2. Some Basic Economics of Information Production and Innovation
Summary of the chapter
Information and culture are fundamentally different from standard "economic goods" such as steel and cars in two ways. First, standard goods are rival or exclusive while information is nonrival. Consider an apple: If I eat it, you cannot also eat it. If you want to eat an apple too, more resources (trees, irrigation, fertilizer, labor, etc.) will have to be expended to procure you that apple. This is not the case with information. You and I can both read this book at the same time. My consumption of the information contained in the book in no way diminishes that information or prevents others from consuming it as well. The second way in which information differs from other goods is that it is both an input and an output of its own production process. Consider a wood table: the inputs of the production process are wood, fastening devices (nails, glue), and labor. A carpenter uses these things to make the output: a table. Now consider an academic paper: The inputs are ideas and labor and these ideas are often grounded in previous academic papers. Academic papers are thus "used" in producing new academic papers.
It is these two properties of information, nonrivalry and input/output circularity, that explain why the law gives people some property-like rights over information but limits those rights in comparison to those given over other types of goods. The standard rationale for giving people property-like rights over information is that people will fail to produce information if they have no incentives to do so. Copyright, which grants the right holder a monopoly over her creation, allows her to turn information into a rival good, creating scarcity and, thus, a market for her work. That is, while the actual cost of producing a second copy of a book is marginal (it is merely the cost of photocopying the pages), copyright prevents people (through threat of sanction) from doing this without the author's permission, and thus allowing the author to control the distribution of her work.
The standard rationale for limiting the monopoly rights created by copyright and patents is that because information is both an input and an output of information, we should not allow the "building blocks" to be captured indefinitely or entirely. Thus we have time limitations on copyright, thus returning old information to its nonrival state, and "fair use" provisions, allow new producers to use otherwise protected information in certain ways and contexts (e.g., criticism, parody, news reporting).
Many economists describe the system of intellectual property as a balance of static and dynamic efficiencies: Intellectual property creates static inefficiency by artificially raising the price of information goods, but this inefficiency is balanced out, however, by gains in dynamic efficiency in form of creating an incentive for new information production.
This chapter explores whether the realities of information production fit well with standard economic theory. It argues that the efficiency of regulation information, knowledge, and cultural production through strong copyright patent is not only theoretically ambiguous but also empirically inaccurate. This chapter posits that innovation and information production is driven mostly by a mixture of (a) nonmarket sources (both public and private) and (b) market actors whose business models do not depend on the protections of intellectual property.
The chapter concludes that while granting exclusive rights-based over information provides some incentives for the future production of information, a comprehensive intellectual property regime that mimics property too closely (such as the system in both the United States and the European Union) cannot work perfectly even under ideal conditions. The upshot is that noncommercial production can often be more efficient and more important commercial production and that noncommercial production is thus moving to a more central role in our information production system.
The Diversity of Strategies in Our Current Information Production System
This section maps the variety of strategies used by information producers. The types of strategies can be divided along two axes. The first is the benefit maximization axis. At one end of this axis are information producers that rely on exclusive rights (such as patents and copyrights) to make money; at the other end, are information producers that give their information away for free in order to gain non-monetary benefits; an in the middle are information producers that aim to make money but not by relying on exclusive rights.
The second axis along which information production strategies are divided is the cost minimization axis. This axis describes the various ways in which information producers minimize the cost of the information inputs they rely on. At one end of this axis are producers who rely on public domain (i.e., free for fall) information; at the other end are information producers who barter with other producers for access to protected information (for example, small companies that form patent pools); in the middle are information producers that a) rely on information they already control, and b) purchase without reciprocity control of new information from other producers.
These two axes of three points each create nine different strategies for the production of information, which Benkler describes as follows:
The Romantic Maximizer: Relies on the public domain to minimize cost and right-based exclusion to maximize benefit. The classic example of the author writing her novel and selling it to a publisher or inventor creating a machine, patenting it, and marketing the invention.
Mickey: Relies on intrafirm strategies to minimize cost and rights-based exclusion to maximize benefit. A salient example is Walt Disney, which owns a vast number of copyrights over cartoon characters such a Mickey Mouse. Disney reuses these characters to produce derivative works or it hires animators to create new characters with the proviso that Disney will hold the copyright over these characters as well.
RCA: Relies on bartering/sharing exclusive rights to minimize cost and rights-based exclusion to maximize benefit. In the 1920s four different companies--RCA, GE, AT&T, and Westinghouse each controlled patents that prevent each other from manufacturing the best radios technologically possible. To solve this problem, these four companies agreed to share their patents with one another but not any outside party. Though no single company had a monopoly, together these four companies could effectively exclude every other competitor from the radio manufacturing market.
Scholarly Lawyers: Relies on the public domain to minimize cost and gives their information products away for free in order to derive income from other sources. For example, legal scholars are not paid when they publish an article in a law journal. But it is not just their devotion to scholarship that drives them to give their information products away for free. They are also trying to acquire status and good reputation in order to get a job or make tenure. Another example is bands that give music away for free in order to attract audiences to their profit-making performances.
Know-How: Relies on intrafirm information production to minimize cost, but does not sell that information to acquire profits. Law firms that have libraries of boilerplate forms minimize their cost because they do not have to "re-invent the wheel" every time a client wants to form a contract or draft a license. However, the law firm is not making its money from selling these forms. It's making it money by serving a larger number clients more efficiently.
Learning Networks: Relies on bartering or sharing information to minimize cost, but does not rely on copyright or patents to maximize benefits. Newspapers that join together to create a wire service share reporting information, but it is not the copyright that makes them money. Rather they wider range of stories they offer (because of the sharing) allows each regional newspaper to attract a greater number of customers.
Joe Einstein: Relies on the public domain to minimize cost and seeks to maximize non-monetary benefits. Joe Einsteins are moved by a wide range of goals. Status in their relevant communities, personal satisfaction from the act of giving, personal satisfaction from the act of creating. Examples of Joe Einsteins include academics who write articles for fame, carolers who sing for free, users who contribute to mailing lists, or programmers who write free software.
Los Alamos: Relies on intra-firm information sharing to minimize cost and produces public goods. Often these public goods are used to secure additional funding (government grants) or increase the status of the agency.
Limited sharing networks: Relies on bartering/sharing protected information to minimize cost and produces public knowledge in order to gain nonmonetary benefits. For example, Academics often join working paper groups where they read each others drafts and provide comments on them so that each member can improve her own work before publication. "Copyleft" conditions on derivative works also fall under this category.
The Effects of Exclusive Rights
It is only after we examine the diversity of information appropriation strategies that we can truly understand the effects of strong exclusive intellectual property-type rights. The standard economic justification for copyright is that it sacrifices static efficiency in order to achieved dynamic efficiency. But the effect of static inefficiency is different for each type of appropriation strategy, making the argued for trade off much more complex and ambiguous than the standard justification, in its simple form, allows.
Consider the following hypothetical: There are ten firms in the "infowidget" business. Two these rely on the Romantic Maximizer model--they produce the goods and sell them based on their patent rights. Six of these firms produce infowidgets on supply-side (Know-How) or demand-side (Scholarly Lawyer) models--they make their Realwidgets or Servicewidgets more efficient or desirable to customers. Finally, the last two firms are non-profits that exited on a fixed, endowed income. Each firm produces five infowidgets, for a total market supply of fifty. Now let's imagine a change in law that increases exclusivity and is efficient under the simple justification of IP rights. We'll say it increases input costs by 10 percent, but increases appropriability by 20 percent (giving an expected gain of 10 percent). The Romantic Maximizers will thus see an increase in production from 10 infowidgets to 11--making the policy change seem like a good thing. Looking at the market as a whole, however, we see that eight firms see an increase of 10 percent in costs and zero gain in appropriability (because the other eight firms do not rely on exclusive rights to sell their products). Thus the productivity of these eight firms declines from 40 infowidgets to 36, meaning that market as a whole declines in production from 50 to 37. When the scope of our inquiry is widened we see that the policy change made the market less efficient.
What is important to note is that the strength of exclusivity will shape the population of business strategies. Changing the level of exclusivity makes some models more desirable than others. This change can also affect the value and attraction of consolidating information production.
When Information Production Meets the Computer Network
Prior to the industrial information age, music was a largely a relational good. It was something people did in the physical presence of each other. Music was produced in the folk way through hearing, repeating, improvising; in the middle-class way through buying sheet music and playing for others or attending public performances; and in the upper class way though hiring musicians. Capital took the form of instruments and performance space and was widely distributed amongst musicians and performance hall (and drawing room) owners. Music, as a market good, depended on the physical proximity of producer and consumer.
The introduction of the phonograph changed all this by transforming music into a product that could be sold in the industrial information economy. Capital became concentrated in the form of studio owners and recorder producers. And with the broadcast of music over the radio, a smaller number of musicians reached a much greater audience. Placing this in the context of the strategies above, we might say that the music industry shifted from a reliance on Scholarly Lawyer and Joe Einstein models to reliance on Romantic Maximizer and Mickey models.
This stylized story of the music industry typifies the mass media more generally. The mechanical press, the telegraph, the phonograph, film, high-powered radio transmitters--all of these evolutions in technology affected the market for information goods. As the industrial information age progressed ownership of newspapers became increasingly concentrated, as was the case with radio stations, cable networks, and film studios. The lesson here is that the technology of information production also affects the kind of business models that are desirable.
Information and cultural production that three primary types of inputs: 1) existing information; 2) mechanical means of producing and sensing information (printing presses and radio transmitters); and 3) human capacity for creativity. In the industrial information the costs of the second type of input was quite high, making it efficient to have a few producers of information that could reach a wide audience. The costs of the second kind of input, though, are drastically declining in the networked information economy. Thus, when information is priced at its marginal cost (zero), human capacity becomes the only truly scare resource in information production. Human communicative capacity, however, is radically different from printing presses and satellites. It's not the kind of capital good that can be transferred or aggregated; rather the capacity of each individual is as unique as the individual herself. The promise of the networked information age is to bring the rich diversity of our individual experiences smack into the middle of our economy and productive lives.
Nonmarket production has always played a greater role in the production of information goods as it has in standard economic goods. Street performances, story telling by the campfire, religious texts that are given away, not sold, folk songs, free galleries, the list goes on and on. What is changing in the networked information age is that means of producing and exchanging information, knowledge, and culture have not only been drastically reduced in price, but also have been widely dispersed throughout the population. Home computers, personal printers, cheap electronic storage, fast transmission capacity over the internet, etc. Nonmarket behavior is thus becoming central to producing our information and cultural environment
Strong Exclusive Rights in the Digital Environment
We now have the basic elements for a clash between incumbent institutions and emerging social practices. The incumbent institutions, acting in accordance with rational self-interest, seek to policies that improve the returns on their current business models, which strong, broad, exclusive rights like copyrights and patents. As we have seen, however, technology has evolved so that strong exclusive rights may adversely affect other actors (both market and nonmarket). Thus, we are at a critical juncture for policy development. We must learn to evaluate the effects of exclusive rights on all the modes of information production. And we must be willing to adjust our institutional environment to make way for the new social practices made possible by the changing information economy.
Sources cited in the chapter
Other relevant readings
Standing-on-the-shoulders-of-giants Effect (37)
Input/Output Circularity (see Standing-on-the-shoulders-of-giants Effect)