DAC Network on Poverty Reduction

Unclassified DCD/DAC/POVNET(2004)6/REV1 DCD/DAC/POVNET(2004)6/REV1 Unclassified Organisation de Coopération et de Développement Economiques Organisation for Economic Co-operation and Development 10-Dec-2004 English - Or. English DEVELOPMENT CO-OPERATION DIRECTORATE DEVELOPMENT ASSISTANCE COMMITTEE DAC Network on Poverty Reduction ICTs AND ECONOMIC GROWTH IN DEVELOPING COUNTRIES This document is sumitted to the POVNET for INFORMATION. It has been prepared and reviewed by Professor David Souter, Managing Director, ictdevelopment Associates Ltd. and visiting professor of Strathclyde Business School, University of Strathclyde. The document reviews the existence of links between ICTs, productivity and economic growth in OECD countrie, set out in "ICT and Economic Growth, Evidence from OECD Countries, Industries and Firms" (2003), and discusses the relevace of these findings for developing countries. English - Or. English Contact Person: Mr. Ichiro Tambo; Tel:(33 1) 45.24.14.25; Email: ichiro.tambo@oecd.org JT00175866 Document complet disponible sur OLIS dans son format d'origine Complete document available on OLIS in its original format

TABLE OF CONTENTS Executive Summary... 3 Introduction... 4 ICTs, productivity and economic growth: evidence from the OECD area... 6 ICTs, productivity and economic growth: evidence from the developing world... 10 Differences between OECD and developing economies... 12 The relationship between economic growth, development policy and the digital divide... 15 Policy recommendations... 19 Sequencing and evaluation of outcomes... 24 Conclusions... 25 Bibliography... 26 Figures Figure 1. Analytical Framework... 5 Figure 2. The share of investment in ICT in total GDP... 7 Figure 3. Value added per sector in 2001 (% of GDP)... 11 Boxes Box 1. The productivity paradox has it been solved?... 6 Box 2. Network Externatilities... 8 Box 3. The Benefits of ICT are not immediate... 9 Box 4. The Digital Divide... 16 Box 5. Innovative solutions to extend access to telecommunications... 23 2

EXECUTIVE SUMMARY 1. This report reviews the evidence for links between information and communication technologies (ICTs), productivity and economic growth in OECD countries, set out in the OECD report ICT and Economic Growth, and discusses the relevance of these findings for developing countries. Its general conclusions can be summarised as follows: 2. There is evidence from the OECD countries that ICTs facilitate economic growth, principally by increasing productivity, though this is a long-term rather than immediate outcome of ICT investment. There is little or no clear evidence that the same outcome is being achieved in developing countries, largely because almost no relevant research has been undertaken. 3. For a variety of reasons including economic structure (for example, the dominance of agriculture in GDP, low incomes) and policy issues (restrictive regulatory environments, low formation of human capital, etc.) developing countries in general, and LDCs in particular, are less well-equipped to take advantage of the potential of ICTs to stimulate growth, and so (to the extent that ICTs do stimulate growth) they are likely to fall further behind OECD economies in relative terms. 4. Developing countries, and development agencies, also have to balance policy and investment options in ICTs against other socio-economic objectives (notably the development objectives set out in the United Nations Millennium Development Goals). Poverty-reduction objectives are nonetheless more likely to be achieved against a background of economic growth, and any synergies between the impacts of ICTs in these two areas of national policy should be exploited. 5. Policy priorities should therefore aim to reduce the factors which inhibit effective use of ICTs (and any gains that may result from them), to take steps to maximise the benefits ICTs can provide, and to integrate ICT policy more effectively into national socio-economic development strategies. 6. The report concludes with recommendations for action by developing-country governments that will facilitate ICT investment and returns on investment, focused on policy-development processes, infrastructure and access, liberalisation and deregulation, and human capital. Acknowledgements 7. The author is grateful for input from and discussions with Mr. J. Dryden, Mr. A. Wyckoff, Mr. D. Pilat, Mr. P. Lindroos and Mr. G. Vickery of the OECD Directorate for Science, Technology and Industry, and for the assistance of Ms Abiodun Jagun. 8. This work was co-ordinated by Mr. Ichiro Tambo, Adviser on Science and Technology, OECD Development Co-operation Directorate, and funded by the Japan International Cooperation Agency (JICA). 3

Introduction 9. Ten years ago most development agencies, analysts and developing-country governments considered information and communication technologies (ICTs) 1 marginal to the achievement of both national economic growth and the reduction of poverty. Today, ICTs are considered so central to development that governments have initiated national e-strategies and donor agencies have made them a mainstream item in national and international programmes. They are now sufficiently important, indeed, for the Information Society to merit a World Summit similar to those on Sustainable Development or Social Development. 10. The speed with which scepticism has given way to enthusiasm has stimulated a good deal of innovative thought, but it also carries substantial risks. Investment in ICTs is expensive, and its impact largely unresearched and easily exaggerated. Many of the assumptions underpinning current thinking on ICTs in development are based on intuition rather than analysis and on limited evidence from a narrow range of pilot projects rather than large-scale impact assessments. The danger is that, without better understanding of the real impact of ICTs on both national economies and community development, the pursuit of over-ambitious, unrealistic goals may mean that resources are misapplied and worthwhile objectives missed. Past disappointments, for example the failure of import substitution industrialisation strategies to transform economic growth, have not destroyed the yearning for a magic bullet for development, and the real capabilities (and limitations) of ICTs must be properly understood if they are to be exploited effectively in both small- and large-scale industrial activity and in their contributions to national economic expansion. 11. Thinking about the role of ICTs in development has focused primarily on their potential for reducing poverty, and especially on the impact they may have on mainstream development objectives in, for example, health, education, providing livelihoods 2 and empowerment 3. Less attention has been paid to the impact of ICTs on national economic growth on productivity and the relationship between the national economies of developing countries and the wider world. That impact is generally assumed to be beneficial, but it has not been seen as the primary aim of the engagement of the development community with the ICT sector. The economic debates about the macro impact of ICTs in industrial countries have largely passed the development community by: hardly any relevant research has been done outside the OECD area, and almost none at all in Least Developed Countries (LDCs). 12. What might be the impact of ICTs on these large-scale economic issues? And what are the implications for the differing circumstances of the developing world of recent research on the 1 The definition of ICTs varies considerably, causing considerable confusion. In the development literature, for example, they generally include old technologies such as broadcast radio and voice telephony. Analysts of the new economy are more likely to mean only new ICTs, based on digital or computer technology. For discussion purposes, this report uses the more generic, less technology-specific definition put forward by Duncombe and Heeks (1999): electronic means of capturing, processing, storing and disseminating information. 2 Livelihoods encompass varied ways of living that meet individual, household, and community need, including social (i.e. networks), human (i.e. skills), natural and physical as well as financial capital. A good description of livelihoods analysis is presented by the UK Department for International Development at http://www.livelihoods.org/info/info_guidancesheets.html. 3 ICTs can help empower people to transform their situations by strengthening their capacities. This is potentially true both for those who already hold power and for those who do not. Actual empowerment outcomes will depend on a number of factors, including the existing distribution of social, economic and political power. 4

relationship between ICTs, productivity and economic growth in OECD countries. First, although economic growth does not necessarily lead to poverty reduction, reductions in poverty are much more difficult to achieve without economic growth a point that seems self-evident but has sometimes been forgotten in redistributive approaches to development. If ICTs do have a positive impact on national economic growth, then their contribution should be factored into general development policies for poverty reduction and the redistribution of economic and social welfare. Second, good policy development depends on an accurate understanding of the nature of the economic impact of ICTs and of the factors which may constrain or enhance it. National e-strategies developed without such understanding will, at best, miss their targets and could even prove counter-productive. 13. This report reviews recent OECD research on the impact of ICTs in the OECD countries and asks whether similar impacts can be expected in the different circumstances of the developing world. It considers the relationship between these findings, the digital divide 4 and the wider aim of development policy the reduction of poverty and sets out a number of recommendations for developing countries and international agencies. 14. An initial caveat is necessary: analysis for the developing world in this context has to be based at present on limited evidence. Even in the OECD countries the links between ICTs, productivity and economic growth have only recently been established, through sophisticated analysis of complex data. No similar analysis has been undertaken in developing countries, least of all in LDCs; and no comparable data are available for most. The best one can do is assess, as a basis for subsequent policy-analysis, how the differences between developing-country/ldc and OECD economies are likely to affect the impact of ICTs. Substantive research is urgently required if investment commitments are to be made by the private sector or development agencies with any real understanding of likely outcomes. Figure 1. Analytical Framework from OECD Publication: A New Economy? (2000) Source: A New Economy?, p. 18. OECD (2000). 4 That is, the disparity in ICT diffusion and use between industrial and developing countries (or, indeed, between rich and poor, men and women, urban and rural areas within individual countries). See also Box 4. 5

ICTs, productivity and economic growth: evidence from the OECD area 15. The uncertainty that surrounds the nature and scale of the impact of ICTs on productivity and growth is best encapsulated in Robert Solow s famous paradox that you see computers everywhere except in the productivity statistics. 5 The beneficial impact economists intuitively expected to find proved elusive, even in the United States, and it was not until the mid-1990s that a new series of studies began to discern evidence to back their intuition. It is now commonly agreed that, in OECD countries at least, ICTs have had a beneficial impact on productivity and growth, and indeed may well play a leading role in maintaining growth, though their impact on national indicators has proved slower to materialise than was expected and is much affected by synergies with complementary factors such as the regulatory environment and the availability of human capital. Box 1. The productivity paradox has it been solved? The Solow paradox, attributed to economist Robert Solow, who once observed that computers are everywhere except in the productivity data, was appropriate during much of the 1980s and early 1990s, when the rapid diffusion of computing technology seemed to have little impact on productivity growth (Solow, 1987). Many studies in the 1970s and 1980s showed negative or zero impacts of investment in ICT on productivity. Many of these focused on labour productivity, which made the findings surprising as investment in ICT adds to the productive capital stock and should thus, in principle, contribute to labour productivity growth. Later studies did find some evidence of a positive impact of ICT on labour productivity. Some also found evidence that ICT capital had larger impacts on labour productivity than other types of capital, suggesting that there might be spillovers from ICT investment. Studies over the past decade have pointed to several factors that contributed to the productivity paradox. First, some of the benefits of ICT were not picked up in the productivity statistics (Triplett, 1999). This is mainly a problem in the service sector, where most ICT investment occurs. For instance, the improved convenience of financial services due to automated teller machines (ATMs) is only counted as an improvement in the quality of financial services in some OECD countries. Similar problems exist for other activities such as insurance, business services and health services. ICT may have aggravated the problems of measuring productivity, as it allows greater customisation, differentiation and innovation in the services provided, most of which is difficult to capture in statistical surveys. Progress towards improved measurement has been made in some sectors and some OECD countries, but this remains an important problem in examining the impact of ICT on performance. A second reason is that the benefits of ICT use might have taken a considerable time to emerge, as did the impacts of other key technologies, such as electricity. The diffusion of new technologies is often slow and firms can take a long time to adjust to them, e.g. in changing organisational arrangements, upgrading the workforce or inventing and implementing effective business processes. Moreover, assuming ICT raises MFP in part via the networks it provides; it takes time to build networks that are sufficiently large to have an effect on the economy. ICT has diffused very rapidly in many OECD countries in the 1990s and many recent empirical studies find a larger impact of ICT on economic performance than studies that were carried out with data for the 1970s or 1980s. A third reason is that many early studies that attempted to capture the impact of ICT at the firm level were based on relatively small samples of firms, drawn from private sources. If the initial impact of ICT on performance was small, such studies might find little evidence, as it would easily get lost in the econometric noise. It is also possible that such samples were not representative of the total population. Moreover, several studies have suggested that the impact of ICT on economic performance may differ between activities, implying that a sectoral distinction in the analysis is important. More recent studies based on large samples of (official) data and covering several industries are therefore more likely to find an impact of ICT than earlier studies. In addition, early studies used a wide variation of data on ICT and ICT diffusion, often of unknown quality. Much progress has been made in recent years in measuring ICT investment and the diffusion of ICT technologies, implying that the range of available data is broader, more robust and statistically sounder than previous data. Source : ICT and Economic Growth, pp. 57-58. OECD (2003). 5 Robert Solow (1987), We d Better Watch Out, Book Review No. 36, The New York Times, 12 July. 6

16. The most substantial evidence for this positive view derives from a major multi-country study undertaken by the OECD and published in its 2003 report, ICT and Economic Growth: Evidence from OECD Countries, Industries and Firms. 6 Using data from thirteen countries, and including extensive analysis of corporate behaviour, this research established clearly that ICTs have acted as drivers of growth in OECD economies firstly in firms themselves and then nationally and, moreover, that considerable differences are evident in the scale of their impact in different OECD countries. ICT investment typically accounted for between 0.3 and 0.8 percentage points of growth in per capita GDP in 1995 2001, with the United States and Canada performing substantially better than other OECD members. Figure 2. The share of investment in ICT in total GDP 17. ICT impact can be found in three main areas: In some countries, such as Finland and the United States, the technological innovation and high volumes of demand generated by an ICT production sector played an important role. But it does not follow that an ICT production sector is necessary to achieve the beneficial impact on the national economy that the study identified, as strong growth rates in other countries indicated: countries with strong ICT service sectors were also at an advantage over those in which the ICT sector as a whole was weak. ICT investment has contributed to capital deepening : it has increased capital input per worker, enabling more efficient production that increases labour productivity. The pervasive use of ICTs throughout the value chain has contributed to improved performance in firms, enabling them in particular to increase efficiency in combining capital and labour ( multifactor productivity ). Networking enabled by ICTs was also important, both within the firm and 6 A useful summary of the report is to be found in Dryden (2003). 7

(as the diffusion of ICTs spreads throughout an economy) by enabling new forms of interaction between firms and other parties such as consumers. 18. There is an important distinction to be drawn here between improvements in the performance of individual firms and in national macro-economic performance. Evidence from firms shows that considerable benefits can be gained from ICT investment, particularly by companies equipped to maximise those gains through adaptation and innovation in their work processes. National gains from ICT investment derive partly from the aggregation of these micro-economic improvements in productivity, but also from ICT-based networking between firms, which reduces transaction costs and accelerates innovation. The importance of networking in unlocking the potential of ICT investment is critical, and has been much increased by the advent of the Internet (which has also made many services much more tradable than before). The externalities of communications networks suggest that the impact of ICTs will increase more swiftly than their apparent rate of deployment may initially suggest. There may also be a threshold effect at play, through which ICTs begin to have a lasting and sustainable national impact only when they achieve a certain penetration of the economy as a whole. 7 The increased value generated by networking is one factor which helps to explain why economy-wide ICT diffusion and use are more important than ICT production in contributing to national productivity and growth. Box 2. Network Externatilities Network externalities are derived from the fact that the value of a telephone line (or similar access point for interactive communications) increases with each new subscriber by the number of potential connections between users rather than the number of potential users: a telephone network with two subscribers has one possible connection, a network with three subscribers three possible connections, a network with four subscribers six possible connections, and so on. 19. However, it takes considerable time for the scale and spread of ICT investment by individual firms and the networking between them in conjunction with other organisational and production changes for them to be translated into macro-economic outcomes. This time-lag is one of the principal reasons for the difficulty earlier economists found in identifying beneficial ICT impact on productivity and growth even within ICT-intensive economies like the United States and it is equally relevant in developing countries today. Understanding this time-lag is potentially very important in the development of national ICT and development strategies, and more research is needed on its variation between countries and economies. 7 This threshold effect is suggested by a number of analysts, including Roller and Waverman (1994), Bedi (1999) and Rodriguez and Wilson (2000). 8

Box 3. The Benefits of ICT are not immediate It takes time to adapt to investment in ICT, e.g. by changing organisational set-ups and worker-specific skills. Firms that adopted network technologies several years ago, notably large firms, have often already been able to make the technology work, whereas more recent adopters are still adapting their organisation, management or skills. Evidence for the United Kingdom, for example, shows that among the firms that had already adopted ICT technologies in or before 1995, over 50% were using electronic networks for procurement by 2000. In contrast, of the firms that only adopted ICT in 2000, fewer than 20% made purchases through networks in 2000. Source : Seizing the Benefits of ICTs in a Digital Economy, Meeting of the OECD at Ministerial Level, Brochure, p. 10, para. The Benefits of ICT are not immediate (OECD 2003). 20. OECD research shows that the extent of diffusion and use of ICTs, and thus their impact on business performance, are also influenced by a number of complementary factors in the business environment. Five of these seem particularly important: the nature of the business in which individual firms are engaged some sectors, particularly services, can make much more extensive use of ICTs to change processes and their relationships with customers and suppliers (for example, through the use of software, call centres and e- commerce); the extent of competition and the nature of the regulatory environment the more competitive and less regulated the business environment, the more likely are firms to take advantage of ICT innovation, and countries to improve their macro-economic performance; the relative costs of ICT deployment, including the costs of hardware and other inputs, including labour, but also indirect costs related to changes in working practices, licensing, standardisation and the usage costs of networking facilities such as telecommunications networks; the amount and quality of human capital available the better skilled the workforce and the better equipped a firm is to upgrade workforce skills to take advantage of ICTs, the more likely it is to achieve higher rates of ICT-related innovation and increased productivity; the ability and willingness of organisations, particularly firms, to restructure and reorganise their working methods to take advantage of the new opportunities made available through ICTs the OECD study confirmed evidence reported elsewhere 8 that adaptability and organisational capital within firms play a crucial part in maximising the value of ICT investment. 21. Without these complementary factors, the OECD research indicated, ICT investment is much less likely to improve business performance, either in individual firms or, through them, nationally. Company managers should therefore focus on steps to maximise the return they achieve on their ICT investments, such as skill upgrading and innovation in organisational management. While this is true of all kinds of investment, it may be particularly important in the case of ICT investment because of the extent to which ICTs transform the intellectual as well as the physical content of work. 8 For example, Brynjolfsson and Hitt (1998). 9

22. The policy implication is that, to retain or enhance their countries relative economic performance, governments and businesses alike should act in ways which facilitate the benefits of ICTs: liberalising markets and reducing regulatory requirements on businesses, promoting access to business finance and facilitating market entry and company growth, encouraging entrepreneurship and innovation, stimulating trust in the efficacy and security of electronic transactions and promoting the development of human capital, chiefly through education and training. ICTs, productivity and economic growth: evidence from the developing world 23. Lack of research means there is much less evidence for the effects of ICTs on productivity and economic growth in developing countries. There may also be intrinsic differences between OECD and developing-country economies, particularly those of LDCs. The experience of developing countries in both the production and diffusion of ICTs is very different from that of most OECD member countries. 24. First, relatively few developing countries have sizable ICT production sectors. Those which do are almost all either middle-income countries transition economies in central and eastern Europe or countries in Asia and Latin America with established industrial/manufacturing sectors or very large countries, such as India and China, whose size gives them substantial domestic markets and skilled workforces requiring remuneration at levels well below those in OECD member countries. LDCs, with very few exceptions, have neither ICT production nor export-oriented ICT service sectors. 25. ICT manufacturing sectors in developing countries also seem likely to have fewer backward and forward linkages into the national economy than ICT production sectors in the OECD area. Much of the investment in ICT manufacturing in developing countries derives from foreign sources rather than from local capital markets, for example, while most of the resulting production leads to improved efficiency in the countries to which products are exported rather than enriching local manufacturing and services. Much the same is probably true of the export-oriented service sectors undertaking software development, data entry or back-office functions which have become established in India and some smaller developing countries with appropriately skilled workforces. 9 26. Second, many developing-country economies, particularly LDCs, are still dominated by commodity production and (often subsistence) agriculture, in which ICT investment obviously has much more limited value. The scale of ICT investment will therefore be much lower as a proportion of national output than in industrialised countries and the benefits for national growth will be slower to materialise. 9 Cf. Joseph (2002). 10

Figure 3. Value added per sector in 2001 (% of GDP) 70 60 50 40 30 20 10 0 Low income Middle income High income Agriculture Industry Manufacturing Services Source: World Development Indicators, 2003, The World Bank, p.192. 27. We shall return to these points later. 10 What they initially suggest is that there are substantial inherent differences between OECD and developing economies whose effect on the outcomes of ICT investment could also be substantial, and that findings from OECD economies should not simply be extrapolated to developing countries without further research, or at least an assessment of the differences between the types of economy concerned. 28. A basic difficulty here is the lack of available historic and current data for cross-country comparisons. Such data are available for a few developing countries, generally because of sectoral analysis. However, this information has generated very little research, and most of that is concerned with middle-income countries and/or those with ICT production sectors; almost none has been done on LDCs, which are the primary focus of the interest of the development community in the value of ICTs. Nevertheless, the research which has been undertaken on middle-income countries and transition economies can usefully indicate some of the factors likely to influence the outcomes of ICT investment in low- as well as middle-income contexts. 29. The findings are less encouraging than for the OECD area. Proven linkages between ICT investment and productivity or economic growth in developing countries are still as elusive as once they were in the OECD: for developing and even transition countries, Solow s paradox still seems to hold true. A recent study of transition economies, for example, finds that the contribution of new technologies to growth [ ] has been minimal, particularly when viewed from a macroeconomic perspective, 11 while an analysis of a substantial group of both industrial and developing countries found a marked difference between the former, experiencing strong links between ICTs and economic growth, and the latter, where no noteworthy impact was identified. 12 30. There is almost no body of research on a single country which can corroborate or challenge these findings, although researchers have suggested a number of factors that might explain them. Some are methodological to do with weaknesses in the data available from developing countries, the age of the data (which makes it more difficult to assess the impact of very recent changes in investment patterns) and 10 Cf. See section Differences between OECD and developing economies p. 12. 11 Piatkowski (2002). 12 Pohjola (2000). 11

the unsuitability of existing national economic indicators for measuring efficiency gains resulting from new technologies with substantial network effects. These methodological problems may well be important, and further theoretical and econometric work to improve the quality of the analytical tools would be useful, alongside more detailed research on individual countries and cross-country studies. The research that is available does indicate that there are substantial social and economic factors which will influence (or are likely to influence) the impact of ICT investment in developing countries in different ways from the industrial countries which are members of the OECD. Differences between OECD and developing economies 31. Obviously, real-world economies cannot simplistically be divided between the industrial countries in the OECD and a broadly homogeneous group of developing countries outside it. There is a continuum in types and degrees of development between OECD and non-oecd countries, with considerable overlap between them. Some countries which were considered developing thirty years ago are now highly industrialised, such as Korea and Singapore. Transition economies in Eastern Europe and the Former Soviet Union and middle-income economies in Latin America and parts of Asia share economic characteristics with both OECD and less developed regions. For example, their domestic markets include both substantial populations with significant disposable income and large numbers of people without. Their response to new economic impulses such as the opportunities represented by ICTs, however, may well be closer to those of OECD members than to LDCs, particularly LDCs with a high degree of donor dependence. There are also substantial differences between the economic behaviour of large low-income economies (such as India and China) - which have major industrial capacity and mass markets in spite of low per capita GDP - and that of smaller, less well-resourced, low-income countries (such as many of those in Africa). 32. The best way to identify possible reasons for different relationships between ICTs, productivity and growth in economies with different degrees of development is to juxtapose the characteristics of industrial/oecd countries with those of LDCs, whose economies diverge from the OECD type most markedly. General economic factors 33. Obviously, ICT investment forms a much smaller proportion of investment in LDCs than in OECD countries, for a number of reasons. OECD economies have large and established service and manufacturing sectors, while LDC economies are dominated by raw material production and domestic or subsistence agriculture. The service sector makes intensive use of ICTs, as does much modern manufacturing in high-income economies, but ICTs add much less value and form a much lower proportion of investment in extraction industries and (particularly small-scale) agriculture. The volume of ICT investment is therefore likely to be much higher in OECD economies than in LDC economies particularly those LDC economies such as Mozambique and other overwhelmingly rural countries in sub- Saharan Africa where subsistence (unmonetised) production forms a sizable proportion of total output. Some OECD economies have substantial ICT production sectors and most have substantial ICT service sectors, which invest directly in ICT products. Very few low-income countries have ICT production sectors of any size, and only those few which can offer both low labour costs and relatively high educational (including international language) skills - for example India or Jamaica, but very few LDCs - can develop significant ICT service sectors (such as software development, call centres and back-office outsourcing). Even where they are established, such service sectors are 12

export-oriented, depend on imported equipment and have few backward and forward linkages into domestic economies. Their networking impact will therefore be much weaker than that of equivalent sectors in industrial countries. OECD economies have large mass markets for products and services, including both ICT hardware (PCs) and usage (telecommunications), which reduce the unit costs of both ICT products and services and facilitate ICT diffusion and use. With the exception of a few very large countries, LDCs lack mass markets for consumer goods and services, including ICTs (although public-access telephony can be regarded as a mass-market service), resulting in higher costs and less efficient use by inexperienced workers and consumers. Labour costs are much lower in LDC economies than in the OECD, and most LDCs have a surplus of low-skilled labour available to undertake jobs which have been automated in OECD countries. The savings firms make by substituting capital for labour in high-wage economies may not arise in comparable firms and sectors in LDCs, regardless of any gains in individual employee productivity. In particular, the costs of adapting the workforce to new technology may outweigh the returns likely to result from higher productivity. ICT-specific factors 34. The duration of the time-lag between ICT investment by firms and any resulting improvement in national economic performance is affected by the extent of ICT diffusion within society and the development of networking between businesses and other organisations: OECD economies have established high-quality communications infrastructures which are geographically universal within their own territories and interconnected with other countries in ways that facilitate high-speed communications and transactions. In many cases these include broadband networks enabling near-universal high-speed Internet access. Most LDCs have poorquality fixed communications infrastructures with limited geographical availability within their own territories and expensive, poor-quality external connectivity constrained by shortages of international bandwidth, thus reducing the value added by ICT investment and detering the development of domestic ICT sectors. ICT investment costs are generally much higher in LDCs where almost all ICT equipment must be imported (often subject to high rates of taxation and non-tariff barriers), and where telecommunications usage charges are generally much higher than in OECD countries (especially for international and Internet connectivity). Regulatory factors such as licence fees often also add to the cost of ICT investment. The net result is that every dollar of ICT investment in an LDC buys significantly less ICT equipment and usage than in the OECD area and is therefore likely to have a significantly lower rate of return. A mass market for ICT equipment in OECD countries also helps to minimise costs of ICT investment and enable firms to benefit substantially from the continually falling prices which characterise the sector. The historic evidence shows a close correlation between teledensity 13 and per capita GDP (or, in effect, disposable income). OECD countries have mass markets in which almost all citizens are accustomed to interacting (networking) for personal and business transactions through ICTs (the telephone and, increasingly, the Internet). A high proportion of LDC citizens have little experience even of telephony. Adoption rates for new ICTs are likely to be much faster in societies in which citizens are used to older ICTs, in terms of both ownership and 13 That is, the number of telephone lines per hundred citizens or households. 13

usage, with effects on the pace with which the price of new ICT products and services falls and the pace with which new networking opportunities become established. The rapid growth of Internet in the United States, for example, took place within a population that had enjoyed residential telephone service far longer than those in Western Europe Complementary factors 35. The OECD report identifies a number of complementary factors within the business environment which influence the ability of firms and national economies to achieve productivity improvements and growth through ICT investment, among them: the extent of competition and nature of the regulatory environment the amount and quality of human capital available the capacity of firms and other organisations to adapt their working processes to take advantage of new technologies. 36. The business environment in each of these areas is more conducive to ICT investment, diffusion and use in the OECD than in LDCs: OECD countries generally have legal and regulatory business frameworks which facilitate or reward entrepreneurship and innovation and encourage inward investment. Licensing and standardisation requirements are generally more straightforward than in LDCs, although there may be more restrictions on labour conditions and employment. Telecommunications markets have generally been liberalised and made subject to competition-oriented regulation, with beneficial effects on the quality, availability and price of services. There are few if any restrictions on the role of external capital. LDC markets are often less open, as well as less attractive, to international trade and investment and have slower and sometimes more burdensome regulatory requirements, especially in areas such as business registration and licensing. Although telecommunications restructuring and liberalisation are now being widely implemented in LDCs, there are still extensive areas of monopoly and regulatory bottlenecks which affect the availability and price of telecommunications services for example, restrictions on the ability of firms to use independent VSAT (Very Small Aperture Terminal) 14 services for international data links. OECD countries have extensive human capital available to use ICTs, including highly skilled workforces benefiting from universal education which generally includes basic ICT skills and widespread specialist training in the ICT sector. ICT investment in such countries is often accompanied by an upgrading in average workforce skills skill-based technological change. Most LDCs have very limited pools of skilled personnel available for ICT work, while their education sectors are under-resourced and ill-equipped to provide even basic ICT training on a selective basis. Lower-skilled personnel often lack the language skills required for effective use of ICTs, especially the Internet, while highly skilled personnel are able to seek more lucrative employment in an international market. Even routine maintenance of ICT equipment is constrained by shortages of expertise, adding to costs and downtime, reducing reliability and the value ICTs can add to productive processes. 14 Very Small Aperture Terminals are a type of ground station used to contact communications satellites for data, voice and video signals (excluding broadcast television). A VSAT has an antenna dish that is smaller than 3 meters, as compared to around 10 meters for other types of satellite dishes. 14

OECD firms benefit from the widespread availability of investment finance, from the confidence of investors in their business expertise and the low-risk business environment, and from extensive experience in restructuring business operations to take advantage of new technological and management techniques. Managing change, which is crucial to firms ability to maximise returns on ICTs, is part and parcel of business culture in the OECD, especially in larger and multinational firms. Companies in LDCs have much less experience of restructuring and much less access to high-quality business advice and venture capital. Although there is considerable dynamism in small and medium-sized enterprises in the ICT sector in many LDCs, it is often under-financed and under-resourced in comparison with similar enterprises in the OECD. 37. These fundamental differences make ICT investment in developing countries more expensive, more difficult to implement or less cost-effective, limiting both its impact on productivity and the ability of ICT-investing firms to gain competitive advantage. The cumulative effect of factors such as these is likely to be substantial, especially in slowing the aggregate increase in productivity achieved by firms across the economy as a whole and it helps explain the difficulties researchers have in identifying improvements in national economic outcomes from ICT investment. 38. Many of these factors are also present, at least to some degree, in transition and middle-income economies and would likewise help to explain the difficulty in identifying improvements in their macroeconomic performance although in both low- and middle-income economies, these factors tend to inhibit, not to prevent, ICT investment. They do not suggest that ICT investment will not happen or that benefits in productivity and economic growth will not arise, but that ICT investment will be slower and that the benefits will be slower to materialise. They also point to ways in which businesses, governments and international agencies can act to increase the pace of ICT investment (where this would be appropriate) and the rate at which benefits in national growth are likely to result. 39. There is no reason to suppose that, with time, the advantages which accrue to firms in OECD countries will not also accrue to LDC firms that similarly invest in both ICTs and the organisational changes required to maximise their value. 15 Policy initiatives that facilitate improvements in the business environment to encourage effective ICT investment are likely, therefore, to be no less valuable for individual firms in LDCs than for their peers in OECD countries and for the sectors in which such firms are congregated. Of course, it will take longer for the impact of productivity improvements to feed through into national outcomes, but that does not alter the fundamental relationship between improvements in micro- and macro-economic performance, nor reduce the importance of positive changes to the business environment. 16 The relationship between economic growth, development policy and the digital divide 40. The interest of development agencies in ICTs is not so much the impact which ICTs may have on business or national economic performance but their potential to address the Millennium Development Goals 17 within their overall focus on poverty reduction. ICT applications clearly have potential to enhance the delivery of mainstream development goals (in health, education and so on), regardless of the effect of the ICT sector on national economic performance. Although the achievement of these development goals is 15 Increased ICT investment, though, may initially tend to give advantage to multinational companies at the expense of domestic firms, since they are more likely to invest in ICTs at an early stage and can make more substantial efficiency gains through intra-company networking than their developing-country competitors. 16 This point is expanded in the section on Policy recommendations, p. 19. 17 The UN Millennium Development Goals (MDGs) of September 2000 are available online at www.un.org/millenniumgoals. 15

likely to be easier in a context of economic growth, the relationship between the two is dualistic: economic growth is also more likely to be achieved in societies that are healthier and better-educated, where individuals and communities have the skills and capacity to fulfil their potential and to develop new business opportunities. ICTs can thus play a part in improving both national economic performance and mainstream social development, a dual potential which should be better understood by policy-makers in both ICT and development fora. Box 4. The Digital Divide A major pre-occupation in the literature on ICTs and development has been the question of the digital divide. It is often illustrated by data on access to particular ICTs. The digital divide is defined as the disparity in ICT diffusion and use between industrial and developing countries (or, indeed, between rich and poor, men and women, urban and rural areas within individual countries). For example, data published in 2002 showed that although the average OECD country has roughly 11 times the per capita income of a South Asian country, it has 40 times as many computers, 146 times as many mobile phones, and 1,036 times as many Internet hosts. 18 In many ways, this digital divide merely parallels similar disparities in access to and use of other development goods health, education and so on which are more readily available to rich than poor, or in industrial countries than in developing countries. A digital divide is to be expected: the key questions for policy-makers are the extent to which it matters (in terms of equity and any secondary effects in other sectors) and to which it is likely to grow or diminish over time, and the identification of ways in which it might be bridged. Source : OECD 41. Opinion on the effectiveness of ICTs in poverty reduction and in delivering mainstream development goals remains divided, not least because here as with its impact on economic growth there is a shortage of impact-assessment research to complement the findings of case studies and pilot projects. Yet there is a growing consensus that ICTs can make a valuable contribution to the delivery of mainstream development-sector goals, provided that they are used appropriately and in circumstances where their potential value has been carefully assessed. There is growing consensus, too, on the impact of the ICT sector in and of itself on poor communities for example, that telephony access is of intrinsic value to citizens of low-income communities as well as within the parameters of specific development projects. Very little research has been done on this latter issue, though what there is suggests that low-income communities use some ICTs, when they become available, much more extensively and dynamically than had been anticipated by policy-makers and suppliers particularly where they do not require significant new skills or resources (radio, voice telephony) and where they bypass or substitute for less effective alternatives (transport, postal services). 19 42. A central part of discussions on the digital divide concerns the role of ICTs in facilitating the availability of information. Modern production is increasingly knowledge-intensive. Knowledge and 18 World Bank (2002): Information and Communication Technologies: A World Bank Group Strategy, p. 5, citing Pyramid Research. The ratio for mobile phones, at least, may now be lower. 19 Cf., for example, research conducted for the Knowledge and Research programme of UK Department for International Development on the use of telephony in low-income communities in Botswana, Ghana and Uganda [McKemey, Scott, Souter et al. (2003)]. 16

access to information are ever-more important assets for individuals, communities, businesses and countries in competitive markets. They also facilitate basic social empowerment and opportunity for individuals and communities. The costs of gathering, processing and distributing information are higher in developing countries, 20 and ICTs have the potential to extend the availability of information so dramatically that some talk of an Information Revolution equivalent to the Industrial Revolution of two hundred years ago. The extent to which this Information Revolution reaches into individual economies, it is argued, profoundly affects their ability to grow economically and to address social challenges to improve their national social and economic indicators both absolutely and relative to other countries. 43. In terms of national economies, where cross-country economic comparisons are required and especially where the relationship between rich and poor countries (broadly equivalent to OECD countries and LDCs) is concerned the question of the digital divide is much more to do with national economic performance than with mainstream development objectives and poverty reduction. Opinion can broadly be divided into two camps: Digital optimists have argued that ICTs offer developing countries, LDCs included, an opportunity to leapfrog stages of technological development and compete in ICT/knowledge areas with industrial countries on more equal terms than they have done in the past. Digital pessimists believe, by contrast, that digital divides are likely to grow over time as ICTs become increasingly pervasive in industrial countries while most developing countries, particularly LDCs, lack the critical mass in terms of expertise and local markets to follow suit. 44. Two characteristics of ICTs seem particularly important here. The first is the importance of network externalities. 21 If, as a result of network externalities, the value of ICTs grows more quickly the more widespread they are diffused within society, the digital divide is likely to increase, at least until societies with less widespread diffusion of ICTs reach any threshold beyond which more rapid growth is likely to occur. 45. The second is the pace of change in ICT development itself, which far exceeds the rate of change in earlier technologies. Driven by the factors underpinning Moore s Law, 22 advances in information technology follow one another at enormous speed, and technologies and infrastructures that seemed advanced five years ago are already giving way to improved alternatives. Disappointingly for optimists, developing countries that seek to leapfrog earlier stages of technological development are likely to find that their new technological base is rapidly eclipsed by further technological advances in industrial countries. By the time most developing countries have rolled out second generation mobile telephone networks, for example, OECD countries will have deployed third generation networks offering much greater functionality. The benchmark for technological advance is constantly moving. While developing countries can, in certain circumstances, leap over intermediate technological stages historically required in industrial countries (moving, for example, from no telephony to mobile telephony in rural areas without 20 Cf. Bedi (1999). 21 Cf.See Box 2. 22 In 1965, Gordon Moore observed an exponential growth in the number of transistors per integrated circuit and predicted that this trend would continue [see Moore (1965)]. His observation has developed into a general observation of the continued rate of very rapid growth in ICT capacity. 17