16 April 2010

Goods/Not Goods

There's a fascinating post (and series of comments) over at Charlie Stross's blog today that asserts that "content" is a public good... and thereby entirely misses the point in its unfortunate mischaracterization of content.

The piece starts off well enough, discussing the accepted definitions of rivalrousness, exclusion, and how they relate to the different types of goods.1 In summary form, it looks something like this:

  Rivalrous Not Rivalrous
Excludable private good club/toll good
Not Excludable common pool good public good

Although Ms Popova's piece doesn't go into the club/toll good distinction (it should, but it's not central to her argument) or much of the bizarre, almost non-Euclidean nonsense surrounding common pool goods, it does do a fairly good job of explaining one aspect of the economics of "content." Unfortunately, her argument falls apart entirely when one tries to discern an internally consistent definition of "content"... because, as for virtually all classical and neoclassical economic arguments, it founders on The Invisible Variable (t).

First, and perhaps most obvious, there's a huge distinction between "content" at the moment of its creation and "content" at the moment we judge rivalrousness and excludability, whatever "content" may be. This is perhaps most obvious with books: The "content" is both what the author wrought (ignoring, for the moment, editorial contributions, which ultimately make my point for me) and the physical package "containing" what the author wrought that can be grabbed from the shelves/e-book source. As I've begun discussing (and Mr Stross himself, in an incredibly useful set of posts such as this one made under the name "CMAP"), the package in which one finds the content is a nontrivial and intertwined aspect of the content. The key point that undermines Ms Popova's piece, though, is this one: It assumes that the same (both in quality and in relative magnitude, and probably in actual quantity) considerations and values apply to "content" at all times and in all potential markets, and that therefore the economic analysis is the same.2

Second, and more subtly, I'm afraid that Ms Popova's analysis of whether content is rivalrous assumes far, far too much. Perhaps the best (and most extreme) refutation of the analysis is implied by Borges's Pierre Menard — best because, in the classic tradition of reflexive refutation, the story itself defies the analysis it implies, and it's just plain fun (even translated into English). My point here is that the content is not a good of any kind for any fixed definition of content; it is only when we allow flux in that definition of content that we can begin to treat it as a good. And that, indeed, is the problem: The rivalrousness and excludability analyses assume that they are dealing with goods, just like Euclidean geometry assumes the Parallel Postulate's validity. Perhaps for most purposes that's a valid assumption; it is not, however, prescriptive or universal... and since you're reading this, and the design of semiconductors depends upon the validity of non-Euclidean geometry, you should be very wary indeed of leaping from "useful in many, or indeed most common, reference frames" to "is therefore a universal law applicable in all reference frames."

Applying the "rivalrous"/"excludable" rubric to "content" is a less-obviously useless exercise than attempting to determine the number of angels on the head of a pin... but it provides no more validity in any given reference frame, let alone the hypothetical objective reference frame assumed by classical/neoclassical economics as the foundation for proper policymaking. Neither does it acknowledge understanding of the distinction between enlightened and unenlightened self-interest that ultimately torpedoes most classical/neoclassical economic theory when it comes time to turn economic principles into policy.

In short, the problem is that "content" is not a good of any kind, and trying to force it into classical/neoclassical economic analysis of goods doesn't just assume a can opener: It assumes field and boundary conditions that are explicitly inapplicable.

  1. Although my colleague Professor Solum and I disagree on what many of these things ultimately mean, and whether — as I'll argue below — any of this is even properly in the realm of economic analysis in the first place, his explanation of the orthdox view is admirably clear. So I'm stealing, albeit I've rotated the table for polemical purposes (and to match the discussion in Ms Popova's initial posting).
  2. This leads into the more fundamental questions of whether content is, in fact, "scarce" (and therefore properly analyzed in a classical/neoclassical economic reference frame at all, because scarcity is the fundamental assumption of classical/neoclassical economics); whether content, if it is "scarce", is a "good"; and whether classical/neoclassical economics provides a valid model of anything that is nonrivalrous at all, much as one must question whether Newtonian physics applies to subatomic particles. And, of course, still lurking behind everything there is t, which remains unaccounted for.
  3. I suppose that I could wade into the morass of behavioral economics; see, e.g., Cass R. Sunstein, Behavioral Law & Economics, for a useful introduction. One need not go quite so far, though; one need only note the failure of classical/neoclassical economics to either explain or develop adequate policies concerning bubbles. And given the recent fashionability of prestige investments in Hollywood, that one isn't far off from being directly applicable to "content"; perhaps, even, it's already there, given what has happened to a certain purveyor of alcoholic beverages that tried to become a big player in recorded music...