Efficiency versus Privacy

An major category of goal directed behavior are market related tasks. One important aspect of information policy is how does it promote economic efficiency. Given our assumptions the question is how can information policy improve the economic efficiency of bounded rational economic agents in a political economy characterized by a high rate of discovery, invention and innovation.

Almost all market decisions involve the future performance of some object or human. The decision maker must forecast this future performance by collecting information and anticipating the future performance based on past experience. Markets generate information concerning the performance of goods and services. Over time this information is usually broadcasted by word of mouth or by more efficient mechanisms such as media reviews. If technology were fixed such markets would gradually converge towards an efficient market equilibrium of classical microeconomics. If the concern were for steady state market equilibrium in an economy without technological change, there would be little need for information policy.

What makes information policy essential is that with a high rate of discovery, invention and innovation markets very rarely reach equilibrium, rather they are constantly adjusting to new technology. For this reason we eschew the traditional concerns for equilibrium conditions and focus our attention on transient performance. We have modeled our political economic agents as bounded rational beings who make mistakes. We wish to consider information policy in terms of reducing the number of mistakes or regret made by our agents. So information policy should be formulated to reduce the number of mistakes. This would mean that the convergence towards a traditional market would be rapid although some change would occur before achieving such an equilibrium.

Even if we provided each decision maker with all possible data to forecast future behavior, the decision maker could not process it. We need to consider the minimum amount of information which is needed. The more experience the decision maker has the greater he can anticipate or intuitively forecast future behavior. The greater the change between the current and the last decision the more he or she has to investigate the current attributes. Much information is required by the decision maker to confirm his or her anticipations. Information which confirms anticipations is not going to change the decision. What is going to change the decision is information which reveals something unexpected either good or bad. In cognitive experiments it has been found that the order data is presented can change a decision. Information which when presented in any order and changes the decision we will call information which makes a difference. Decision makers should be entitled to information which makes a difference. This will lead to the fewest mistakes.

The number of mistakes also depends on the amount of knowledge possessed by the decision-maker. For data to be considered information, the decision-maker has to have the knowledge to interpret it. There are three aspects of knowledge which would affect the decision. One is the knowledge possessed by the individual decision maker. A second is the knowledge possessed by society. The third are the discoveries made as new technology is used in society.

The problem with this criterion is that much of the information which makes a difference is in the possession of others. In some cases when it is in their interests individuals will freely give up information. Information policy concerns the cases when individuals should give up information which may not be in their interests. This raises the question to what extent should informational efficiency be sacrificed for the sake of privacy.

In this book we shall use privacy as a broad term covering individuals, firms and government. Privacy of the firm is usually covered under the expression `proprietary rights' and privacy of government is usually considered as `secrecy.’ To define privacy we shall partition it into five attributes: access, observe, disseminate, verify, and use.

One dictionary definition of access is ``Permission to approach, enter or use.'' Since the topic of the book is information policy for an informational society, we need to define access relative to various aspects of the social nervous system as well as the traditional forms of access to a home, firm or government facility. Thus access would cover permission to read a datafile as well as permission to enter an office.

Under the attribute observe we include both to collect scientific data and to collect data in a file for an administrative purpose, such as a sales transaction, to be used later for analysis, such as determining which customers are most likely to respond to a promotion of a new product.

Disseminate covers under what conditions a political economic agent can transmit data to another economic agent. This covers traditional media dissemination of information in the news and newer electronic forms of dissemination such as the electronic sales of data as well as electronic bulletin boards and E-mail transmissions.

A basic issue which relates to all of the growing information provider industry is who should be responsible to verify that the data collected is accurate.

The issues of the use of information will pertain to information collected by an agent or purchased by an agent. For example, after seven years personal bankruptcy information must be removed from credit files so that it can not be used for determining credit.

It is obvious to most readers that a certain level of privacy is in the public interests. Let us consider the various attributes of privacy.

If their were no restrictions on access, then individuals whether at home or in a private or public office could be constantly bombarded with requests for access their performance would seriously deteriorate. There are numerous business situations in which observations would result in a reduction in economic performance. Currently internet firms are working hard to provide secure communications to facilitate commercial transactions. Without secure communications credit card numbers and other vital information will be stolen by hackers. Business negotiations require no observations for effectiveness. The same is true of many government negotiations. The military needs secrecy to maintain a tactical advantage over potential enemies. Personal intimacy would be threatened if such behavior could be observed at will. The uncontrolled dissemination of information from a firm would result in a loss in trade secrets and would provide rivals a possible advantage. The uncontrolled dissemination of military secrets would aid the enemy.

In the interests of racial, sexual and religious equality, such information is not supposed to affect employment decisions. After a time period personal bankruptcy is supposed to be removed from credit files. In the future an increasingly controversial subject will be the use and nonuse of genetic information.

The issue on who has the responsibility for verifying the accuracy of datafiles should fall under a Hand type rule asking the question is it cheaper for society to have the collector be responsible for accuracy or is it cheaper for the subject.

Obviously there is a conflict between the concept that decision-makers should have all information which makes a difference and the various attributes of privacy. What has totally upset the social balance between these two concepts is the advance in information technology which we will study in the next chapter.

 

*********

As was pointed out previously with a high rate of discovery, invention, and innovation decision-makers and analysts are constantly facing new situations in which previous experience may be of little value. For example, a consumer buying a durable, such as a new refrigerator will need considerable information to evaluate the advances and alternatives in new technology. Similarly, a firm will need much information to upgrade information technology because efficient upgrading usually requires a complete reorganization of the flow of work through the office. With an increasingly rapid introduction of new products, firms need information to rapidly locate potential customers. The greater the change the more they must be able to analyze the alternatives, which has increased the demand for information.

What has vastly increased the information supply and demand to analyze change has been the advance in information technology. In this section we shall assume that no restrictions are placed on the demand and supply of information other than market considerations of value and cost. It is important to gain this perspective prior to considering the conflict with privacy. We shall consider the current state of the supply and demand for information in markets and forecast future developments based on projected advances in information technology. Again it is important to emphasis we are forecasting the developments if there were no restraints on market activity in regard to data collection and processing.

In the precomputer age all operations dealing with paper media were labor intensive. Observations had to be recorded by hand on forms. These forms had to be filed by clerks. If a file was needed for analysis, a clerk had to retrieve it manually for the storage facility. To determine a subset of files had a particular set of characteristics, a clerk would have to laboriously search through all the files page by page. Analysis of the files would usually require a transfer of data from the files to another paper form. The amount of data collection was limited by the cost and the amount of analysis by lack of tools to aid decision makers.

The advance in information technology has lead to the creation of enormous databases called data warehouses or datamarts with terabytes of data created for the purpose of analysis [Ref: Oct 95 Byte]. A terabyte could store about two million books which would take humans millions of manhours to pursue. In order to create a datamart, institutions, usually firms, first decide what data and format to store the data. Data must be corrected. For example, a customer file might have my name stored as A. Norman, A. L. Norman, Alfred Norman, Fred Norman, Alfred L. Norman or Alfred Lorn Norman. All entries must be corrected to the same form. To ease the burden on production computing the datamart is usually stored on separate disk storage. To process such large data files institutions are turning to parallel processors. With the fall in storage costs and the increased power of computers the size of these datamarts should grow rapidly.

What type of data is stored in datamarts depends on the application. GTE stores all data relating to the health of its employees in a datamart. Marketing organization develop datamarts storing all available information on potential customers. Mutual companies create datamarts to store all available information on financial assets. Astronomers have a datamart to store information on the cosmos. All information on global weather is stored in such a database. Genetic decoding of DNA is stored in data warehouses. The information on technical processes such as a plant can be stored in such an arrangement.

What makes these datamarts useful have been the creation of tools to analyze massive datasets. For example, it is relatively simple for a program to search through a large life and pick out the subset which has a particular set of characteristics such as all households with children who own their own home and have a combined income of greater than $50,000. Search and query programs have been improved for such operations. Semi-intelligent software such as knowledge-based system provides tools for analysis such as determining how much credit such a firm should grant a customer. Statistical packages enable researchers to find correlations and test hypotheses, and the new neural network programs allow analysts to search for patterns in the data. For example, the financial assets in Fidelity's Asset Manager funds are picked by a neural network program. While such software may not have the capacity to think it embodies the knowledge of the creators. To use such software the user needs to know how to apply it and interpret the results. Thus his or her knowledge is generally much less than the creator of the software. Software allows bounded rational individuals to turn vast databases into information.

Without the software much data would remain meaningless, either because the decision-maker could not interpret it or does not have the processing capacity. Nevertheless, it is a grave mistake to assume the information technology provides bounded rational individuals with anything near the assumed computing power of the mythical rational decision maker, which is infinite. Convention computers solve tactable problems which are no more that polynomial in the growth factors of the problem. Even if physicists and engineers develop a quantum computer, which should be much more powerful than current computers, it will still be finite.

The rate of return in creating datamarts can be very high. For example, one firm which invested $20M paid for the initial investment. However, there are numerous failures in creating datamarts. Frequently, the datamart has to be redesigned because the data quality is poor or the organization is not conducive to analysis.

One of the fastest growing type of datamart is sales' datamarts containing information concerning customers. This growth can be attributed to changes in production. To be more competitive in applying rapidly advancing technology firms have reorganized to reduce the product cycle, which results in a stream of new or improved products constantly entering the marketplace. Increasing automation has allowed firms much more flexibility in production and has decreased the production cost of small production runs in comparison to previous mass production with fixed production technology. Consequently firms are increasingly trying to promote new and improved products in niche markets.

To promote products in niche markets firms need more efficient marketing procedures than trying to promote products in mass markets. Thus, firms are beginning to collect voluminous socioeconomic files in order to direct their product promotions towards customers most likely to respond. Some of the information is obtained from production registration forms. Some of the information is obtained by third parties from public files such as wedding announcements and purchase of homes. As business transactions shift to electromagnetic media, it becomes less and less expensive to maintain a file of all previous customers purchases. This information facilitates forecasting what type of products the customer might want in the future. If the market were unrestrained, entrepreneurs would create giant databases with complete transactions histories of every household and firm.

In terms of mass mailings this development represents a tremendous increase in efficiency. Rather than send mass mailings to every household, firms can target them to households most likely to respond based on analysis of socioeconomic information and past purchases. As television becomes interactive, this means that advertisers can send individual ads into each household. Advertisers will want to send ads into only those households which are most likely to respond. Political ads will also be individualized so that politicians can tell each household what they want to hear.

This advance will represent a significant increase in efficiency. Many, if not most, households would actually like the change because they would not be bombarded with ads to which they are unlikely to respond. For example, manufacturers of paper diapers would only send their ads to households with babies and toddlers.

Advances in information technology and other technology has also changed the ability of firms to monitor their employees. Market considerations place limits on monitoring. For example, productive, self-directed employees who could easily find another job would quickly leave a firm which created an atmosphere of big brother is constantly watching you. Because E-mail, unlike paper mail, has no privacy, employers can read employee messages at will. Second, advances in surveillance technology enable employers to constantly monitor their employees. Third, employers can monitor employees behavior while not at work by tests, for example drug tests. How much monitoring is in the long run interests of a firm various greatly depending of the type of work and employees.

For decisions made by a large number of firms, such as credit, there are tremendous economies of scale in a third party collecting the information and selling it to interested parties. As knowledge accumulates these third party firms will increasingly sell value added services such as analysis programs rather than raw data.

Let us look at information for decisions made by firms. One highly developed third party collectors of information industry is individual credit rating industry composed of such firms as TRW, Equifax and Union. As the techniques of electronic interchange of document expand, these firms will receive their reports automatically concerning how promptly their subjects pay their bills. As all types of communications merge into a broadband social nervous system clients of this industry will be able to obtain their credit reports more quickly.

Another collector of information is the consortium which collects medical information for medical insurance companies. The amount of information which this consortium collects could dramatically increase as medical files on all patients are maintained in electromagnetic form. Second with the increased knowledge concerning the human genome leads to inexpensive tests to determine propensities towards various diseases, such information becomes very important to determine insurance rates based on market risk.

In an unrestricted market for information, firms which self-insure their workers' medical costs would also have a powerful incentive to form a consortium to collect voluminous medical files on both the household members of current workers and the household members of potential employees. This would be efficient because firms could greatly reduce their medical costs by not hiring workers who were medical risks. As DNA research isolates the genetic basis for various types of diseases and defects, firms will have very intense incentives to collect this information also. The firm would also have incentives to fire workers determined to be medical risks; however the need to maintain a productive work force with high morale places some restraints on firing workers for medical reasons.

Now let us consider the supply and demand for information for buyers whether households buying goods and services or firms buying raw materials or intermediate products. Sellers to promote sales have great incentives to provide potential buyers with information which portrays their product in a favorable light compared with rival products. Firms which buy in large quantities on a regular basis have economic incentives to test such claims and determine the quality for themselves. For example, computer makers test memory chips from suppliers. Households, on the other hand, do not buy in such large quantities to create incentives for systematic testing. Rather by trial and error they test frequently purchased items such as foodstuffs. For durables which consumers buy at infrequent intervals, consumers have little opportunity to test which the possible exception of test driving an automobile.

For household purchases there are large incentives for third parties to test products and sell the test information. Computer magazines test hardware and software and software. Many magazines test automobiles. Some magazines specialize in travel opportunities. The Consumer Union tests a wide variety of household purchases and reports the results in Consumer Reports. Nevertheless, the Consumer Union only tests a small fraction of all the products on the marketplace.

As the social nervous system develops third party testers of consumer goods and services will gravitate to the social nervous system. Already consumer reports are available on America on Line. Some computer magazines are available on the Net. As household markets develop in the social nervous system shoppers will easily be able to access evaluation reports while considering there alternatives. For example, today on the Internet Shopping Network consumers can access software and hardware evaluation reports while considering their alternatives.

As household shopping on interactive networks expands, organizes of catalogues will have strong incentives to provide their potential customers with options to see videos explaining how the product works and access to all positive evaluation reports. Interactive networks would also greatly assist consumer services such as Consumer Union gather data. Currently Consumer Union sends its clients each year a survey so that each household can provide information on their durables such as automobiles. The problem is this is a costly labor process on the part of both Consumer Union and their clients. As information technology advances consumers will eventually obtain their repair records in digital form. Such records could be transferred from the shop to the customers home computer automatically. Clients of Consumer Union or any other evaluation service could automatically transfer these repair reports on to their evaluation service. Surveys which appeared in E-mail instead of paper would be much cheaper for the evaluation service and much less trouble for the client to fill out. Thus as the social nervous system advances household product evaluation services would find the cost of collecting and analyzing market generated information greatly lowered. Entrepreneurs would create numerous new evaluation services for both firms and households.

A very large component of the increased demand for information on individuals is a result of bounded rational decision makers utilizing applications of advances in information technology to become more efficient. As information technology continues to advance, this striving for greater efficiency will transform all markets.

The reduction in the cost of acquiring, storing, retrieving, and analyzing information increases market efficiency by enabling bounded rational decision makers to rapidly screen large numbers of alternatives and to analyse decisions in much greater detail. Thus the advances in information technology have fueled a tremendous increase in the demand for data which software tools convert into information. Now we need to consider how the increased demand and supply of data to convert to information has exacerbated the conflict between efficiency and privacy.

**************

In this section we shall discuss many conflicts between privacy and efficiency which we will resolve in the subsequent chapters. Evolving information technology has undermined current resolutions between privacy and efficiency. In some cases privacy is needed to make markets work. In other cases the new information technology offers the possibility of new improved resolutions between privacy and efficiency. In this section we shall explore the range of possibilities to be resolved in subsequent sections.

Before deciding what current problems will require information policy or a resolution by government we need to forecast the possibility of a market solution by new technology evolving to resolve the problem. For example, in an unrestricted market, advances in cryptography can provide secure communications for users of the Net. But, because the police organizations want to be able to access encrypted communications with warrants, encryption still creates an important problem in information policy.

Certainly physical access in granting permission enter a physical space such as home, firm, or government installation would not change much from current society which requires permission of the owner. As electronic technology creates new ways to identify individuals, the electronic means of granting individuals access to physical spaces will become more sophisticated. Increasingly physical access will be controlled by software rather than by individuals. For example, there are numerous ways to identify humans using fingerprints, hand measurements. Access will be determined automatically by those on file. Permission by a human operator would be necessary for those not on file. Advances in electronics are unlikely to change the current resolutions in the right to physical access to physical space. Technological advances will offer cheaper alternatives for screening people.

Similarly information technology will offer increasingly sophisticated alternatives for screening mail and phone calls whether voice of video. Currently paper mail must be screened by hand. As mail shifts to E-mail it will become progressively easier to screen E-mail by software. With the advent of caller identification it will become progressively easier to screen phone calls. Given the volume of junk mail and phone calls, which can be generated by computers, the ability to deny access for mail or phone calls using software is not likely to become a controversial issue

Another important aspect of access in informational society is access to files stored in computer networks. In most computer networks the creator has some control over who has access to files, although the network administrator can override the access codes. The policy for access to files created by members of an organization is controlled by that organizations access policy.

The real issue concerning access policy is under what conditions should the subject whether an individual, private or public organization should have control over access to files maintained by other individuals or organizations. For example, as genetic testing becomes more prevalent, an individual would have increasing concern over who has access to his or her medical file. Again most individuals would like to have some control over who has access to their credit information files. Should individuals have the right to restrict access to their credit cards files to those to whom they wish to apply for credit. As the volume of information collected on each individual explodes, the need for some control over access by the individual increases. The economic issue is can individuals be granted more control over access without seriously decreasing economic efficiency.

Now let us consider access for the perspective of under what conditions should the state of the individual be able to control access to various types of materials through the SNS. The issue which is currently controversial and is likely to become even more controversial is the issue of pornography on the internet. What makes this an interesting issue is the fact that the Supreme Court has decided that the definition of pornography is a community issue. On this basis a Tennessee community is prosecuting a California firm for sending X rated material over the internet to the Tennessee community. One might assume that this issue can only be resolved at the national level given the nature of the internet. We will argue that advancing technology will provide a large number of possible market solutions. To the extent that market solutions are possible the need to have a purely legal resolution of the problem is lessened.

Let us consider this an access problem and consider the issue from the level of the household. To what extent will advances in the social nervous system enable households to control, that is deny access, to unwanted intrusions whether they be in the nature of pornography, violence, or even advertisements. Currently a household only has control over the media in the sense that encountering an ad the viewer does not like the viewer can change channels, go to the kitchen or bathroom, or mute it out. The same is true for programs which from the value of the household constrain excessive violence or sex. Currently, in some cases parents have passwords to restrict access of their children to certain channels.

With the convergence of communication into a single interactive broadband system, the surviving media/communication companies will be able employ even more sophisticated billing systems than the phone companies currently employ. This means that media companies will be able to individualize their program offerings to the desires of the household. For example in regards to sexual content households will be able to order programming with a G, PG, PG-13, R, or NC17 rating. Furthermore, parents using passwords will be able control the access of their children. A similar rating structure is likely to be constructed for violence which the viewer will be able to control.

Moreover with an interactive social nervous systems viewers will watch their programming on demand rather than be programmed by the networks. Because the networks will have sophisticated billing systems so that they will be able to bill customers by the cost of individual programs. Current network programming is free to the viewers and the networks make their money by selling ads to companies. Network providers will be able to offer the viewers a choice between pay for view without ads or free with ads. The viewer could allow the network to select the ads the view might want to see or the viewer could select the types of products of interest. The network under these circumstances would be able to change more for the ads shown because the viewer is much more likely to be interested than under the current system.

The problem of differential standards of pornography and violence on the internet could be resolved by similar measures. If standards are established for labeling multimedia materials, then software at the viewers station can control whether they can be accessed. Parents could control the access of their children through passwords controlling the level of sex in what their children can access. One legal question about community standards of pornography is whether the community can control what takes place in each residents home. This controversial issue will be considered in subsequent chapters because advancing technology will offer a wide range of combinations of market-legal alternatives.

For individuals, markets or governments to functions there is a conflict between to observe or not to observe.

Advances in information technology will make telecommuting and teleconferencing ever more effective. With the capacity for multiple video links in an organization, telecommuters will not be isolated in the organization because these links will define the organization. From a social perspective promoting teleconferencing and telecommuting is in the public interests to reduce pollution. Yet for these two to flourish communication must have security.

With vast amounts of personal and corporate information both in databases and being transmitted through communication systems there is a pressing need for security in computer-communication systems. However, the more secure you make a computer-communication system from hackers the harder access becomes for legitimate users. Computer-communication systems can never be made completely secure against unauthorized users. Military style encryptography, the translation of computer-communication information into code, is a current trend in computing and communication. Is all information being used by computers or communicated between people or machines going to be encrypted? Currently there is a major battle between police organizations and business concerning cryptology. The National Security Administration and the FBI want to be able to tap into any communications (with a warrant, of course) for the purpose of catching criminals. They propose a cryptology system developed by the NSA. Computer experts are opposed because they believe the NSA has a backdoor to easily listen to any communication message. Also, organized crime would have the resources to break such a system and steal bank transfers. Businesses generally want unlimited encryption to protect vital corporate secrets. Recently the Clinton administration rejected the NSA encryption scheme.

In order to make decisions, individuals, households, firms or governments must acquire information to analyze the alternatives either through direct observations or indirectly through previous observations collected in files. Since obtaining information is costly, economic incentives limit the amount of information decision-makers want to acquire. In addition, voters need information from direct or indirect observations to hold government officials accountable for their actions. Firms need information to judge the performance of their workers. Finally social scientists and the allied business disciplines need to make observations to understand the various aspects of human behavior.

In contrast the need for no observations is equally as important for many aspects of human behavior. For example, most individuals need freedom from the possibility of observations to express their innermost thoughts of behavior. Intimate behavior between couples is severely restrained by the prospect of third party observations. Much economic behavior requires no observations by outsiders for proper functioning. If the interactions of groups within a firm were to be observed by rivals the firm would be exposed to the possibility of loss of trade secrets and any tactical advantage in formulating its plans. Many delicate market negotiations such as mergers and takeovers would be much more difficult of observed by rivals. Market transactions require sensitive information such as credit card numbers not be observed by third parties. Government in pursuing military policy needs no outside observations to gain a tactical advantage. Likewise policy need secrecy in investigating criminals so that they will not change their behavior.

The advances in information technology are upsetting the balance between observe and no observe based on paper. One example is the consequence of markets moving into the internet. The growth of markets on the internet is currently stymied because of a lack of security. Computer knowledgeable thieves can obtain credit card numbers and other vital information being transmitted through the internet. Currently various internet firms are creating standards for secure encrypted communications. This compromise should be completed in 1995. As a consequence buyers, who currently buy from catalogues, are likely to shift to internet market as the bandwidth increases and the multimedia catalogues on the internet become more sophisticated.

The advance of electronics is also advancing spy technology. With the end of the cold war, spy agencies, such as the CIA, are shifting their energies to terrorists groups and regrettably spy on commercial ventures in rival economic powers. With such technology criminals can spy on the investigations by government into their activities. In addition, firms are increasing their industrial espionage to obtain trade secrets from their rivals. Also, stalkers can use such spy technology to pry into the personal lives of their victims. The resolution of this problem of outsiders being able to use spy technology to observe behavior of government, firms or individuals which previously was much more difficult to observe can not be stopped by a prohibition of the public sale of spy technology. While such action will slow down the legal sale it will create a prosperous black market.

The advances in surveillance technology will necessarily require increased security by government agencies, firms and individuals likely to be spied upon by outsiders. The incentives will create a escalating race between spy technology and preventions. As organizations use WAN to create work groups and have critical communications worldwide, they will need better cryptography to encode message. Many conversations among individuals will also require cryptography. There will be a continual conflict between the need for better security by firms and individuals and the police's desire to be able the tap the phones of suspected criminals.

Advances in information technology have created a major problem in the workplace in how employers monitor the actions of their employees. Consider first E-mail which unlike paper mail does not have any criminal sanctions against being read by third parties without permission of the sender. The protection of E-mail would best not be served by criminal sanctions but rather by promoting encryption. The only problem which this might not solve is the fact that in most systems the system administrator has access to all files and programs so that employees would still not be guaranteed privacy of their E-mail. As digital assistants become inexpensive workers would have to download their E-mail to their personal digital assistants to ensure privacy. This solution is likely to gradually evolve.

With better surveillance technology employers could keep their employees under constant surveillance. Employers wishing to maintain worker morale would need to establish clear procedures will such methods were going to be used. For example, in firms where workers were mobile, managers would be reluctant to engage in surveillance activities which might encourage their workers to seek employment elsewhere. The issue is whether market incentives are strong enough or whether workers need specific rights in the workplace.

For example, prior to the advances in information technology transactions histories were used by salesman in some cases: however, they were almost never used for systematic analysis of customers previous purchases because the collection and processing cost were too high. As the costs of collecting, organizing and analyzing large files have constantly been dropping, marketing analysts have increased their analysis of customer previous purchases in order to focus their marketing strategies on the customers most likely to respond. As the advances in information technology make the collection, processing and analysis of any routinely collected data easier and easier, the amount of analysis of such data grows accordingly. For example, as medical records become routinely digitized, they can easily be analyzed in order to provide the best care, administer insurance, or research into the effectiveness of treatments. With advancing information technology any recorded data becomes an observation which can be used in any type of decision or research task.

In some instances individuals can maintain anonymity from having their identities revealed in market transactions. For example, customers can pay with cash. Advances in smart cards could provide customers with a cash card with no individual identity. Even though some police organizations object to these forms of money because they would facilitate drug dealers laundering money, they are likely to become commonplace. On the internet individuals will be able to pay with forms of money which do not reveal the identity of the buyer. Nevertheless, anonymity in market transactions is limited because in many cases the product must be delivered to an address. Also, the individual usually must reveal his or her identity for products with warranties.

The resolution of observe or no observe will be broken in two parts. The first is operational information policy which defines what information should be observable to make decisions whether public or private. The second is scientific information policy which defines what information should be observable to advance science. The reason for breaking this into two components is the two aspects of information policy are very different. The dissemination of information in the media has been established by the common law reaction to yellow journalism of the 1890s. The media can not release non newsworthy facts about private individuals. This restriction could be applied to individuals making mass e-mail broadcasts on the internet.

The new problem which arises from the advance of information technology is the creation of giant data files containing information about individuals, firms, or government agencies. The evolution of such files is likely to follow the pattern of the creation of credit histories of individuals and medical files of individuals. Moreover, as the social nervous system grows the transfer of data will take place automatically. To create ads for individual households for interactive television, marketers are going to want extensive data files containing transactions histories and socioeconomic data on each household. There are economies of scale for third parties to provide this data. Much like the current credit history industry a new industry will arise to provide the data. As informational society advances there will be a whole series of information industries providing specialized data to clients.

There are two aspects to dissemination in these information industries. First the industry collects data from public and private sources. Second the industry sells data to interested parties. In considering whether some limits should be placed on either the collection of data or its release to clients we need to consider the limits created by market incentives.

Firms now collect and analyse much more data than previously because the collection of data is a byproduct of administrative tasks such as sales paperwork. To complete their datafiles on their customers a firm might purchase data from various information provider firms. Firms do not have unilateral incentives to sell data about their customers to information provider firms. First a firm might find it has a competitive advantage in telling its customers that all their purchases and other information presented on warrantee forms will be held in strictest confidence. Second if a firm has acquired an excellent database upon which to make marketing decisions it simply may not wish to sell this data to a third party because it could decrease its competitive edge over its rivals. Market incentives do not create unilateral incentives for the sale of information to information provider firms.

The fact the economic incentives place some limits on the sale of data by firms raises the question of whether laws are needed to create specific prohibitions on the sale of particular types of information. One example might be the results of genetic tests. Currently there is a restriction on the federal government creating a national database.

The creation of various information markets also creates a problem with the care with which the information is handled. For example, medical data is very sensitive. If carelessly handled, leaks could have tragic consequences to some individuals. The issue here is who has the responsibility to make sure that information can not easily be obtained by unauthorized individuals. For example, do the medical information providers use encryption and pass words to eliminate leaks in hospitals. This subject is closely linked to who should have access to information about subjects.

These problems indicate that in addition to market incentives restricting the flow of data their may need to be some laws protecting individuals.

The final problem which will grow in proportions as information advances is the accuracy of information objects especially images. In the next chapter we shall propose the advance of markets in computer networks. Household markets will not expand until the consumer can obtain images of the products, the consumer is evaluating. As it becomes inexpensive to manipulate the images presented to the consumer. The problem is much more complex than a black and white issue of fraud because manipulation can enhance the quality of the image and under what conditions should subtle changes be admissible.

Data on individuals when inaccurate can have negative consequences such as the loss of credit or job opportunities. Are market incentives for accuracy by the clients of information providers sufficient to achieve accuracy or do individuals need common law protections against inaccuracies.

As the information provider industry grows the treatment of inaccuracies obtained by the credit history industry under the Fair Credit Reporting Act requires careful analysis. Is this a model for other information provider industries or is it a major mistake in policy. This law absolves the credit industry from common law responsibility for the accuracy of credit reports on individuals and supersedes all state law. Without the protection of this act individuals who were denied credit because of incorrect information in their credit files would be able to sue the credit card companies for damages. The act place the responsibility for accuracy of files on market incentives for accuracy and the subject who must review his or her own file and take the credit companies to court if they refuse a subjects request to correct a mistake.

For markets which move to the social nervous system to work properly they require accurate information. A trend in hiring college graduates is to have an information company verify the information on their resumes. Consider two labor markets on the social nervous system. The first allows all interested parties to list their resumes without verification of any information and the second assesses the parties to be listed the cost of verification. From the employers perspective the second market would be greatly preferred to the first.

Verification of the data in information providers files is important both for the proper operation of markets and the protection of individuals from the consequences of false information. Whether the market's need for accurate data is sufficient incentive to be in the public interest is an open question.

While the amount of information used to make decisions has greatly increased due to the advances of information technology over what was previously used with paper as a medium the amount is by no means infinite. Under market conditions decision-makers would only demand useful information. The issue is whether the needs of information for an efficient market are socially desirable.

Consider medical insurance. As scientist discover more and more about the human genome this information can be used to predict the likelihood of individuals acquiring various diseases. Under an efficient market each individual would pay medical insurance based on his or her previous medical history and his or her genetic makeup. While many individuals would agree with the logic of an efficient market many others would believe that genetic information should not be used in determining insurance rates. Such an information policy constitutes a pooling of risk. If such pooling is considered politically desirable then prohibitions would have to be made concerning access, observe, dissemination and use.

Medical information and insurance rates is one of the more controversial conflicts between market efficiency and other social concerns such as equity. The question which needs to be asked is what is the tradeoff between efficiency and equity. Also, if restrictions are to be imposed what is the most effective method of imposing them.

Where information is gathered for the purpose of harassment or political intimidation, the need for restrictions is much clearer. Here the primary issue is effectiveness of the proposed restrictions.