Economic and Game Theory
Why Napster is Right
Competition with Sunk Costs
by Michele Boldrin and David K. Levine
top | general | research | classes | links | DavidWhat is the economic basis of our claim that the copyright law should be changed to encourage downstream resale and distribution of copies?
We have argued in the main essay that the current controversy over Napster and the more general discussion about keeping, modifying or altogether abolishing the copyright law is mostly about distributive issues. We are not arguing that our proposal leads to a pure efficiency gain, since recording and distribution companies would loose under it. However, we propose that there would be efficiency gains in the usual sense that it would be possible to compensate them for their losses while making everyone better off. More to the point since consumers and creators of intellectual products constitute about 99.9% of the population, we claim that dramatically reducing downstream copyright protection to increase competition is socially beneficial in the less technical sense. Put another way: current restrictions on downstream usage do not serve a socially useful purpose and, contrary to the commonly accepted view, are not necessary to achieve a socially efficient outcome. The main function of these restrictions is to protect some powerful monopolies. A socially efficient outcome in the production and distribution of cultural products can be achieved by market competition and competitive pricing without the wasteful monopolies generated by current copyright restrictions. In this essay we elaborate on this point further, without getting too technical.
The arguments here are not restricted to either the case of Napster, or to music, or even to the copyright law. The essential ingredients are that the primary input into production of the good is the good itself and that demand for the good is elastic. Whether the low cost of reproducing the good is because of cheap electronic duplication or inexpensive reverse engineering does not matter. These arguments can apply to patents as well as copyrights: medicines seem a likely category, for example. The general case for (against) patents, though, is more complicated than the case for (against) copyrights. This is because reverse engineering may be expensive, while copying is cheap. To avoid claiming too much and generating unpleasant misunderstandings, the arguments here will be restricted to copyrighted products. We to forthcoming work the appropriate extension to the case of patents. ("Forthcoming work" may mean a few months ... come back often!)
We accept as a fact that for ideas, books, paintings, music and so forth, two half-baked ideas are not the same as one fully baked. Technically this means not that there is a fixed cost, as it is often argued, but that there is an indivisibility in the form of a minimum scale for any particular idea. Or more precisely, for any particular quality of a particular idea. This is relevant, because half-baked ideas (that is: different qualities of the same idea) have positive social value. The initial design of the idea we call "car" would be classified as a very low quality idea by today's standards. The same goes for radios, TVs and what not. This implies, on the one hand, that the indivisibility of ideas should not be exaggerated and that, on the other hand, the incremental aspect of "technological discovery" should be emphasised when modeling innovative activity. Technological breakthroughs, after all, almost always consist of some thousands of "half-baked" ideas fitting together just right! But enough digression: let us get back to copyrighted ideas.
Before getting carried away with patentable ideas, we were saying that we accept that, for any given quality of a new product, there is a constraint that no less than one full unit has any value. However, as we have point out in our example, this need not matter in many cases: the constraint may simply not bind. The cost of coming up with (at least) one sample of the idea may be a lot smaller than the quasi-rents the same producer will collect when his creation is priced competitively. Think, for example, of the opportunity cost of recording a new lyric, and the rents needed to compensate for it. The mere presence of sunk costs, in other words, does not justify giving a government monopoly to the creator of the first copy. This is because under competitive pricing the creator of the first copy may earn enough rents to recover her sunk cost. However, it is perfectly possible, especially for material that has only limited social value, that the constraint could bind, that is, under competitive pricing the quasi-rents are not enough to compensate for the initial cost and, foreseeing this, the potential creator would abstain from producing the first copy. Hence, under our proposal, art products of very little social value may not be produced without giving copyright protection to their creators. However this is not so bad: the products in question have little social value anyway.
Is it the case, however, that as the technology for replication and distribution improves, the minimum size constraint eventually binds? This would reduce the force of our arguments, because it would suggest that while strong copyright protection might not be needed now, the advent of cheap reproduction will make it so. The answer to this question is a firm: it depends. If demand for the product is elastic, meaning that a 10% reduction in price will increase demand by 10% or more, then reducing the distribution cost will not reduce the competitive rents earned by the originator. In this case reducing production and distribution costs would make the minimum size constraint even less binding. It is reasonable, however, to presume, that when distribution costs become low enough, demand becomes inelastic. That is, at some point the market is so saturated, with every individual in the world having copies of a particular song on every musical device he owns, that reducing the price further calls forth little if any more demand. From a practical point of view, though, the latter case should not worry us too much: if most people already have access to the product at a low enough price the social gains from reducing its cost of distribution are relatively small. We are also willing to argue that, for most products covered by the copyright law, this is not the case. Their demand is quite elastic and the social gains from reducing production and distribution costs of additional copies are likely to be large. Competitive pricing achieves such reduction. Notice the empirical evidence that monopoly pricing does not: the advent of the CD, which lowered production costs, led to an increase in the price of a musical album.
All this aside, have we not just said that under certain assumptions (inelastic demand) a technology like Napster may make it impossible for competitive pricing to recover costs? Allowing downstream sales effectively eliminates the originator's natural monopoly: As soon as the originator has sold just a few copies, she faces a competitive market in which copies of the original product are sold at much lower prices. Initial copies may not command a very high premium since the purchasers know that, in a short span of time, they will also have to compete with each other, the originator, and their own customers. Is this not an argument for allowing the originator to use licensing agreements that enable her to maintain her original monopoly? For, only through the power of monopoly can she hope to recover her costs, and so be willing to produce the good in the first place?
The answer again is a firm: not necessarily. The radio and television industry, which have substantial sunk costs, give their products away for free. In the case of creative activity, we might wish to look at the market for art products. A signed original painting commands an enormous premium over a very good imitation. People eagerly seek out authors at book signings; autographed baseballs sell for prices that seem to us to be absurdly high. Consider that the same digital technology that lowers distribution costs, makes the sale of personalized copies practical. That is, I can buy my music/book/movie, with a signed personal message from the author to me. As long as I am willing to pay a premium for this, the author has an opportunity to recover his unique production costs.
But isn't this a contradiction you say? Why can't the purchaser resell his signed copy? Well he can - but what he should not be able to do legally, is to change the inscription so that the book is personalized to the new purchaser rather than to himself. So which do you prefer? A CD from your favorite artist personalized to you, or to me? You can have the one personalized to me very cheap! Notice also that while copy protection is technologically difficult, making personalized copies that cannot be forged is relatively easy. Your personalized copy is encrypted with the private key of the author. Anyone can decrypt it and play the music, but they will see your name. Of course they can change the name, but they cannot reencrypt it with the private key of the author, so it will be evident that the copy was not really personalized to them. This technology is sufficiently reliable that electronic signatures now have the same legal validity as written ones.
Finally: why not considering the fact that advances in the technology for production and distribution of intellectual products may allow the original creators to produce and directly distribute the objects in which their creativity is contained? By doing so, they can avoid the parasitic gatekeepers and earn much higher rents (on average) than under the current arrangement. In these circumstances, creators would have an incentive to figure out quite precisely how successful their product may be and try charging quite a lot for the first few copies they sell, or for those personalized copies demanded by aficionados. As a result, we believe more art will be produced and distributed to consumers. Which is what matters from the social viewpoint.
So it is not obvious that creators of original work need to have government grants of monopoly power to recover their costs. Ultimately the issue is an empirical one: what is the elasticity of demand? What are the costs of production? What clever methods can authors use to command a premium over downstream rivals? The fact is that those in favor of more copyright protection for producers do not generally give empirical evidence to support their point of view. They merely make theoretical arguments. Since their theoretical arguments are either weak or wrong, it is up to them to produce empirical evidence that the monopoly power they crave has some social justification.
So far, they have been able to produce one piece of empirical evidence: copyright law was eliminated for 3 years during the French revolution, with apparently bad consequences for literature and news reporting. However, one may imagine that potential authors may have had more immediate concerns than copyright - keeping their heads on their shoulders, for example. We hope that some stronger justification for a strong government supported monopoly can be provided than this. The rest of existing evidence points the other way. Copyright protections for producers have been strengthened without doing anything but reducing supply and enriching the monopolists. In addition the opportunity cost of authors and singers is low, and they get almost none of the revenue from their works: It is difficult to imagine that we cannot cover their production costs in a competitive system.
What then is the policy we are suggesting? We are simply suggesting that the copyright law be changed to prohibit downstream licensing agreements. This would not mean that you would be free, for example, to change the author's name to your own and redistribute the modified work: that would be profiting from a lie, and is and should be considered to be fraud. Similarly, you would not be free to repersonalize an inscription certifying you bought the material from the originator to show that someone else did. Nor would authors be prevented from trying to restrict downstream consumption by technological means: they would still be free to sell you copies that explode in 24 hours, or phone home for permission to play, or whatever. On the other hand, you, as the purchaser, would be legally entitled to reverse engineer or crack whatever silly thing they did, and redistribute the result for a profit, so if authors were foolish enough to think that their product was so special that people would put up with all sorts of silly restrictions .... they would quickly and legally be disabused of this notion.
If you are interested in pursuing these issues, you might want to read the example showing how a competitive market operates in the presence of sunk costs. If you are more patient, you should read our paper "Perfect Competitive Innovation." Although a bit technical, it is pretty nice. And it tells you also about a bunch of other interesting things, such as when it might really be an advantage for a country to have a technological disadvantage.
Details