Next: Operational information policy Up: Information Policy Previous: Introduction
With advances in information technology, the most important aspect of information policy will become the policy for decision-support systems. Increasingly decision support systems will be used in all types of decisions whether personal, economic, or political. The problem to resolve is what information should the decision maker, seated at his terminal, be allowed to access, given the premise that he would be technologically capable of accessing all information. This problem has an operational and a scientific component. In ranking alternatives the decision maker uses criteria to rank alternatives. The operational question is whether he can obtain the information in order to use the criteria to rank the alternatives. The scientific question is whether he can obtain the data to evaluate and improve the criteria.
The starting point to resolve this issue is the question of how well will economic incentives alone achieve the operational and scientific goals. Decision makers demand information to rank new alternatives created by constantly advancing technology. Such information, however, is often costly to acquire and to process. Accordingly, to rank his alternatives each decision maker must consider the costs and benefits of acquiring and processing information. From experience he learns to forecast the benefits and costs of obtaining alternative information and as a bounded rational man he employs rules of thumb to obtain the correct amount of information for a decision. The cost of collecting and processing information have been dramatically reduced by the advances in information and communication technology; hence, most decision makers in informational society will collect and process much more information than prior to the introduction of information technology.
Many decisions have common features, a fact which creates great economies of scale in collecting and processing information. This situation provides an economic incentive for entrepreneurs to create databases to which they can sell access to interested decision makers. Given the scope of the potential markets, entrepreneurs will tend to specialize in particular types of common problems. While the decision makers using rules of thumb may err toward collecting too much information, to the extent that his behavior approximates rational man, he has no interest in collecting extraneous information. The entrepreneur, then, in creating a database has an incentive to collect only information considered valuable by potential clients.
For comparing large numbers of alternatives the amount of information considered valuable depends on the decision technology employed. Using unaided intuition, a decision maker would not want to process more than a few variables on each subject. In evaluating alternatives analytically the demand for information depends on the criteria employed for numerical processing. To analytically search to find the lowest price requires a single variable, whereas to employ an expert system to evaluate loan applications requires numerous variables. As more complex criteria are demonstrated to be more effective in evaluating alternatives than unaided intuition and single attribute evaluations, the demand for information will increase.
Economic incentives will induce database services to focus increasingly on selling value-added services as opposed to raw data. Database owners may have property rights to the information they have created; however, these rights are difficult to enforce, especially with regards to the giving away of information free to friends and associates. Because information can be reproduced and disseminated at very low cost, database services need to provide more than simply raw data to ensure a return. For this reason database entrepreneurs will focus on creating value-added services such as criteria software to evaluate alternatives and, as new decision rules and expert systems are created to analyze alternatives, database managers will offer them as software packages in order to analyze their databases. Such entrepreneurs will make most of their profits from software processing accessed data.
Before databases can promote economic efficiency, database managers must be able to obtain the requisite information. Consider the consequences of granting each person-that is, each legal entity-an absolute right to the release of his own information. Such a policy has a positive bias since persons will provide at their expense information which reflects positively on their interests and would restrict access to that information which reflects negatively. To be sure, persons do have incentives to make a public relations release of negative information which will become public knowledge in order to minimize the adverse publicity. But in general, negative information will be more costly to acquire than positive.
For example, competition in advertising does provide the consumer with some useful information concerning alternative market choices, but market competition will seldom reveal negative information concerning products. A pharmaceutical company whose drug has a negative side effect in 1 out of 100 cases has no incentive to advertise superiority over a drug which has the same negative side effect in 1 out of 10 cases. The negative publicity resulting from such advertisement would probably decrease the demand for both products and might prompt action from the Food and Drug Administration. An example of a public relations release of negative information would be a politician who, knowing he could not keep the fact that he once smoked a joint of marijuana secret, admits this fact as a mistake of his youth.
The objective of operational information policy is to incorporate both positive and negative externalities into decision making. In repeating-type decisions, such as purchasing a product or performing repeating tasks, negative information will be revealed and will become public information over time. Thus, if technology were static over time, both positive and negative externalities would be incorporated into decision making. The more rapid the rate of change in technology, however, the less time negative information has to accumulate before the next change. Consequently, the amount of bias is proportional to the rate of change. Economic incentives alone are insufficient to ensure that decision makers can acquire negative information to rank alternatives.
Similarly, economic incentives alone are unlikely to ensure that decision makers can obtain the information to evaluate and improve criteria. Consider, for instances, an expert system for evaluating loan applications. Ideally an investigator would like to conduct a controlled experiment comparing the performance of alternative expert systems with loan officers of various intelligence, training, and experience. Baring this, the investigator would like to obtain a representative sample of the past performance of the various alternatives. To use past data the investigator would have to predict the consequences of how a loan applicant who was actually denied credit, would perform if he were granted credit under a different evaluation procedure.
Scientific data and operational data have very different characteristics. For operational purposes the decision maker needs to identify the individual and needs the specific information required to apply the criteria. For scientific purposes the investigator needs only a representative sample and has no need to identify the individuals. For many scientific purposes the investigator, however, would like a large number of variables to test alternative hypotheses. As long as individuals fear that scientific information might be used in some operational capacity against their interests, they will have little incentive to provide potentially damaging information. Firms will never have an incentive to provide information which might benefit competitors. Economic incentives alone are unlikely to produce representative data samples.
To achieve the objectives of information policy decision makers and scientists need information rights to obtain information, and the basis for these information rights will be the creation of a right to learn for all legal entities. Specific operational and scientific information rights will be derived from this general right to learn through legal precedent and legislative statute. For operational decisions the right to learn provides the basis for creating an information policy which enables decision makers to analyze alternatives. To promote innovation the right to learn implies an information policy which provides observations on all phenomena. This observation right is the basis for empirical science in the political economy.
Information rights granted under the right to learn must, however, be balanced against other social concerns. If information rights were absolute, there would be no trade secrets, private decisions, or national security concerning the military. Taken absolutely, an observation right would give an individual the right to observe his neighbor's bedroom behavior at will. Obviously, compromises must be made between information rights and other social concerns, such as property rights, national security and personal privacy.