Circulation Title: An Information Framework for the Planning and
Design of "Information Highways"

THE CHARACTER OF INFORMATION:
Characteristics and Properties
of Information Related to Issues Concerning
Intellectual Property

February 10, 1985
(Revised September 2, 1985, October 1, 1994)

In Support of the Office of Technology Assessment
Project on Intellectual Property
Contract 533-1510.0

(c) Center for Information, Technology & Society, 1985, 1994

Parts may be copied in derivative works with full attribution
under the Doctrine of Fair Use.

Author:
W. Curtiss Priest, Ph.D.

Submitted to:
Congress of the United States
Office of Technology Assessment
Washington, DC 20510

Table of Contents

1. An Operational Definition of Information ................................................ 1
   1.1 The Function of Information in Achieving Outcomes .................. 1
   1.2 Information as Intellectual Property ............................................. 3
   1.3 Conclusion ..................................................................................... 5
2. Characteristics and Properties of Information in Commerce and Transactions ... 6
   2.1 Market Related Characteristics of Information as a Commodity ........ 6
      2.1.1 Intrinsic Co-production ..................................................... 6
      2.1.2 Time Constrained Consumption of Information ................. 7
      2.1.3 High Investment to Reproduction Cost Ratios .................. 7
      2.1.4 Relevance of Information More Variable Across Consumers .. 8
   2.2 Market-failure Related Characteristics of Information ................... 10
      2.2.1 Public Good Characteristics ............................................. 11
         2.2.1.1 Inappropriability ................................................. 12
         2.2.1.2 Non-depletability ............................................... 13
      2.2.2 Externalities ..................................................................... 13
      2.2.3 Indivisibilities (of supply) ................................................. 15
      2.2.4 Economies of Scale and Scope ......................................... 16
      2.2.5 Inherent Uncertainty and Risk in Information Production .. 17
      2.2.6 Information/Knowledge About the Information ................ 19
      2.2.7 Intangibility ..................................................................... 19
      2.2.8 Transaction Costs and Information ................................... 21
   2.3 Non-market Related Characteristics of Information ....................... 23
      2.3.1 High Intrinsic Relationship to Human Welfare ................... 24
      2.3.2 High Intrinsic Relationship to Freedom and Privacy .......... 26
3. Relationship of Information to Products and Services ................................... 28
4. Rewards, Incentives, and Funding Sources for the Creation of Intellectual Property ................................................ 36
   4.1 Rewards, Incentives, and Funding for Science ................................ 36
   4.2 Rewards, Incentives, and Funding for Culture ................................ 39
   4.3 Rewards, Incentives, and Funding for Products and Services .......... 40
5. Structural Role of Intellectual Property in Maintaining a Viable Economy ..... 43
6. Summary .................................................................................................... 45
Appendix A – Bibliography ............................................................................. 46

EXECUTIVE SUMMARY

To comment on issues concerning intellectual property it is useful to identify what distinguishes information from other forms of property. Fourteen distinguishing characteristics of information are identified and discussed in relation to property rights.

These characteristics, however, are not only useful for considering issues concerning intellectual property but for more generally inquiring into the nature and purpose of information in society. Summarized below are brief definitions of the fourteen characteristics and introductory remarks about the general importance of the distinguishing characteristic.

Definitions and Introduction to Fourteen Characteristics of Information

As a Commodity:

1. Intrinsic Co-production
   The character of information to be instrumental in achieving other goods and outcomes. This character makes information inherently more valuable than goods that are not instrumental in character.
2. Time Constrained Consumption
   The character of information to occupy more consumer time per dollar expenditure than other commodities. This characteristic combined with the relatively low reproduction cost characteristic (3) has long run employment implications.
3. High Investment to Reproduction Cost Ratios
   The creation costs of information divided by the cost of reproducing one unit of the good. The implications of this characteristic are economies of scale and scope, and resulting market structure.
4. Relevance, More Variable Across Consumers
   The character of particular information to be acquired usually only once. The result is high variability in consumption by each consumer. This characteristic tends to work in the opposite direction of low reproduction costs, since it implies that information will become more and more customized and particularized.

Market-failure Related Characteristics:

5. Public Good
   The same information can be used by many consumers without interference.
      Inappropriability – The difficulty in receiving full market compensation for the creation of information due to the problem of exclusion. The result is under-production and under-compensation.
      Non-depletability – Information does not dissipate with use. Producers must compete with past producers but society benefits with an accumulation of knowledge.
   Goods with substantial public good characteristics such as national defense, recreational parks, and safety facilities such as lighthouses are usually supplied by the government to reduce the "free rider problem" associated with inappropriability.
6. Externalities
   The effects of information, usually positive, that are not accounted for in its price. The effects of information, especially as education, have considerable positive externalities in terms of reducing unemployment and increasing general social welfare.
7. Indivisibilities (of supply)
   Information must be purchased in lumps; these lumps may be vastly greater than the information actually sought. This characteristic along with the variable relevance characteristic will contribute to utilization of information technology that reduce indivisibilities and permit customization.
8. Economies of Scale and Scope
   1) Decreasing unit costs when the scale of operation is increased; and 2) decreasing costs associated with joint production. Historically, information distribution such as telegraph, telephone, radio, and television have exhibited sufficient economies of scale and scope as to require government regulation to reduce problems associated with natural monopolies.
9. Uncertainty and Risk in Production
   The inability of firms to produce information when risks and uncertainties are present. A problem, in particular, in the generation of basic knowledge that requires substantial investment in research.
10. Information/Knowledge
    Information about information is less likely to be available because of appropriability problems. This leads to under-consumption of information due to problems of search.
11. Intangibility
    The character of the value of some information to be non-monetizable. Information is the basis of education, communication, and other activities which are difficult to value because the contribution of these activities to the welfare of society is largely intangible.
12. Transaction Costs
    The additional costs incurred by the producer in appropriating the value of information. [Transaction costs, in the economic sense, are those costs associated with negotiation, contracting, and enforcement, and does not refer to the general costs related to distributing or transmitting information.] Transaction costs are a major contribution to indivisibilities in the supply of information since contracting and enforcement costs are difficult to reduce below a certain minimum.

Non-market Related Characteristics:

13. Intrinsic Relationship to Human Welfare
    Human welfare is a product of individuals and groups achieving desired outcomes. Thus, information is intrinsically related to human welfare in that it inherently facilitates the achievement of outcomes.
14. Intrinsic Relationship to Freedom and Privacy
    Freedom – Information affects the range of choices available to the individual. Freedom is a lack of restriction on choices. Thus information leads to greater freedom.
    Privacy – Incomplete information may result in defamation of character. Therefore, information must be selectively made private to reduce the probability of defamation.

1. An Operational Definition of Information

The common-use definition of information refers to the act of informing or the condition of being informed. But how can we really judge if some arrangement of bits, words, etc. really does inform? To answer that we should look at what it means to inform – we should see what is different about a person after information has been communicated to him or her.

1.1 The Function of Information in Achieving Outcomes

As the section title implies, it is useful to think in terms of the function of information to define it. Indeed, the common-use definition states a function – "the act of informing." Once "informed", what is different about the person? The choices that a person might make are different. For example, if I inform a shopkeeper that it may rain, the shopkeeper may choose to cover-up outside merchandise or may choose to close the shop's windows. By informing the shopkeeper I have provided "... a basis for choice – that is, a belief in the greater efficiency of one choice compared to another" (Ackoff, 1972, p. 153). The concept of "efficiency" is a very key one. It is tied with a fundamental purpose of information and human activity.

Nearly all human activity is directed toward achieving outcomes. Outcomes include getting to the bank on time, relaxing, getting an A in a course, and cheering a friend. These outcomes can be achieved with more or less efficiency. Now there are two basic attributes about a person that affect the efficiency of achieving an outcome – 1) The ability to choose among alternative ways of achieving the outcome, and 2) The capability to make sure the choice succeeds. Returning to the shopkeeper, the outcome is keeping rain off the merchandise. Informing the shopkeeper of the possibility of rain changes his choices – he can now know to protect the merchandise before it rains. But having chosen to, say, cover an outside table with plastic, his capability to keep the plastic in place (using lead weights or whatever) also determines the efficiency of achieving the outcome. For if the plastic were to blow away the efficiency of protecting the outside merchandise has gone to zero. If information affects the basis for choice, what affects the capability of ensuring success? Success comes from practice, instruction, and other activities that affect capability.^*

Thus, the fundamental function of information in human activity is to affect the efficiency of achieving outcomes. The effect need not always be positive for if I am poorly informed, I will make less efficient choices and my efficiency will decline.

By defining information in terms of outcomes, the value of information follows immediately. Take the expected value (EV) related to achieving an outcome – this is the sum of the probability of each particular choice, times the efficiency of each choice, times the relative value of the outcome (cf. 2.3 for relative value). Considering the expected value (EV₁) without the information and the expected value (EV₂) with the information, the value of the information (EV_i) is:
    EV_i = EV₂ – EV₁

In economic terms, EV_i = P (price), but only:
1) When there are no market imperfections, and
2) For those consumers where the marginal utility (the utility of the last unit) equals the market price.

Examples of where the price is not a proxy for value include situations where the information provides benefits to other than the purchaser (a positive externality) or where the purchaser would have paid more than the market price. In the latter case, with no market imperfections, we know that the value to the purchaser is at least the market price and we could attempt to use a "willingness-to-pay" measure (how much the person says or indicates he might pay) for the information as a better proxy for its value (cf. discussion in 3).

This choice based approach to the value of information differs from Machlup's approach where he uses the opportunity cost of the productive factors as a shadow price for the value of information (Machlup, 1980, p. 208). The approach is more like Hall's – "[the] value of a specific message is equal to the utility gained from shifting to a better choice among terminal actions" (Hall, 1981, p. 161) and Taylor's description of information value in decision contexts (Taylor, 1985).

We can now define information as:
    Information: A communication that produces a change in the tendencies to choose certain actions over other actions (where a change in tendency can be observed by noting changes in the probabilities of choice). (cf. Ackoff, 1972, p. 144).

Mackaay (1982, p. 107) states, "[i]nformation is the essential ingredient in choice..." Cherry (1966, p. 170) states, "Information can be received only where there is doubt; and doubt implies the existence of alternatives – where choice, selection, and discrimination is called for."

At the recent Elsevier Symposium on Information Policy and Scientific Research, Manten and Timman examine "external" and "internal" definitions of information. An external definition of information presented by Vinken (Manten & Timman, 1983, p. 154) defines information as something someone is willing to pay for. Internal definitions of information are typically: "Information consists of data, that change the world view of the receiver" (Manten & Timman, 1983, p. 7). The internal definition is rejected because "it is impossible to say which data are information and which are not."

The above definition, however, is not an internal definition but a different and more fundamental external definition. The problem with the strictly economic external definition is that it excludes all information for which pecuniary (monetary) measures are inapplicable and all information which a person wouldn't pay for. For example, if I inform my daughter that she must stay in at night, I doubt she would pay me for that information. Manten & Timman, similarly, reject the external economic definition.

We can also define communication as:
    Communication: A message between sender and receiver that affects 1) the choices of the receiver (information, as above), 2) capabilities of the receiver (instruction), and/or 3) affects the receiver's values (motivation).^*

1.2 Information as Intellectual Property

To what extent is intellectual property a piece of information? We should investigate some particular forms of intellectual property and see how they meet the definition. Some forms are:

• An invention – An invention is the embodiment of many new choices. The invention combines parts and processes in new combinations to produce the desired effect. A description of the invention, such as a patent, or the invention itself in practice is informing since it conveys to others how a desired effect can be attained with new choices or different combinations of choices. For example, to create a letter printer, the invention may relate to spraying ink rather than transferring ink by impact. (The idea of spraying ink, per se, cannot be patented, rather the embodiment of the idea in a specified set of choices for making the printer is what is patentable.) Thus, invention clearly is information.
• A song – The music and score are copyrightable under law. Songs can evoke past feelings and evoke new feelings by providing a vicarious experience and stirring empathy as well. They can be symbols and images to identify with and emulate. A sense of catharsis is common as difficult situations are experienced at-a-distance.

Since much of the effect of a song is affective (affecting feelings) the relationship among feelings, outcomes, and information should be identified. A feeling is the sense of satisfaction or dissatisfaction associated with a particular outcome or potential outcome. For example, the feeling of anxiety is a sense of dissatisfaction with the uncertainty of a potential positive outcome. (If I sense that the actions of a co-worker will jeopardize my getting a promotion, I will feel anxious about my job.) Thus, feeling is an evaluation associated with an outcome (cf. Ackoff, 1972, p. 100).

A song provides the listener with a synthesized experience that temporarily replaces (or becomes) "reality." The experience can be mainly vicarious, mainly empathetic, or both. If it is vicarious, the experience substitutes for one's own; if it is one of empathy, the experience becomes like a close relationship to another. There are two behavioral effects of these experiences 1) The positive and negative feelings can be "consumed", for example, a stirring piece of music can evoke exhilaration and that can simply be felt, and 2) The feelings evoked can influence future choices. Feelings influence choices mainly by becoming attitudes. For example, marching music can be used to build up strong attitudes of nationalism. A strong nationalistic attitude will then result in loyalty to country and other forms of patriotism. Thus a song is partly information but not obviously so. A song, by creating synthesized experiences, creates feelings and attitudes that influence choices to achieve outcomes and is, therefore, by definition information. Songs also are motivational (i.e. produce changes in a person's values) and are consumptive goods – to be enjoyed for the immediate sensate.

• A Textbook – Textbooks are subject to copyright. The material in the textbook directly changes the reader's available choices through the knowledge that it brings the reader. For example, if the reader needs to show a correlation between crop growth and rainfall there are varying degrees of efficiency he can achieve in obtaining the outcome – a statement of the correlation. He might have no notion about how to go about making this statement. A textbook on statistics would provide information about the technique of regression analysis and the R² term that expresses correlation as a number from 0 to 1. With this new information, the reader would choose that analysis and his efficiency would significantly increase. (His capability may still be low and require practice and instruction to be very efficient in achieving a correct outcome.) Thus, a textbook squarely fits the definition of information.

In summary, while most intellectual property is mainly information, there are exceptions. In particular, music, movies, and similar products that create synthesized experiences are part informational, part motivational, and part consumptive. Further, the proportion of these three components depends on the response of the viewer or listener.

What remains is the question of what is it that intellectual property rights are really given for. As technology continues to change the medium of expression, it is important to distill what is fundamentally being protected by patent law, copyright law, etc.

It is commonly stated that what is being protected are not these ideas but the form of expression. While this loosely pins things down, it is too vague to apply to many new situations resulting from technological change. Let us consider the textbook again.

Paul Samuelson's Economics is an excellent textbook. Why do I say that? Because it is comprehensible, anticipatory, progresses logically, is well illustrated, and is comprehensive. These features make it easier for me to improve my knowledge. This comes back to a notion of efficiency.

As indicated above, what makes information valuable is how it changes my efficiency in achieving an outcome. Thus, what is valuable about a piece of intellectual property is how effective it is in changing my efficiency. The list of positive features on Samuelson's textbook all relate to how effective it is in changing my efficiency. It is this aspect of the work that is unique and can be protected.

As a short cut, however, copyright protects the work by preventing copying. Of course, a good copy would have the same efficiency characteristics as the original, and by preventing copying the work is "protected." Can we, however, generalize the notion of efficiency to get a more fundamental protection approach? Particularly when "bits and pieces" of various works can be combined on wordprocessors, in film editors, etc. problems of copyright and royalties become difficult.

To accomplish this, the concept of efficiency would have to be generalized to include not only information but motivation and consumption so we can account for synthesized experiences. With these in place, we could deal with issues of imitations, replicas, and partial copies. As a guide, it would be useful to compare court decisions on copyright and patent infringement with the evolving theory.

1.3 Conclusion

The definition of information permits us to identify the extent that information is embodied in intellectual property. The value of information follows readily from the definition. The definition, by tying messages to outcomes, is not susceptible to the problems associated with definitions of information that are defined in terms of bits. The definition is also a reasonable extension of the common use meaning of the word.

2. Characteristics and Properties of Information in Commerce and Transactions

Information in commerce has different characteristics from most economic commodities. In a society whose commerce is based increasingly on buying and selling information these characteristics will have increasing significance. Also, there are inherent characteristics of information that relate to non-economic transactions such as the role of information in maintaining human relationships. Fourteen different characteristics of information are identified and their implications to commerce, transactions, and intellectual property are described.

The characteristics are grouped in three categories. The first category, "market related," pertains to commodity characteristics that set information apart from other goods in market activities. The second category, "market-failure related," pertains to characteristics of information as a commodity that cause markets to behave inefficiently (or not at all). The third category, "non-market related," pertains to characteristics of information in non-economic transactions – primarily those comprised of human relationships.

2.1 Market Related Characteristics of Information as a Commodity

Drawing on the definition of information in Section 1, information as a commodity can be examined. Fundamentally, producers of information are in the business of increasing the efficiency of choices made by the purchaser of information. Their products are the instruments by which the purchasers increase their efficiency by "consuming" the product.

2.1.1 Characteristic I – Intrinsic Co-production

Information is not typically the object of consumption as are apples, coal or umbrellas. In contrast, information has an instrumental value – it aids in attaining desired outcomes. We thus refer to information as an intrinsic co-producer. Information inherently is a co-producer of outcomes. New information works in tandem with the knowledge already possessed by the person and with the capabilities of the person to achieve outcomes. The same applies to groups, families, firms, and other organizations.

There are many classes of commodities that are co-producers. While we do consume pencils as they wear down, the pencil is a co-producer of the written word. Cosmetics are co-producers of romance and cars are co-producers of transportation. But all of these co-producers are dissipated as we use them. Perhaps the closest parallel to information is the chemical catalyst. The catalyst is not consumed in the reaction and yet accelerates the process. But even most catalysts become contaminated after a while and need to be replaced.

Some writers have termed information a resource (Levitan; 1980, 1982, Cleveland, 1982). A resource is something that can be drawn on for help. Most resources are depleted with use – oil, iron ore, etc. A non-depletable resource such as solar power is a closer analogy to information.

Cleveland (1982, p. 36.) describes two characteristics of information that are directly related to intrinsic co-production. Cleveland refers to the "expandability" of information. "[I]nformation expands as it is used." Information is co-productive of more information – knowledge builds upon existing knowledge and news often repackages older information with newer information. The proliferation of information has been referred to as the information explosion and has resulted in "information overload." Indeed, the production of information for information's sake is not necessarily co-productive of other desired outcomes. Thus there is a need to manage information to keep it relevant and make it efficiently co-productive.

Secondly, Cleveland refers to the substitutability of information (Cleveland, 1982, p. 37). "It can replace capital, labor, or physical materials." This primarily emphasizes the co-productive capacity of information through invention. As information contributes to the transformation and automation of manufacturing, capital will further substitute for labor. To refer to this as information replacing labor is perhaps euphemistic.

In summary, we can define the intrinsic co-productive capacity of information as:
    Intrinsic Co-production – The character of information to be instrumental in achieving other goods and outcomes. This character makes information inherently valuable.

Since information is inherently valuable it is significant whether the information can be held as property. Property rights will enable the owner to gain some return from its creation (cf. 2.1.1.1 – Appropriability).

2.1.2 Time Constrained Consumption of Information

For a message to be information it must affect the choices of an individual (see def. Section 1.) To affect choices, the receiver must take time to assimilate the information. This process involves reception, memorization, thinking, and reflecting. (Even the process of synthesized experiences, such as watching a movie, takes considerable time.)

In relation to many other goods, more consumer time is occupied by information than most other goods per dollar of expenditure. For example, $2.00 of beef can be consumed in ten minutes while a $2.00 movie can last two hours (and $2.00 of basic cable service can occupy days of time).

The economic significance of this characteristic of information has not been specifically studied. In general, it would operate to reduce the demand for information. Reduced demand would work against the potential for economies of scale (cf. 2.1.4).

In summary, we can define the time-constrained consumption of information as:
    Time-Constrained Consumption – The character of information to occupy more consumer time per dollar expenditure than other commodities.

Reduced demand would result in reduced prices, making the value of information as property less valuable to its owner.

2.1.3 High Investment to Reproduction Cost Ratios for Information

The initial costs for creating information are high relative to the costs of reproduction. This contrasts sharply to most other goods where the production of multiple units of a good require considerable factor inputs of capital, labor, and materials.

Returning to the definition of information, the reason for this large difference relates to the necessary facilities needed to affect the choices in behavior. The informing process is mainly one of communication. Communication primarily consists of transferring audio and visual images. For the written word printed paper is quite effective and very low in cost. For motion picture images, the VCR cassette is less than $5.00 plus recording or pre-recording costs.

Equipment for accessing information such as VCR's, television monitors, computer terminals, and communication lines are not low in cost. If one amortizes the costs of these devices over the media costs, the reproduction costs are noticeably higher for some types of information. For example, the cost of a basic VCR over five years would be around $700 including maintenance and the cost of capital. Allocated over a film library (and rentals) of 200 tapes, this would add another $3.50 per tape. While appreciable, the additional cost is not large.

The literature abounds with accounts of dropping prices for terminals, minicomputers, and other hardware. Though this is not expected to continue to drop as steeply in future years, even the costs of the more expensive devices will probably not swamp the situation. Nonetheless, there is still considerable room for improvement in picture resolution, 3-dimensional display, and the instant accessibility of large quantities of information. The question remains, how much more effective would the communication be with these types of enhancements and what will be the costs of these improvements?

High investment costs put pressure on the industries to search for the means to reduce these costs. The development of information is labor intensive. Some film footage can cost $2000 a second to produce. Because of the skill and intellect required to produce information, these costs are not likely to drop soon. Thus the investment/reproduction ratio is likely to remain high.

In summary, we can define the investment to reproduction cost ratio of information as:
    Investment/Reproduction Cost – The creation costs of information divided by the cost of reproducing one unit of the good.

The implications to issues of intellectual property are important and one of the main problems today. Since investment costs are high and it is so relatively costless to reproduce the product, it is extremely important to protect the owners of the investment (cf. 2.1.1.1 – Appropriability).

2.1.4 Relevance of Information More Variable Across Consumers

There are various factors that determine the relevance of a good to a consumer. The most common factor is the consumer's preferences (or values). Consumer goods that are purchased out of preference are generally consumed again and again. Further, while many goods are slightly differentiated by styling, the demand for many basic goods is consistent across consumers. Combs, pencils, shovels, lumber, etc. are demanded consistently by many consumers.

Information, by contrast, is peculiarly different. The basic goods of information, such as the alphabet, vocabulary, etc. need only be consumed "once." The consumer does not generally go back for a second helping of learning arithmetic. He does go on to learning more arithmetic. Thus at any time, each consumer of information has a stock of knowledge that he or she is adding to and makes selections for more information based on an assessment of what further information is useful. Only when "memory fails" might the consumer go back and then usually to a familiar reference already on the shelf.

The "synthesized experiences" (cf. Section 1), if they are consumed again and again (like seeing a movie many times), become less and less information each time and more a simple consumptive good like food.

Thus, a second factor that determines the relevance of a good is its utility in building up a stock of knowledge. If the person reads something that is completely redundant to what he or she already knows, the material is not information at all – for there will be no changes in the future choices of the individual as a result of the message. If the person persists in re-reading it we can only assume there is a consumptive value in doing so.

This characteristic of relevancy produces high variability in the "consumption" of information as a commodity. What I require for information today I will probably not require tomorrow. Only if we look at information in the aggregate, over time, does its variability substantially decrease. Everyone will learn the alphabet at some time; everyone will probably read Romeo and Juliet.

In summary, we can define the variable-relevance of information as:
    Variable-relevance – The character of particular information to be acquired usually only once. The result is high variability in consumption by each consumer.

The implication of this to property is striking. Basically a producer of information gets "one shot" at his consumer. The market segments for information as a commodity are much more tightly drawn than for other goods. The success of a "reader" for five year-olds is very dependent on previous learning environments and the age-group.

But because of the variability in prior information, it encourages product differentiation – information can be packaged and repackaged to fit the particular circumstance of the consumer. (Computer-assisted instruction, with the ability to branch as required by the student, is a good example.)

But product differentiation is problematic to establishing property rights. I must continue to copyright each variant to try to protect it and if someone else produces a variant for a slightly different market segment, that producer might be able to take most of my product without compensating me.

2.2 Market-Failure Related Characteristics of Information

The failure of markets has been a subject of public finance study for many decades. Market-failure and the role of government are closely entwined since a major role of government is to provide what the market cannot.

In the early part of the twentieth century, market-failure problems associated with transportation (railroads, trucking and airlines), communications (telephone, radio, and eventually television), public utilities (water, gas, etc.), and food, drugs, and cosmetics were closely studied and the associated industries were subjected to "economic regulation." In the mid-part of the twentieth century market-failure problems associated with health, safety and the environment were addressed with "social regulation."

Economics, as a discipline, primarily concerns the operation of markets. Market-failure, therefore, represents exceptions to the rules of the discipline. The basic concepts of Pareto optimality and cost/benefit analysis, however, transcend market operation and are thus useful economic concepts in the analysis of market-failure. (But even these concepts seldom address the form of market-failure associated with the distribution of wealth.)

In a short space it is not possible to describe all of the economic concepts related to markets and market-failure. To provide some common reference, however, the following terms are germane:

• Allocating Resources – Using labor, raw materials and capital in production
• Efficiency – Allocating resources to their most valuable use (also called allocative-efficiency)
• Pareto Optimality – The condition where all resources are allocated to their most valuable use
• Pareto-move – a change in allocation that increases the net value to society
• Market-failure – a situation where the operation of markets will not result in pareto-optimality or the fair distribution of wealth
• Perfect Competition – many producers and many buyers, where no individual producer or buyer can affect price; the absence of market-failure in a competing market
• Imperfect Competition – competition under forms of market-failure
• Good – Something bought and sold
• Consumptive Good – A good that is acquired mainly to be consumed
• Instrumental Good – A good that is acquired mainly for its co-productive capacity
• Marginal – the last unit of production or consumption
• Marginal Cost Pricing – Usually a firm operates where: marginal cost (the cost of producing the last unit) equals the market price for that good (Stated as MC=P)
• Marginal Utility – The utility derived from the last unit of a good
• Rate-of-Return – the income (or benefit) derived from an investment (Typically expressed as a percentage of the original investment)

Information as a commodity is subject to many forms of market failure. The recognition of some of these in economics date back to the mid-1920's when Pigou (1924), and Marshall before him, recognized that knowledge provided benefits to society as well as to the purchaser (cf. 2.2.2 – externalities). The recognition of externalities of information and knowledge was continued by Viner (1931), Meade (1952), Scitovsky (1954), and Bator (1958).

In 1962, Arrow (1962) wrote a seminal piece "Economic Welfare and the Allocation of Resources for Invention," in which he recognized three forms of market-failure associated with information: indivisibilities, inappropriability, and uncertainty. This was one of the first treatments of "information as a commodity." Others who have addressed information as a commodity and the subject of market-failure include Braunstein (1976, 1981), Levitan (1982), Mackaay (1982), Cooper (1983), and Priest (1984a, 1984b).

2.2.1 Public Good Characteristics

Public goods include goods like national defense, parks, and lighthouses. They are goods that multiple consumers can enjoy at the same time (sometimes referred to as jointness-of-consumption). Information has the characteristic of a public good – the use by one consumer does not prevent the use by others.

Public goods are often provided by government. Since consumers cannot be excluded from enjoying the benefits of a public good, use of tax-payer money to supply the good can be often justified. Economists and public administrators are quick to point out that government is not obligated to provide all public goods – there are situations where the waste and distortion of government in trying to counteract the market failure results in more inefficiency than the original inefficiency (cf. Machlup, 1984, p. 157).

The market failure associated with public goods relates to the concept of marginal-cost-pricing described above. For maximum profit and maximum social efficiency, the firm should set the price of the good equal to the cost of producing the last unit (P=MC). But for a pure public good the cost of the last unit is zero and the firm should charge nothing at all. Thus for public goods social efficiency and private efficiency diverge. Pareto-optimality would be achieved only if the good were given away but then no firm could stay in business to produce it. (Baumol and Ordover suggest that the analysis is not this simple. "Resource allocation may benefit by the imposition of positive prices for public goods in cases where such prices are collectable" [1975, p. 1]). One common prescription for the situation is to have government subsidize the firm to an amount equal to the cost of creating the information. Another is to have government supply the good.

In summary, we can define the public good characteristic of information as:
    Public Good Characteristic of Information – The same information can be used by many consumers without interference.

2.2.1.1 Inappropriability

The term inappropriability is from the word appropriate. To appropriate something is to take possession for use by oneself. In the economic sense, it means to be able to take possession of the worth of something – typically by selling it at the market price. Inappropriability is the difficulty in taking possession of the worth of something.

As indicated in the public good discussion, there is difficulty in excluding people from using information who have not paid for it. This occurs in two basic ways:

• Exclusion 1 – Materials such as books and films are often passed along. This is particularly the case when they are provided by libraries or rentals. Fifty percent of computer software is copied (Frazier, 1985).
• Exclusion 2 – The information contained in any one material often represents the embodiment of prior work. For example, while I give credit to those writers from whom I have drawn for this piece, I have not financially compensated them.

The major result of inappropriability is under-production of information. If the producer of information cannot appropriate its full market value the producer will only produce information where he can recover costs and make a profit. This results in fewer books, movies, information systems, etc. than is socially efficient (all other things being equal). Further, the types of information that are produced are distorted toward those types that can be subject to exclusion.

In summary, we can define the inappropriability characteristic of information as:
    Inappropriability of Information – The difficulty in receiving full market compensation for the creation of information due to the problem of exclusion. The result is under-production and under-compensation.

Problems of appropriability have been recognized by Arrow (1962, 1971), Boulding (1971), Braunstein (1976, 1981), Cleveland (1982), Gordon (1982), Hall (1981), Macklup (1962, 1984), O'Brien (1980), and Priest (1984a, 1984b), and in inventive activity by Dutton (1984), Nordhaus (1969), Mansfield (1977), and Scherer (1977).

Cleveland (1982) refers to information as diffusive, easily shared, and transportable. He relates Colin Cherry's comment that information by nature cannot give rise to exchange transactions, only to sharing transactions (p. 37). Some of the positive effects associated with ease-of-sharing and transportability are discussed in Section 2.3 (Non-market characteristics).

2.2.1.2 Non-depletability

Information is a very durable good. It does not dissipate with use. It may become less valuable when "better" information is generated but the original information is still present.

Information and knowledge can therefore accumulate over centuries and millennia. From society's standpoint this is a very positive characteristic but from a firm's standpoint this is a negative characteristic. A producer of information must not only compete with other present producers of information but also with all past producers. During periods like the twentieth century this is not an insurmountable problem because of the enormous increase in scientific and technological information, and the significant changes in media technology.^*

In summary, we can define the non-depletability characteristic of information as:
    Non-depletability of Information – Information does not dissipate with use. Producers must compete with past producers but society benefits with an accumulation of knowledge.

In relation to property and property rights, the public-good nature of information makes holding information as property very difficult. Thus the assignment of rights in the form of patents, copyrights, trade marks, and mask works, directly addresses the public-good nature of information.

Appropriability is a serious problem for the producers of information. The 1976 Copyright Act tackled Exclusion 1 (above) by establishing the Copyright Clearance Center to collect royalties for photocopies (McDonald, 1983, p. 19). Also, the Act established the Copyright Royalty Tribunal to collect payments from cable operators for the importation of distant signals. Valenti (1982) notes how low these royalties are in relation to other income. Besen (1978) points out that royalties set by a tribunal are unlikely to resemble real market valuations and will result in social inefficiency.

Solutions to Exclusion 2 are more difficult. The fair-use provision of the copyright law generally assures that exclusion will not occur. (For a most comprehensive discussion of fair-use and market-failure see Gordon [1982]).

Non-depletability can result in dwindling markets for information, particularly if new information cannot be readily generated. It is generally believed that extending the seventeen years of patent protection will not provide the inventor significantly more return on the invention's value. Within seventeen years newer inventions tend to take the older's place, but with notable exceptions such as xerography and the instant camera.^*

2.2.2 Externalities

If there is an effect associated with a good – positive or negative – that is not accounted for in its price, there is an externality present. For example, the polluting of a stream by a pulp and paper mill represents a negative externality. The production negatively affects others and the good is more cheaply produced than if the pollution were controlled. We refer to internalizing the externality if there is a way to make the producer pay for the cost of the externality.

Information, especially education, has large positive externalities. A more educated workforce is more productive, pays higher taxes, commits fewer crimes, and generally adds to the welfare of society. Because the producer of educational information cannot capture all of these benefits, the market will under-produce education relative to the socially optimal level.

2.2.3 Indivisibilities (of supply)

Information must be purchased in lumps; these lumps may be vastly greater than the information actually sought. A textbook contains far more material than any one reader will use. A database may contain millions of records when the user needs only a few. This characteristic, combined with variable relevance, drives the development of technologies that let users extract exactly what they need (search engines, hypertext, etc.).

2.2.4 Economies of Scale and Scope

Decreasing unit costs with increased scale and the ability to produce many information products jointly (e.g., a newsroom producing print, radio, TV) have historically led to natural monopolies in information distribution (telephone, broadcasting). Government regulation or public provision has been the typical response.

2.2.5 Uncertainty and Risk in Production

Creating new information, especially basic research, is inherently risky. Private firms under-invest in high-risk, high-social-return research; government grants and tax credits are common remedies.

2.2.6 Information/Knowledge About the Information

Because of appropriability problems, there is often little reliable information about the quality or existence of information goods, leading to under-consumption (search costs are high).

2.2.7 Intangibility

Many benefits of information (education, culture, inspiration) are difficult to quantify in monetary terms, making market valuation incomplete.

2.2.8 Transaction Costs

Negotiating, monitoring, and enforcing intellectual-property rights are costly. These costs add to indivisibilities and can deter small-scale creation.

2.3 Non-market Related Characteristics of Information

2.3.1 High Intrinsic Relationship to Human Welfare

Human welfare depends on achieving desired outcomes. Information is the primary means of expanding the set of feasible actions, hence it is intrinsically tied to welfare.

2.3.2 High Intrinsic Relationship to Freedom and Privacy

Freedom – More information enlarges the choice set, thereby increasing freedom.
Privacy – Selective withholding of information protects reputation and autonomy.

3. Relationship of Information to Products and Services

Information is embedded in virtually every modern product and service. Software controls machinery, databases drive financial services, and entertainment is pure information. The value-added chain from raw data → knowledge → product is the core of the information economy.

4. Rewards, Incentives, and Funding Sources for the Creation of Intellectual Property

4.1 Rewards, Incentives, and Funding for Science

Government grants, university overhead, prizes, and prestige drive basic research.

4.2 Rewards, Incentives, and Funding for Culture

Patronage, foundations, public broadcasting, and copyright royalties support artistic creation.

4.3 Rewards, Incentives, and Funding for Products and Services

Licensing, venture capital, and market sales fund applied information goods.

5. Structural Role of Intellectual Property in Maintaining a Viable Economy

Intellectual property rights correct market failures (appropriability) while creating temporary monopolies. Balancing access and incentive is the central policy challenge.

6. Summary

This paper has shown that fourteen characteristics of information make it necessary to treat information different from other commodities. In particular, there are many ways in which information will be under-produced without legislation or government support.

The assignment of rights to intellectual property is one mechanism to solve problems of appropriability, transaction costs, and encourage more information about information. However, the assignment of rights exacerbates problems when there are high social returns to the easy and free access to information. Further, problems of positive externalities, indivisibilities, and risk/uncertainty are not addressed by the rights approach.

Alternatives to the assignment of rights include various forms of subsidies and tax incentives from government (and other sources) – for example, the support of science through grants; research through tax credits. When the producer is motivated from mainly non-pecuniary (non-monetary) rewards, the abuse of these systems can be low.

The development of an information product or service is part of an information life-cycle. Information sources become information resources. These are tapped in the creation of products and services, and the results disseminated. Along this cycle, value is added to the original information. The value-added data from the Department of Commerce was found to have limited application. More fruitful are industry level studies of value-added activities.

Finally, little is still known about the political economy of the information sector. What are the employment effects? Are various segments of the industry seriously concentrating? To what extent can and should property rights be extended to those who add value to a basic piece of information?

Appendix A – Bibliography

(contact CITS for a KWOC index)

• Abelson, Philip H. (1982). "Essential Federal Information Services," Science, vol. 216, May, p. 937.
• Abernathy, William J. and Utterback, James M. (1976). "Innovation and the Evolution of Technology in the Firm," Harvard University, June.
• Ackoff, Russell L. and Emery, Fred E. (1972). Choice, Communication and Conflict (On Purposeful Systems). Seaside, CA: Intersystems Publications.
• Adler, Reid G. (1984). "Biotechnology as an Intellectual Property," Science, vol. 224, April 27, pp. 357-363.
• Aines, Andrew A. (1984). "The Grace Commission's View of Federal Research and Development," American Society for Information Science, February, pp. 8-10.
• Aldrich, Robert F., et al. (1981). Issues in Information Policy. Washington, DC: U.S. Department of Commerce, February.
• Allen, Thomas J. (1977). Managing the Flow of Technology. Cambridge, MA: MIT Press.
• Arrow, Kenneth J. (1971). "Economic Welfare and the Allocation of Resources for Invention" in Economics of Information and Knowledge (ed. D. M. Lamberton). Baltimore: Penguin Books.
• Arrow, Kenneth J. (1979). "The Economics of Information" in The Computer Age: A Twenty-Year View (ed. M. L. Dertouzos & J. Moses). Cambridge, MA: MIT Press, pp. 306-317.
• Back, Frank E. (1968). Economics: An Introduction to Analysis and Policy. Englewood Cliffs: Prentice-Hall.
• Baran, Paul A. (1957). The Political Economy of Growth. New York: Monthly Review Press.
• Bator, Francis M. (1958). "The Anatomy of Market Failure," The Quarterly Journal of Economics, vol. 72, August, pp. 351-379.
• Baumgarten, Jon A. (1983). "Will Copyright Survive the New Technologies," American Society for Information Science Bulletin, August, pp. 20-22.
• Baumol, William J. and Ordover, Janusz A. (1975). On the Optimality of Public-Goods Pricing with Exclusion Devices. New York University, November.
• Bell, Daniel (1979). "The Social Framework of the Information Society" in The Computer Age, pp. 163-211.
• Ben-Ner, Avner (1986). "Nonprofit Organizations: Why Do They Exist in Market Economies" in The Economics of Nonprofit Institutions (ed. S. Rose-Ackerman). New York: Oxford University Press, pp. 94-113.
• Besen, Stanley M., et al. (1978). "Copyright Liability for Cable Television," The Journal of Law & Economics, vol. 21, no. 1, April, pp. 67-95.
• Bikson, Tora K., et al. (1984). Scientific and Technical Information Transfer: Issues and Options. Santa Monica: Rand Corporation, March.
• Bobbitt, Philip C. (1974). "Cable Television and Copyright Royalties," The Yale Law Journal, vol. 83, no. 3, January, pp. 554-578.
• Boulding, Kenneth E. (1971). "The Economics of Knowledge and the Knowledge of Economics" in Economics of Information and Knowledge.
• Bowman, Ward S., Jr. (1973). Patent and Antitrust Law: A Legal and Economic Appraisal. Chicago: University of Chicago Press.
• Branscomb, Lewis M. (1983). "Improving R&D Productivity: The Federal Role," Science, vol. 222, October, pp. 133-135.
• Braunstein, Yale M. (1976). "Information as a Commodity," AAAS Meetings [Boston], February.
• Cherry, Colin (1966). On Human Communication. Cambridge, MA: The MIT Press.
• Cleveland, Harlan (1982). "Information as a Resource," The Futurist, December, pp. 34-39.
• Cooper, Michael D. (1983). "The Structure and Future of the Information Economy," Information Processing & Management, vol. 19, no. 1, pp. 9-26.
• Cornish, W. R. (1981). Intellectual Property: Patents, Copyright, Trade Marks and Allied Rights. London: Sweet & Maxwell.
• Dertouzos, Michael L., Moses, Joel, eds. (1979). The Computer Age: A Twenty-Year View. Cambridge, MA: MIT Press.
• Gordon, Wendy J. (1982). "Fair Use as Market Failure: A Structural and Economic Analysis of the Betamax Case and its Predecessors," Columbia Law Review, vol. 82, pp. 1600-1657.
• Hall, Kent (1981). "The Economic Nature of Information," The Information Society, vol. 1, no. 2, pp. 143-166.
• Machlup, Fritz (1962). The Production and Distribution of Knowledge in the United States. Princeton: Princeton University Press.
• Machlup, Fritz (1980). Knowledge: Its Creation, Distribution, and Economic Significance, Volume I. Princeton: Princeton University Press.
• Mackay, Ejan (1982). Economics of Information and Law. Hingham, MA: Kluwer-Nijhoff.
• Nelson, Richard R. (1987). "Roles of Government in a Mixed Economy," Journal of Policy Analysis and Management, vol. 6, no. 4, Summer, pp. 541-557.
• Porat, Marc Uri (1977). The Information Economy: Definition and Measurement. Washington, DC: U.S. Department of Commerce.
• Priest, W. Curtiss (1985). The Character of Information: Characteristics and Properties of Information Related to Issues Concerning Intellectual Property. Lexington, MA: Center for Information Technology and Society, Feb. 10.
• Taylor, Robert S. (1985). Value Added Processes in Information Systems (in press). Norwood, NJ: Ablex Publishing Company.