Tom Lenard's excellent post on differential pricing in the drug industry has broad application throughout high fixed cost, low marginal cost industries, as Jim DeLong notes. Thus, industries from communications to airlines to drugs to software to movies and recording all engage in differential pricing as a second best means to recover their high fixed costs, even though their incremental costs might be very low or even zero.
This differential pricing creates a big problem though, arbitrage. Arbitrage is the practice (and wasteful investment) of efforts to avoid the differential pricing; that is, the efforts of high value customers to pay the prices low value customers are paying. Technology has incidental arbitrage effects. TiVo and VoIP, for instance, allow arbitrage play around the broadcast advertising model and the access charge pricing model, respectively. Thus, these new technologies impede the ability of companies to differentially price.
Arbitrage is neither per se good or per se bad, but a fact of economic behavior. If I can easily and cheaply avoid being the high-fixed cost payor, I will do it. My VoIP calls then are a cheap and easy was of avoiding access charges. My mother-in-law thinks she has a absolute right to Canada's price-controlled drugs. In both cases, though, the destruction of the ability to differentially price causes a crisis for the existing model, and for future pricing models. If I collapse the access regime through VoIP, then rural phone companies especially will be hard hit and investment in the PSTN will decline. Likewise, my mother-in-law's drug buying habits will eventually: a) bring down the Canadian price control system; b) force pharmaceutical companies to stop selling to Canada; and, c) decrease investment in new drugs (or as the Thalidomide article shows, investment in new uses for old drugs).
In all cases of successful arbitrage then the subsequent problem is the prisoner's dilemma, where all of us want to get the marginal price and none of us wants to pay the price of the higher fixed costs.
All of this is to say, differential pricing is necessary and pervasive in network industries -- but it will remain forever problematic. High fixed cost industries need to engage in differential pricing to survive, and technology should make it easier for them to do so with greater precision. At the same time, technology enables consumers to engage in arbitrage more easily. So, you end up in a technological and legal arms race over differential pricing, all the while with the big prisoner's dilemma problem looming in the background.
A solution to this? There aren't any pat ones, and manifold stupid ones. Starting with the Robinson-Patman Act -- thank goodness for the legal doctrine of desuetude -- and continuing to the "government should subsidize the high fixed costs so the firms sell at marginal costs," there is no good legal or regulatory response. So, I end with the challenge: You gotta' better idea?