WSIS Thematic Meeting on Measuring ICT Indicators for Development 7-9 February 2005, Geneva

3 February 2005 Original: English

International Telecommunication Union CORE INDICATORS: BASIC ACCESS AND INFRASTRUCTURE1

Ensuring universal service and access to information and communication technologies (ICTs) is a top national objective in many countries, often enshrined in laws that govern the sector.2 Despite this, few governments presently track accessibility or connectivity on a regular basis. Those governments that do measure and monitor access, do not always use the most appropriate indicators. Furthermore, given the different approaches taken by different countries, the different indicators used worldwide are not always compatible. These factors have made it difficult to measure ICT development accurately and to elaborate targeted plans for enhancing access. Access to ICTs can be quantified in various ways, with indicators based on different categories. For instance: Individual. Indicators that measure accessibility in terms of people. This includes indicators such as fixed telephone lines per 100 inhabitants or the percentage of the population that uses the Internet. This also includes spatial indicators that measure accessibility in terms of coverage or distance from ICT facilities. Some indicators in this category are useful for tracking universal access, or the percentage of the population that could theoretically use an ICT device or service (e.g., mobile cellular coverage). Per capita measurement is the traditional method of illustrating individual access to ICTs. One reason for this is that virtually all ICT service providers compile administrative records for operational and billing purposes. It then is a simple mathematical exercise to divide an ICT device or service by the population to derive a per capita indicator. Household. Indicators that measure the availability of ICTs in the home such as the percentage of households with a telephone. Household measurements determine the level of universal service. Business. Business use of ICTs is particular important in understanding the evolution of the information economy and the diffusion of innovations, such as e-commerce, more widely in society. Through the work of the partnership for Measuring ICT for Development, which includes ITU as well as other international and regional organisations (such as UNCTAD, UIS, OECD and the Regional UN Economic Commissions), it is planned to develop a core set of indicator for tracking the development of the Information Society. This document focuses on core indicators of infrastructure and access as they correspond primarily to individual use, including individual use of community facilities. The proposed definitions correspond to the basic core indicators A1 to A10 and to the extended core A11 and A12, as they appear in the draft core set. The full list is provided in Table 1.

Table 1: Proposed core set of indicators for infrastructure and access

Basic core A-1 A-2 A-3 A-4 A-5 A-6 A-7 A-8 A-9 A-10 Fixed telephone lines per 100 inhabitants Mobile cellular subscribers per 100 inhabitants Computers per 100 inhabitants Internet subscribers per 100 inhabitants Broadband Internet subscribers per 100 inhabitants (fixed and mobile) International Internet bandwidth per inhabitant Percentage of population covered by mobile cellular telephony Internet access tariff (20 hours per month), in US$, and as a percentage of per capita income Mobile cellular tariffs (100 minutes of use per month), in US$, and as a percentage of per capita income Percentage of localities with public Internet access centres (PIACs) by number of inhabitants (rural/urban)

Extended core A-11 A-12 Radio sets per 100 inhabitants Television sets per 100 inhabitants

A1. Fixed telephone lines per 100 inhabitants

ITU has been publishing data on telephones since the 19th Century, and since 1972 this has been formalised in the annual Yearbook of Statistics. The indicator has evolved with market trends and technological development. Initially, the number of telephone sets was compiled. This proved less useful as an indicator over time given the increasing number of telephone sets in the home, or attached to private branch exchanges (PBXs) in companies. Also, liberalization of equipment markets in many countries allowed consumers to choose their own sets, which meant that incumbent operators no longer knew how many there were. This led to a preference for lines in operationalso referred to as main or direct exchange lines (DELs) as the primary indicator for measuring telephone access. A fixed line has traditionally referred to the connectiontypically a copper wirefrom a subscriber to the telephone companys switching exchange. Technological change has since blurred this definition. For example, in some countries, telephone service is provided via coaxial cable over pay television networks. In others, wireless local loop (WLL) technology severs the traditional concept of the physical line represented by a copper wire. The emergence of integrated services digital networks (ISDN) has also dramatically impacted the concept of the main line. ISDN converts a single physical line into virtual channels. Basic rate ISDN provides two channels while primary rate provides many more (e.g. 30 in Europe and 23 in North America and Japan).3 This led to the practice, particularly in Europe and Japan of including ISDN channels in main line statistics. More recently, the use of fixed line broadband technologies (such as Digital Subscriber Line or Cable Modems) has meant that a single subscriber connection can support much higher data connections, including data but also, potentially, several simultaneous telephone conversations (for instance, using Voice over IP) In order to enhance comparability, all countries should provide a breakdown of how their main telephone line figure is computed. While it is important to measure the total number of lines in a country, for comparable statistics, it is more conventional to express this in terms of lines per 100 inhabitants (teledensity). However, in the case of an increasing number of countries, this is not particularly meaningful now that mobile phone users outnumber fixed lines. An alternative would be to measure mobidensity, or mobile 2

phones per 100 inhabitants (see indicator A2), but this would disadvantage those economies, where the fixed-line network is well established, or others, where the mobile network is still in its infancy. . For this, and all the other core set of indicators, a working definition is indicated below:
A1. Fixed telephone lines per 100 inhabitants Fixed telephone lines refer to telephone lines connecting a customer's terminal equipment (e.g., telephone set, facsimile machine) to the Public Switched Telephone Network (PSTN) and which have a dedicated port on a telephone exchange. Fixed telephone lines per 100 inhabitants is calculated by dividing the number of fixed telephone lines by the population and multiplying by 100.

A2.

Mobile phone users per 100 inhabitants

Indicators of the usage of cellular mobile phones are critical for analysing access to telephone service given that most countries now have more mobile phones than fixed telephone lines. Mobile density, or the number of mobile subscribers per 100 inhabitants, has surpassed 100 in some nations. This phenomenon of above 100% ownership may arise due to a number of factors, including the practice in some countries of issuing car phones as standard and because of the inactive use of prepaid accounts. Statistics regarding mobile users should include both post-paid subscriptions and prepaid accounts, and should distinguish between the two.
A2. Mobile cellular subscribers per 100 inhabitants Mobile cellular subscribers refer to users of portable telephones subscribing to an automatic public mobile telephone service using cellular technology, which provides access to the PSTN. Users of both post-paid subscriptions and prepaid accounts are included. Mobile cellular subscribers per 100 inhabitants is obtained by dividing the number of mobile cellular subscribers by the population and multiplying by 100.

A3.

Computers per 100 inhabitants

Access to a computer (PC4) is important not only because is it an information device in its own right, but also because it is the leading gateway to Internet access. Furthermore, PCs are useful for developing basic computer skills prior to navigating the Internet. Despite their importance, only a few countries publish reliable data on the number of PCs. Unlike television sets, which are basically found in homes or hotel rooms, collecting data would involve surveying all the places where there might be a PC: schools, homes, offices, libraries, Internet cafs, etc. Therefore most estimates regarding the installed base of PCs are derives from data on shipments (e.g. the number of PCs sold each year) in a given country. Annual shipment data can be multiplied by an estimated average useful lifespan before replacement to obtain an approximation of the number of PCs for the country. The life of a PC will vary depending on various factors such as wear and tear and obsolescence, and replacement rates differ between developing and developed nations with the former hanging on to PCs for longer.5 Though there is no precise methodology for determining PC replacement rates, a general rule of thumb is that they are changed every three to five years. Apart from wear and tear, computers also become obsolete, as software updates require faster machines. In light of all these factors, an overall country figure for the number of PCs could be estimated by adding up the last five years sales.6 It is a major drawback that, as with so many other statistics, reliable data on the number of PCs sold is not available for many developing nations. A surrogate for sales is PC import figures, data that is sometimes available from customs departments of national governments. However, use of import data has limitations. Often, only value rather than volume data is available. Also, if PCs were assembled in the country from 3

imported parts, they would not be counted. Customs data would also not include undeclared imports. Additionally, some of the imported PCs may be later exported.

A3. Personal computers per 100 inhabitants Computers measures the number of computers installed in a country. The statistic includes PCs, laptops, notebooks etc, but excludes mainframe and mini-computers that are primarily intended for shared use, and personal digital assistants or smart-phones that have only some, but not all, of the functions of a PC (e.g., they may lack a full-sized keyboard, a large screen, an Internet connection, drives etc). Computers per 100 inhabitants is obtained by dividing the estimated number computers in use by the population and multiplying by 100.

A4.

Internet subscribers per 100 inhabitants

Since the early days of Internet use, there have been estimates of the number of users. This statistic is represented elsewhere in the core set, and is based on census data, surveys or other estimates. The number of Internet subscribers -- those paying for access to the Internet -- is a more precise indicator of access than users. Subscription suggests a certain intensity of use since it is unlikely one would be paying for Internet access unless it is being utilized regularly. It should be noted that the number of subscribers measures all those who are paying for Internet use, including those who pay via the cost of their telephone call (sometimes referred to as free Internet, those who pay in advance for a given amount of time (pre-paid) and those who pay for a subscription (either flat-rte or volume/usage based. Many Internet users obtain access without paying directly, either as the member of a household, or from work or school. There fore the number of Internet users will always be much larger than the number of subscribers, typically by a ratio of 2-3 in developed countries or more in developing ones. The number of subscribers thus sets a minimum threshold for the number of users in a country. Data about the nature of the Internet subscription, such as paid or free, business or consumer, fixed or mobile, and the access speed (e.g., dial-up or broadband) and the access device (e.g., PC, mobile phone, game console) also are useful7.
A4. Internet subscribers per 100 inhabitants An Internet subscriber is someone who pays for access to the public Internet (a TCP/IP connection). The statistic is measured irrespective of the type or speed of access, or the type of device used to access the Internet, or the method of payment. Internet subscribers per 100 inhabitants is obtained by dividing the number of Internet subscribers by the population and multiplying by 100.

A5.

Broadband Internet subscribers per 100 inhabitants (fixed and mobile)

The growing importance of high-speed Internet access means that it is worth collecting the number of subscribers in a country that have broadband access, as a subset of the total number of Internet subscribers within the country. Broadband may be defined as technologies that provide speeds of at least 256 kbit/s, where this speed is the combined upstream and downstream capacity.8 It refers to the number of active registered Internet accounts including: all fixed network broadband access technologies: Digital Subscriber Line services (DSL), Cable Modem services, Satellite broadband Internet, Fibre-to-the-home Internet access, Ethernet LANs, Fixed wireless subscribers; all wireless broadband access (at speeds at, or greater than, 256kbit/s) including, for instance, public Wireless LANs, WiMAX, Fixed wireless access (FWA) and some forms of 3G cellular network but excludes 3G subscribers.

A5. Broadband Internet subscribers per 100 inhabitants (fixed and mobile) A Broadband Internet subscriber is someone who pays for high-speed access to the public Internet (a TCP/IP connection). High-speed access is defined as being equal to, or greater than 256 kbit/s, as the sum of the capacity in both directions. The statistic is measured irrespective of the type of access, or the type of device used to access the Internet, or the method of payment. Broadband Internet subscribers per 100 inhabitants is obtained by dividing the number of Broadband Internet subscribers by the population and multiplying by 100.

A6.

International Internet bandwidth per inhabitant

Bandwidth refers to the range of frequencies available to be occupied by signals. In analogue systems it is measured in terms of Hertz (Hz) and in digital systems in bits per second (bit/s): the higher the bandwidth, the greater the amount of information that can be transmitted in a given time. High bandwidth channels (equal to, or greater than, 256 kbit/s) are referred to as broadband. International Internet bandwidth refers to the capacity which backbone operators provision to carry IP traffic measured in bits per second. This indicator is intended to represent the quality of the experience of Internet users within a country. If the experience of an Internet user in a country is poor, because of slow speed, then either people will not use ICTs, or they will not be able to use them effectively and creatively. In many developing countries, most Internet access is to sites abroad and therefore the amount of international bandwidth has a major impact on performance. In many developed countries, people visit domestic sites so that international bandwidth is not as important as last mile bandwidth. A useful derivative of this indicator or International Internet bandwidth per Internet subscriber, which takes out of the equation the differing levels of usage in different countries. One weakness of this indicator is that countries with a large domestic market (like China), or which are major generators of local content (like USA), or which have languages that are now widely used outside the country (like Japan), may have relatively low levels of International Internet bandwidth, in that domestic bandwidth fulfils most of their needs. A further weakness is that countries that much International bandwidth on leased lines is paid for unilaterally (ie, the costs are borne by users at one end of the link) even though users at both ends use link. Again, this tends to underestimate the real level of bandwidth available to users in the United States. However, this statistic is valid for all developing countries as well as for most developed ones.
A6. International Internet bandwidth per inhabitant International Internet bandwidth refers to the capacity which backbone operators provision to carry IP traffic measured in bits per second. International Internet bandwidth per inhabitant is obtained by dividing the amount of bandwidth by the population.

A7.

Percentage of population covered by mobile cellular telephony

One of the most useful indicators of universal access is the percentage of the population covered by a mobile cellular network. Inhabitants who are covered by a mobile cellular signal have the potential to subscribe to the network whether or not they actually do so. Where there is a large gap between population coverage and penetration, it suggests that bottlenecks in access are more due to affordability than to infrastructure shortcomings.

A7. Percentage of population covered by mobile cellular telephony Percentage of population covered by mobile cellular telephony refers to the percentage of a countrys inhabitants that live within areas served by a mobile cellular signal, irrespective of whether or not they choose to use it. This should not be confused with the percentage of the land area covered by a mobile cellular signal or the percentage of the population that subscribe to mobile cellular service. Note that this measures the theoretical ability to use mobile cellular services if one has a handset and a subscription.

A8. Internet access tariff (20 hours per month), in US$, and as a percentage of per capita income Affordable access to the ICTs is an important part of building an Information Society. For this reason, two of the draft core set of indicators (A8 and A9) are devoted to tariff comparisons. It is to be noted that there are hundreds of candidate ICT tariffs that could be compared, and a thorough benchmarking would require a tariff comparison basket. However, collecting a wide basket of tariffs would be too demanding for the purposes of the core set of indicators. For that reason, it is better to focus on a straight comparison of individual tariffs. The indicator includes the tariff components of monthly line rental, line usage charge and Internet access charge. However, it excludes the cost of equipment purchase or rental as well as connection charges. It may be possible for the consumer to avoid one or more of the tariff components (for instance, the cost of line rental may be waived for mobile Internet access; or the cost of line charge may be bundled with the price of Internet access for free Internet services). The package chosen would normally be the cheapest, but not necessarily the one offering the highest performance (for instance, a dial-up offering may be cheaper than broadband for 20 hours usage per month). In the longer term, it would be preferable to replace this indicator with an indicator that includes reference to the quality of the service (for instance, price per Mbit/s per month). But until broadband is more widely available, a comparison based on simple Internet access is more directly comparable.
A8. Internet access tariff (20 hours per month), in US$, and as a percentage of per capita income The Internet Access Tariff includes the tariff components of monthly line rental, line usage charge and Internet access charge, plus any tax that may be levied (as this is a service used by both residential and business consumers). The tariff chosen for a particular country would be the package for 20 hours per month that is the cheapest, that is widely available (or, in the case of regional service providers, is available in the capital city) and is available to the general public without restriction (e.g., excluding in-company or limited time offers, and excluding offers that are bundled with some other service). The price comparison is expressed in a commonly used currency (such as US$), which could be converted either at the average exchange rate, or at Purchasing Power Parity (PPP) rates. The indicator should be compared, as far as possible, for the same date between countries. As a percentage of per capita income involves dividing the Internet access tariff by the average monthly Gross National Income per capita of the country.

A9. Mobile cellular tariffs (100 minutes of use per month), in US$, and as a percentage of per capita income In addition to indicator A8 on the access price for basic Internet, it is worth also adding an indicator or voice service to the core list. There would be a strong case for taking fixed line voice. However, there are fewer subscribers of fixed line voice than mobile and, in any case, fixed line charges are already incorporated as a component of the Internet access charge. So, in the interests of future proofing the list, and of avoiding double counting, it is preferable to include a tariff comparison for mobile. The chosen indicator, comparing the cost of 100 minutes of use per month, is intended to represent an average use basket, which is applicable to individual consumers. Mobile users are split between post-paid subscribers and pre-paid accounts, and for some countries, and some operators, the former would be cheaper for 100 minutes use whereas in others the latter would be cheaper. 6

A9. Mobile cellular tariffs (100 minutes of use per month), in US$, and as a percentage of per capita income The Mobile cellular tariff includes the tariff components of monthly service rental (if relevant), 50 minutes of peak time calling and 50 minutes of off-peak calling, plus tax. Differences in the distance of calls, which may be applicable are not taken into account, nor are international calls or SMS messages. The possible one-time charge for connection is not taken into account, excepting where this is bundled into the costs of a pre-paid account. The price comparison is expressed in a commonly used currency (such as US$), which could be converted either at the average an exchange rate, or at Purchasing Power Parity (PPP) rates. The indicator should be compared, as far as possible, for the same date between countries. As a percentage of per capita income involves dividing the Mobile cellular tariff by the average monthly Gross National Income per capita of the country.

A10. Percentage of localities with public Internet access centres (PIACs) by number of inhabitants (rural/urban) The final indicator in the basic core set of infrastructure and access indicators is a measure of community access available to individuals. This is distinct from the measures available to households and to businesses measured elsewhere in the core set. The vast majority of households in developing countries do not have modern ICTs such as computers and the Internet.9 Current data suggest that community access is and will continue to play an important role in providing the citizens of most developing countries with access to ICTs, a prerequisite for participating in the information society and reaping its benefits. While developed countries tend to have lower public Internet facility access rates, some have used public facilities to spread the Internet to difficult-to-connect regions. This highlights the importance of measuring access to community ICT facilities. The question is: how? There are different ways, for example, through the inclusion of a specific question in an Internet user survey. A specific question should address the location of Internet access and allow respondents to chose between different options, such as at home, at work, at educational institutions, or at public Internet access facilities. This would provide information on the number/or percentage of people using public Internet facilities. Some countries, such as those in the European Union, have started collecting data on the number of public Internet access points.10 One major limitation with this indicator is that it does not show the distribution of facilities. Nor does it provide a basis for a recommended value since this would be a function of how necessary public Internet access centres are (which in turn depends on the underlying level of ICT ownership). To identify and agree on a set of indicators the ITU organized the Global Indicators Workshop on Community Access to ICTs in November 2004. ITU took the initiative to organize the workshop as part of its overall responsibility to track access to and use of ICTs. The workshop came up with a set of indicators for measuring access to community ICT facilities. The main approach taken by the ITU is to look at the percentage of localities with public Internet access centres (PIACs) by number of inhabitants (rural/urban). This indicator has been added to the core list of indicators to measure the information society. Here the availability of at least one facility in a locality is what is important rather than the total number of facilities. This is the weakest of all indicators proposed in the basic core list because very few countries collect it, and there is little standardisation between countries in how PIACs is defined, nor even how localities and degree of urbanisation is defined. Nevertheless, it might be useful as a way of indicating a countrys commitment to extending the benefits of ICTs to all its inhabitants.

A10. Percentage of localities with public Internet access centres (PIACs) by number of inhabitants (rural/urban) A public Internet access centre (PIAC) is a site, location, centre of instruction at which Internet access is made available to the public, on a full-time or part-time basis. This may include digital community centres, Internet cafs, libraries, education centres and other similar establishments, whenever they offer Internet access to the general public. All such centres should have at least one public computer for Internet access. Localities refer to a countrys villages, towns and cities. The percentage of localities with public Internet access centres (PIACs) is computed by dividing the number of localities with at least one PIAC by the total number of the country's localities and multiplying by 100. The indicator should be broken down by range of inhabitants.

Extended core set of indicators The extended core set of indicators covers, additionally, basic measures of ownership of radio and television broadcasting devices. Radio and television broadcasting is the predominant means of electronic information and entertainment in all countries. Time use surveys for most developed nations show that watching television is the activity people devote the most time to after work and sleep. Broadcasting is also important to monitor because of its fusion with other ICTs.11 For example, it is possible to make telephone calls and access the Internet over cable television networks. Broadcast technologies also have a role to play as a development tool particularly in developing countries. Radio is being combined with Internet technologies to overcome literacy and language barriers. In this situation, radio stations download information from the Internet and re-disseminate it orally to the surrounding community, in local languages.12 The conventional indicators for measuring broadcast penetration are the number of radio and television sets.13 Few countries collect the number of devices and thus most data are estimates.14 These are derived from sales of sets or estimates based from surveys asking households whether they have a television. Some countries with licensing regimes collect data on the number of licences. This statistic is often used as a proxy for household availability. However not all people pay the licence fee so the true figure is underrepresented. This is apparent when licence data is contrasted with census or household surveys on the number of homes with a broadcast reception set. For this reason, and because of the underlying unreliability of the data, these indicators are left in the extended set rather than in the basic core.15
A11. Radio sets per 100 inhabitants A radio set is a device capable of receiving broadcast radio signals, using popular frequencies, such as FM, AM, LW and SW. A radio set may be a standalone device, or it may be integrated into another device, such as a Walkman, a car, or an alarm clock. Radio sets per 100 inhabitants is obtained by dividing the number of radio sets in use by the population and multiplying by 100. A12. Television sets per 100 inhabitants A television set is a device capable of receiving broadcast television signals, using popular access means such as overthe-air, cable and satellite. A television set may be a standalone device, or it may be integrated into another device, such as a computer or a mobile phone. It may be useful to distinguish between digital and analogue signal delivery and between TV sets receiving only a limited number of signals (usually over-the-air) and those that have multiple channels available (e.g., by satellite or cable). Television sets per 100 inhabitants is obtained by dividing the number of sets in use by the population and multiplying by 100.

Parts of this document have been extracted from Chapter two and Technical Notes of ITUs World Telecommunication Development Report. Access Indicators for the Information Society. ITU, 2003.

There is no shortage of references to universal service/ access being the main goal of telecommunication policy. For further information see ITU. (1998). World Telecommunication Development Report: Universal Access. Available from: http://www.itu.int/ITU-D/ict/publications/wtdr_98/index.html; accessed December 1, 2003 and ITU. (2003). Trends in Telecommunication Reform: Promoting Universal Access to ICTs Practical Tools for Regulators. Available from: http://www.itu.int/publications/docs/trends2003.html. Variations on basic and primary ISDN exist in some countries, sometimes referred to as fractional ISDN. For example in Denmark a variant known as Flex-ISDN provides 12 channels per line. Personal Computer (PC) is a computer designed for use by one person at a time as opposed to main frames or mini-computers, which are shared by many users. According to some researchers, the PC replacement rate in the US is as high as 70 per cent. On the other hand In more developing regions, PC replacement rates are much lower. CyberAtlas. (2003). PC Market

headed for geographic shift. Available from: http://cyberatlas.internet.com/big_picture/hardware/article/0,,5921_988841,00.html ;.

Prince and Cooke. (1998, December). Mercado Informtico. Available from: http://www.spkrsbr.com/biblioteca/htm/resultados.htm.
6 7 8

http://www.isoc.org/inet2000/cdproceedings/8e/8e_1.htm# s6 For more on broadband developments, see ITU. (2003). Birth of Broadband. Available from:

www.itu.int/birthofbroadband.
9

Even basic infrastructure needed to use modern ICTs, including electricity, is often not available to households and thus excludes home access. European Union. eEurope 2005: Benchmarking Indicators: http://europa.eu.int/comm/lisbon_strategy/pdf/655_EN.pdf

10

11

By the same token, new technologies can substitute for older ones. Radio and television stations provide audio and video streaming over the Internet while some mobile phone models have built-in radios. Minges, M. (2002, April). Mixed Media in the LDCs. Available from:

12

http://www.itu.int/osg/spu/ni/ipdc/index.html;
The broadcast industry uses other metrics such as universe estimates (e.g., potential television audience). See FAQ About Ratings at the Nielsen Media Research website: http://www.nielsenmedia.com.
13 14

The United Nations Educational, Scientific and Cultural Organization (UNESCO) had published the number of radio and television sets in different countries but stopped with its 1999 Statistical Yearbook.

15

This lack of data may be a problem in the future, as countries shift towards digital radio and television broadcasting. Important policy-decisions on when to turn-off analogue broadcast channels may be delayed due to lack of reliable data on homes with radios and televisions.