"Marginal Costs and Average Costs"
Nate Anderson at Ars Technica has an excellent writeup of the little dustup between me and Sonny Bunch about whether copyright policy should try to serve the interests of consumers or the interests of incumbent producers. I did want to reply, though, to a criticism Anderson made of my post since the general issue of marginal costs and average costs is something that comes up a lot:
One quibble with Yglesias comes in his original post, when he claims that “under conditions of perfect competition, the price of a song ought to be equal to the marginal cost of distributing a new copy of a song. Which is to say that the marginal cost ought to be $0. That’s not a question of habit, you can look it up in all the leading textbooks.”
This is clearly not true; the price of a good in a perfect world would be some fraction of its upfront cost plus its marginal cost. If I write a novel, and it takes me a year of full-time work, but it can be distributed digitally at a marginal cost of $0, that hardly implies that I should price it at $0. I may price it that way, especially if I make money in other ways—through speaking fees, perhaps, or other paper editions of the book. But a year of my time is a fixed cost that needs to be considered when setting the price.
I think I introduced unnecessary confusion to this issue with the use of the word “ought” which seems to imply an ethical claim about the price of goods. The point, which really is in all the intro microeconomic textbooks, is just that if you had a market that features perfect competition and perfect information goods would, in fact, be sold for their marginal cost rather than for something based on the average cost. That’s because if the market price for widgets is higher than the marginal cost of making another widget, then it makes sense for firms to ramp up widget-production. This process will either drive the price of widgets down to the point where it doesn’t make sense to make any more or else firms will hit some kind of production chokepoints where the marginal cost starts to spike. Fixed costs associated with widget-production aren’t really relevant to this. Whether or not building the factory in the first place was a good idea, that’s in the past and it makes sense to keep making widgets at the margin as long as doing so is profitable at the margin—i.e., as long as the sale price exceeds the marginal cost.
This seems wrong to people because in many cases (if not most) average costs will exceed marginal costs and firms will lose money if they sell goods at the marginal cost. But of course it’s a feature of the real world that lots of firms do, in fact, lose money. It’s not rare for a startup to fail nor is it rare for a once-profitable firm to be driven out of business by changes in market conditions or new competition. But of course the other point is that actual markets generally don’t feature perfect competition and perfect information. As per the widget example, a market like that isn’t a very good line of business. This is why people don’t get rich operating storefront dry cleaning operations—you’ve got a generic product and tons of competition.
To return to the intellectual property example, the idea behind giving people patents and copyrights is precisely that it might be a bad idea to force people to sell things (books, songs, prescription drugs) for the marginal cost of distributing them. It might be better to give someone a monopoly on the sale of Abbey Road, The Merchant of Venice, or Propecia. That way you can sell things for much higher than marginal cost and recoup initial investments. But any time you grant a monopoly, you also create some deadweight loss and inefficiency. If we retroactively extended copyright back to works created in the 16th century and nobody could produce any Shakespeare plays due to difficulties in identifying the rights-holder, or Tolstoy’s heir decided he wanted to charge $100 per copy of Anna Karenina the loss to society would be gigantic. So there’s a need to strike a balance through both limited scope of copyright, limited duration of copyright, and to some extent through imperfect enforcement of copyright.