The economics of information applies to any idea, expression of an idea, or fact that is known by one person and is potentially of value to another. Examples include research that discovers or produces new knowledge about nature or a new product design; creative expression by artists, musicians, poets, and novelists; expository prose that provides a new explication, explanation, or interpretation of facts and ideas that are themselves not new; or "signals" (such as trademarks and logos) that identify a product, company, or organization to others. For the purposes of economic analysis, the narrower and more precise definitions of information that are used in other disciplinesnotably computer science and mathematical statisticsdo not provide economically meaningful distinctions and, therefore, are not addressed here.
The usefulness of information to others can be either as an instrument to an end or as an end in itself. For example, research may lead to new facts about nature that can be used to make new or improved products, but it may also be valuable in its own right as a kind of consumer good. Likewise, some people may use newspapers and magazines to inform their business and personal decisions, while others may read newspapers and magazines simply because they enjoy doing so. For present purposes, the source of value to others is not a crucial distinction. All that
NOTE: Other accessible background works include Noll (1993), Shapiro and Varian (1998), and Besen and Raskind (1991).
matters is that someone other than the creator of the information will value having access to it.
Information Creation, Distribution, and Consumption
The single most important feature of information products is that the creation of the original information content is what economists call a pure public good: The cost of generating new information is independent of how many people eventually gain access to it. The information creation stage involves all the activities necessary to develop a new information product to the point that it can be distributed to others. For example, in written expression, the creation stage might include conducting necessary research; writing the various drafts of a book, story, or essay; and preparing the final manuscript for publication (whether as printed or electronic text). The costs of carrying out these activities do not depend on how many people will eventually read the product. In publishing, these costs are called "first-copy costs" and refer to all the costs incurred in preparing to print and distribute copies of the original expression.
Distribution costs are the costs of delivering the information product to consumers. Historically, this activity entails two main steps: reproducing the physical embodiment of the expression and then delivering it to the consumer. For some popular entertainment, such as music CDs, novels, and magazines, the duplication stage consists of manufacturing a large number of copies of the physical embodiment of the information for distribution by mail or retail stores. With the widespread use of digital networks, however, information products are increasingly being distributed (or otherwise made accessible) electronically. In the case of live theatrical performances or broadcasts, the product need only be produced once but is simultaneously delivered to everyone in the audience. For theatrical performances, the duplication and delivery costs consist of the rental cost of the theater facilities and the salaries of nonperforming theater personnel. For broadcasts, the duplication and delivery costs are the costs associated with distributing the program to television or radio stations around the country and then transmitting the program in each community to its audiences.
Distribution costs vary widely according to the nature of the medium. In many cases, such as printed material and audio and video recordings,
distribution costs are more or less proportional to the number of people who receive the product. In other cases, such as broadcasting (whether from television stations, radio stations, or direct broadcast satellites), distribution costs depend only on the size of the geographic area to which the signal is transmitted rather than on the number of people who receive the signal. In still other cases, distribution, although not exactly free, is nearly costless. A prominent example is distribution of documents, music, and pictures over the Internet, where the actual costs of the transmissions are often well below one cent. These costs are so tiny that they are typically not worth monitoring and billing on a per-transmission basis.
A consumer incurs costs to use an information product, in addition to the purchase price. Such costs refer to the materials and equipment that a consumer must buy to make use of an information product. Examples include a television set, a radio, an audio system, a computer with modem, and transportation to the theater. Although these costs can be regarded as part of the distribution system, separating them from the distribution costs is useful because the consumer, not the information producer or distributor, decides whether to incur them.
The important feature of these consumption costs is that they are highly variable across different types of information products. For example, a consumer must pay much more to use the Internet to read the newspaper at home than to read the hard-copy version. (In the latter case, the only consumption cost is either storing or disposing of the used copy.) An important feature of electronic distribution of information products is that it reduces distribution costs (and, where relevant, storage costs) but usually increases other consumption costs.
Information Supply: Compensating Information Creators
Although some people create new information as an act of pure self-expression, perhaps seeking recognition but expecting and receiving no compensation (some Web sites and almost all personal conversations are of this type), most participants in the business of information creation expect compensation. They will not produce new information products unless the amount of compensation is sufficient to justify spending time and resources in this endeavor.
An appropriate conceptualization of the first-copy cost mentioned above is the amount that the creators of new information products must be paid to induce their creative effort. If compensation is required to
induce effort to generate new information products, information creators can be paid by:
• Setting the price of the information product above the cost of duplication and distribution so that a surplus above the costs of those activities can be paid to the creator of the informationfor instance, an author or inventor can be paid a royalty based on the sales revenue of the information product;
• Using payments from a granting agency to defray the first-copy costs of creating the productfor instance, a government research grant to a university to undertake basic research or to a defense contractor to develop the knowledge for a new weapon; or
• Combining an information product with some ancillary product (frequently another information product) and selling the two in combinationfor instance, the news and entertainment content of media products is combined with advertising which, in turn, may pay for all or part of both first-copy and distribution costs.1
Each of these approaches is likely to be economically inefficient in important respects.
Payment of Royalties
The first approachinto include first-copy costs in the final price of the information product to cover a royalty for the creatoris economically inefficient because it causes the price of a product to depart from the marginal social cost of producing and distributing it. For example, a recording or a mystery novel typically has built into its price royalties for artistic creation of a few dollars per copy. Consequently, the price is two or three dollars above the social cost of producing and distributing one more copy of the product. This higher price will cause fewer copies of the product to be sold, because some potential customers will be willing to pay one dollar more than the cost of manufacturing and distributing one more copy, but not two or three dollars more. Excluding these potential customers from enjoying the product is economically inefficient. Economically efficient production requires that everyone be allowed to buy the product who is willing to pay at least the social cost of providing his or her copy.
Thus, a fundamental property of a royalty-based system of paying for the creation of new information products is that these products will be
1In some cases, advertising itself can be a product. For example, some magazines are purchased in part for their advertisements (such as some bridal magazines).
less widely disseminated in society than economic efficiency dictates, and hence the new information leads to less improvement in economic welfare than the information is capable of producing. On the other hand, an advantage of this approach is that it guides producers of information products toward products that consumers or users want and are willing to pay for.
The presence of royalties creates an incentive for the creator of intellectual property (IP) to promote the product, because the information creator can derive an increase in income by increasing sales. Individuals who engage in the promotion of IP must expect an increase in revenue that exceeds the cost of promotion in order to engage in promotional activities. These promotional activities are themselves a form of information productthey provide information to consumers that, presumably, increases the likelihood that they will buy the product.
Use of Grants
The main source of distortion in a grant process arises in deciding which activities to reward and then calculating the appropriate payment. A grant system transfers risks and financial responsibility away from the producers of new information to the granting agency. The effect of this transfer is to attenuate the incentive for the creator to produce the new information that the marketthe users of information productswould value most highly unless, of course, that user is the granting agency.
In addition, the essence of new knowledge is that the form it will take cannot be predicted in advance. Consequently, grants are typically based on the principle of cost reimbursement rather than on the value of the output. Cost reimbursement weakens the incentive of the contractor to manage the project efficiently to minimize cost. As a result, cost-plus contracts are typically accompanied by an expensive system for monitoring and auditing performance in an attempt to make project management more efficient.
Grants to support the creation of information products can lead to a second type of distortion. If the granting agency is the government (which it often is), then tax revenues must be raised to finance the grants. Taxes (such as sales, income, and property) are like royalties in that they drive a wedge between the social value of the taxed item and its market price, thereby discouraging the taxed activity and creating a loss of economic efficiency.
Combining Information and Ancillary Products for Sale
Media products are the most common examples of combining information with ancillary products. The principal source of support for some commercial Web news and magazine sites is advertising. In effect, viewers are being attracted with the information or entertainment product and then their attention is being sold to advertisers. Advertising revenue then provides the means to cover first-copy costs.2 In media products, subscription charges to viewers or readers, when they are imposed, rarely cover more than a modest part of the cost of creating the information product.
Bundling an information product with other revenue-generating activities is in some ways similar to a royalty system. The revenues from the ancillary product must cover both the costs of the ancillary product and the primary information product, which means that the price of the ancillary product must contain a markup over its average cost that is similar to a royalty. In some cases, the connection between the intensity of consumers' preferences and the attributes of these information products can be considerably weakened. In the case of "free" television, for example, programming is designed to be just good enough to induce viewers to watch, thereby delivering the attention of viewers to advertisers. Programming that viewers might find more satisfying but that would not increase the number who are watching is not of any greater value to either broadcasters or advertisers, and so, if it is more costly, will not be produced.3
2Patents can be thought of in similar terms. Often a return to the patent holder is generated not by selling the patent itself but by producing and selling the patented product.
3The discussion above neglects the theoretical possibility of a perfectly discriminating monopolist who succeeds in pricing the product to each user in each use at exactly the value placed on the product by the user. From a formal standpoint, perfect price discrimination solves the efficient pricing problem, although at the cost of generating a redistribution of wealth from users to suppliers. It can be argued that licensing schemes for IP (instead of outright sale of IP) offer more opportunities for a price-discriminating monopolist and, therefore, may offer gains in economic efficiency. Experience with software products seems to confirm this possibility, including, apparently, the effects of income redistribution.
In practice, perfect price discrimination is impossible. One reason is that a seller cannot possibly know how much each person is willing to pay for IP. The best that sellers can do is to categorize buyers into groups that generally are willing to pay more versus groups that are generally willing to pay less. To the extent that people within these categories differ in the value they place on the product, the problems discussed in this section will arise within the groups. A second impediment to implementing extensive price discrimination is that sellers often cannot prevent arbitragethat is, the circumstance in which the buyer who places a low value on the product simply resells it to the buyer who values it highly.
The first important lesson from this discussion is that no system for compensating those who create new information can provide a perfect solution to three central economic problems:
• Adequately compensating those who create new information products;
• Maximally disseminating and using the new information in the economy; and
• Selecting the most valuable information products that will be produced.
Hence, information policy inevitably involves trade-offs between these three objectives. The royalty approach is effective in selecting the information products that users value, but it fails to distribute the products as widely as needed to satisfy economic efficiency criteria. The grants approach is capable of solving the distribution problem efficiently, but it lacks an incentive to ensure that users obtain the information products that generate the greatest net benefit to them. Finally, the ancillary product solution is better than the royalty approach but worse than the grants approach on grounds of economically efficient distribution, but is better than the grants approach and worse than the royalty approach when it comes to selecting the products that users want for production.4
The Role of Intellectual Property
Intellectual property is the area of law and policy that determines the solution to the trade-off between fostering incentives to create new information and diffusing its benefits throughout society. Intellectual property law inevitably is a two-edged sword in that it both grants and, simultaneously, limits the rights of the producer of new information in order to articulate how this trade-off will be made.
Patents, copyrights, trademarks, and trade secrets are all types of IP protection that give those holding them some degree of control over the use of information. By creating these rights, the government is conferring the power on the rights holder to refuse to allow others to make use of the
4This appendix focuses on information products that could be sold plausibly in the marketplace. The reader should not infer that the marketplace is the only or optimal mechanism for stimulating the production of socially desirable and valuable works. Works whose creators or rights holders are unlikely to be able to recapture their costs (e.g., some forms of basic research) will not be pursued in the marketplace and will require funding by government or other institutions.
information under certain circumstances. By granting this power, the government is relying on financial incentives to induce the rights holder to allow dissemination of the information, usually through a financial arrangement that is some form of a royalty system or by means of combining its sale with that of an ancillary product.
Of course, under such a system the producer of the new information rarely will be compensated in a manner that precisely equals the cost of producing the information. However, if the number of potential creators of new information products is large, suppliers of new information will continue to produce it as long as the average payment (combining the successes and failures) equals the information's cost. The rationale for limiting the control inhering in these rights is that doing so creates more benefits, in terms of greater use of the information, than harms, in terms of weakening of the incentive to produce the information.
Two useful examples of such limits are the prohibition against lever-aging and the requirement to permit ''fair use." The prohibition against leveraging means that a property right in information can be used to monopolize the products emanating from the direct use or duplication of that information but cannot be used to acquire a monopoly in some other product. For example, the creator of a new method for making steel is entitled to monopolize the steel industry for the duration of the inventor's intellectual property right in the new method. However, he or she is not entitled to monopolize the production of automobiles as well, which could be achieved simply by refusing to sell steel to all auto makers.5
The principle underpinning this prohibition is that the rewards to the creator of knowledge should be limited to the direct-use value of that knowledge, even though granting broader rights might well induce much more innovative effort in the quest to monopolize the entire economy. Various antitrust actions against Microsoft are illustrations of an attempt to clarify and enforce the boundary between legitimate and illegitimate uses of the market power inhering in intellectual property (in this case, copyrighted operating system software).6
5This doctrine is by no means unanimously endorsed. Some scholars of antitrust and intellectual property law believe that no limits should be placed on the use of a patent or copyright to create as extensive a monopoly as possible, to induce maximal innovative effort. In essence, these analysts are advocates of a different solution to the trade-off between innovation and dissemination than is presently embodied in law and practice.
6For example, in a complaint against Microsoft filed by the Antitrust Division of the U.S. Department of Justice, the government argued that Microsoft had used its monopoly power in operating systems for personal computers to disadvantage competitors in applications software, such as Internet browsers, in part to monopolize these markets as well and in part to prevent other browsers and office applications suites from becoming integrated products that would threaten Microsoft's operating system monopoly. Regardless of the merits of
(footnote continued on next page)
In the second example, the fair-use doctrine has many facets, but one that is instructive is the right to characterize and quote copyrighted publications in other publications that evaluate or extend the first. Thus, an author is entitled to decide who can publish the entire text and to prevent others from incorporating part of the text into another publication without attribution. However, he or she is not entitled to prevent another author from making references to or quoting from the text in a work that evaluates, extends, or corrects the first. In addition to other motivations, this doctrine also has an economic rationalethat the value of criticism and extension arising from fair use exceeds the value of a greater incentive to produce new text that might arise from preventing others from criticizing or correcting it.
Given that society has elected to limit intellectual property rights, the economics of information can shed additional light on exactly how one should approach making the trade-off between the conflicting objectives of generating new information and disseminating existing information as widely as possible. In particular, a crucial component in deciding how to make this trade-off is to quantify the magnitude of the effect of incentives on the creation of new information.
If increasing the strength of intellectual property and, hence, the rewards that come from it generates a large supply response, then the case for strong intellectual property protection is great; however, if the amount of effort put forth in creating new information is not sensitive to the rewards derived from it, the case for strong IP rights is weaker. Unfortunately, the amount of hard evidence about the quantitative significance of IP rights is extremely limited. The reason is that changes in IP rights are infrequent and, typically, are limited to a specific type of information, leading to questions about the generality of the lessons learned from any specific case (see the example discussed in Box D.1).
The Enforcement Issue
Socially beneficial systems of intellectual property rights necessarily create costs as well as benefits for the simple reason that such rights are costly to enforce. For instance, the holder of a copyright or patent needs to pay the costs of negotiating an agreement with a licensee to exploit the information that the right protects. Then, to protect the value of the right, either the creator or the licensee must make certain that others do not engage in unauthorized use.
(footnote continued from previous page)
this complaint, the issue is clear: to what extent should the holder of a valuable intellectual property right that confers substantial, presumably legal, market power (here, Windows) be able to use that IP to gain market power in related markets (here applications software)?
Obviously, the value of many IP rights exceeds the private costs of enforcement or else license agreements would not exist. Of course, if these enforcement costs are borne by the creator or licensee, they must be recovered in the final price of the product that uses or embodies the information. Hence, like royalties, these costs drive still another wedge between the price and the social cost of the product and, like royalties, are another source of potential economic inefficiency in the dissemination of the product.
Like all forms of property, some part of the cost of enforcing IP rights is paid by taxpayers through government enforcement. Examples are the costs of running the court system (to adjudicate disputes over rights) and the police system (to investigate the failure to honor enforceable property rights). And, as is the case for all forms of property, enforcement is not entirely public. Just as companies pay for private security, they also pay part of the cost of protecting their IP.
An important point about enforcement is that the holder of a private property right has an incentive to maximize the extent of government enforcement, regardless of the efficiency of public-versus-private enforcement. A public enforcement mechanism, as long as it works, is superior to private enforcement from the perspective of the rights holder because the cost of the former can be spread among all members of society through the tax system.
One example of cost shifting was an attempt in the 1980s to outlaw videocassette recorders (VCRs) because they could be used to create unauthorized copies of motion pictures and television programs. Outlawing VCRs would have created two costs. First, the government would have assumed responsibility to ferret out and capture VCRs, much as it bears responsibility for finding and confiscating illegal drugs. Second, consumers would have been forced to bear the cost of forgoing legal and legitimate uses of VCRs, and the VCR manufacturing industry (including holders of VCR patents) would have been forced to abandon the income generated by a product that sells for hundreds of dollars, to protect a product that could be sold for much less and, in the case of television broadcasting, is given to consumers for free.7
The important points here are these:
• Like any other kind of property rights, intellectual property rights can be costly to enforce;
• Enforcement, in and of itself, adds nothing to the social value of IP (although enforcement may be important to induce others to create and distribute IP);
• The cost of enforcement is a balance of public and private costs, seldom entirely one or the other; and
• The balance of enforcement costs between public and private, as well as the overall level of cost, must be considered in designing the legal and social institutions for managing IP.
7As a result, popular feature films usually are not released to the videotape market until after their first theatrical run. These releases on videotape have a negligible effect on revenues from theaters and can be profitably priced low enough so that pirates have little incentive to engage in extensive unauthorized commercial copying.