Journals as Shared Goods: Comment
评论Ordover和Willig的期刊共享模型，指出其隐含假设过于严格，忽略交易成本，导致模型与现实脱节，对政策讨论无实际意义。
In a recent issue of this Review, Janusz Ordover and Robert Willig present a model of the provision of journals to libraries and individuals. While the provision of journals is itself of interest to academics, the concepts presented have ready extension to a broad class of quasi-public goods. Ordover and Willig treat journals as shared which means that some copies are consumed privately, while others are consumed collectively in libraries. Libraries face one subscription price, individuals another. It is readily recognized that libraries are not necessarily harmful to publishers' profitability, since collective consumption means that more people can be served at any given output. If publishers can appropriate a sufficient share of the value created by library use, they will benefit from collective consumption. Ordover and Willig's treatment of journal pricing incorporates increasing returns to scale in publishing, and the time or inconvenience costs that some readers face in using the library as opposed to using private subscriptions. They argue that because the amount of library use follows from a subscription price that is not equal to marginal cost, library user fees will enhance efficiency in this setting. While the concept of sometimes shared goods is an interesting one, and one which may be useful in some contexts, the model presented by Ordover and Willig imposes implicit and explicit assumptions which cannot be maintained for discussions of policy. This is especially important since their paper makes strong claims of policy relevance and practicality. Even without those claims, these restrictions preclude correspondence of the model to anything in the real world. The restrictions imposed by Ordover and Willig are of two sorts. First, they impose severe restrictions on the demands facing libraries. Second, they ignore all transactions costs except the one which is responsible for all of their results. We maintain that the ignored costs could well be larger than the one that they recognize. We proceed as follows: Section I outlines elements of the model which are essential to our presentation. Section II presents some simple analytics of the type of user fees considered by Ordover and Willig, introduces the policy implications that they draw, and provides the first indication that their results might depend on rather stringent assumptions. Section III demonstrates that the model requires an assumption that the ranking of libraries by willingness to pay is invariant over the domains of a number of key parameters and that this invariance assumption imposes severe restrictions on consumer demands. Section IV discusses restrictions which are fairly explicit in the model, but which are as important as the hidden ones in establishing policy irrelevance. The problems considered in this section are further restrictions on demand and Ordover and Willig's inconsistent treatment of transaction costs.