E-COMMERCE - TOWARD AN INTERNATIONAL DEFINITION AND INTERNATIONALLY COMPARABLE STATISTICAL INDICATORS Bill Pattinson Information, Computer and Communications Policy Division, OECD The development of appropriate statistical indicators about the information economy and information society are essential for an adequate understanding of the mechanisms and trends occurring in this field and to permit an adequate base for policy development and monitoring. This premise led, in 1997, to the formation of an ad-hoc statistical experts group (the ICCP Statistical Panel), the aim of which was to establish a set of definitions and methodologies to facilitate the compilation of internationally comparable data for measuring various aspects of the information society, the information economy and electronic commerce. Late in 1998 this ad-hoc group was afforded the status of a Working Party and the first meeting of the Working Party on Indicators for the Information Society (WPIIS) was held in April 1999. Dr Fred Gault, Statistics Canada, chairs the WPIIS, with Sweden and France providing the vice-chairs. The WPIIS meets only once a year; however a number of the issues that it addresses are also of interest to the Voorburg Group, which meets annually to discuss matters associated with Services sector statistics. The presentation of papers and the subsequent discussions at the most recent meetings of that Group have enabled a number of the issues outlined below to be advanced more rapidly than would otherwise be the case. The work has also benefited from its inclusion on the agenda at various Eurostat meetings on the Information Society. This paper aims to summarise the work of this Group insofar as it relates to electronic commerce. In the first instance it might be appropriate to present the work of the measurement of e-commerce within the overall parameters of the WPIIS agenda, i.e., the definition and measurement of the information society. The WPIIS Agenda and E-commerce The WPIIS agenda relates to the definition and measurement of all aspects of the Information Economy and the Information Society. To do this it has developed a model in which it sees the Information Economy as comprising the Information and Communication Technology sector, the Content sector, the outputs of these sectors, and their application to normal business processes. The initial task of WPIIS was to develop an internationally agreed definition of the ICT sector. The activity-based OECD definition for the ICT sector was approved by the WPIIS in June 1998 and declassified by the Information, Computer and Communication Policy Committee (ICCP) in September 1998. The OECD has subsequently moved to implement this definition by developing a set of statistical indicators for the sector. 1 It is also important to define a set of commodity outputs emanating from the ICT sector, and the OECD is working to produce such a definition for its April 2001 meeting. This set needs to be built around existing standard commodity classifications. Those relevant for this exercise are the Central Product Classification (used for goods and services of local production) and the Harmonised System (used for data on international trade). While these are both structured classifications, they do not specifically identify ICT goods and services; hence, they require modification before they can be used for measuring ICT sector commodities. More importantly, they are both relatively old classifications and hence they do not uniquely identify the new goods and services about which users require information. Thus work to date on this part of the project has been somewhat slower than would be liked, but it is hoped to draw it to a successful conclusion at the April 2001 meeting of WPIIS. 1. A copy of this publication is freely available on http://www.oecd.org/dsti/sti/stats/index.htm or on request from OECD. Statistics.net 113
The activity-based ICT sector definition is limited to those industries which facilitate, by electronic means, the processing, transmission, and display of information, and it excludes the industries which produce the information, the so-called content industries. The content sector is a rapidly growing sector of the economy and of intense interest to policy-makers. The work being undertaken is looking at the issue of a content sector, both in terms of industries and commodities. The initial step is to form a rationale for the goods and services that we are considering to fall into these categories and to proceed from there to the industrial and commodity definitions. We need to measure the size and rate of diffusion of electronic content and the impacts that are being felt within the more traditional industries and commodities and more generally across the economy and society. To fully assess the Information Economy and the Information Society further requires an elaboration of the way in which the outputs of these sectors are diffused across the economy. To this end the OECD is working with the groups referred to earlier to develop model surveys of the use of ICT goods and services across all sectors of the economy, in the Business Enterprise sector, the Government sector and the Household sector. The most developed of these surveys is that in respect of the Business Enterprise sector where the Nordic countries have taken a lead role in the development. It is envisaged that the initial development of this survey will be adopted at the 2001 meeting of WPIIS. Surveys on the Household use of ICTs are also being developed, with Australia taking the lead country role. A proposal for a set of model questions to be used internationally as seen fit by Member-Country Statistical Agencies will be presented at the 2001 meeting of WPIIS. Work on a model survey for the Government sector is less well developed as very few countries have undertaken ICT Use surveys for this sector with Australia and Canada being two notable exceptions. Definition and measurement of electronic commerce Electronic business processes are carried out using ICT technologies and applications. In this respect, the measurement of e-commerce is one of the components of ICT use and so can be considered in the context of the model ICT usage surveys referred to above. However, because of the extreme policy and media interest in e-commerce, it has been developed as a separate stream within the WPIIS work program. The impetus for this work comes from the OECD E- commerce Action Plan endorsed by Ministers at the 1998 Conference on Electronic Commerce (Ottawa, October 1998). This included an item to Initiate work on defining and measuring electronic commerce as one of its seven action points. Subsequently, OECD organised a Workshop (21 April 1999), which brought together the expertise of business, policy-makers, researchers and statisticians to discuss issues of definition and measurement of e-commerce. At its meeting which followed this Workshop, WPIIS decided to create an Expert Group on Defining and Measuring E- commerce with a mandate to compile definitions of electronic commerce which are policy relevant and statistically feasible. To do this, the Expert group worked on the three inter-related aspects of this problem a framework for user needs and priorities, definitions, and statistical measurement and to report back to the 2000 WPIIS meeting. The results of that work were reported at many meetings and conferences over the 12 months between the 1999 and 2000 meetings of WPIIS and are summarised below for information. In respect of user needs, work concentrated on the development of a model based on one presented by Industry Canada to the April 1999 Workshop referred to above. In that model, there are three basic sets of indicators required for policy purposes: those relating to Readiness, to Intensity and to Impacts. These are defined in the box below with the overall model expressed diagrammatically in Figure 1. The policy needs and the indicators will vary depending on the degree of e-commerce activity that is occurring. Initially there will be a demand for indicators about how ready an economy is for participating in e-commerce; subsequently there will be a set of indicators required to measure the intensity of the use of e-commerce; at later stages of development of e-commerce there will be a need for indicators about the impact of e- commerce on the economy and society generally. Readiness - measuring the infrastructure in place to allow for e-commerce activities to occur, Intensity - measuring the extent to which users are utilising e-commerce to undertake their normal business and social processes, and Impacts - measuring the impact of e-commerce on the economy and society 114 Insee Méthodes
As depicted in Figure 1, there is a degree of overlap between the various components. In addition some of the indicators that might be developed can serve as indicators for more than one of the levels of e-commerce activity. In respect of definitions, there have been many used throughout the world and these differ significantly. Some include all financial and commercial transactions that take place electronically; other limit e-commerce to retail sales to consumers via the Internet. The first type has existed for decades and results in trillions of dollars worth of activity every year. On the other hand, the second type is fairly new and is not yet very large. A survey of business views on the definition of e-commerce undertaken on behalf of Statistics Canada drew a distinction between e-business and e-commerce, with e-commerce only including commercial transactions and e-business all other forms of business processes. It was essential for the transaction to be conducted over some form of computer network for it to be included in either of these categories. For more details see Statistics Canada (1999), A reality check to defining e-commerce. Policy-makers definitions are often very broad, as they wish to understand the impact of e-commerce across the economy and over all business processes. On the other hand, they also have a need for data at a very fine level as there is a need for measuring different e-commerce segments as the drivers, technological solutions, impacts, and policy implications may be different. Figure 1. Translating Policy Needs into Indicators for E-commerce Level of electronic commerce activity Readiness "potential" usage access technology infrastructure socio-economic infrastructure Intensity transaction/business size nature of transaction/business Impact efficiency gains employment, skill composition, work organisation new products, services, business models contribution to wealth creation changes in product/sectoral value chains Source: OECD (1999), Defining and Measuring E-Commerce: A Status Report Time Given the range of user requirements, it became fairly clear that one definition would not be suitable for all needs. Thus the Expert Group has looked to develop a set of definitions which fit together coherently and can be used to provide indicators for all purposes. From all the definitions that have evolved, it is clear that these definitions will have to include three different dimensions: the communication networks on which the e-commerce is based, the applications which are being viewed as being part of e-commerce, and the economic activities (or business processes) being undertaken. In terms of statistical measurement, there has been considerable activity to date although not all of it has been fully coordinated. There are many measures that show the readiness of various economies for e-commerce. There are many indicators about the telecommunications infrastructure in member countries and information on the skills of the workforce to handle e-commerce activities. In many countries, there are also indicators of the diffusion of computers and the Internet throughout industry and society more generally and a number of indicators on the intensity of this use have been developed. As far as measuring the value of electronic sales, particularly over the Internet, the conduct of probability-based samples has up till now not provided very reliable results because of the small number of businesses and households that are undertaking such transactions. As the number and value of transactions occurring over the Internet increases, this problem will become less serious and the degree of accuracy of the data will improve. Many countries have now started to make attempts to get a measure of such transactions. The measurement of the impacts of e-commerce has so far received little attention and more work is needed in this area. Statistics.net 115
In April 2000, OECD organised a joint meeting of the Working Party on the Indicators for the Information Society (WPIE) and WPIIS to help bring these matters to a conclusion. The meetings came to the following conclusions: The use model (readiness, intensity, impacts see above) was appropriate. Two definitions of e-commerce transactions were appropriate a broad definition and a narrow definition. The broad definition covers the sale or purchase of goods or services over computer mediated networks. The sale and purchase is defined by the action of ordering, not the payment or delivery. The narrow definition related to the sale or purchase of goods or services over Internet-protocol based networks. A provisional lists of indicators for the business and Government sectors were agreed as a suitable starting point for a more detailed examination of their suitability for providing internationally comparable data. It was appropriate to form an ad-hoc Expert Group to provide better (more detailed) specifications of the terms used in the definitions and to review and refine the list of indicators based on a study of Member-Country statistical practices. The OECD therefore convened an ad hoc Expert Group meeting in November 2000 to consider the issues raised at the April meetings. The Experts supported the previous decision that that there should be two definitions adopted and only proposed to change the wording for the narrow definition so that it only related to transactions over the Internet and not to Internet-protocol based networks. The Experts further proposed that countries should show what was meant by the definitions of Internet and computer-mediated networks on their survey forms by giving examples of what is to be included. It therefore proposed the operational definitions shown in boxes 1 and 2 below. Box 1. An operational definition of Internet transactions The narrow definition Orders received/placed on a Web page, over Extranets and other applications that run over the Internet, such as EDI over the Internet, Minitel over the Internet, or any other Web enabled application regardless of how the Web is accessed (e.g. through a mobile phone or a TV set). The payment and the ultimate delivery of the goods or services may be conducted on- or off-line. Box 2. An operational definition of electronic transactions The broad definition Includes all Internet transaction as defined above plus orders received/placed over EDI or any other online applications used in automated transactions (e.g. interactive telephone systems). Orders received/ placed using facsimile, telephone or non-interactive e-mail should not be included. The payment and the ultimate delivery of the goods or services may be conducted on- or off-line. The ad-hoc Expert Group meeting was also asked to consider the issue of establishing a list of specific business processes for which indicators might be compiled. In the initial list adopted in April, only the generic term specific business processes has been used, other than the recognition of orders as required by the definitions adopted. Individual OECD Member Countries have measured specific business processes particularly relevant for their circumstances and the business-sector statistical collections that they are undertaking. Based on an inspection of these lists and the discussion which occurred at the meeting, it is proposed that the set of business processes shown in Box 3 below should provide the initial set of internationally comparable e-commerce indicators. Box 3. Business processess to be used as an initial set for international comparisons - Ordering of goods or services both sales and purchases - On-line payment for goods and services both receipts and payments - Conduct of financial transactions other than online payments e.g. banking - Digital delivery of services as a supplier and a customer - Marketing and promotion/ information search - Employee recruitment/job search - Use of online databases as a provider and a user Having established operational definitions of e-commerce and the processes about which indicators might be required, the ad-hoc Expert Group meeting was then able to consider the range of indicators based on their own national 116 Insee Méthodes
experiences and the findings of a data availability and methodology survey conducted by OECD. They concluded that the indicators listed below should be considered for inclusion in an initial OECD data compilation exercise, subject to the indicators being able to be compiled on an internationally comparable basis. The OECD was encouraged to complete the task of determining how to arrive at a uniform set of the indicators shown in Box 4 below. The Expert Group meeting also recognised the importance of compiling internationally comparable indicators for the Household sector. While not discussing in detail the activities (or processes) to be measured, they were able to arrive at a set of indicators for the Household sector. In this it recognised that some indicators were required in respect of the household and some in respect of individuals. Based on international experience gained to date, the Experts agreed to the core set of indicators for the Household sector shown in Box 5 below. Box 4. Core business sector indicators for measuring E-commerce 1. No/Proportion of businesses with computers 2. Employment (level and share) of businesses with computers 3. No/Proportion of businesses with access to the Internet 4. Employment (level and share) of businesses with access to the Internet 5. No/Proportion of businesses undertaking specific business processes on the Internet - Ordering of goods and services ( both sales and purchases) - Online payment for goods and services both receipts and payments - Conduct of financial transactions other than online payments e.g. banking - Digital delivery of services as a supplier and a customer - Marketing and promotion/ information search - Employee recruitment/ job search - Use of online databases as a provider and a user 6. No/Proportion of businesses with Web sites 7. No/Proportion of businesses which plan to use the Internet 8. No/Proportion of businesses receiving orders over computer-mediated networks 9. Value of orders of goods and services received over the Internet 10. Value of orders of goods and services received over computer-mediated networks 11. Proportion of orders of goods and services received over the Internet 12. Proportion of orders of goods and services received over computer-mediated networks Box 5. Core Household-sector indicators for measuring E-commerce 1. No/Proportion of households with computers 2. No/Proportion of households with access to the Internet 3. No/Proportion of households recognising specific barriers to e-commerce 4. No/Proportion of individuals undertaking specific activities on the Internet - purchasing goods and services (for private use) - paying for goods and services (for private use) - undertaking work activities at home - undertaking games activities at home - undertaking banking activities 5. Value of orders placed over the Internet by individuals Statistics.net 117
The future The conclusions of the Expert Group meeting above is only provisional at this stage and is currently being referred to participants for their verification and further consideration. It will then be circulated more widely to delegates of WPIIS for their comments. The views of the WPIE delegates will be sought at their December 2000 meeting being held on 6 December. It is perhaps worth pointing out that at this stage the largest point of conjecture relates to the treatment of orders that take place via e-mails. As can be seen by the definition, the intention is to include only e-mails where there is an automated action of some sort. Practically it is not certain if this sort of definition could be implemented in statistical collections. Further debate is likely to occur before a final decision can be reached. Having reached agreement on a set of definitions and indicators it is important to take the next steps in this process the compilation of indicators. As part of that process, a good deal will be learned about the extent to which indicators are comparable between countries, or can be modified so that they are comparable. The OECD therefore proposes to commence with a compilation of these indicators to the extent possible early in 2001. That compilation will provide the real evidence of just how easily the agreements can be implemented. Should the indicators prove to be comparable, the OECD will proceed to publish them towards the summer of 2001. The final step in the chain will then be to draw up a provisional methodological manual that can be used by Member countries to enable them to compile statistics that can be used in this internationally comparable way. No timetable has been scheduled for this task as it depends upon the degree to which the earlier steps prove to be effective. The April meeting of WPIIS is likely to consider the priority that should be given to this task. Finally, it should be noted that all the OECD work to date has been in respect of readiness and intensity indicators as defined in the earlier work. There is still much work to be done to extend the indicator set to enable measurement of indicators of the impact of e-commerce. 118 Insee Méthodes
Reference documents OECD (1998), Summary Record of the second Ad Hoc Meeting on Indicators for the Information Society, DSTI/ICCP/AH/M(98)1/Rev 1 OECD (1999), A Presentation by Richard Simpson, Industry Canada, to the OECD Workshop on Defining and Measuring E-Commerce, April 1999. http://www.oecd.org/dsti/sti/it/ec/act/agenda_ecworkshop.htm OECD (1999), Defining and Measuring E-Commerce: A Status Report, DSTI/ICCP/IIS(99)4/FINAL. http://www.oecd.org/dsti/sti/it/ec/act/paris_ec/paris-ec_docs_1.htm OECD (1999), The Household Use of ICT Goods and Services, Towards a Framework for Internationally Comparable Statistics, DSTI/ICCP/IIS/RD(99)3 OECD(2000), Summary Record of the Joint Meeting of the Working Party on the Information Economy (WPIE) and the Working Party on Indicators to the Information Society (WPIIS). Statistics Canada (1999), A Reality Check to defining e-commerce. A report prepared by CGI for Statistics Canada. http://www.statcan.ca/english/ips/data/88f0006xib99006.htm Statistics.net 119