[1]More sophisticated possibilities might also be created. Though each individual might prefer to be insular, he might agree to receive contrary views if others would do so as well, leading to a general agreement on a system that provides some contrary exposure to all.

[2]A mistrustful user might subscribe to several Better Bit Bureaus that follow different filtering rules, much as medical patients often ask for a second opinion from another doctor.

[3]Vote selling offers an interesting analogy. I am hurt when others who are like me sell their votes, but whether they do or not, it is still worthwhile for me to sell my own. Similarly, my privacy is lost if others like me sell theirs, yet I might sell mine. In either case, a like-minded group might agree to preserve their votes or their privacy, even though each would sell without an agreement.

[4]Economists would label this two-part pricing, with the dues being a fixed cost that enables individual items to be priced closer to marginal cost.

[5]A weaker relative of SASP, in which bids are considered actions, is met by both Groves-Clarke and expected externalities approaches. Individuals who report the same preferences and take the same actions face the same price.

[6]By the same reasoning, AT&T, as a long-distance carrier was prohibited from providing local telephone service. The vertical integration argument is also at the center of current cable debates. Should cable services operate merely as a common carrier, or should they also create and buy programs?

[7] Consider one monopolist who sells to another, who then sells to a market. If the first raises its price, this reduces the profits of the second, and vice versa. But since the separate monopolists do not take into account these reductions, the ultimate price is greater than it would be if they integrated vertically. The additional deadweight loss due to double marginalization may occur in the form of lower quality, not just lower quantity and higher cost (Economides, 1994).

[8] This set of issues has also arisen in an array of other arenas. For example, Microsoft has been accused of selling its operating systems at lower prices to computer manufacturers that agreed not to offer any other operating system. Government policy towards competition will be sound, we suspect, if appropriate analogies are drawn, say from the telephone or airline industries, about the success of existing methods. The market may also play a helpful role: in the computer industry in recent years, "open systems" strategies, in which downstream firms are encouraged to make complementary products, have fared better than proprietary strategies. By contrast, efforts to gain or maintain predominant control over upstream software resources have fared poorly. The disasters SONY and Matsushita have suffered in the Hollywood movie business are perhaps the best examples.

[9] Similarly, information carriage could be broken into two pieces, actual carriage and value-added brokering services. For example, if accounting and bill collection were separate from carriage, consumers would likely have the choice of telephone bills with calls itemized at a somewhat higher charge than un-itemized bills, and an even wider array of pricing packages than is currently available. Other brokerage services might identify the cheapest carrier for a particular phone call or even for a single data packet. Brokering of information carriage, however, is beyond the scope of this paper.

[10] Technical standards must be set with great care. Once a system is in place with even a moderate number of users, there can be considerable technological inertia. An outmoded interface in current use may win out over a much better one just being introduced since it will be very costly for users to switch to the new interface without the guarantee that others will be switching as well.

[11] Experience with stock exchanges is instructive, although the data are unclear about the advantages and disadvantages of centralization. The New York Stock Exchange long argued for its monopoly on the theory that gathering all bids and asks in one place provided for the most liquid market. Parallel markets, however, are now well established, and are often quite innovative.

In the pure information exchange realm, there are now services that enable employers to compare medical costs against those of other employers on the system, standardized by such factors as age and occupation. The more employers on a single system, the richer and more informative will be the comparison data. The virtue of aggregating information provides one argument for a natural monopoly, assuming data would not be readily exchanged between services.