Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

The overall scope of R&D statistics among developing countries

The major problem in international R&D comparisons that include the developing countries is not the intricacies of statistical methodology but the simple question of reliable sources.

The evolution of R&D and innovation indicators among the developing countries does not follow the same trajectory as that of the industrialized part of the world. Several developing countries were among the pioneers in national R&D statistics and, since the late 1960s, many more of them have begun producing. Until now, however, only a few developing countries have followed the more ambitious R&D statistical path set by the OECD countries. Useful summaries of S&T indicators for a whole developing region have been provided by GRADE, a centre for development studies in Peru (see ref. 42, with the revised figures for the early 1980s in ref. 41).

For a number of national statistical agencies, R&D and innovation are not among the top priorities. In dozens of developing countries the status and quality of R&D statistics have even deteriorated. During the 1980s some governments simply stopped publishing internationally comparable R&D data, while others provide only fragmented R&D statistics on an ad hoc basis. In fact, Dedijer's characterization earlier of national statistical priorities is not outdated, although many more statisticians have been trained to produce R&D data. These professionals have had few chances to improve their skills.

Today, the universe of R&D statistics in Africa, Asia, and Latin America ranges from crude summaries of the total number of formally trained scientists in a given country to very detailed data, describing in full-time equivalents the number of scientists engaged in R&D activities and the current expenditures available. Most developing countries have low statistical ambitions in the field of R&D and innovation.

Given the relatively poor state of R&D statistics in developing countries, no general diagnosis can be made of the more than 130 "national R&D systems" in that part of the world, neither of their efficiency nor of their economic and other effects. And no demarcation lines can be drawn between scientific research, technological development, and other innovative work. Anyone interested in detailed international comparisons ought to be concerned about the deplorable fact that "science and technology," "applied research," "experimental development," and similar notions refer to slightly different activities in different countries and are usually performed by different organizations with different objectives. Using currently available R&D statistics, little can be said about the strength and relative performance of R&D activities in what was once called the third world.

For a developing country it is not enough to count the stock of resources currently available for innovative activities. It has become increasingly important to measure the R&D potential and also - by way of statistics - to reveal immediately available R&D resources such as highly qualified manpower not involved in R&D activities. In order to lay a basis for change, the flow of resources over time should be included.

Even if statisticians of all countries can agree on the feasibility and usefulness of a detailed world handbook of R&D and innovation statistics such as Dedijer proposed [14], many elementary problems of a common methodology remain to be solved. Unesco experts in the developing countries do not lack work opportunities. Due to the differences in the availability as well as the quality of data, one cannot be careful enough in drawing empirical conclusions from international R&D statistics, especially when comparing industrialized and developing countries.

My intention has not been to criticize R&D statistics in general, but to underline the differences in quality and, hence, the problems of reliability and validity in international comparisons. Everyone should be aware of the elementary state of the art. Later in this chapter I shall discuss the opportunities. With limited resources and relatively simple means, much more professional statistical work could be done in the developing countries and useful comparative data on both R&D and other innovative activities would become generally available.

In the following pages, I concentrate on basic input data such as resources devoted to research and experimental development and tackle a much-debated issue: Is the relative position of the developing countries in science and technology improving? Are the developing countries actually spending more on R&D than they did 20 or 30 years ago? As a key indicator, I use their expenditures on R&D, comparing them with the industrialized countries.

Has R&D spending by developing countries increased?

During the first two "development decades" proclaimed by the United Nations, one set of figures was frequently used to describe the international division of labour within world science and technology. The numbers 70 - 28 - 2 were quoted in most documents dealing with global science and technology policy. The figures showed that 70 per cent of total R&D funding was spent by the USA, 28 per cent by the other market economies, and, so it was claimed, only 2 per cent by the developing countries. This set of figures was widely quoted by the United Nations and others, simply because there were no other data available.

This uneven relationship in R&D between the three major groups of countries was a fact in the first half of the 1960s. But the figures of those years did not project a global picture of all R&D resources. The centrally planned economies - what were then the socialist countries at different levels of development - were not included. Moreover, R&D statistics did not even exist for a number of developing countries and were of poor quality in some developed market economies. But still, for many years, these figures for R&D funding were the best available.

Table 1 Distribution of R&D expenditures (estimated in billion US dollars [current prices] and in percentage of annual totals) among selected industrialized market economies and developing countries in 1963/64, 1977, and 1988

  1963/64 1977 1988
United States of America 70 46 40
Other industrial market economies (except Australia and New Zealand) 28 51 56
Developing countries (sample) 2 3 4
Total (%) 100 100 100
Total (billion US$) 29 97.5 340

Sources: The statistics for this chapter, including the data presented in this table, are drawn from four major sources: (1) An OECD Development Centre study of the world distribution of R&D resources (refs. 1,2), which was subsequently up-dated (ref. 3); (2) R&D data collected and processed by regional organizations such as the OECD (refs. 26-28, 31, 47. 6, 42, 41, 37); (3) Unesco's Statistical Yearbook and R&D data processed for comparative purposes (refs. 49, 50, 54, 55); (4) National R&D statistical publications and selected independent studies (refs. 8, 11. 12, 18, 23, 22).

Nearly US$29 billion was spent on R&D by the countries included in the 1963/64 UN comparison (for further details see ref. 2). Thirteen years later, in 1977, the same country grouping spent about US$97.5 billion. In 1988, nearly a quarter of a century after the first estimate, their R&D spending was about US$340 billion (all dollar values in current prices). Not only has the magnitude of the R&D efforts changed, but so have the relations between the country groupings and within the group of developed market economies.

During the past 25 years, as shown in table 1, the share of global R&D spending by the USA has decreased from 70 per cent to 40 per cent, while the other Western industrialized countries, which were included in the original sample, have doubled their common share to 56 per cent of the grand total. Since 1980 the redistribution of the financial R&D inputs among the industrial nations has continued. In particular, Japan, France, Germany, Italy, and some of the small, highly industrialized countries like Sweden and the Netherlands have expanded their gross domestic expenditures on R&D more than the average OECD member country.

During the late 1970s and in the 1980s some of the western European countries did not expand their gross expenditures on R&D as much as the largest industrial countries. In relative terms this was a decline in European R&D activities.

During the same 25 years, the R&D position of the developing countries evolved a little differently. (Here the "developing countries" are those in the original sample from 1963/64 only.) In the 1960s their share of total R&D expenditures grew from 2 per cent to 3 per cent. In current US dollars, R&D spending went up more than three times over a 10 year period, to US$1.7 billion in 1973. Towards the end of the 1970s, it is estimated that the same grouping of countries spent a little more than US$3 billion on their R&D activities. This was 3.1 per cent of the grand total. Mainly because of a decline in R&D spending among the OECD countries, the relative share of R&D funds available among the countries in Africa, Asia, and Latin America grew to about 5 per cent of the grand total during the first half of the 1980s.

None the less, improvements in the relative position of the developing countries in the late 1970s and early 1980s were not the result of a general increase in R&D spending. First and foremost the countries with the largest R&D systems in Asia and, to some extent, in Latin America increased spending on R&D. So did a few, but not all, of the Arab states. These enhancements were big enough to affect the position of the developing countries as a group. Then came a relative decline. By 1988-1989, the developing countries had a little less than 4 per cent of total R&D funds, and preliminary estimates for the early 1990s show a further decline.

Yet, before concluding that the North-South division of labour in R&D proved to be rather stable up until the late 1970s, and that changes then appeared that implied a better, but far from enduring, position of the developing countries, we should delve a little deeper into available statistical data. A more representative global picture, based on both national and regional statistics, is presented in table 2.

In which regions are the world's R&D resources concentrated?

If all countries - and not just the previous sample of countries - that have recently produced R&D statistics are included in a world total, they devoted about US$435 billion to R&D in 1988. More than 96 per cent was spent by the industrialized countries, while the developing countries accounted for the remaining 3.9 per cent of global R&D finance. In current prices, the developing countries - still taken as a group- had a 1988 R&D budget that was nearly six times that of 1973. Their R&D growth rate was significant in the late 1970s and continued to grow for a few years in the early 1980s, but it has since declined.

Table 2 Distribution of world R&D expenditures (estimated in billion US dollars [current prices] and in percentage of annual totals) among major groupings of countries in 1973, 1981), and 1988

  1973 1980 1988
Developing countries 2.8 6.5 3.9
  Latin America and the Caribbean 0.8 1.7 0.7
  Africa (except the Arab states and South Africa) 0.1 0.3 0.1
  Arab states 0.2 0.4 0.1
  Asia (except the Arab states, Japan, and South Korea) 1.7 4.0 3.0
Industrial countries 97.2 93.5 96.1
  Japan and South Korea 7.9 10.2 19.3
  Australia and New Zealand 1.2 0.9 0.9
  USSR and Eastern Europe 33.0 27.2 17.3
  Western Europe 21.6 24.2 25.8
  North America 33.7 31.1 32.8
  World total (%) 100 100 100
  (in billion US$) 97 218 435

Sources: Table 1; percentages are calculated and rounded within each grouping of countries.

In a global perspective, until the early 1980s, the R&D position of the developing countries greatly improved. Using only 2.8 per cent of all R&D money in 1973, and even a somewhat smaller share during the rest of the 1970s, the share came to 6.5 per cent in 1980 and higher in the next few years. However, during the rest of the 1980s most developing countries did not expand their R&D budgets relative to the industrialized countries; they contracted. The industrialized countries, particularly those with large market economies, regained some of their lost positions from the developing countries.

It must be remembered that the developing world is a heterogeneous entity. In 1980, nearly two-thirds of their R&D dollars were spent by countries in Asia, particularly those with relatively large R&D systems such as China and India, but also by Indonesia, Taiwan, and Thailand. Other countries with small or medium-size R&D systems, e.g., Pakistan and Malaysia, have also expanded their R&D finance, although not as much as the largest of the emerging industrial countries of Asia. For 1988, it was estimated that 3 out of 4 R&D dollars in the developing world were spent by East and South-East Asian countries. Today, more than six out of ten developing country researchers are Asians. Latin America and particularly Africa lost their previous strengths in R&D finance during the 1980s.

As a region with more than a quarter of R&D funds among the developing countries in the early 1980s, Latin America with the Caribbean has lost ground to other regions. But the countries with the largest R&D systems, e.g. Brazil, Argentina, and Mexico, seem to have kept relatively high rates of expansion even in the years with severe fiscal problems. In the early part of the 1980s, two of the three countries mentioned were spending more on R&D relative to other economic activities than the average developing country. In the last few years, however, the situation seems to have changed. By the early 1990s, the average Latin American country was not following the pace set by leading Asian countries; the continent as a whole is now lagging behind in R&D spending.

If the Republic of South Africa is excluded from comparisons, Africa south of the Sahara is still very much part of the old third world R&D desert. In all but a few African countries, R&D resources are comparatively scarce. Still, as a region, there are signs of change. Measured in percentages, sub-Saharan Africa's share of global R&D expenditures more than doubled between 1973 and 1980, then dropped to the earlier level by the end of the decade. There was no growth of expenditures; in fact, in the average African country there was a decline during most of the 1980s.

Among the industrialized regions of the world, the country grouping with the highest growth rate is Japan and South Korea. During the last seven years of the 1970s their annual R&D budgets nearly trebled (in current US dollars). By 1980, these two countries in East Asia together accounted for considerably more than the whole third world R&D spending. According to more recent statistics, their yearly R&D inputs have grown even further. As shown in table 2, their gross domestic spending on R&D in 1980 represented a tenth of total R&D finance in the world; eight years later it was more than 19 per cent.

In the 1970s, the western European countries were nearly as fast-growing spenders on R&D as Japan. Their investments in R&D grew by nearly 2.5 times in current US dollars and their international position went up from 21.6 per cent in 1973 to 24.2 per cent in 1980. More recent statistics show a steadily high average growth rate for many of the European countries. But now, the growth is not as significant in relative terms. Western Europe kept its strong position during the 1980s, but only with some difficulty.

On the other side, the eastern part of Europe and the USSR - here still presented as a single bloc of countries - have weakened their position as big spenders on R&D. In 1973 their gross national R&D expenditures were estimated to represent a third of the global total, while they spent only 27 per cent in 1980. With much less than a fifth of the world total by the end of the 1980s, the decline has continued.

In the aftermath of the 1989-1991 revolutions in eastern Europe and the former USSR, it has been revealed that R&D spending in the 1960s and 1970s may have been estimated and officially reported higher than it actually was. So far, however, there are few data available for better founded international comparisons [22].

In relative terms, the North American region, primarily the USA, also declined during the 1970s and into the 1980s. In seven years the share of world R&D fell from about 34 per cent to 31 per cent, while in absolute terms, in current US dollars, the R&D budgets doubled. The following seven years proved to be a period of stabilization, even growth. By the end of the 1980s the two countries accounted for nearly a third of the world's R&D expenditures.

In conclusion, financial resources devoted to R&D in the 1970s grew substantially faster among the developing countries than among the industrialized ones. But for the 1980s, it is fair to say that the highly industrial countries of the North regained nearly all of the lost ground. To the average industrialized country, R&D has become a much more significant element in the build-up of innovative capabilities.

Because of fluctuating exchange rates, variations in local purchasing power, and other problems of measurement, international comparisons of R&D expenditures do not always reflect the magnitude of available resources. Our North-South picture changes somewhat, though not dramatically, if we look at R&D manpower data.

By the end of the 1980s, the developing countries employed 18-19 per cent of the world's researchers (scientists and engineers engaged in R&D). This is a much larger share than for R&D expenditure, but differences between industrial and developing countries remain great. For instance, in Africa, Asia, and Latin America, average overhead costs for laboratory equipment are relatively small. The typical third world researcher in the 1980s did not have similar or the same working conditions as a scientist or engineer in an industrial country. The productivity of the R&D activity is affected.

Table 3 Distribution of the world's researchers (scientists and engineers engaged in R&D; estimated in full-time equivalents) among major groupings of countries at the end of the 1980s

Developing countries 18.6
  Latin America and the Caribbean 2.9
  Africa (except the Arab states and South Africa) 0.6
  Arab states 1.3
  Asia (except the Arab states, Japan, and South Korea) 13.8
Industrial countries 81.4
  Japan and South Korea 11.7
  Australia and New Zealand 1.0
  USSR and Eastern Europe 26.6
  Western Europe 16.7
  North America 25.4
    World total (% ) 100
    (in full-time equivalents) 4,130,000

Sources: Calculated from Westholm [56] (which is based on Unesco, OECD, and national manpower statistics), supplemented by selected national data. Percentages are calculated and rounded within each grouping of countries.

Table 3 provides a global estimate for 1988-1989 of 4.1 million researchers in full-time equivalents [56]. The industrialized countries (including countries like South Korea) employed about 82 per cent of these. Leaving behind both the manpower indicators and the crude and simple R&D expenditure data, a science and technology-related typology might be more useful in discriminating among the developing countries. Their strengths and weaknesses will be more visible if we look at their relative position in R&D-related technical change. Such a typology could reflect several criteria, such as the economically active population; the sectoral distribution of specialized manpower in relation to science and technology; and the size and structure of the domestic product (GDP), including the share of R&D (see, for example, ref. 20, pp. 55-77). Given the current state of international statistics, such a worldwide typology could not relate resources and capabilities in science and technology directly to the country's economic performance nor to the competitiveness among its main industries. Anyhow, by grouping all countries according to a set of available indicators, it will be easier to identify countries with (a) no science and technology base, (b) fundamental elements of a science and technology base, (c) a science and technology base well established, and (d) an economically effective science and technology base, notably in relation to industry. The last grouping is identical with the highly industrialized countries, while the three others belong to the developing world.

The first grouping of developing countries numbers about 55, including most African countries. These are countries with no science and technology base, still in the initial stage of development, with low GDP per capita, low science and technology manpower potential, and a low share of manufacturing of total production.

The second grouping of countries, which have essential elements of a science and technology base, are in the process of industrialization. With moderate GDP per capita, they have developed a limited endogenous industrial production. Some of them may have a relatively high percentage of science and technology manpower that could be activated in R&D, but the potential is low in absolute terms. This second group represents nearly 40 developing countries and includes Algeria, Ghana, Indonesia, Iraq, Malaysia, Paraguay, and Sri Lanka.

The third group of countries, with a high percentage of potential science and technology manpower, have a solid science and technology base and a functioning industrial system. Their GDP per capita is relatively high. This grouping covers about 40 developing countries, including the "newly industrializing countries" (the NICs) in Asia and some Latin American countries such as Argentina, Brazil, Mexico, and Venezuela.

Two developing countries are difficult to fit into any of the above categories or groupings of countries. China and India have to be treated separately: they both have a low GDP per capita; at the same time, due to their size, they have a huge science and technology manpower potential in absolute terms, but low as a percentage of total population or in relation to the economy. However, manufacturing represents a large share of their total production.

This science and technology-related typology of countries does not necessarily correlate with economic performance. As discussed earlier, innovative capabilities may develop from many different sources. And the linkages between the economic actors, including government agencies, in a national, regional, and international setting might prove crucial for industrial competitiveness.

In the final sections of this chapter, I examine the intensity among these linkages or couplings at the regional level of the world economy and consider how they are being measured.

Science, technology, and new economic patterns

Following the emergence of not just one but of several major centres of science and industrial technology, the world economy has become much more integrated and interdependent. But there seem to be limits even to the processes of internationalization

Over the last two decades, data on R&D, innovation, and trade patterns in high-tech products make a distinction possible between three dominating regions of the world economy. Each of them forms a separate supply base for industrial development and production, although related to the other two and to other regions of the globe. There is an East Asian industrial space with Japan at the centre; a North American one with several industrial zones in the USA as its core; and a western European economic space with a handful of technologically important national economies. Based in these regions, about 1,000 major corporations control more than half of the world's manufacturing and almost two-thirds of international trade, much of which is in fact intra-regional trade. The three regions have been further consolidated in the last few years.

R&D and related economic statistics, which reflect both the diversity and integration of these regional supply bases, are available but not always beneficial for detailed comparisons. "The data collected systematically at the international level has until now primarily addressed international trade, patent applications and, to a lesser extent, capital movements (in a relatively aggregated form). In addition, private data banks have recently been developed for data concerning different categories of inter-firm cooperation agreements" [34, p. 10]. More is being done, especially through the OECD, to further develop data into sets of comprehensive indicators that could better describe the changing regional and national conditions for innovation.

From a statistical point of view, little is detectable of the "global reach" of large industrial corporations that operate from the three major industrial supply bases. There are no indicators at the firm or at the branch level of industry revealing the role of science, technology, and innovation in the expansion into other regional markets for products and services. Nor are there statistics of the transfer of technology and other knowledge in their quest for foreign supplies and new sources of production. The operations by corporations on patent protection, licence agreements, and royalty issues are not recorded in any databases.

In the polycentric economic context - with corporations based in three high-technology regions as the principal players - "globalization" refers to a set of emerging conditions in which value and wealth are produced and circulated by way of both regional and worldwide communication networks. The transnationally managed firms operate concentrated, even oligopolistic, supply structures by way of modern technology. For banks and other credit institutions as well as for large manufacturing firms and many service producers, modern communication technologies make it easier than before to manage intracorporate information networks on a global level. Directly or indirectly, some of the small and medium-size firms in the three regions are linked to the same systems. As product and technology life cycles become shorter, this helps in the functional integration as well as in the economic fortification of the three dominating industrial regions.

Whether big firms or small, technological strengths and competitiveness are not determined solely at the level of the industrial firm, but also by the economic environment in which firms operate. Today, more than earlier, managers and entrepreneurs need to combine indicators in the economic environment that influence technical change and industrial innovation.

The developing country firms have similar needs, but face different challenges at home and, more importantly, in the international market-place. The pressure on them from abroad has increased tremendously. Technical change of their industry is very much needed. But the deterioration of the terms of trade, in particular the overall decline of prices of primary products over the past 10 years, the ups and downs of energy costs, the worsening balance of payments caused by the rise of interest rates on loans and credits, the repatriation of foreign investments, etc., have forced countries in the developing world to eliminate research projects, reduce experimental development, and downgrade or close R&D laboratories and related institutions.

To develop policies that could avoid further marginalization in foreign investment and technology transfer, the developing countries need much more detailed and statistically grounded analyses of the role of science and technology in the globalization process. We should not forget that most of the currently available R&D and innovation indicators were created in a specific national or regional context. Although they have been further developed out of broader policy needs, they do not take into account the current internationalization of the national economies. Available R&D statistics do not capture well the new forms of technology-based competition or contemporary economic interdependencies.

Innovation indicators in the making

Over the past three or four years, R&D statisticians in the industrialized countries have speeded up the improvement of indicators describing the role of technology in industrial innovation, human resource development, industrial performance, and international competitiveness. Similar work is done to reshape indicators of scientific research and of R&D performed within the public sector with government objectives. The reasons are obvious.

At present it is not possible to quantify such important tendencies as direct international investments by sector or product area, international flows of technology (licences, patents, know-how), interenterprise and intergovernmental technical cooperation agreements, and the international diffusion of high technology incorporated in goods. Data are available on various forms of localized R&D, but the level of aggregation is usually too high for detailed analyses. Typically, changes concerning different forms of relocated R&D are not very well described in current statistics. The same is true for detailed data on transborder flows of researchers or, more generally, of scientists and engineers.

For a developing country, these kinds of R&D and innovation statistics may become strategic for situating the country's economy and its industrial firms in the changing regional and global contexts. Data on technical standards and the protection of these, access to specific technologies, intellectual property rights, and competition policies in different regions and countries could be fundamental ingredients in government policies and corporate strategies.

The uneven performance of national economies sometimes leaves room for doubt as to the possibility of a balanced sharing of gains from trade among the industrialized and the developing countries. To shed light on specialization patterns and the changing international division of labour, more statistics in the form of economic and technical intelligence is needed. Subsequently - as in the highly industrialized countries - models for interpretation should first be constructed and, following this, the most relevant indicators be defined. Then the corresponding primary and secondary data should be processed with the most significant policy objective or corporate strategy in mind.

Data should be functionally organized to form indicators that could help describe dynamic situations involving a cluster of firms and other organizations involved in innovative activities. A "techno-economic network" to take one illustration out of this context may be defined as a coordinated set of heterogeneous actors such as public laboratories, technical research centres, firm laboratories, financial organizations involved in industrial investment, intermediate and final users, as well as public authorities that participate actively in the design, development, and production/distribution of production processes, goods, and related services, some of which may entail a commercial transaction. The developing relationships between these actors in the innovation process may centre around the following three poles of attraction.

The first may be labelled the scientific pole. It consists of research centres, university laboratories, and company research units, where knowledge is generated. Here, activities are measured by bibliometric indicators, contracts between firms and research centres, training of personnel, migration of skilled labour, etc.

The second pole of attraction could be called the technical. Here new goods and related services are produced, i.e. prototypes, pilot projects, models, patent descriptions, etc. Activities are measured by surveys of major innovations, patent applications, the creation of new high-tech or science-based firms, licence agreements, and other forms of cooperation between technically advanced firms.

Thirdly, there is a market pole that focuses on the demand for goods and services by users and customers. Indicators here should measure and describe the main characteristics of the distribution system, provide information on user participation in the design of goods and services, particularly quality control and definition of standards.

In some developing countries, and definitely in most of the highly industrialized countries, the tendency now is to see R&D and innovation indicators as advanced tools for evaluations and assessments as well as for analysis and policy formulation. At the same time, there is widespread agreement that the traditional "factors of production" measured in relation to science, technology, and innovation - do not help much in explaining the dynamic interplay between technical change, industrial competitiveness, and economic growth. Recent advances in economic theory have to be clearly manifested in new R&D and innovation indicators, particularly concerning the place of science and technology in both macroeconomic and microeconomic models, and in the interaction between tangible and intangible investments.

Among intangible investments, R&D is by far the best measured economic activity. But existing data on patents and licences, design and engineering, manpower training, information flows, and organizational structures are seldom defined in such quantitative terms that would allow for a more detailed, comparative understanding of the preconditions and driving forces of industrial innovation. A more de tailed breakdown - by product rather than by branch or product group - is usually needed.

The complexity of such measuring tasks should not be underestimated. Already, rather simple quantitative analyses give rise to a variety of interpretations. And it is not enough to settle for existing indicators; the combinations of old and the creation of new innovation indicators must be placed at the top of the agenda, and substantial work is needed to reach a generally accepted quality in the statistical analysis [29].

Contents - Previous - Next