would be for copyright to change from an exclusive right to one more like a right to equitable remuneration. I know that such a scheme has been advocated by several prominent people in recent years, most notably Terry Fisher of Harvard in his books ‘Promises to Keep’.
On the question of term, it is not immediately obvious to me whether such a ‘remuneration right’ would be time-limited or not. After all, with a discount rate of e.g. 6% one dollar for ever is equal to an up-front payment today of ~$18 thus a regime with a one year copyright that paid someone $18 for that one year would be equivalent to a perpetual copyright which paid $1 for each year.
Finally, while agree with the general direction of your comments I do have some reservations on some specific points. For example you say:
But this is a standard fallacy. Knowledge goods (and I include music in this category) are nonrival — if you tell me an idea or play me a song for me you still have the idea (or the song), but that is not true if you lend me your car or your shoes. This difference is of crucial importance. It means, among other things, that creating exclusive rights in intangibles (and remember when talking about property right in a CD you don’t mean in the CD itself but in the song stored on that CD) is per se inefficient for society in a way in which creating exclusive rights in your house is not. To see this point forcefully put (it may even go too far) see the excerpt from the work of the famous Austrian economist Friedrich Hayek presented in this post: http://www.rufuspollock.org/archives/169.
July 16th, 2007 at 11:56 am
I see no problem with a pay per use system - as long as the system is sufficiently sophisticated. I also have no problem with Mozart’s estate being credited every time his work is exploited. Of course this payment might be split over multiple individuals or organisations - but this is a matter of better technology and cheaply administered micro-payments.
I read the Hayek extract and take issue with this statement;
I can think of a lot of great records that would never have been made without the artists having made a significant income from their previous work. And an even greater number of films that would never have been made without a secure and long term copyright system in place to cover the studio’s costs.
I really don’t understand the term ‘inefficient for society’. What happens when more than fifty percent of individuals making up the society earn their livelihoods through their connection with the knowledge based economy? The fact that I can exploit my intellectual property while also keeping it is simply a feature of copyright ownership and the ability to duplicate I.P. assets - which only have value if people want to pay for them. It is of crucial importance, it creates value without depleting resources - isn’t that efficient for society?
I’m sorry if my arguments seem superficial, I’m not an economist, I’ve just spent twenty years turning intellectual property into an income for the artists that create it. The suggestion that copyright terms should be shortened makes no sense to me at all and I have serious issues with the assumptions you’ve made in your optimal copyright document.
July 16th, 2007 at 4:51 pm
Technical point: Mary Bono’s “forever less one day” remark was made in a speech on the House floor, not in hearings as stated in footnote 1.
July 16th, 2007 at 5:53 pm
I’ve only skimmed through the paper, but it’s an excellent read and I’m recommending it to our readers. It’s nice to see an even handed, scientific approach to the subject. This is the kind of thing we need more of to cut through the rhetoric on both sides.
July 18th, 2007 at 11:08 am
I think I found a typo: on page 11, “consider what would occur if NU >= 0″ should have “NU <= 0″
July 20th, 2007 at 9:24 am
[…] Last week I was at the 2007 Society for Economic Research on Copyright Issues (SERCI) Annual Congress (also acronymed under the label SERCIAC). The event was a nice size with a good mix of people, well organized (a big thank-you here to Christian Handke) and with many interesting presentations (some of which I was able to take notes on — see below). I also had the chance to present my paper on optimal copyright and get some useful feedback. I’ve already mentioned this in a previous post but for those interested you can find the slides from my talk here: […]
July 20th, 2007 at 9:30 am
Timothy Phillips: You are quite right. I did know this was a speech in the house (having read the transcript) and had simply meant to indicate was taking place during debate on the bill. Of course ‘hearings’ have a quite specific meaning and I should therefore correct this as you suggest.
StefanoC: thanks for the ‘bug report’ and I will check that inequality out.
Thanks to everyone else who has made comments. They are all much appreciated.
July 24th, 2007 at 12:11 pm
[…] If you think this article lacks substance, anecdotes or hard data, please read Against Intellectual Monopoly which covers all the mentionned topics in detail. It even goes further than that, it also argues that the other intellectual monopoly — copyright — should be abolished as well. I’m not quite sure whether this is a good idea, but copyright has grown like fungus in the last few decades, and there is certainly something very wrong with it as well. Forever Minus a Day? argues that copyright should last 14 years after creation, no more. And I’d be happy with these 14 years and throw in a “renewable once for another 14 years” as a tradeoff, but I won’t agree to more than that. With patents, I won’t give in to anything else besides complete abolishment. […]
July 29th, 2007 at 8:31 am
Steve Baker wrote, If I buy land and build a house, I own it in perpetuity. There should be no economic difference between a tangible and an intangible asset.
LOL!
You clearly know absolutely nothing about economics.
Land is the classic example of a factor of production giving rise to economic rent. In fact, for reasons of justice and efficiency, people should never be able to “own” land in the sense of being able to reap the fruit of the land itself for free, because (a) they didn’t create the land, and (b) they’re excluding others from using the land. (For more, see “Are you a Real Libertarian, or a ROYAL Libertarian?.)
The reason that copyright should be limited is that it, like other so-called intellectual property rights, is a severe constraint on freedom—that’s one of its most important costs. On the other hand, its putative benefit (the one recognized in the US constitution, for example)—that of allowing the creation of works of art by allowing creators to recoup sunk (non-marginal costs)—can be realized without recourse to absurd measures like perpetual copyright.
August 13th, 2007 at 9:56 pm
[…] Boing Boing post is here. Rufus Pollock’s blog post is here. Via BOL, here. […]
September 14th, 2007 at 1:04 pm
[…] Economics PhD student Rufus Pollock has written a paper on the optimal term length for copyrights. (Found via BoingBoing and Ars Technica.) Using economic models, he calculates the optimal term at 14 years, much less than the current “Sonny Bono” term of author’s life plus 70 years. Specifically, we demonstrate that (a) optimal copyright falls as the costs of production go down (for example as a result of digitization) and that (b) the optimal level of copyright will, in general, fall over time. This work, unless otherwise expressly stated, is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 3.0 United States License. […]