Saturday, January 17, 2009

Intellectual Property and Price Discrimination - iTunes

There's a lot of debate about whether allowing or encouraging price discrimination - charging different people or classes of people a different price for the same good - is good policy with respect to goods that embody copyrighted works of authorship or patented inventions. The general population seems to have different reactions to different kinds of price discrimination schemes. On the one hand, most people don't seem to have a problem with senior citizen discounts at movie theaters, even though some seniors are quite wealthy. But, if Amazon chooses to customize the price of DVDs for each consumer, an uproar ensues.

I am very interested to see what happens with the recent deal between Apple and the recording companies to allow variable pricing on music files distributed by iTunes. This may or not be price discrimination depending on whether the good is "music" or particular songs because everyone will still pay the same price for particular songs.

Here, I just want to make two related theoretical points about the way that economists and legal scholars who use economic models talk about price discrimination.

Point 1. Most economic or law-and-economic analysts who talk about price discrimination say that the policy goal is to maximize people's welfare. I'm at the Penn Law Review symposium on the Foundations of Intellectual Property, and just watched Christopher Yoo make this point about his paper with John Conley on intellectual property and impure public goods.

For these analysts, whether price discrimination is good or bad for society depends on who wins and loses and by how much. But they model this trade off using people's willingness to pay for intellectual property goods as a signal for how much having a copy of a song or book or movie is going to improve their welfare. And these models get pretty complicated quickly, but they all are built on this foundation.

The problem is the gap between ability-to-pay and willingness-to-pay. If you care about welfare, ability-to-pay is a poor proxy for utility or welfare because the marginal value of a dollar depends on how many dollars you have. When iTunes charged you 99 cents a song, how does a consumer decide when it is worth paying that price? Imagine two people who value owning a legal copy of a particular song exactly the same, but one is wealthy and the other is struggling. At certain values, the wealthy person buys the song and the struggling person won't even though they both would get exactly the same amount of pleasure from the song because the relative cost to each person is quite different. So their respective decisions to buy or not are not really telling you how much they value owning a copy of the song.

So, in my view, analysts need to defend the proposition that their models tell us something about the effect of variable pricing on people's welfare when they have not accounted for the gap between wealth and welfare in the model. (Of course, this is a more general point about neoclassical economics, but it has particular salience in this context.)

Point 2. Those who have a reflexive antagonism to price discrimination need to be careful about form and substance. Some folks who have this reflex are reacting to the underlying market power that gives a seller the ability to engage in variable pricing without losing all of its customers to a competitor. I agree that market power is something to watch.

But firms with market power also engage in uniform pricing, as Apple has done with iTunes. In such a case, uniform pricing functions as a form of value discrimination or cost discrimination. Because the marginal value of a dollar varies across people, the price may be the same, but the relative cost of the song from the buyer's perspective varies. So if you want greater equality in the market place, you may actually want to encourage variable pricing if it has the economic effect of equalizing relative cost. I agree that this is very difficult to do in practice, but remember this is a theoretical point. In practice, though, this point about variable cost is what progressive taxation is all about.

So, if copyright is a tax on readers, should it be a progressive tax?

No comments: