The Committee on Indicators of Internet Impacts on Development defined its role in this study as a steering panel that would draw on the expertise of other researchers and analysts working in the areas of development, information technologies, and regional affairs. The committee is grateful for the assistance and advice it received throughout the course of the study from a number of individuals and organizations, many of which are cited in the text. In particular, the committee members would like to thank the Network Startup Resource Center, the Academy for Education Development, and the American Association for the Advancement of Science. The committee members would also like to thank the Bellanet Initiative, which hosted an on-line discussion of Internet indicators after the committee presented its initial findings at the Global Knowledge í97 conference.

  During their field research in Ghana, Kenya, and Senegal, the committee members received invaluable assistance from the U.S. Agency for International Development missions and Leland Initiative coordinators. Lane Smith (Leland Initiative), who accompanied the committee in Ghana and Senegal, deserves special thanks. Thanks are given also to the many individuals in Ghana, Kenya, and Senegal who took time out of their busy schedules to discuss the Internet during the committeeís visits.

  The committee gives special thanks to two individuals who contributed greatly to the research and writing of this report. Kelvin Wong (Center for International Development and Conflict Management, University of Maryland) provided much of the extensive background material for the team and did the data collection and analysis in Kenya. Jon Eisenbergís (National Research Council and AAAS Fellow) help as a member of the team was invaluable in adding a disciplined and expert technical perspective on all the committeeís work.

  Finally, the committee would like to thank the staff of the National Research Council. They especially wish to thank Inta Brikovskis for her tremendous energy and the intellectual contribution she made to drafting the report.

It is easy to be excited about the opportunities offered to developing countries by the Internet and its associated technologies, but it also is necessary to better understand the kinds of impacts one desires and those already observed. This report begins to provide a framework to analyze and better understand the impacts of this technology on Internet service providers and users and on various sectors of society.

Developing countries are rapidly acquiring information technologies, and use is growing in many sectors. Some organizations already consider the Internet a necessity rather than a luxury, and on its site visits to Africa in August 1997 the Committee on Indicators of Internet Impacts on Development heard from business leaders in Ghana and Senegal who know that to compete in international markets they must have Internet access. Many obstacles exist, however, to widespread Internet use in Africa, including economic barriers, poor infrastructure, restrictive policy and regulatory environments, and lack of effective leadership and constituencies.

With use of the Internet increasing at a rapid rate, there is an expressed need among development agencies, policymakers, Internet providers and users, and others for some framework to evaluate impacts of the Internet. At the same time, the committee found a number of challenges to gaining the information and knowledge it sought in Ghana and Senegal, where the Internet is still a new phenomenon.

The impacts of Internet access and use depend on who communicates what to whom. Some impacts of introduction of the Internet are direct and measurable; others are only indirectly related to the Internet and less easily measured. Since penetration of the Internet in Africa is still very limited in most areas, this report proposes an approach that seeks to infer the indirect impacts of the Internet from extrapolation of the direct impacts on users and organizations. However, readers must be aware of the perils in such extrapolations. The indicators presented here can be helpful in analyzing impacts of the Internet, but the indicators themselves do not tell the whole story of Internet use, diffusion, and impacts. To be useful in Internet analysis, the information collected must be used with other instruments that simultaneously collect data on gender, geographic location, and other indicators to allow analytical disaggregation of impacts.

To understand "Internet supply," one must consider the political and social environments in which suppliers and users operate as well as technical and organizational aspects of the suppliers themselves. Thus, indicators are presented here to measure the competitiveness of the market and provide clues to the sustainability of Internet development and growth in a country. Also, indicators are presented to measure the quality and quantity of Internet service.

"Internet use" refers to the quantity and type of use. Indicators are presented to analyze the number of users and subscribers and the reasons for which they are using the Internet. For example, indicators are presented to analyze Internet use by different segments of the population, whether the Internet is being used primarily for electronic mail or more sophisticated uses, and whether use is increasing or decreasing. Indicators are presented to analyze the types of impacts on institutions. The indicators measure whoñfor example, what level of employeeñwithin an organization is using the Internet; what types of functions it is being used for; and what the perceived benefits are. Indicators are also presented for impacts of the Internet on organizational decisionmaking and in institutions such as markets and communities.

The Internet can have impacts on sectors and on the developmental goals central to those sectors. Indicators are presented for impacts on education and on the broader developmental goals associated with educationñfor example, increased literacy and employment. Indicators are also presented for impacts on the private sector and broader economic developmental goals, such as increased employment, production, and capital formation. And the government sector and civil society are discussed. Indicators are presented to measure Internet use by the government and nongovernmental organizations as well as the impacts on related developmental goals such as increased citizen participation in government and increased flow of information. Finally, Internet "diffusion" or paths of impacts are discussed and relevant questions are posed.

This report is intended to stimulate discussion on the impacts of the Internet. It is recognized that much study is yet to be done. Based on the committeeís findings, some areas for further research should include issues of causality, country-specific case studies of Internet use and diffusion, impacts of nonuse of the Internet, institutional use and diffusion of the Internet, and public policy and Internet use.

The Internet is touted as a revolutionary technology that will change many aspects of our lives. What began as a small computer network for the Defense Advanced Research Projects Agency, the research arm of the U.S. Department of Defense, has expanded to a global infrastructure. It has gone from service to research and academic communities to serve a broad spectrum of industrial, commercial, and governmental users, bringing together people from different cities and countries and from some of the most remote areas of the earth. Services over the Internet have also expanded from e-mail to the World Wide Web, and Internet software is increasingly being used for communications within organizations (Intranet).

In the developed world an ever-increasing share of research, communication, business transactions, and entertainment takes place via the Internet. For the developing world the Internet is considered by many to be a powerful tool for development that will allow countries to leap-frog ahead, economically and socially. Supporters point to the small farming cooperative in an agricultural country, using information technologies to learn about new farming methods and obtain reports on product supply and demand, thereby increasing local exports severalfold and enabling entire villages to prosper; or to the doctor in a rural village, diagnosing and treating a patient with the help of an Internet connection. Supporters would likely agree with the basic premise that economic growth of nations and quality of life of people are dependent on communications. As nations progress, so too will economic growth and quality of life be dependent on the broad use of telephones and in time on electronic mail. Therefore, it is assumed that for Africa to be competitive, the broad development and use of the Internet will evolve.

It is easy to be excited about the opportunities offered to developing countries by the Internet and its associated technologies. But hard decisions remain, particularly for policymakers and others decisionmakers faced with many competing priorities. Should expenditures for local-area networks and wide-area networks outweigh investments in roads and rails? In the money-strapped health care sector, where tough tradeoffs must be made, should clinics buy Pentiums or penicillin? Still other projections point to deleterious impacts the Internet is having, and will continue to have, on developing countries. Some see the Internet as a route by which Western values compete and come to substitute for indigenous ones, and they fear the consequences of increased communications ability and access to information coming at the cost of face-to-face, personal interactions. Most people agree that those individuals and countries not connected to the Internet will become isolated and backward.

The different scenarios and arguments, the positive and the negative, lead to many questions. Whatever the argument or scenario, the importance of being able to identify, analyze, and quantify the impacts of the Internet on the development of a country is critical to policy design and implementation. Are there enough data to warrant investment in the Internet as distinct from the more basic day-to-day needs of rural or urban people? We need to better understand the kinds of impacts new information technologies are having on our own country and others.

The technology and development issues will continue to interest analysts and policymakers for years to come. The findings are likely to be that the Internet has many of the impacts suggested above and many more unintended and unanticipated ones. The important point is that the impacts can be better understood and, to be understood better, an analytical framework for measuring them will be useful.

This report begins to provide such an analytical framework. The framework contains indicators that can be used by development agencies, national policymakers, individual users, potential investors, and other interested parties to help assess the impacts of the Internet on various aspects of a countryís or a regionís development. The indicators are also intended to provide continuity for researchers who are further analyzing and quantifying impacts of the Internet. Which of the indicators are used will depend on who is conducting the analysis and for whose benefit it is being conducted. The framework can be used to shape policies and behaviors with regard to building up new information services. While the framework was developed in the African context, in particular for Senegal and Ghana, it is intended as a broad tool to be applied, with limited customization, to other developing countries as well.

The framework discussed in this report considers the Internet as a communications mechanism and suggests how to analyze the impacts of its introduction and use. It considers but does not focus on the impacts of the Internetís content. That is, for the most part, the indicators are not intended to measure the impacts of any information communicated via the Internet.

Getting answers in West Africa to the initial research questions posed by the committee did not prove to be easy. Committee members found a number of challenges to gaining the information and knowledge they sought. For example, in Africa and most other developing countries the Internet is a very new phenomenon, with most people and organizations who use it having at most a year or two of experience (with the exception of basic store and forward electronic mail, which has been used by some for several years). Given their limited experience, people are only now beginning to develop a systematic means of measuring the impacts. Even the most sophisticated Internet service providers (ISPs) and users in Ghana and Senegal are facing more basic issues (e.g., obtaining adequate technical expertise and equipment; navigating the new and often confusing regulations and administrative requirements; putting pressure on the government to deregulate, liberalize, lower prices, etc.). And nonusers may not yet have noticed any significant impacts of not having Internet access. Data about the most basic elements of Internet use were not easy to obtain. Some ISPs or government agencies proved reluctant to share the information they did have.

While it is premature to expect systematic evaluations to be in place in Ghana and Senegal, people are certainly conducting informal evaluations. The committee found a lively interest and recognition among Internet users in the utility and importance of some sort of evaluative framework for looking at the Internet. Furthermore, some of the individuals and organizations visited were beginning to pose to themselves the same issues of measuring impacts. For example, one organization that the committee interviewed was contacting its member organizations to ask how the Internet will help them and what they would like to do with it. Indeed, committee members were impressed with the openness and enthusiasm they found among organizations to share ideas about this issue. International organizations, for example, have reported decreased communications costs and increased access to materials and information. This for them justifies the investment in Internet access. It is likely that some organizations have done a similar analysis but found the investment too great to offset any potential gains.

In 1996 the U.S. Agency for International Development (USAID) asked the National Research Councilís (NRC) Office of International Affairs to conduct a series of activities to highlight applications of information and communications technologies to development and to examine ways in which those technologies can help USAID and other development assistance organizations better achieve their goals. The USAID had a specific interest in tools to help assess the Leland Initiative, a five-year $15 million program to extend full Internet connectivity to approximately 20 African countries in order to promote sustainable development.

The NRC convened an initial meeting of experts in development, African studies, and Internet and related technologies in March 1997. Following that meeting, the chairman of the NRC appointed a small steering committee to guide the effort. That committee, the Committee on Indicators of Internet Impacts on Development (referred to as the committee here), defined its goal as developing a broad and detailed set of indicators that could be useful for development activities and then sought the views of African and other experts in Ghana and Senegal on the utility of information technology evolution in general and how the Internet has impacted them.

The NRC had already been involved in a series of projects designed to bring information and communications technologies, in particular the Internet, to Africa. Under a project funded by the Carnegie Corporation of New York, a 1996 volume of case studies documented some of these efforts, especially individual ones, at bringing information and communications technologies to sub-Saharan Africa. The NRCís involvement in these and other activities served to further support the view that development organizations, local policymakers, and indeed many others involved in information technologies were interested in a means of assessing the impacts of the Internet.

The committee was aware of other important work in this field, as well as the limited time and resources at its disposal but concluded that it could contribute to the small existing body of actual field research. The committee considered this a serious and intensive, if short, research project.

The committee drew upon the large and growing literature on the Internet, its real and anticipated benefits, and Internet connectivity in Africa. For background on the general state of Internet connectivity in Africa, particularly Ghana and Senegal, the committee relied heavily on work by others. There is also an increasing literature on the impacts of information and information technologies. The Internet Society, Canadaís International Development Research Centre and its Acacia Initiative, the World Bank, the University of Marylandís Center for International Development and Conflict Management, and many other organizations are involved in related studies and projects. This report builds on that work and focuses on how to use indicators to analyze impacts of the Internet. It is not intended to be either the first or the final word on the topic but aims to broaden the existing discussions and lead to more in-depth studies.

Because of time and resource constraints, committee members did not try to incorporate explicitly all Internet-related technologies and services; nor did they exclude any. Unless otherwise noted, in this report the term Internet is used in its broadest sense and refers to Internet technologies and networks, including store-and-forward and e-mail-only networks as well as the World Wide Web and its associated technologies. The committee recognizes that, particularly in Africa where wider applications of the Internet are less common, the term Internet is often used interchangeably with electronic mail.

Believing that this research topic is in its early stages and will benefit from ongoing discussion, the NRC and the steering committee strove, from the projectís early stages, to disseminate its research directions and findings widely and to gain input from as many knowledgeable people in the field as possible. The committee developed an initial set of draft indicators and then shared them with a larger group of experts, including scientific organizations, U.S.- and African-based Internet service providers, indicator specialists, and others. To gain further input, the committee presented its preliminary indicators at the Global Knowledge í97 conference in June 1997 in Toronto. The tremendous interest in the topic at that meeting led to the establishment of an ongoing, on-line discussion of Internet indicators.

In August 1997 members of the committee traveled to Ghana and Senegal for field research. For the reader unfamiliar with Internet connectivity in Africa, specifically in Ghana and Senegal, background information can be found in Appendix D. In each country committee members arranged meetings with senior government officials and met with representatives from ISPs, nongovernmental organizations, academia, the private sector, and others. A separate trip to Kenya provided additional input for this report. A list of the institutions and individuals visited in Africa is contained in Appendix E. The committee believes that the number of its site visits and meetings in Senegal, Ghana, and Kenya was extensive and provided the committee members with an accurate, albeit somewhat superficial, overview of the situation with respect to the Internet in those countries. The committee emphasizes that the recommended next steps in research on Internet impacts will require studying a wider range of developing countries in Africa and elsewhere and at a more in-depth level.

In the countries visited, committee members solicited comments on the preliminary indicators, looked for case studies, and requested answers to the following questions from people they met:

• How did you (individual/institution) get interested in the Internet?
• Who within the institution is pushing Internet use?
• How is the Internet used?
• What are the tradeoffs of obtaining Internet access?
• What are the changes in your institution?
• How is/was the Internet introduced into your institution?

The committee was sensitive to the differences in the scale of Internet usage and diffusion between the United States, which has thousands of ISPs and tens of millions of users, and African countries such as Senegal and Ghana, which have three and nine ISPs, respectively, and only several thousand users each. The committee attempted not to import Western-based models and assumptions into the discussion of developing countries.

The committee reached five main conclusions from its meetings, interviews, and field research that it believes merit the attention of those interested in the diffusion of information technologies in developing areas and the associated impacts:

1. There is an expressed need among development agencies, policymakers, Internet service providers and users, and others for some evaluative framework for analyzing Internet impacts.

2. In the countries studied, many individuals and organizations recognize the benefits of having Internet access and consider it a necessity rather than a luxury, even after only a year or two of use, but obstacles remain to widespread use.

3. Barriers to more widespread Internet use can include the high costs of access and poor telephone infrastructure.

4. Patterns for Internet diffusion can be identified and models to describe those patterns proposed. Institutional leadership seems to be a key element in early diffusion.

5. There is an enormous amount of research yet to be undertaken on this topic.

This chapter describes the types of impacts and indicators with which this report is concerned. The importance of recognizing and anticipating positive as well as negative impacts of the Internet is noted, and the differences between direct and indirect, and intended and unintended, impacts are presented. The discussion then proceeds to the limitations and qualifications of indicators in and of themselves and as used in this report. For the subject at hand, the committee found that the most useful method of looking at impacts of the Internet is to begin with the most direct and move to the more indirect. Finally, the chapter notes the importance of using qualitative measurements in addition to quantitative indicators to develop enough understanding of the processes under way to guide the development of appropriate indicators.

In this report, direct impacts are those that are an immediate consequence of the introduction or use of the Internet. Introduction of the Internet in Africa and other countries has some direct impacts that are relatively easy to understand and measure, such as the reduction in communications costs for some uses. One can easily identify and quantify the reduced costs of sending e-mail versus a fax. There are other direct impacts that are easily identified, but somewhat less easily quantified such as the increasing ease of some communications and improved timeliness of communications.

Still other impacts are less obvious and more difficult to measure but perhaps more important in the long run. For example, how do we assess the value of an individual or organization increasing international contacts and participating in activities that previously were unknown or out of reach? In this report, indirect impacts are those that are not an immediate and exclusive result of the Internet. They may depend on who communicates what to whom and why. Indirect impacts may also depend on changes that occur in programs and institutions as a result of the technological opportunities made available by the Internet. Indirect impacts can range from increased cost effectiveness of a small nongovernmental organization (NGO) to restructuring of an economy as a result of changes in the structures of comparative advantage induced by the Internet.

POSITIVE AND NEGATIVE, INTENDED AND UNINTENDED IMPACTS

It is unlikely that all impacts of the Internet are or will be positive. Using television as a comparable technological innovation, some studies have concluded that TV is responsible for making violence more acceptable in U.S. society, for increased emphasis on consumerism, and for reducing literacy by substituting television viewing for reading as a pastime for large numbers of people. There is also growing concern about criminal activities, dubbed "cybercrime" by some, on the Internet. Any complete analytical framework for the Internet must provide for both positive and negative impacts. Moreover, any single stakeholder, be it an individual Internet user, development agency, or policymaker, will likely see both positive and negative impacts from Internet use and will have to balance those impacts according to his or her individual needs and expectations.

In addition to positive and negative impacts, one can distinguish between intended and unintended impacts of introduction and use of the Internet. These can be either positive or negative and depend on the perspective of the analysis. For example, decreased international communications costs, and hence decreased costs of sending data to another country, may be the positive impact an NGO desires upon acquiring Internet access; but when that NGO outsources analytical work that was previously done inhouse and incountry to analysts in another country, it is an unintended and negative impact from the perspective of a development agency striving to increase local employment and Internet experience.

Impacts of the Internet occur at several levels, including the individual, organizational, and sectoral levels and, as noted earlier, the impacts may be positive at some levels or for some stakeholders and negative at other levels. An individual user of the Internet will feel personal impacts, in terms of greater access to information and people, the cost of acquiring information, time budgets, and so forth. But the cumulative impact on an organization is probably greater than the sum of the impacts on individual users. That is, an organization will likely change structure and process to take advantage of the new opportunities provided to individuals by the Internet. Similarly, the impact on an industry may not be simply the sum of the impacts felt by the companies in the industry, where the industry restructures and modifies its processes in light of the opportunities and risks it perceives. Governments and donor agencies are also interested in the impacts of introduction of the Internet on society as a whole. Thus, an analysis of the cumulative impact is difficult unless one is able to analyze individual impacts together and look for the resulting balance.

The approach taken below deals with indicators of the penetration and use of the Internet and indicators of its impacts at the level of various institutions. On the basis of their studies of West Africa, the committee members suggest that an appropriate strategy to measuring the complex and indirect impacts of the Internet in countries where it is not widely used would be to trace the impacts from the most direct to the increasingly indirect. Thus, while many changes within organizations will take much time to be visible and measurable, the impact on organizations can be partly estimated from knowledge of the impacts on the users and nonusers of the Internet in the organization, as well as changes in the structure and processes of the organization made as a result of introduction of the Internet. Similarly, the impact on a market can be partly inferred from the impacts on the companies that buy and sell in that market, the organizations that regulate or operate the market, and changes made in the structure and processes of the market based on introduction of the Internet.

Development agencies and others typically use indicators to define how performance or impacts will be measured along a scale or dimension, without specifying a particular level of achievement. Such indicators are an increasingly important management tool.

There are a number of useful classifications of indicators. For example, below we differentiate between quantitative and qualitative indicators. The following discussion includes

quantitative indicators, such as the number of Internet services provided, and qualitative indicators, such as the quality of the services, as well as quantitative indicators on the amount of content available on the Internet and qualitative indicators of the quality of that content.

While quantitative indicators are not necessarily more objective, their numerical value tends to lead to more agreement on interpretation of results data, and they are often preferred by program managers. However, qualitative indicators can supplement the numbers and percentages and, thereby, add a real-life perspective to the evaluation of a program. Indeed, it will hardly be possible to find quantitative indicators for some of the indirect impacts of the introduction of the Internet in large and complex socioeconomic systems, and qualitative indicators may be the only ones available that are relevant to key policy and management issues. A further challenge is to devise a model to explain the qualitative indicators and their significance.

Indicators that measure numerical values are commonly divided into such groups as ratios, intervals, ordinal values, and nominal values. The division is important in terms of the analysis that the indicators will support. Another classification is input, output, and outcome indicators. In the following text, many indicators are provided that measure ratios, such as the percentage of information and data an institute obtains via the Internet and the percentage of business organizations with Internet access. Also suggested are ordinal indicators such as whether users find electronic mail to be of no use, marginally useful, useful, very useful, or indispensable.

There are a number of limitations and qualifications that must be borne in mind when considering and using the impact indicators presented below. Given the wide variety of direct and indirect impacts of introduction of the Internet, there will be a large number of potential indicators related to such impacts. An immediate concern is simplicity, as busy people could be overwhelmed by the demand for data and overworked analysts could be overwhelmed by an excess supply of data. Moreover, data beyond very basic supply and use statistics may be considered confidential or be otherwise hard to obtain.

The indicators are intended to provide a framework for local and foreign users, providers, and policymakers for collecting, organizing, and analyzing information. The indicators are intended as guidance for people who will choose areas and indicators to meet their own needs: a development agency evaluating its Internet project in an urban setting will use different indicators than will a local education minister evaluating the costs and benefits of investing in computers for Internet access in a rural school. Appropriate subsets of the indicators could also be used for large cross-national studies that allow comparison of two or more countries, provided adequate explanation is given as to differences between the countries in all factors related to the Internet.

Indicators and the analyses based on them are decision aids. The value of indicators stems from the analytical purposes for which they are used, and ultimately from the improvements in policy and management to which they contribute. Some analyses of impacts of the Internet in Africa will require complementary indicators that are not addressed in detail in this report, such as social indicators for measuring economic well-being and development. They are not addressed here because there is a huge and excellent literature on the subject, and that literature is too extensive and profound to summarize adequately in this report. However, instruments that collect data on indicators of Internet penetration, use, and impacts must also collect corresponding data needed for proper analysis of the impacts. For example, based on the committeeís research in Senegal and Ghana, it is clear that males have greater access to the Internet and use it more extensively worldwide than do females. Gender information will be required to consider equity issues. Comparably, geographic indicators will be required to differentiate urban from rural impacts as well impacts by region of a country. Information on language of use and ethnicity of users will similarly be needed for many analytical purposes.

Causality

Another concern is the ability to sustain causal attribution through lengthening chains of indirect impacts. While causality is implied by the term impact, the key analytical problem in application of the indicators will be attribution of the changes observed in the impact indicators to introduction of the Internet. Thus, for example, the Internet may contribute to restructuring of the economy in some advanced countries, but it is not possible to definitively identify how much of the observed change is due to the Internet and how much is due to other factors.

There is also a need to look at those individuals and institutions in Africa who are not using the Internet and are likely to feel certain impacts or repercussions as a result. Nonusers may be affected indirectly by the uses others make of the Internet. For example, a farmer in Ghana may find a larger market for his produce if the intermediaries who buy that produce and sell it to others successfully use the Internet to expand their output markets. Alternatively, if exporters from other countries are sufficiently more successful in using the Internet to market to Europe than those buying from the Ghanaian farmer, he may suffer the consequences in lower sales and less income. It is important that, in planning for the development of impact indicators and in analyzing the impacts of the Internet, the impacts of nonuse at all levels be considered as well.

Qualitative Studies

Although the framework of this report and most of the indicators proposed below suggest emphasis on quantitative approaches to analysis of numerical data, the committee strongly endorses the need for use of ethnographic techniques and the analysis of qualitative information to understand more fully the impacts of greatest interest. However, a discussion of such analysis methods is beyond the scope of this report.

There are, of course, many impacts of Internet use and impacts that the indicators presented in this report will not measure. As new uses and applications of Internet technologies are developed and as Internet use increases globally, its impacts will also increase and expand, sometimes faster than our ability to develop appropriate means of understanding and analyzing them. It is important to keep the limitations noted above in mind and to realize that any list and any measurements will be incomplete and likely undervalue the true impacts in any one country, region, or institution. The limitations do not, however, invalidate the value of a framework for analysis.

This chapter begins by describing a framework, presented in terms of Internet supply and demand, in which a discussion of the impacts of the Internet can be understood. Indicators are then offered to help analyze possible direct and indirect impacts of the Internet. The indicators are intended as tools, with different indicators useful for different analyses. Thus, a development agency, an investor, or a local policymaker likely would each use different sets of indicators. As noted in the Introduction, the indicators suggested in this report were developed in the African context and in particular based on the committeeís field research in Senegal and Ghana and to a lesser extent Kenya. However, the indicators are intended to apply to other countries as well. All of the indicators are presented in tabular format in Appendix C.

The supply side refers to the Internet service providers (ISPs) themselves and the services they offer. In Ghana and Senegal there are both commercial providers and nonprofit organizations providing Internet access to individuals and organizations. The supply side also refers to the environment in which ISPs operate. This environment includes governmental policies and regulations, the countryís telecommunications infrastructure, and the general environment that affects the offer and use of Internet services, such as economic conditions, level of education and literacy, and population characteristics.

The demand side includes those who use Internet servicesófor example, individuals and organizationsóand the way in which Internet services are used and the factors that affect their use. The demand side also refers to impacts on formal organizations such as schools, government agencies, and private enterprises. Each of these organizations serves other societal goals (e.g., improving the level and quality of education, governance, private-sector development), which in turn can be affected by Internet use by the organizations.

In very simple terms, the relationship between the supply and demand factors can be shown as:

Environment ´ Internet ´ Institutions ´ Sector Impacts

In this simple relationship, the environment influences the supply and demand of Internet services; those services are used by individuals and institutions, which in turn affect sector development. As more details are considered, the components of the categories noted above and the number of interactions between the categories will increase and become increasingly more complex. For example, Internet suppliers deliver content, which in turn influences the organizations and institutions using the Internet. But impacts also affect the institutions, which look for different content and service supply and then change their use, which in turn infers or requires changes in the environment. Thus, the interactions go in both directions. One suggested model that begins to illustrate the complexity of the components and interactions is shown in Figure 1.

Figure 1. Model of the impacts of the Internet.

The reader may only be interested in the impacts of the introduction of the Internet at some specific level in this model, such as the business organization or the education sector. This report, however, is suggesting that analytically the reader should consider all of the impacts in the model to the left of the one of interest. Thus, in Africa at this time, it may be helpful to understand the availability of Internet servicesñthe way they are being used (or not used) by individuals, the ways that organizations are adapting to the Internet, and their success in using the Internetñin order to adequately understand the larger sectoral impacts of the Internet.

The penetration, use, and impacts of the Internet depend on the environment into which it is introduced. At a minimum, in order to connect with and use the Internet, one needs a computer, a telephone line, and electricity to power them. Lack of access to these fundamentals limits Internet access in many African countries. Obviously, the ability to invest in computers, communications, electricity, and so forth depends in part on a countryís wealth and the rate of generation of new wealth. Equally, the ability of a country to introduce and utilize the Internet will depend on the human resources available, including the availability of technically trained people to install, maintain, and manage the technology and the ability of managers and workers to adapt to and utilize the Internet. Similarly, introduction of the Internet and its impacts will depend on the institutions in a country and their capabilities to acquire and utilize the technology. Finally, the path of introduction for the Internet is contingent on the policy environment in the country involved.

Such assertions are essentially truisms. The reality is that many factors combine in complex ways to influence the rate of introduction of the Internet, the rate of Internet use, the rate of adaptation to the opportunities and risks related to introduction of the Internet, and the ultimate impacts of the Internet. Over the past half century a literature has developed on technological innovations and their diffusion in developing countries that should be the basis for analyzing the environmental factors. A summary of this literature is beyond the scope of this report. However, selection and utilization of the indices discussed below should be done in light of a plan of analysis, which in turn is based on a broad understanding of the theories of technological innovations and diffusion and the social impacts of new technologies.

Supportive Economy and Infrastructure

Conventional wisdom suggests that a correlation exists between Internet use and such variables as gross national product (GNP), GNP per capita, and teledensity, or the number of telephones per capita. In fact, preliminary research undertaken by the committee shows this to be the case in a set of African countries. In other words, the more phones and money a country has, the better and more likely its access to the Internet. It would be expected that there are minimum and maximum levels of GNP and GNP per capita at which that relationship holds. Moreover, a more complex picture is likely to emerge if one considers more variables, such as variations of Internet connectivity (full Internet protocol, or IP, dial-in access, store and forward only). However, for the purposes of this report, it is assumed that an individualís or a countryís access to the Internet will be impacted by these and other variables. The relevant indicators for a supportive economy and infrastructure then include:

• GNP per capita
• number of telephones
• number of telephones per capita
• indicators of penetration of telephone service in rural areas
• indicators of penetration of electrical power in rural areas
• density of population in rural areas
• percentage of population in urban areas
• indicators of the strength of markets for personal computers, modems, and related technologies

Policy and Regulatory Environment

A supportive role for a countryís government is to advocate and provide an open and competitive policy and market environment, ensuring that there are multiple connections to the international information highway as well as opportunities for intranets in the country. Taxes and tariffs on imported equipment and other goods and services can assist or be a significant disincentive to Internet use and growth. In the countries visited by the committee in Africa, the governments can be described as passive with regard to the Internet, only sometimes initiating reforms. While no formal barriers to market entry exist, some problems were cited by Senegalese ISPs with regard to laws and rules governing Internet supply and use. There is little aggressive leadership in the government to push liberalization of laws and regulations. A common view of the local Post, Telephone and Telegraph (PTT) administration seems to be that by controlling Internet service it will increase its control of the infrastructure and be able to control all of the profits. The Ghanaian government has liberalized telecommunications policy some, but more remains to be done for it to be considered truly supportive. In Ghana and Senegal, only now are a few government officials beginning to use the Internet themselves. Representatives from the Ghanaian government recognize the need for an overall policy framework for communications policies, but the ministries have done little to date in terms of developing or implementing any such framework, whether because of other priorities or lack of expertise.

As noted above, the overall telecommunications infrastructure in Africa is relatively poor. In Ghana, problems include the absence of routing between ISPs and the resulting inefficiencies and lack of or poor-quality telephone lines. One ISP reported significant improvements in service in the past year, with waiting time for phone lines having decreased noticeably. Some relevant indicators of a supportive policy and regulatory environment are:

• estimated cost, time, and rate of success in establishing an ISP
• estimated cost, time, and rate of success in establishing an ISP account
• nondiscriminatory access to Internet service
• modem and/or computer tariff
• waiting time for a telephone line
• cost for installation of a telephone line
• waiting time for a leased line
• cost per minute to access points of presence (POPs)
• commercial availability of modems and computers
• local service for modems and computers

ISPs connect individual users to a national backbone or multiple backbones, which in turn connect to other backbones globally. An ISP can be considered an "onramp" to the Internet. ISPs can be commercial profit-seeking organizations or organizations providing Internet access to a limited target constituency, such as universities that provide access to students, staff, and alumni or nongovernmental organizations (NGOs) that provide access to their members and member organizations. In some countries, universities or NGOs may start Internet access for the public on a nonprofit basis and then reduce their services as other, usually for-profit, ISPs come into the market. As user sophistication and the availability of Internet applications increase, ISPs are offering more content-providing services, such as Web hosting, development and design, and electronic commerce.

ISPs in Senegal and Ghana range in size and operation from Telecom Plus and Africa Online, each with approximately 1,000 subscribers, to constituency-based providers such as ENDA, an NGO in Senegal that provides Internet access to other NGOs and has approximately 100 subscribers. As noted earlier, the total number of ISPs providing full access as of August 1997 in Senegal and Ghana was nine and three, respectively. Some are looking beyond electronic mail to offering other Internet-based services. To date, Internet access in Senegal and Ghana is limited primarily to the capital cities, though several ISPs in each country reported plans to expand to other cities.

Quantity of Internet Service

The quantity of Internet service supplied within a country, and the area reached by ISPs, can be measured using basic indicators such as the:

• total number of ISPs
• total bandwidth to outside country (kilobytes/second)
• total number of modems connected to ISP servers for dial-up access
• total number of leased lines to customers
• total number of points of presence (POPs)
• total number of secondary-city POPs
• percentage of population within local calling area of POP

Quality of Internet Service

The quality of Internet service that is offered can be measured by such indicators as the:

• percentage of send failures (messages that fail to reach their destination)
• average delivery time of e-mail/data transfer from each ISP to every ISP
• average delivery time of messages
• average time to check an empty mailbox
• mean connect speed of subscribers
• call failure rates for ISPs (the percentage of calls that fail to connect to the Web)

Indicators can of course include further derivatives of these indicators, such as the rate of change of the number of modems connected to ISP servers or acceleration (second derivative) of the average time to check an empty mailbox.

The number of ISPs in a country and the degree of competitiveness in the ISP market are important indicators of the likelihood of fair pricing and quality service. In Senegal, for example, there are few formal barriers to new ISPs entering the market, and competition has helped reduce monthly access prices for simple telephone call-up service to 10,000 Francs (approximately $20) from twice that or more. But at the same time, small ISPs complain about policies that they think limit true competition. ISPs and related organizations may benefit from working together to influence government policies and regulations. The potential impact of such organizations suggests another indicator:

• number of members in an information industry association or ISP association

A related issue is the profitability of ISPs and the degree to which they are able and willing to invest earnings in expansion. With a few exceptions, the provision of Internet services in Africa, as in the United States, is not yet profitable or financially viable. In the United States this is largely due to the unusual expansion of Internet use and a rush for market share, which results in high expenses for ISPs. In a more "normal" or developed industry not expanding so rapidly, many companies would likely be posting profits. In Ghana and Senegal, relatively high telecommunications costs and small markets have resulted in high access charges (relative to GNP per capita or to average monthly salaries) compared to Western countries. In this situation, ISPs reported financing their Internet operations through other business activities, such as computer hardware or other telecommunications services. As one individual noted to the committee, "Itís a losing business." But clearly the organizations and businesses that are providing Internet services are doing so in the expectation of future profitability.

It is clear that any analysis of the competitiveness of an ISP market should reflect the unique situation in African countries. An analysis is nonetheless important, and suggested indicators include the:

• number of ISPs offering full Internet service
• percentage of nonprofit ISPs
• number and percentage of profitable ISPs
• prices charged by ISPs for Internet access
• total funds invested by ISPs in expansion
• total ISP revenue

Sustainability

In addition to financial indicators, such as the cost of Internet service to providers and consumers or the profits to providers, there are more general economic considerations. It seems likely that the cost per user to provide services in Africa may be reduced as networks become more densely populated. Similarly, the benefits per consumer will likely increase as the number of users in the network increases. One may also suggest that the Internet will result in positive and negative externalitiesóbenefits and costs that will not be appropriated by, or fall on, Internet providers or users. Firms obtaining new business through the Internet may create jobs in suppliers, and competing communications technologies may lose business.

Thus, there is a need for economic indicators that go beyond those directly associated with costs and profits for providers and users. For example, analysts may wish to look at the subsidy flow that will be required to sustainably institutionalize the Internet in African countries. Unfortunately, the complexity of the system makes it very unlikely that models providing such indicators will be affordable or accurate. Still, it may be useful to record the estimates of experts on the amounts needed and the duration required for such subsidies.

Related to the issue of having a supportive environment for Internet use is the notion that merely importing a technology without developing the skills to sustain, tailor, and expand it, particularly when the technology is changing rapidly globally, is of limited benefit to a country. Therefore, the degree of indigenous skill and competency in all aspects of the Internet is one of many indicators for sustainability of the Internet and thus for its potential to affect a particular country.

One limiting factor on ISP expansion in Africa is the lack of technical capacity and human resources. Most ISPs in Ghana and Senegal reported to the committee that finding and retaining qualified technical staff is a significant problem. In Kenya this is a particular problem for the public sector because it cannot compete with private-sector salaries. This situation, too, complicates measures of competition because companies that might otherwise be making a profit, investing that profit, expanding and growing, and thereby contributing to the industryís expansion and growth, are instead not seeking to expand.

For ISPs in Ghana and Senegal the primary source of technical skills, to date, has been the West. In Ghana one of the largest commercial ISPs is Africa Online, a subsidiary of U.S.-based Prodigy, Inc. In Senegal many of the small ISPs operating alongside Telecom-Plus, the Internet arm of the national PTT, are owned by expatriates or recent immigrants. The number of foreign-owned or -operated ISPs is not intrinsically good or bad. The issue is whether skills are being transferred to the local population, a more difficult measurement to make. It is reflected in part in the sophistication of Internet use and the degree to which the Internet is used for increasingly advanced applications.

Relevant indicators include the:

• number of foreign- and domestic-owned ISPs
• number of local technical staff
• number of ISPs offering user training
• number of institutions that monitor their own traffic, use, and number of hits on pages
• average number of years of schooling of adult population
• literacy rate
• number of information technology courses offered in universities
• average salary of Web designers and other ISP employees

Related to the issue of development of indigenous skills is the development of indigenous information sources in the form of Web pages, databases, news sources, electronic commerce hosting, and others. These more sophisticated applications generally involve two-way transfers of information and content and thus allow an individual or organization to be a provider, not just a consumer, of information, goods, and services. Measurements of participation and skills in specific applications will show the extent to which an individual, organization, or country participates or hopes to participate in the target activity of that application, an example being the:

• percentage of ISPs offering Web hosting, Web design, and other services

The amount of indigenous content developed in a developing country, and the degree to which that indigenous content is used domestically, regionally, and internationally, can be shown by the:

• ratio of national, regional, and international traffic to total traffic (both coming into and going out of the country)
• number of home pages on domestic servers
• ratio of national, regional, and international participation in listserves and news groups

A typical measure of the quantity of Internet use is the number of users and the number of subscribers and any change in those numbers. But these numbers can lead to a false picture of Internet use if not qualified. For example, several hundred computers may have Internet access through a local-area network (LAN), or a single ISP subscriber may be several computers with several users who take turns connecting through each computer. A news agency that connects to newspapers and radio and television stations through the Internet may get news to vast numbers of individuals. Thus, as elsewhere, the number of users in Africa is expected to be larger than the number of personal computers with access to the Internet, and the number of PCs is expected to be larger than the number of ISP subscribers.

According to The World Information Report:

In Africa a finer division of "users" may be necessary. Interviews by the committee in Senegal, Ghana, and Kenya suggested that there may be hundreds of people who access electronic mail by traveling physically to a facility (school or university, information center, cybercafe, business center, etc.) and picking up paper copies of incoming messages and leaving paper copies of messages to be sent out. ISPs in Africa may have several distinct levels of use, each with a specific tariff structure. If data are to be useful analytically, categories of users will have to be disaggregated sufficiently so that there is less diversity within categories than between them.

An accurate measure of a user base must include the:

• total number of subscribers by category of user
• average number of workstations per subscriber
• average number of people with access per workstation

Simple derivatives of the number of usersñthe change and rate of change in the number of Internet subscribersñare basic, yet important, indicators of the technologyís current and potential impacts. In very simplistic terms, individuals and organizations pay for what they believe to be financially or otherwise beneficial to themselves and their organizations. The more subscribers there are, the more important the technology can be considered. In Senegal, according to some ISPs interviewed by the committee, the Internet subscriber base was estimated at 2,000 in August 1997 and growing at a rate of approximately 10 percent per year. As noted in the background section on Senegal above, that rate is likely to fluctuate significantly. In fact, one ISP reported 15 and 20 new subscribers in June and July, respectively; Telecom-Plus claimed 50 new subscribers per month. It also may be useful not only to measure the gross rate of change but also to take into account the rate of turnover, or the number of subscribers who cancel their subscriptions. Thus, the relevant indicators are the:

• rate of change in the number of subscribers
• turnover rate

Related to the rate of change of the user base and the rate of change in total Internet use is the change in use reliance by subscribers. Changes in reliance on the Internet versus other communications tools and information technologies and increased value found in the Internet will be reflected in increased time spent using it. Of course, a direct relationship cannot be assumed, as time may be spent "surfing" or traffic may include graphics, pictures, or other files that use many kilobytes but contain relatively little information. However, value found in the Internet can be measured to a degree with the following indicators:

• total traffic (kilobytes per day)
• total connect time per day
• total number of e-mails per day
• average subscriber connect time
• average subscriber connections per day
• number of subscribers using leased lines

It is also possible to develop quantitative indicators of what the Internet is used for. Surveys have estimated Internet use for:

• communication
• downloading software
• interactive discussions
• noninteractive discussions
• use of another computer
• real-time audio or video

and for:

• searches for product/service information
• purchases based on Web information
• searches for company/organization information
• searches for other information on the Web
• browsing/exploring
• seeing what is new at a favorite Web site
• business purposes

As noted above, interviews in Africa indicate that, to date, the Internet is often equated with electronic mail and considered an alternative means of long-distance communications. Respondents in a recent study reported that two-thirds or more of their messages are sent outside Africa. That same study asked respondents in Senegal to classify the purposes for their use of electronic communications, under such categories as personal messages, exchange of ideas, providing or receiving technical advice, arranging meetings, managing projects, and computer conferences. Similarly, survey data have been used to estimate the number of Internet and Web users according to intensity of use: high users, low users, and nonusers.

Another helpful measure related to types of Internet use is the point of access location, for which indicators include the:

• percentage of users who connect from their own homes
• percentage of users who connect from an office
• percentage of users who connect from both home and office
• percentage of users who connect from an Internet cafe or business center

Costs of Internet Use

The costs of Internet access, and the cost variances within a country, while not impacts themselves are an important issue in evaluating Internet supply and demand. General economic knowledge would suggest that the lower the cost of Internet services, the higher the usage would be, all other things being equal, and, by extrapolation, the greater the impacts. What does one include in a cost estimate? The answer may not be as easy as expected. Costs can be divided into several categories, including fees paid to an ISP for access, fees paid to a telephone company, hardware and software necessary to access and utilize the Internet, and staff and personnel costs and other organizational costs necessary in an organization to enable people to use the Internet well.

The prices charged by ISPs for Internet access depend on many factors, including, as mentioned earlier, the competitiveness in the ISP market, the countryís GNP or general economic environment, and other telecommunications costs. But to the degree that price influences demand, a more important issue may be the price of a telephone call to dial the service provider, as this can be a major cost in many countries and a barrier for much of the population to gain access to the Internet, particularly when local calls are billed on a per-minute basis in addition to monthly line charges. An individual at the University of Ghana, for example, noted to the committee that the cost of a local phone call is 6 million cedis per month (approximately $3,000) for 100 people to work one hour per week on the Internet. When people get the phone bill, the university official fears, they will be disillusioned with the technology.

In developing countries, international calls may be required from many locations to reach the nearest ISP, and international rates must be paid. For small enterprises that simply use a modem and connect to a local provider through an existing phone line, there may be no additional telephone cost for Internet connectivity. On the other hand, in countries with inadequate telephone services, enterprises that require new telephone lines to connect to the Internet may find the cost of telephone service exceeds that of the ISP. Where high-capacity lines are required, total costs may be quite significant.

Determining hardware and software costs also is difficult. In developed countries, connection to the Internet is generally accomplished using hardware and software that is available but underutilized and thus is accomplished with minimal marginal cost. In developing countries, because the Internet may be the initial use of the equipment, the hardware and software costs allocated to Internet use are likely to be relatively high.

Finally, staff and organizational costs resulting from restructuring are difficult to determine. Often, because optimists assume that the Internet increases efficiency, staff costs such as training are generally ignored. There is, however, an investment needed for a person to be able to use the Internet efficiently and effectively, and this cost at least should be considered. Also, in larger organizations there will be a need for personnel to maintain local-area networks and software and otherwise service the facilities through which the Internet is accessed. In addition, enterprises are increasingly contracting for services such as Web page development, and such services should be treated as Internet-related costs.

Descriptive research is needed to identify the range of such costs for a developing countryís organizations using the Internet, since such information is needed in terms of policy to make the Internet more affordable. Some suggested indicators include the:

• price elasticity of demand
• fees paid to an ISP for leased-line, dial-up service, and other services
• installation fee(s)
• fees (fixed and/or usage dependent) paid to the telephone company
• price of a phone call per minute to connect with the ISP
• costs to the organization of Internet training courses and staff salaries paid during training

Quality of Internet Use

In many countries there is concern about the quality of Internet use, which will determine to a degree the impacts the Internet may have. Users are largely self-taught and tend to gain facility over time by using trial-and-error methods. Problems with effective use of the Internet can be quite basic, such as lack of basic literacy or typing skills that limit even the use of electronic mail. Given that much of the content of the World Wide Web is in English (and, to a lesser degree, Spanish, French, German, and Japanese), lack of fluency can also limit a userís quality of Internet use. The ability to search the Internet to find information that is needed efficiently and effectively is perhaps the most frequently discussed element regarding the quality of use. This ability involves an understanding of various search engines and their advantages and limitations, as well as an ability to use the hypertext aspects of the World Wide Web effectively. One research project, for example, has estimated the ability of U.S. users to navigate the Web, classifying them as "very skilled" and those for whom navigating the Web is "very challenging." Comparable indicators can be derived to measure the effectiveness of Internet use by African users.

In Africa, where the costs of Internet use are relatively high compared to personal incomes and where access to computers linked to the Internet is limited, the ability to use the Internet efficiently is especially important. However, in Africa one may expect users to have less experience with other information tools, compared with people in the United States where libraries have been equipped with on-line search engines for years. Interviews by committee members in Ghana and Senegal indicated that there was considerable concern about the quality of use of the Internet and the lack of training and experience of users.

One possible source of indicators of the quality of Internet use in Africa would be from help desks. A help desk is a service that provides (usually on-line) assistance to Internet users in how to fulfill specific information demands quickly and effectively. Help desks have been supported by donors, governments, professional associations, and other groups. As a result of their close daily contact with Internet users, their staff may provide useful qualitative data and their records useful quantitative data on quality of use.

As noted in the Introduction, this report focuses primarily on the impacts of the Internet as a mechanism itself, not on the content that is contained in Internet communications and data. However, the quantity, quality, and relevance of material on the Internet deserve some discussion, since they determine in large part the degree and types of Internet impacts. There is room for concern about the content of electronic mail transmitted through the Internet, as when corporations and government agencies are concerned with personal use of electronic mail facilities by employees and its possible deleterious effect on productivity or costs. More importantly, there is considerable interest in Internet commerce, which may be electronic mail mediated and which has given rise to considerable effort to develop appropriate indicators. This section focuses primarily on the content of Web pages.

With regard to quantity, it has been estimated that as of August 1997 there were 180 million Web pages on the Internet. However, many Web users in Africa and in the West have commented on the dearth of information generated in Africa and on the more general lack of distribution of scientific, technological, and other information produced in Africa. Scientists and others have hoped that using the Internet would make it possible to counteract this general deficiency and make scientific and other data more generally available in the countries of origin and regionally in Africa. In counting Web pages in Africa, search engines and software agents can play a critical role. It will be important to categorize Web pages to obtain meaningful indicators, differentiating by language, purpose (governmental, commercial, scientific, etc.), subject matter (agriculture, health, education, etc.), and other criteria.

With regard to quality, users in developing countries are already complaining (as are users in developed countries) that in using the Internet they often find more information than they can easily digest and that much of that it is of dubious quality. There are public policy reasons for which countries would seek to make and implement policies to restrict material they find morally or ethically unacceptable from being placed on or transmitted over the Internet, such as sexually explicit materials or materials that might exacerbate ethnic or racial tensions. Other countries may wish to make policies on the language of postings on the Internet. For such purposes it is, in principle, possible to construct indicators of the degree to which Web pages appear likely to include objectionable content, and some such indicators now exist. (Of course, using such indicators as tools, quantitative indicators can be created based on estimates of the number of Web pages containing such objectionable content.) The EDSITEment Web site, for example, seeks to provide links to "the best humanities-related educational content on the Internet." To do so, it asks people to classify candidate Web sites according to:

• intellectual quality, or the quality of the material posted on the Web
• user friendliness, or the ease of use and visual attractiveness of a site
• user impact, or the ability of a site to serve multiple audiences

Conceptually, it would be possible to assign an ordinal scale to each such criterion, with values for not acceptable, poor, average, good, and excellent. This example comes from the education sector and suggests that it would be equally possible to qualify sites for a wide variety of educational purposes. Similarly, without doubt the approach could be generalized to other sectors, and sites could be categorized according to quality in terms of private-sector development and/or democratization and civil society.

Finally, with regard to relevance, it is possible for a Web site to provide large quantities of information of high quality but still be of little relevance to Africa. Comparably, it is possible for a Web site to be very relevant to health care practitioners but of little relevance to educators or business people. Consequently it may be useful to develop indicators of relevance of the content of the Internet by sector, country, region and other categories.

One possible use of the Internet is control of computers and automata at a distance. In some situations, networks of personal computers have been used in off hours by researchers to do complex calculations, such as those needed for factoring large numbers; in other situations, supercomputers have been controlled via programs or instructions sent over the Internet. The degree to which Africans are using the Internet to enhance the very limited computing power on the continent through the use of computers abroad is certainly worthy of investigation, and indicators could be developed fairly easily of the amount of computer power and the time of use of that power through the Internet. While these would not be "content" indicators as normally construed, the amount of computational depth underlying certain classes of Internet communication is indeed a measure of content. Thus, a weather prediction sent via e-mail has more content if it is the result of several hours of analysis by a supercomputer based on large amounts of remote sensing data than if it is the result of a simple guess.

Perceived Benefits of the Internet

Diffusion of the technology, of course, depends on individuals and organizations finding value and an economic benefit in using the Internet. The most commonly stated advantages of the Internet by current users in Dakar, Accra, and Nairobi are time and cost savings on communications and ease of communications with others in the field. Increasing use of the Internet by an organization, decreasing use of other communications means, and cost savings in communications costs are generally considered the most straightforward and obvious impacts of using the Internet. Those and more sophisticated analyses can be made with the following indicators:

• number of messages/transactions to/from/by an organization per day that are domestic versus regional versus international in source or destination
• reported relative importance of the Internet versus other means
• cost savings on communications
• time savings on communications
• ratio of Internet to other channels in obtaining information

Also important to organizations increasing their access to outside data is that organizations increase dissemination of their information to others. As emphasized above, individuals, organizations, and countries must become information producers, not just consumers. An outward flow of informationñfor example, information about products, valuable scientific datañwill help ensure that a country can becomes a leader in the information economy. Indicators in this regard include:

• percentage of an institution's dissemination through the Internet
• percentage of total public information made available through the Internet
• number of Web server hits or requests fulfilled per month from domestic versus regional versus international sources
• number of electronic newsletters or bulletins produced
• number of subscribers to newsletters and/or bulletins

The likely impacts that the Internet will have on the operations of any given organization are a function of how that organization and the individuals within it use the Internet and the importance they place on it. The pattern of Internet use in an organization, not merely the quantity of use, is important. For example, the types of impacts in an organization will differ according to whether users within that organization are part of an LAN or have individual stations and access to the Internet. In Africa it is likely that, at present, there is little transfer of benefits to the organizational level because individual use is more common than group use and there is little sharing within organizations.

Further, in general, there is likely to be a progression of Internet use: e-mail is the first use, surfing is the next, the design and use of a home page is third, and then the development of an intranet. Several indicators may be useful in this respect:

• number and percentage of people trained in using the Internet
• number and percentage of subscribers with a LAN
• number of top/middle/lower-level users in an organization with access to the Internet
• relative importance placed on the Internet by top/middle/lower-level staff
• number of networks and "virtual organizations" of which an institution is a member
• investments in computer and other telecommunications facilities
• approximate number of users who (1) use e-mail, (2) "surf," (3) maintain own (individual or organization) home page, and (4) use an Intranet
• presence of a distinct information strategy as part of an organizationís overall organizational strategies and plans

Organizations are likely to change structure, linkages, and processes because of the Internet, and such changes can be measured as impacts of its introduction. Thus, it has been suggested that organizations using Internet technology are becoming flatter and less hierarchical. Processes such as "just-in-time manufacturing," electronic commerce, and interfirm collaboration on product design are facilitated by Internet communications technology and are likely impacts of that technology (at least in technologically advanced countries). Similarly, it has been suggested that firms are tending to downsize and out-source more in advanced technological countries, substituting interfirm linkages for intrafirm linkages in part because of the improved interfirm communication made possible through Internet technology. While such changes may be hard to measure quantitatively, it would be possible to develop indicators to discern such changes in at least larger organizations and perhaps smaller organizations, for example, by noting changes to the organizational chart, the abolishment of existing positions, and the establishment of new positions (e.g., chief information officer or its equivalent.) This is especially important, as most studies show that organizations cannot capture substantive efficiencies from information technologies unless they are willing to make serious structural and staffing changes.

Higher-level impacts can be sought, such as the utility of the Internet in carrying out the functions of an organization. In business, for example, one study of Internet applications in productive enterprises focused on applications in:

• production and operation
• marketing
• finance
• accounting
• human resources management

A recent study by the International Finance Corporation used ordinal values to ask users if the Internet had been not useful, slightly useful, useful, very useful, or indispensable in carrying out specific functions and how useful it would be in these functions within a year. A similar approach has been used to study Internet use by financial services firms in carrying out the functions of:

• advertising and public relations
• on-line customer services
• on-line financial transactions
• on-lines tools (e.g. credit assessment)
• human resources management

Such an approach can obviously be extended to other kinds of organizations. Thus, one can identify key functions of education and governmental organizations as well as NGOs and develop indicators relevant to the impacts of the Internet on those functions. To use such an approach, it is critical that organizations be disaggregated so that those within a specific class are comparably dependent on the same functions for their overall success. Thus, one might not be able to classify educational organizations into a single group but would be likely to have higher-educational organizations that are separate from general secondary schools, organizations specializing in nonformal education, trade schools, and so forth.

Of course, the availability of the Internet to the rest of society must be taken into account. That is, African NGOs cannot easily communicate with clients and members unless they have similar communications technologies.

Organizational Decision Making

One view of decisionmaking within an organization is as a network, with people and their available information technologies (books, files, etc.) corresponding to the nodes and their channels of communication as the linkages among nodes. With the introduction of the Internet and other information technologies, one can consider that the node is the individual augmented by more complex information technologies, including his or her personal computer and the linkages to include electronic communications lines as well as phones, faxes, and the more traditional channels for verbal and written communication. Introduction of the Internet can improve the calculating power of each individual decisionmaker, the speed of access to data and the amount of data that can be accessed, and the rate of communication among people in a network. By embodying decision rules or decision support within hardware and software, the decision processes can be improved or at least made more nearly standard and predictable. Examples of relevant indicators are:

• change in number of people involved in an institutionís decisionmaking
• relative importance of the Internet versus other means of gathering data and information in decisionmaking

The extent to which the decisionmaking process in an office is made more or less inclusive, or hierarchical, as a result of the Internet depends, of course, on the number and types of employees who have access to the technology. A focus on the use of the Internet by different categories of employees (upper-management, middle-management, foremen, etc.) would help to illuminate the impacts of the Internet on organizational decisionmaking.

Ultimately, people will be interested in the impacts of the introduction of the Internet on the overall performance of organizations. Are firms becoming more profitable, better able to serve their customers with high-quality goods and services produced at low cost, more able to adapt to changing markets and environments and protect the interests of their owners and employees? Are schools better able to help students to learn, to provide educational services of high quality efficiently, to assure the relevance of their curricula to changing social and economic circumstances, and to meet the needs of the community they serve? Are government agencies, political parties and civil society organizations better able to involve citizens in governance, to provide high-quality services to citizens at low cost, to meet the changing needs of their electorates and institutions? Each such question can lead to an indicator of organizational performance. The Committee found exactly these kinds of concerns in their interviews in Senegal, Ghana, and Kenya.

Nonutilization of the Internet in the face of utilization by other organizations may result in a competitive advantage by the latter, as well as negative impacts on the former. Failure to adopt and utilize the Internet may assure that an organization will not benefit from the Internet, but adoption of the technology and relative success in applying it to some functions will not assure that the organization will be notably more successful in achieving its overall objectives. It is the pattern of use and success with the Internet that will determine overall success or failure of an organization in the Internet world. Even where indicators of the overall success of organizations are available, such indicators for a large organization will be sensitive to a wide range of causal variablesóthe overall state of the economy, the quality and morale of the staff, the overall policies of the organizational leadership. Thus, the problem of finding an indicator of the impacts of the Internet on the overall objectives of large organizations becomes methodologically complex. In general, one will seek qualitative or proxy indicators at this analytical level, such as the consensus opinion of a panel of experts. There exists a considerable literature and body of experience on how to organize such panels and create such indicators for different types of organizations; the existing theories and methodologies could, and should, be further developed to apply to the relationship of Internet use with an organizationís effectiveness or success.

There are many kinds of institutions. The previous section dealt with formal organizations. Anotherñthe marketñis discussed in a subsequent section on the private sector. All societies, however, include many other cultural, social, economic, political, and religious institutions. Critically important to all these institutions are their underlying communications and information networks. Thus, as Internet use in Africa results in institutionalization of new communications and information links, it changes the information and communications networks underlying a wide variety of institutions. This change in information and communications networks may induce other important changes in the institutions themselves. Indeed, many of the most important impacts of the Internet in Africa may occur in the cultural, social, religious, and other institutions which are neither formal organizations nor markets. (ISPs in Africa reported that many subscribers were people who used the Internet to communicate with family members abroad. Thus even so basic an institution as the family is affected by the Internet.)

Some generalizations about indicators may be possible. Indicators of the geographic dispersal of people affiliated with an institution and the relative authority of information and analysis sources may be useful, and gradients in institutional affiliation with income and education may be examined as secondary impacts of the Internet. That is, as the Internet relieves geographic constraints to communication, one may expect people to participate in institutions that are more geographically dispersed. And as people find new sources of information and analysis available over the Internet that are superior to local sources, the authority attributed to those traditional local sources and institutions may diminish. Also, as richer and more educated people have more access to the Internet, and to the extent access to the Internet leads to changes in relative affiliation to distant or disperse institutions, so then the introduction of the Internet may add to class differences in institutional affiliations.

However, it seems likely that appropriate indicators of the impacts of introduction of the Internet will usually depend on the specific nature of the institutions in question. It is beyond the scope of this report to seek to inventory every institution in Africa and to define every appropriate indicator for each institution. The general approach used throughout this report seems widely applicableñthat is, measuring the penetration of the Internet into the group of people involved with the institution, seeking indicators of direct effects of this penetration, and then seeking indicators of indirect effects on the institution.

The preceding sections considered indicators of Internet impacts related to the supply and use of the Internet by individuals and institutions; this section takes the discussion up one level to sectors and sectoral goals. First, the diffusion of the Internet among sectors is considered. The discussion then proceeds to impacts of the Internet on certain sectors and on the developmental goals associated with those sectors.

Sectoral Use and Diffusion of the Internet

If one considers the Internet as a tool (and outcome) of development, perhaps the largest challenge faced by ISPs, governments, and othersñbut of crucial importance if the Internet is to help move the developing countries into the information economyñis to ensure that the Internet reaches a diverse user base, an indigenous user base, and that it is used for development purposes. All sectors of the population, including schools, universities, libraries, and small organizations, not just the rich businesses, must have access.

Related to the question of which sectors are using the Internet is that of sectoral flow. That is, which sectors, if any, seem to dominate and propel the Internet? While electronic mail and the Internet originated in the United States with the Department of Defense, the academic sector has played a large role in its diffusion. Private-sector use has followed, and recently individual use has skyrocketed. But in Africa, in the countries visited, the committee found that Internet diffusion has been driven largely by the private sector, foreign companies, and development organizations. NGOs have played a smaller but significant role. The academic sector was among the earliest users of FIDOnet and other precursors of the Internet, and in some cases university computer centers have played a key role in obtaining and maintaining access to the Internet. However, in the African countries it visited, the committee found that the academic sector has had limited influence, with universities still struggling to gain access. The government has had little influence in promoting diffusion of the Internet, either passively allowing it to develop or, in some cases, taking an active role to control and restrict it; and there is very little individual use. Individual use is far less common than organizational and institutional uses, largely because many individuals lack the capital for a telephone line and computer.

More so than in the United States, the Internet in Africa seems to develop only within sectors where visionary leadership emergesñin the NGO community, private sector, or government (less so). In Ghana and Senegal, in each sector, identifiable individuals have pushed to move the technology into acceptance and use. These individuals are then in a position to (1) create new institutions, such as the Ghana Internet Society; (2) lead, restructure, and invigorate existing institutions, such as the Ecole Polytechnique in Senegal; and (3) gain public attention, for example through the press, which leads to public education.

Some of this difference can, of course, be explained by economic differences, as discussed more in other sections, but there are nonetheless different forces pushing Internet use. The relevant indicators in the study of Internet diffusion are:

• total number of subscribers per sector
• increase in the number of subscribers per sector
• percentage of Internet use per sector for (1) e-mail, (2) "surfing," (3) maintaining a (individual or organization) home page, and (4) use of an Intranet

In addition to reaching all sectors, for the Internet to be fully effective it will have to reach rural areas. To date, globally and in Africa the Internet is largely an urban technology. This is a particular challenge in Africa, where countries like Ghana are predominantly agricultural and rural communities have minimal or no access to the technology. Some efforts are under way in this respect. For example, in Senegal, in a project partially funded by the Food and Agriculture Organization of the United Nations, the nongovernmental organization Conseil National de Concertation et de Cooperation des Ruraux is planning to link radios in rural areas via Internet to help provide market information to farmers and fishermen. The relevant indicator here is the:

• number of subscribers in primary city and secondary cities

A sector is composed of formal organizations, markets, and other institutions. For example, the education sector is composed of schools, training facilities, and teachers and students. It also involves markets for the labor of educators and for educational supplies such as books and professional associations for education professionals. The Internet can impact the organizations and institutions in the education sector as, for example, when teachers use it to enrich their curricula and increase student participation in classroom activities or when teachers, colleges, and schools use it to improve the recruitment and placement of new graduates as teachers.

There also exist broader developmental goals associated with sectors. The Internet, and through it networking, information gathering, and information sharing, is a tool for supporting initiatives that can improve and further these development goals. For example, for isolated countries and communities, Internet access can permit increased communication and interaction with the international community and thereby facilitate technological and economic development. Another common goal of development programs is to increase access to education, health care services, employment opportunities and decisionmaking influence for women. A proven tool in this endeavor is networking. Similarly, the developmental goals associated with the education sector include increased literacy, a more educated populace, and higher national technical competency, which in turn lead to a more productive work force, job creation, capital formation, and a variety of other values.

The committee noted a specific example of how the Internet is supporting broader development goals in Ghana. The Ghana Association of Women Entrepreneurs provides its 250 members in the 10 regions of the country with Internet exposure, links to buyers, training on basic business principles, and other moral support. A project funded by the U.S. Agency for International Developmentís Leland Initiative will provide computers and link the 10 regional chapters in order to increase communication and otherwise assist in its services. On an individual scale, women manufacturers will be better served by the organization and may increase their own activities, income, and influence in local decisionmaking. Ultimately, the networking assistance may help members increase sales and gain exposure overseas, thus potentially increasing exports. However, the committee also found that many of the business associations in both Ghana and Senegal have only recently obtained Internet access and use it rarely.

Importantly, the impacts of the Internet on a particular sector can be seen as an aggregate of the impacts on the components of the sector. However, in so complex a system, impacts may emerge as a result of intended or unintended synergies. Unexpected impacts may emerge as a result of the overall pattern of impacts of the Internet on the components of the sector. Impacts may also result from nonuse of the Internet, as when the competitive position of enterprises in a sector deteriorates with respect to enterprises in other sectors in the same country, or enterprises in the same sector in other countries as a result of their use of the Internet. Similarly, impacts may be positive, reflected in enhanced performance of the sector as judged against its goals, or negative, as reflected in deteriorating performance.

It has been suggested, in fact, that one of the most important functions for government and international donor agencies in dealing with the Internet is to analyze the overall sectoral pattern of penetration and utilization of the Internet and to intervene with policies, subsidies, regulations, and demonstrations that will improve the overall impacts of the Internet on sector performance. For such purposes it will be difficult to develop definitive indicators. Indicators would have to not only identify changes in the overall performance of the sector but would also have to do so in such as way as to establish that they were results of causes stemming from introduction of the Internet. This report again suggests the utility for such purposes of qualitative indicators, based on the judgment of experts provided with comprehensive information about the sector.

In the following three sections, impact indicators are suggested for three sectorsñeducation, the private sector, and government and civil society. The impacts on the developmental goals central to the sectorsñthat is, wider access to, and better, education; private sector development; and democratization and civil society developmentñare also discussed. The report contains no judgment on the appropriateness of the developmental goals themselves. That is, the potential impacts of the Internet on the development goals are considered but not the development goals themselves.

With the proper skills, technical support, and financial resources, the Internet can become a catalyst for important changes at schools and universities. In terms of sectoral goals, one would want to measure the impacts on students, schools, and educators and the extent to which the Internet improves the learning environment, enriches curricula, and expands access to education. For example, a major contribution of the Internet is the availability of learning and teaching material through digital libraries and networked learner support services.

Relevant indicators would include the:

• number of schools/universities with Internet access
• number of students with Internet access
• average time of student access
• number of teachers with Internet access
• number of training courses on the Internet offered to teachers
• quality of training courses on the Internet offered to teachers (accreditation)
• number of new courses offered since the Internet was introduced
• number of schools/universities utilizing distance education via the Internet
• number of courses that supplement conventional teaching methods with distance education or other Internet-dependent technologies
• number of students enrolled in distance education
• number of nonuniversity institutions offering distance education

It is also possible to ask what the impacts of the Internet are on functions, such as classroom teaching, distance teaching, curriculum development and maintenance, preparation and dissemination of teaching materials and texts, educational finance, teacher training, supervision and continuing education, and design and maintenance of school facilities, etc.

Other indirect impacts of the Internet in education include the career experiences of students who have had access to the Internet in primary or secondary school. For example, are students who had access to distance learning or other advanced teaching techniques that utilize communications technologies able to secure better and higher-paying jobs than their peers who had no access to the Internet or other communications technologies? Such information could likely be obtained by monitoring job advertisements that mention Internet skills. Some relevant indicators in this regard include the:

• ratio of job placement of students with Internet experience/training in school to overall placement
• ratio of average starting salaries of individuals with Internet experience/training in school to overall starting salaries

Inasmuch as universities are also centers of research, they strive not only to attract and educate the best students but also to attract the best scholars and researchers, often on a temporary basis. And as researchers increasingly are communicating, publishing, and conducting research via the Internet, they are attracted to research at institutions that have Internet services. Universities in developing countries will likely find that Internet access is key to attracting scholars and researchers from abroad. A relevant indicator would be the:

• number of scholars/researchers attracted to a university/country (in part) because of Internet access

 Development Goals

The development goals associated with education in Africa clearly depend on the specific context but include literacy, the ability to work competitively in a global marketplace, preparation to work in todayís workplace and adapt to that of the next century, knowledge and understanding necessary for effective citizenship in a participatory government system, social competence to safeguard the health of families and communities and to evaluate and implement alternatives for family planning, and the ability to live and participate peacefully in a multiethnic community. Each such global educational objective can be seen as inducing the development of indicators as to whether changes are occurring, their desirability, and the degree to which those changes are impacts of the changing information and communications environment typified by introduction of the Internet.

The keystone institution in the education sector is clearly the school: primary, secondary, and tertiary (or higher). This sector also includes distance education and nonformal education programs and institutions. It includes other formal organizations such as government agencies that regulate education, research organizations that conduct educational research, and financial organizations involved in school financing or in administering fellowships and scholarships. It includes other institutions, such as community institutions that relate to the schools (e.g. parent-teacher organizations), professional associations among teachers, and institutions that mediate the selection of schools by students and teachers by schools. Thus, the impacts of the introduction of the Internet on the education sector may be many and varied.

Businesses and foreign investors increasingly see Internet access as a necessity for doing business. In Ghana various chambers of commerce and business organizations are recognizing that, as one organization representative noted to the committee, "you need to get connected in this competitive world." Other organizations, such as the West African Enterprise Network, are able to get information on domestic and foreign markets with which to help their members. But many other businesses, small and large, do not yet have Internet access.

The impacts of the Internet on the private sector can be measured by seeking to understand the impacts on firms, especially firms engaged in manufacturing and trade; on markets for raw materials, intermediate goods and services, and final products; on such associations as trade associations and chambers of commerce that participate in the sector; on community institutions that relate to the industrial and trade sectors; and on the government agencies that serve and regulate the sector.

Communications technologies would appear to be especially important in market institutions that depend heavily on information on supply and demand and on the capacity to analyze changes in a situation rapidly and accurately. Anecdotal evidence suggests that the private sector can benefit, and already is benefiting, from increased information about suppliers, products, and markets; from communications with customers or clients and with other firms; and from increased product differentiation. As one example, the Kenyan automobile market, which imports approximately 90 percent of its vehicles used from the Gulf region, has experienced a decline in per-unit cost to the consumer. In committee interviews in Nairobi, one observer attributed this to use of the Internet by both consumers and competing dealers to compare prices from suppliers, thereby driving prices down.

The availability of the Internet should thus reduce some transaction costs, increase the speed with which information is transmitted through a market, and increase the quality of some of the information used. Unequal access to the Internet would seem likely to create market advantages for some and disadvantages for others. The Internet seems likely to increase the size of many markets, insofar as it serves to reduce geographical barriers to obtaining and providing market information. Indeed, some have credited the developments of comparable communications technologies with the explosive growth of international financial markets in recent decades. Market processes may be modified to take advantage of new potentials and to avert new risks created by the Internet. Stock market procedures have been modified to reduce the risk that very rapid changes in market value created by computerized stock trading might result in market instability.

Particular interest in Africa will probably be directed at such indicators as the:

• rates of participation of African firms in international markets
• rates of participation of foreign firms in African markets
• numbers of Web pages providing information on a market
• numbers of persons communicating about a market on the Internet
• volume of transactions in a market using the Internet

Comparable indicators can be derived to measure the impacts of the Internet on other African institutions.

Increased information for buyers and sellers, in turn, has the potential to level the international playing field for developing countries by helping them to increase regional trade and their exports to Western countries. Thus, at the sector level, concern is with the quantity and quality of goods and services produced, their relevance to the needs of the consumer, the affordability of the goods and services to the consumer as well as the employment created, the return on capital invested, and the externalities such as pollution or benefits to the community.

One should be concerned with changes in indicators as to whether societal objectives of these kinds are being met. Appropriately selected individuals can make reasoned judgments as to whether the impacts of introduction of the Internet have been supportive or destructive to these objectives. Modest information should be possible on the magnitude of the sectoral impacts observed or inferred through such indicators as the:

• number (percentage) of chambers of commerce with Internet access
• number (percentage) of other business organizations with Internet access
• number of small and medium-sized enterprises with Internet access
• number of small and medium-sized enterprises posting products and prices on the Internet
• rate of change in the value of an enterpriseís exports (imports) since acquiring Internet access
• rate of change in the value of a countryís exports (imports) since acquiring Internet access
• rate of change in the value of a firmís exports since acquiring Internet access
• number of companies reporting growth since availability of the Internet
• number of firms engaged in electronic commerce
• value of sales via the Internet
• funds allocated by private companies to Internet-related training
• growth rates of private telecenters that provide Internet services

The development goals related to private-sector expansion are like those discussed above: goods and services relevant to the needs of the population that are of good quality and equitably available at reasonable prices, with the creation of good employment opportunities, a good return on capital, and a minimum of negative externalities. The goals might well involve improvement of performance in terms of the rule of law (such as when one seeks to substitute legal activities for those involved in the manufacture and distribution of narcotics) or improved affiliation of workers and customers with a national culture. Again, such macro socioeconomic objectives can, in theory, indicate Internet impacts.

Africa is faced with a scarcity of financial and human resources needed to take advantage of this potential and has an unfortunate recent history of lack of economic progress. Maximizing the beneficial economic impacts of the Internet there will require not only introduction of the technology in millions of enterprises but also the training of millions of users and institutional reforms. Neither "parachute drops" of computers and software or showpiece computers in executive offices will do much good. To maximize benefits, the scarce resources will have to be used to introduce the Internet to reduce key bottlenecks to economic progress. Connectivity will have to focus on key enterprises and regulatory agencies, on key market constraints, and on key trade and commercial organizations, and the evolving pattern will have to create synergies among applications. Some of this orchestration can be planned by governments. Governments in other developing countries such as Malaysia and India have recently taken steps to meet long-term development needs through information technologies by joining with private sector interests to design incentives and regulations to support investment in the sector. But to achieve the full potential of the Internet, decentralized decisionmaking on its acquisition and use will have to be encouraged and done well and in ways that benefit the enterprises as well as the total economy. Data on the impacts of introduction of the Internet will have to be available then not only to government decisionmakers, but also to the large numbers of people making these decentralized decisions. Ideally, data should be sufficient to inform decisions on the priority for allocation of resources among alternative applications of the Internet. To achieve this, top government officials must be much more proactive in extolling the economic benefit of, for example, electronic commerce, to their populations.

Negative economic impacts of the Internet are a real possibility and must be monitored carefully. The Industrial Revolution brought not only great benefits to the world but also great suffering to millions, and it is increasingly clear that the impacts were not intrinsic to the inventions per se but contingent on the way in which they were disseminated and used. The impacts were contingent on the policies of governments and economies. Certainly, most of Africa has fallen behind in other technologically based revolutions, and history suggests that one must consider the possibility that African countries will lose even more market share in international markets to countries that overcome distance or resource constraints by using the Internet more effectively.

Indeed, one must consider the degree to which foreign enterprises and governments may utilize their greater mastery of the Internet to put Africa at a further disadvantage. The Economic Commission for Africa recognizes these risks and benefits in its Africa Information Society Initiative, which urges governments to get actively involved so as not to be left further behind. Moreover, the transfer of the Internet from the United States, Europe, and Japan will entail transfer of parts of developed countriesí economic culture to Africa. These cultural aspects may be all but invisible as they occur, but the impacts on African economies and cultures may be important. It seems likely that the economic impacts of the Internet in Africa will be powerful and will fall unevenly on the continentís population. Those impacts will be contingent on the policies adopted by African institutions in each country. It seems important therefore to monitor the economic impacts of introduction of the Internet in real time, using the information provided by these indicators analytically to guide policy and to monitor the functioning of institutions. While it will be easier to monitor expected and quantifiable economic impacts at the level of the firm or enterprise, one must not miss the forest for the trees.

Government and Civil Society

Governments, political parties, and other political entities can use the Internet to disseminate information, facilitate government responsiveness to citizens, and increase (or hinder) access to information by the entire electorate. Citizens and interest groups can use the Internet to facilitate open debate and discussion and increase their participation in governance.

Relevant indicators include the:

• number of ministries/departments with a presence on the Web
• number of ministries/departments with e-mail reply addresses on the Web
• quality of Web site content in the above-described classes of Web sites
• percentage of ministries/departments who use it for dissemination of information about governmental actions or policies
• number of political parties with a presence on the Web
• Internet access to government policy papers and pending and existing legislation and regulations
• number of organizations using Internet networks, user groups, etc., to influence government
• number of list servers, news groups and conferences holding on-line discussions of public policy issues
• number of NGOs with Internet access
• number of publicly-available sites with free or low-cost Internet access, such as kiosks, post offices, community centers, or libraries

The Internet can also increase (or decrease) the transparency of an electoral process. For example, in Ghana, Network Computer Services monitored its Internet traffic into and out of the country during the recent elections. NCS reported that, whereas the normal ratio of incoming to outgoing traffic is 4:1 or 5:1, it was 2:1 or 3:1 during the elections. NCS attributed this increase in incoming traffic to people in other countries seeking information about the elections.

The indicators above generally are intended to measure actions by governments that are considered "positive" in terms of democratization and civil society building. But it is also probable that some governments, particularly authoritarian ones, will take actions that would be considered negative by Western democratic standards, such as imposing sanctions on Internet access for political reasons. Indicators specifically designed to measure such negative actions might be developed.

Related to government and citizen use of the Internet is the mass media. Effective media, independent from government and lobby pressures, can contribute to effective governance and civil society. Relevant indicators are the:

• number of independent sources of information and news provided via the Internet
• number of newspapers, radio stations, and TV stations using the Internet to collect news
• number of newspapers, radio stations, TV stations, and other media with Web sites
• percentage of domestic and foreign readers

When considering the government sector, development goals center around democratization and civil society building in addition to the services governments typically provide, such as health services, law enforcement, and so forth. Democratization implies a transparent political process and the participation of informed citizens in the process of governance, especially in the election of governmental leaders, development of government policy, and oversight of governmental executive and judiciary functions. Clearly, it is possible to conduct sample surveys to ascertain if people believe that participation is increasing or decreasing because of Internet diffusion, whether it is more or less informed, and whether these functions are better or more poorly conducted now than in the past and, further, to discover the degree to which introduction of the Internet is believed by the public to be responsible for changes or for maintaining the status quo. According to the criteria used for selection of persons to be included in the survey, different indicators would be developed. Thus, a survey of government bureaucrats might result in different opinions than a survey of social scientists, politicians, or citizens at large.

Inasmuch as the Internet has the potential to help a government exercise legitimate authority, it can also help illegitimate forces undermine authority. Also, what information is provided or used and to or by whom is important. Whereas a government offering information on a proposed law and providing a venue for citizen input is seen as a positive impact on democratization, an authoritarian government providing or "bombarding" the public with false information or propaganda designed to further its own aims would be considered a negative impact.

Comparably, the development of civil society involves the development of a network of noncommercial, nongovernmental organizations, with memberships of informed and active people, that are able to articulate the positions of their members to the public at large, to elites, and to the various branches of government. Surveys of informed persons could be used to obtain data on whether the Internetís impacts are perceived as beneficial or detrimental and on the magnitude of the impacts of the Internet on direct communication between government and citizens, media-intermediated contact between citizens and government, and on the role of civil society organizations in promoting public awareness and participation in government decisionmaking.

The Internet is a critical element of an information and communications revolution that is likely to prompt profound changes in interpersonal, intergroup, and interinstitutional relationships in Africa. Whether the overall impacts of the Internet are to stabilize or destabilize governance, to increase participation or to support coercive efforts, will be crucially important. Yet the complexity of the network of effects of the Internet on society and the ramifications of these effects through policies and institutions are extraordinarily complex. Ultimate judgments may have to be made about the impacts of the Internet, and these without doubt will be made by the population, using qualitative techniques. Such would be the ultimate indicator of the impacts of the Internet on democratization and civil society.

Having offered indicators for various aspects of supply and demand for the Internet, in this chapter, the committee deviates from indicators slightly to discuss how the Internet is diffused within countries and sectors. Diffusion processes have been studied in a number of circumstances, and there is some understanding of the diffusion of the Internet in developed countries. Are there key differences between the diffusion process in developed countries and that in Africa? Can models of diffusion or paths of impact within countries be developed to describe actual and anticipated patterns?

In the United States, scientific and technological organizations led in the development of the Internet and its earlier networking versions. The military developed computer networking early to provide cross-networking for military information in the event of nuclear attack; early users of alternative computer networking protocols included airlines, the travel industry, banks, and others in the financial services industry. The academic sector quickly saw the advantages that the Internet offered in networking and information exchange, and growth of Internet use in the academic community has grown steadily in the 1990s. Commercial use of the Internet in the United States did not occur until the restrictions on commercial use were removed and the National Science Foundation moved to privatize Internet services. But commercial domains have now come to dominate the Internet, and firms involved in business-to-business transactions dominate commercial-sector use (as opposed to firms using the Internet to communicate with private consumers).

The committee found that in Senegal the introduction of electronic mail came through international development organizations; but, unlike in the United States, the academic sector to date has not played a large role in further development of the Internet. Internet diffusion in Ghana initially has largely been driven by the local private sector, foreign companies, and development organizations. In both Ghana and Senegal, individual use is far less common than organizational and institutional uses, largely because many individuals lack the capital for a telephone line and computer. Thus, what is evident today is a high rate of users in nongovernmental (NGOs) and international organizations users as compared with domestic academic and commercial users.

In the United States the development of the Internet seems to have been driven by key technical changes, including the development of e-mail and file transfer protocols, development of the World Wide Web, and development of browser software. In Africa the lag in diffusion of Internet technology has reduced the immediacy of the impacts of these innovations. The Internet is much more affordable in developed countries than in Africa, since not only are prices lower but per-capita incomes are much higher. Therefore, diffusion of the Internet in Africa appears likely to lag that in developed countries and to be much more limited to economic elites than it has been in the United States. Given the economic status of foreign expatriates in Africa as compared to Africans, and given the experiential advantage that earlier introduction of the Internet in their home countries provided the foreign expatriates in Africa, the diffusion process also appears much more likely to provide comparative advantage to foreigners in Africa than was the case in the United States. General diffusion theory suggests that leaders in the adoption of a new technology appear to be better educated than average, and data for the United States suggest that education is a predictor of Internet use. Developed countries are much more homogeneous educationally than are African countries, and diffusion of the Internet in Africa can be expected to be slower in part because most citizens have had little schooling.

In the United States the initial emphasis was to stimulate communication within the country, while the fact that developed countries are now heavily networked as Africa begins the process suggests that the effect of the Internet in Africa may be to increase international communication rather than internal communication. The fact that international telephone connections are often of far better quality than domestic ones in Africa will strengthen this tendency toward increased international communication relative to internal communication. Similarly, the availability of significant digital collections of information outside Africa might tend to increase reliance on external (foreign) sources over indigenous sources.

One might also look at the Internet path of impact within an institution or organization. For example, a typical pattern of Internet diffusion in a U.S. organization may be:

an enterprise acquires Internet access ñ>

communication within the enterprise improves ñ>

the enterprise talks to its clients via the Internet ñ>

the enterpriseís productivity increases

But the Committee research suggests that in Africa interorganizational communications may precede intraorganizational communications. That is:

Field research by the committee in Senegal and Ghana suggested a similar pattern for diffusion, particularly for NGOs:

an organization acquires Internet connectivity à

communication with (foreign) counterpart organizations improves and increases à

organization-to-client and organization-to-organization communication increasesà

sophistication of use increases à

the organizationís activities increase and become more effective à

communication with (domestic) sectoral colleagues improves and increases

Under this latter model, connectivity is established by enabling individuals within an organization to communicate more effectively with similar organizations, often in other countries. An example of this would be increased communications, and hence collaboration, between NGOs working on health improvements speaking with other NGOs, both domestically and with sister organizations internationally, working on the same issue. Following the model, that organization would then increase communications with foreign funders and beneficiaries of its services. This would be followed by increasingly more sophisticated use and by the NGO increasing its activities and effectiveness. For example, the NGO will begin to use the Internet for more than just electronic mail, perhaps designing its own Web page, providing information about its activities, and so forth. Communication with other domestic NGOs in the same sector follows.

Within a country there are certainly a variety of models or paths that can be described. For example, while appropriate for NGOs, this model may be less accurate when describing the government sector or the private sector. The committee found that, among government ministries in Ghana and Senegal, communication with other governmental organizations is likely not the first use of the Internet.

Clearly, more study is required to develop any diffusion models or to suggest patterns in the paths of impact. The committee believes that diffusion and impact pathways are topics requiring significant further research, as noted in the next chapter.

In the preceding chapters the Committee addressed the issue of analyzing impacts of the Internet. The discussion began with Internet suppliers and users, and the environmentñsocial, technical, and regulatoryñin which the Internet is supplied and used. Quantity and quality of Internet services and content were discussed. The report then looked at some of the potential impacts of the Internet on such organizations as businesses, government agencies, or private voluntary organizations. Finally, the discussion proceeded to the impacts of the Internet on the education, private, and government sectors and on the developmental goals central to those sectors.

In their research in the United States and their many site visits and interviews in Africa, committee members were impressed with the recognition among Internet users and providers and policymakers of the need for a better analytical framework within which introduction and use of the Internet can be studied. Even those individuals in Ghana and Senegal who were just beginning to use the Internet in their organizations were looking for a way to evaluate its utility.

The committee members found that much can be determined about Internet use in a country and by institutions and individual users. The information collected in Ghana and Senegal, and to a lesser extent Kenya, is a sampling of the available data and anecdotal evidence of the Internetís impacts, perceived benefits, and unexpected consequences.

Some measurements are very straightforward and can be monitored regularly with very little effort. For example, an Internet Service Provider (ISP) can easily use a proxy server to monitor the amount of Internet traffic and the source and destination of that traffic. Other measurements would require a significant amount of effort. Any interested organization or individual must decide what information is needed and which measurements would best serve his or her needs, and ISPs and other organizations that have the data must be convinced of the need for steady monitoring and sharing of data at least in part with researchers.

The committee believes that this report is a first step in providing an analytical framework for evaluating the Internetís impacts. Much work remains to be done in the topics included in this report, in identifying indicators for other sectors and other uses, and of course in-country coverage for testing the indicators. A number of researchers and research organizations are studying various aspects of information technologies and their effects in both industrialized and developing countries. As this body of research grows, there is a need to share the data with those conducting the research and those who might be impacted by it. To this end, the committee suggests that a standing international network of research and policy formulation would be useful. Additional areas beyond the scope of this report that deserve further study are noted below.

• Causality. As noted in Chapter 2, the use of indicators is complicated by the difficulty in proving a causal relationship. Some direct impacts of the Internet are very specific and relatively easy to measure, but as one moves to more and more indirect relationships (e.g., increases in productivity and speed of work) external influences increase and causality becomes increasingly difficult to ascertain. In such cases, ascertaining causality will require minimizing external influences and using surrogate indicators. While indirect impacts may be difficult to identify, measure, and modify, they may also be of great social and economic importance; indeed, the more global the impact, the more indirect and difficult it is to measure. These difficulties of determining social impacts are not unique to the Internet. A study commissioned by the Rockefeller Brothers Fund found similar difficulties for other media in other regions. More research is needed in both the methodology and application of such techniques.

• Communication and development. This report began by stating that the committee considers the Internet to be a mechanism for conveying information and that the framework or indicators presented here are intended to help analyze the impacts of its introduction and use. But how does communication per se lead to development? The links between communication, decisionmaking, and development are contained implicitly in much of the above discussion. They have also been the topic of related research. But more can be done to understand more fully the impact of information and, by association, information technologies on decisionmaking and development.

• Country-specific case studies. The committee very strongly recommends an in-depth, longitudinal case study approach to compliment the approach used in this report to study the impacts of the Internetñthat is, quantitative and qualitative indicators. A study over time of individual and institutional users and the benefits and limitations they anticipate, perceive, and realize would provide additional useful information on Internet use and diffusion. The committee is aware of some efforts of this type (e.g., work by the Network Startup Research Center) and believes the approach should be further employed.

• Institutional use and diffusion of the Internet. This report offers a preliminary model for institutional use and diffusion of the Internet among sectors. However, much work remains to be done in this area. Is there a "typical" path of Internet diffusion within countries? Is it dependent on the work of one or very few strong individuals, or do gross national product and other environmental factors play a larger role? What is the effect on diffusion of privatization of the Internet and removal of limits to commercial traffic? Also, inter-networking was an indigenous development in the United States and to a lesser degree in Europe, but largely a transferred technological system in Africa. What effect has this had? These and other questions related to diffusion merit further research.

• Cost of the Internet. Organizations in Africa will find that the monthly fee to an ISP is only a part of the cost of the Internet, indeed perhaps a small part of the cost of using it effectively. In addition to hardware and software costs and others noted in previous sections, effective use of the Internet entails support costs, which in the United States include help lines and information officers in the organization. In fact, little is known about the true costs to organizations of the Internet. ISPs have complex pricing policies and are reluctant to disseminate information on income from clients. Typically accounting processes in firms do not provide reports on costs (or benefits) of Internet use. (Indeed, measuring the contribution of information technology to productivity is a major challenge for economists today.) Development of methods would be needed to make such reports possible. Yet in Africa the costs of the Internet will be critical in determining the rate of diffusion of the technology and the priority that should be assigned to it. Case studies of the costs of Internet use in typical enterprises, NGOs, and government agencies should be given priority.

• Use versus nonuse of the Internet. An area requiring additional research is the effects of growth of the Internet on nonusers. Penetration of the Internet is occurring so quickly and is already so great that it will affect even those who cannot or choose not to connect. People in countries who do not have access to the Internet when large numbers of others do or who live in a country that does not have the Internet when competing nations do will feel certain effects of that nonuse.

• Public policy and Internet development. A topic the committee finds to require further research is the role of public policy and what policies seem to work best in supporting the development of the Internet. The rush of new communications technologies is forcing governments, employers, and users to alter their perceptions and policies on a host of public policy issues. For example, increased attention is being paid to issues of privacy and the boundaries between public and private information, and the shift away from governments as producers and owners of broadcasting and telecommunications and sometimes the press.

Ernest J. Wilson III (Chair) is Director of the Center for International Development and Conflict Management and Associate Professor of Government and Politics and Afro-American Studies and a Faculty Associate in the School of Public Affairs at the University of Maryland. Previously he served as Deputy Director, Global Information Infrastructure Commission, and has worked on information issues at the U.S. Information Agency and the White House. He was also an expert advisor to the African Information Society Initiative of the Economic Commission for Africa. Dr. Wilson is the author or editor of several books and studies on the political economy of international oil, the global information revolution, and economic reform in developing countries. His recent report, prepared for the Center for Strategic and International studies (CSIS), "The Information Revolution Comes to Africa," also combines his interest in Africa and the information revolution. Wilson has degrees from Harvard University and the University of California at Berkeley. His web site can be found at http://www.cidcm.umd.eduwilson/.

John A. Daly is a self-employed consultant working on issues of science, technology, and development. He served for 25 years with the U.S. government foreign assistance program, including a term as director of the Office of Research of the U.S. Agency for International Development. His overseas experience includes several years with the World Health Organization in Colombia, and several years in Chile as a Peace Corps Volunteer and Ford Foundation Consultant. In the 1960s he worked as an engineer in computer related research in the aerospace industry. Dr. Daly has maintained a long interest in the applications of information and communications technology to development. He holds a Ph.D. in Operations Research, Public Administration from the University of California at Irvine and an M.S. in Electrical Engineering from the University of California, Berkeley.

José-Marie Griffiths is Chief Information Officer of the University of Michigan, and Executive Director of the Information Technology Division. She has overall responsibility for all aspects of information technology, including telecommunications, networking, computing and infrastructure. Additionally, she is a full professor in the School of Information. Dr. Griffiths is a Commissioner of the United States National Commission on Libraries and Information Science (NCLIS), nominated by President Clinton and confirmed by the U.S. Senate in 1997. Previously, she was at the University Tennessee-Knoxville where she filled several positions, including Director of the School of Information Sciences. Dr. Griffiths has extensive experience in information systems and networking. Her degrees include a Ph.D. in information science and a B.Sc., with honors, in physics. Both degrees are from the University College of the University of London.

AAAS American Association for the Advancement of Science

GNP Gross national product

IP Internet protocol

ISP Internet service provider

LAN Local-area network

NCS Network Computer Services

NGO Nongovernmental organization

NRC National Research Council

POP Point of presence

PTT Post, Telephone and Telegraph (administration)

USAID United States Agency for International Development

Supportive Economy and Infrastructure

• GNP per capita
• number of telephones
• number of telephones per capita
• indicators of penetration of telephone service in rural areas
• indicators of penetration of electrical power in rural areas
• density of population in rural areas
• percentage of population in urban areas
• indicators of the strength of markets for personal computers, modems, and related technologies

Policy and Regulatory Environment

• estimated cost, time, and rate of success in establishing an ISP
• estimated cost, time, and rate of success in establishing an ISP account
• nondiscriminatory access to Internet service
• modem and/or computer tariff
• waiting time for a telephone line
• cost for installation of a telephone line
• waiting time for a leased line
• cost per minute to access points of presence (POPs)
• commercial availability of modems and computers
• local service for modems and computers

• total number of ISPs
• total bandwidth to outside country (kilobytes/second)
• total number of modems connected to ISP servers for dial-up access
• total number of leased lines to customers
• total number of POPs
• total number of secondary-city POPs
• percentage of population within local calling area of POP

Quality of Internet Service

• percentage of send failures (messages that fail to reach their destination)
• average delivery time of e-mail/data transfer from each ISP to every ISP
• average delivery time of messages
• average time to check an empty mailbox
• mean connect speed of subscribers
• call failure rates for ISPs (the percentage of calls that fail to connect to the Web)
• number of members in an information industry association or ISP association
• number of ISPs offering full Internet service
• percentage of nonprofit ISPs
• number and percentage of profitable ISPs
• prices charged by ISPs for Internet access
• total funds invested by ISPs in expansion
• total ISP revenue

Sustainability

• number of foreign- and domestic-owned ISPs
• number of local technical staff
• number of ISPs offering user training
• number of institutions that monitor their own traffic, use, and number of hits on pages
• average number of years of schooling of adult population
• literacy rate
• number of information technology courses offered in universities
• average salary of Web designers and other ISP employees
• percentage of ISPs offering Web hosting, Web design, and other services
• ratio of national, regional, and international traffic to total traffic (both coming into and going out of the country)
• number of home pages on domestic servers
• ratio of national, regional, and international participation in listserves and news groups

• total number of subscribers by category of user
• average number of workstations per subscriber
• average number of people with access per workstation
• rate of change in the number of subscribers
• turnover rate
• total traffic (kilobytes per day)
• total connect time per day
• total number of e-mails per day
• average subscriber connect time
• average subscriber connections per day
• number of subscribers using leased lines
• Internet use for:
• communication
• downloading software
• interactive discussions
• noninteractive discussions
• use of another computer
• real-time audio or video
• searches for product/service information
• purchases based on Web information
• searches for company/organization information
• searches for other information on the Web
• browsing/exploring
• seeing what is new at a favorite Web site
• business purposes
• percentage of users who connect from their own homes
• percentage of users who connect from an office
• percentage of users who connect from both home and office
• percentage of users who connect from an Internet cafe or business center

Costs of Internet Use

• price elasticity of demand
• fees paid to an ISP for leased-line, dial-up service, and other services
• installation fee(s)
• fees (fixed and/or usage dependent) paid to the telephone company
• price of a phone call per minute to connect with the ISP
• costs to the organization of Internet training courses and staff salaries paid during training

Perceived Benefits of the Internet

• number of messages/transactions to/from/by an organization per day that are domestic versus regional versus international in source or destination
• reported relative importance of the Internet versus other means
• cost savings on communications
• time savings on communications
• ratio of Internet to other channels in obtaining information
• percentage of an institution's dissemination through the Internet
• percentage of total public information made available through the Internet
• number of Web server hits or requests fulfilled per month from domestic versus regional versus international sources
• number of electronic newsletters or bulletins produced
• number of subscribers to newsletters and/or bulletins
• number and percentage of people trained in using the Internet
• number and percentage of subscribers with a LAN
• number of top/middle/lower-level users in an organization with access to the Internet
• relative importance placed on the Internet by top/middle/lower-level staff
• number of networks and "virtual organizations" of which an institution is a member
• investments in computer and other telecommunications facilities
• approximate number of users who (1) use e-mail, (2) "surf," (3) maintain own (individual or organization) home page, and (4) use an Intranet
• presence of a distinct information strategy as part of an organizationís overall organizational strategies and plans

Organizational Decision Making

• change in number of people involved in an institutionís decisionmaking
• relative importance of the Internet versus other means of gathering data and information in decisionmaking

Sectoral Use and Diffusion of the Internet

• total number of subscribers per sector
• increase in the number of subscribers per sector
• percentage of Internet use per sector for (1) e-mail, (2) "surfing," (3) maintaining a (individual or organization) home page, and (4) use of an Intranet
• number of subscribers in primary city and secondary cities

• number of schools/universities with Internet access
• number of students with Internet access
• average time of student access
• number of teachers with Internet access
• number of training courses on the Internet offered to teachers
• quality of training courses on the Internet offered to teachers (accreditation)
• number of new courses offered since the Internet was introduced
• number of schools/universities utilizing distance education via the Internet
• number of courses that supplement conventional teaching methods with distance education or other Internet-dependent technologies
• number of students enrolled in distance education
• number of nonuniversity institutions offering distance education
• ratio of job placement of students with Internet experience/training in school to overall placement
• ratio of average starting salaries of individuals with Internet experience/training in school to overall starting salaries
• number of scholars/researchers attracted to a university/country (in part) because of Internet access

• rates of participation of African firms in international markets
• rates of participation of foreign firms in African markets
• numbers of Web pages providing information on a market
• numbers of persons communicating about a market on the Internet
• volume of transactions in a market using the Internet
• number (percentage) of chambers of commerce with Internet access
• number (percentage) of other business organizations with Internet access
• number of small and medium-sized enterprises with Internet access
• number of small and medium-sized enterprises posting products and prices on the Internet
• rate of change in the value of an enterpriseís exports (imports) since acquiring Internet access

• rate of change in the value of a countryís exports (imports) since acquiring Internet access
• rate of change in the value of a firmís exports since acquiring Internet access
• number of companies reporting growth since availability of the Internet
• number of firms engaged in electronic commerce
• value of sales via the Internet
• funds allocated by private companies to Internet-related training
• growth rates of private telecenters that provide Internet services

Government and Civil Society

• number of ministries/departments with a presence on the Web
• number of ministries/departments with e-mail reply addresses on the Web
• quality of Web site content in the above-described classes of Web sites
• percentage of ministries/departments who use it for dissemination of information about governmental actions or policies
• number of political parties with a presence on the Web
• Internet access to government policy papers and pending and existing legislation and regulations
• number of organizations using Internet networks, user groups, etc., to influence government
• number of list servers, news groups and conferences holding on-line discussions of public policy issues
• number of NGOs with Internet access
• number of publicly-available sites with free or low-cost Internet access, such as kiosks, post offices, community centers, or libraries
• number of independent sources of information and news provided via the Internet
• number of newspapers, radio stations, and TV stations using the Internet to collect news
• number of newspapers, radio stations, TV stations, and other media with Web sites
• percentage of domestic and foreign readers

There are several alternative systems for communication of digital data, and systems using alternative formats predate the Internet in Africa and continue to provide service to specific classes of customers such as airlines and banks. In the past several years, Internet connectivity in Africa has spread to over three-quarters of the countries and thousands of users. A recent status report states that as of November 1997, 47 of the 54 African countries had some form of Internet access and that only five countries remained without full access or any known plans for achieving full access. The U.S. Agency for International Development and other Western development agencies have allocated millions of dollars to increasing and supporting connectivity, and an increasing number of commercial enterprises, both Western based and indigenous, are providing Internet services to businesses and individuals.

Internet access is available in most African countries. However, the relatively poor telephone infrastructure in Africa is a limiting factor to Internet growth, and, as a result, in many cases connectivity is limited to capital cities. Research indicates clearly that the relative absence of modern technologies in Africa results from a variety of factorsñfinancial constraints, personnel constraints, weak basic telecommunications infrastructure, and especially public policies that too often preempt rather than promote new information technologies and the services they can provide. The relatively high telecommunications costs and small markets have resulted in high access charges compared to Western countries (relative to gross national product (GNP) per capita or to average monthly salaries), which in many countries limits Internet use to the elite few. Of course, new technologies initially demanding a high price is a phenomenon not unique to African countries or to the Internet. Moreover, prices do vary from country to country, and as more providers enter the market, Internet use likely will spread among all sectors of the economy and access charges should drop.

The committeeís research confirms the importance of policies that promote information technologies and strong institutions with experienced capable people. In Senegal, Ghana, and Kenya the committee heard from Internet service providers (ISPs) who were able to decrease access prices and improve service as a result of recent government policy reforms.

While Africa as a whole can be considered to have a relatively poor telecommunications infrastructure compared to the United States or Western Europe, differences certainly exist among African countries and cities, and marked differences exist between major urban areas and the rest of the continent. The following sections provide some country-specific information on Senegal and Ghana, the two countries visited by the committee in August 1997. Boxes 1 and 2 give a brief summary of the history of Internet development in Senegal and Ghana, respectively.

Senegal is located on the west coast of Africa and has a population just under 10 million. It is one of Africaís oldest democracies and has an independent press and judiciary and a flourishing private sector. Its economy, which is dominated by agriculture, is showing signs of improvement after years of stagnation and falling GNP per capita. A devaluation of the currency in 1994 and a new government reform program have led to growth in both GNP and GNP per capita, decreased inflation, and increases in agricultural production. In 1995 the GNP per capita was $570.

Senegal has received loans and other financial assistance from the International Monetary Fund and the World Bank. Banks and other institutions in France, Japan, and the United States have provided significant foreign direct investment. The government is gradually reforming restrictive import policies, increasing privatization, and dismantling monopolies to encourage investment, increase competition in the marketplace, and improve the countryís economic infrastructure.

SonatelñSenegalís Post, Telephone, and Telegraph administrationñis a government-owned agency that holds a monopoly in the telecommunications sector. Sonatel provides a link to the global Internet via a satellite circuit to the United States. It has established Telecom-Plus as its value-added provider, including Internet service. Telecom-Plus operates a full dial-up and leased-line Internet service and leases circuits to other organizations, which can then sell Internet access themselves.

At the time of the committeeís visit, there were nine providers, including commercial providers and a number of nongovernmental organizations (NGO) and other constituency-based providers. The most commonly stated estimate of subscribers in Senegal was 2,000, approximately half of whom obtain their access through Telecom-Plus. An estimate by Telecom-Plus representatives put the annual growth of total Internet subscriptions in the country at 10 percent, but this figure likely varies greatly over time and among ISPs and likely does not accurately reflect the different modes of Internet use, such as dial-up access and local-area networks with multiple users. The average price of unlimited Internet access per month as of August 1997 was 10,000 CFA, or approximately $20.

There are no formal legal or regulatory barriers to other ISPs entering the market in Senegal, though some ISPs complain of Sonatelís favoritism to Telecom-Plus. Two other common problems noted by ISPs were the poor quality of the telecommunications infrastructure (i.e., the phone lines) and the difficulty of finding employees with the required technical expertise.

Of particular interest in Senegal is the availability of Internet access from anywhere in the country at the cost of a local call. Sonatel has established a special number service that can be used by any Internet service provider, encouraging the use of its service and eliminating the need for local points of presence in other cities. This makes up in part for the lack of direct Internet links between the countryís 10 regions.

In most cases, providing Internet access is not (yet) profitable, nor is providing it the primary activity of most ISPs. Rather, they cover their Internet operating costs with fees for Internet access and make a profit through other (usually) telecommunications products and services. ISPs in general seem to have the technical expertise to track some use information, and in fact at least some do track information on traffic coming into and going out of the country, though they do not know what percentage of that is simple electronic mail versus the transfer of files and other information flows. One organization the committee met with reported monitoring the source of hits to its Web page.

Ghana, located on the west coast of Africa, has a population of approximately 18 million. It is an agricultural country and its economy has been based on the export of such primary products as cocoa, timber, and gold. After gaining independence in 1957, Ghana went through a period of state-controlled development during which the economy declined significantly and the countryís infrastructure virtually collapsed. International assistance programs and the economic recovery program initiated by the Ghanaian government in 1982 have helped stabilize and liberalize the economy and reduce rural poverty. The business environment and private sector are growing. Today, Ghana has a democratic government and free press. In 1995 the GNP per capita was $390.

Ghanaís infrastructure, particularly the telecommunications infrastructure, is still far below Western standards. The countryís telephone network has less than 100,000 phones for a population of nearly 20 million, although the network has undergone an expansion since 1994 and is developing further with the appearance of a new carrier. The government has recently encouraged such development, and in 1995 it adopted policies to encourage private participation in the telecommunications sector.

  Despite its poor telecommunications infrastructure, Ghana has a relatively well-developed Internet sector compared to other African countries, due in part to a large number of international assistance projects, the shift toward policies more open to local and international providers, and the leadership of strong individuals.

The number of Internet subscribers in the country as of August 1997 was estimated at 3,000, with about 2,000 getting their access through one provider, Network Computer Systems (NCS). Also as of that time, there were three full Internet service providers offering access to the World Wide Web, FTP, and other Internet servicesñNCS, Africa Online, and Internet Ghana. NCS was planning to establish points of presence in regional capitals, starting with Kumasi, Takoradi, and Tamale and was also pursuing satellite technologies.

Ghanaís ISPs are sophisticated providers, able to project and monitor traffic, although currently the focus seems to be less on tracking and more on securing the necessary technical equipment and expertise and expanding the user base. According to some individuals in Ghana, the driving force behind Internet development in the country is the private sector, with the government playing only a limited role. The Ghanaian government does not have any concrete plans to interconnect all the government ministries, nor is the Internet used in daily operations. But nor is the government trying to control or limit Internet use, and the telecommunications sector in general is much more liberalized in Ghana than in some other African countries. Several government officials noted to the committee the need to develop an overall policy framework for Internet regulation and use.

  It is clear to the committee that Internet use in both Senegal and Ghana is growing noticeably in most sectors. Individuals, both providers and users, reported frustrations with infrastructure, quality of service, and technical expertise but were also optimistic that the difficulties were temporary. Many individuals were particularly interested in building in to their activities a means of studying the impacts of the Internet during this early stage of its development. The committee believes this work is especially timely because many African governments are liberalizing their telecommunications sectors and hence expanding opportunities for information technologies to impact society. Creating empirical baselines early in the diffusion process is important to serve both policy and analytic purposes.

In each country that the Committee visited, it met with representatives of Internet service providers (ISPs), both commercial and constituency-based; nongovernmental organizations (NGOs) involved in development-related activities that use the Internet; representatives from a university or other higher-education institution; private-sector Internet users; government officials; and others who are involved in Internet-related activities. The complete list is below. In most cases the site visits and meetings consisted of informal discussions based on the general questions outlined in the Introduction. The committee members considered, but did not follow, a predetermined, specific line of questioning or a prepared survey. The committee recognizes that further research on this topic would benefit from such complementary research tools.

Tse was managing director of Africa Online Ghana (see above) until August 1997.