Absolute levels of R&D expenditures are indicators of the breadth and scope of a nation's S&T activities and are a harbinger of future growth and productivity. Indeed, investments in the R&D enterprise strengthen the technological base on which economic prosperity increasingly depends worldwide. Findings from a study of 25 countries by Porter and Stern (1999) indicate that human talent and R&D spending are among the most important factors contributing to nations' innovative capacity. Consequently, the relative strength of a particular country's current and future economy-and the specific scientific and technological areas in which a country excels-is further revealed through comparison with other major R&D-performing countries. This section provides such comparisons of international R&D spending patterns. It examines absolute and relative expenditure trends, contrasts performer and source structural patterns, reviews the foci of R&D activities, and looks at government priorities and policies. Although R&D performance patterns by sector are similar across countries, national sources of support differ considerably. In nearly all OECD countries, government has provided a declining share of all R&D funding during the past decade, whereas the industrial share of the funding total has increased considerably. Foreign sources of R&D have been increasing in many countries.
The worldwide distribution of R&D performance is concentrated in relatively few industrialized nations. Of the $500 billion in estimated 1997 R&D expenditures for the 28 OECD countries, 85 percent is expended in just 7 countries (OECD 1999d). These estimates are based on reported R&D investments (for defense and civilian projects) converted to U.S. dollars with purchasing power parity (PPP) exchange rates. (See appendix table 2-2.)
The United States accounts for roughly 43 percent of the OECD member countries' combined R&D investments; U.S. R&D investments continue to outdistance, by more than 2--to--1, R&D investments made in Japan, the second largest R&D-performing country. Not only did the United States spend more money on R&D activities in 1997 than any other country, it also spent as much by itself as the rest of the G-7 countries-Canada, France, Germany, Italy, Japan, and the United Kingdom-combined. (See appendix table 2-63.) In only three other countries-the Netherlands, Australia, and Sweden-do R&D expenditures exceed 1 percent of the OECD R&D total (OECD 1999d).
In 1985, spending in G-7 countries other than the United States was equivalent to 90 percent of U.S. R&D expenditures that year. The non-U.S. total climbed steadily to peak at 105 percent of the U.S. total in 1993. Since then, however, non-U.S. G-7 R&D expenditures have slipped back to an amount equivalent to about 98 percent of the U.S. total. (See figure 2-27.) Initially, most of the United States' relative improvement vis-à-vis the other G-7 countries since 1993 resulted from a worldwide slowing in R&D performance that was more pronounced in other countries than in the United States. That is, although U.S. R&D spending stagnated or declined for several years in the early to mid-1990s, the reduction in real R&D spending in most of the other large R&D-performing countries was more striking. In Japan, Germany, and Italy, inflation-adjusted R&D spending fell for three consecutive years (1992, 1993, and 1994) at a rate of decline that exceeded similarly falling R&D spending in the United States. In fact, large and small industrialized countries worldwide experienced substantially reduced R&D spending in the early 1990s (OECD 1999d). For most of these countries, economic recessions and general budgetary constraints slowed industrial and government sources of R&D support. More recently, R&D spending has rebounded in several of the G-7 countries (though not in France or the United Kingdom, according to the latest available statistics), as has R&D spending in the United States. Yet since annual R&D growth generally has been stronger in the U.S. than elsewhere (see figure 2-28), the difference between the U.S. and the combined other G-7 countries' R&D spending has continued to narrow.
Concurrent with the relative increase in the U.S. share of the G-7 countries' R&D performance has been a reduction in the U.S. R&D share of all OECD countries' R&D spending. In 1986 the United States accounted for 48 percent of the R&D reported by OECD countries; by 1997 the U.S share had dropped to less than 43 percent of the OECD R&D total. Part of this share reduction (perhaps up to 2 percentage points) resulted from the addition of several countries to OECD membership (thereby increasing the OECD R&D totals); worldwide growth in R&D activities, however, was a greater contributing factor to the loss of R&D share experienced by the United States. If actual "world" R&D totals were available (rather than for the OECD countries only), the decline in the U.S. share would likely be more pronounced.
Comparisons of international statistics on R&D are hampered by the fact that each country's R&D expenditures are denominated, obviously, in its home currency. Two approaches are commonly used to normalize the data and facilitate aggregate R&D comparisons. The first method is to divide R&D by GDP, which results in indicators of relative effort according to total economic activity and circumvents the problem of currency conversion. The second method is to convert all foreign-denominated expenditures to a single currency, which results in indicators of absolute effort. The first method is a straightforward calculation, but it permits only gross national comparisons. The second method permits absolute-level comparisons and analyses of countries' sector- and field-specific R&D investments, but it entails choosing an appropriate currency conversion series.
Because (for all practical purposes) there are no widely accepted R&D-specific exchange rates, the choice is between market exchange rates (MERs) (available from IMF 1998) and purchasing power parities rates (PPPs) (available from OECD 1999d). These rates are the only series consistently compiled and available for a large number of countries over an extended period of time.
At their best, MERs represent the relative value of currencies for goods and services that are traded across borders; that is, MERs measure a currency's relative international buying power. Sizeable portions of most countries' economies do not engage in international activity, however, and major fluctuations in MERs greatly reduce their statistical utility. MERs also are vulnerable to a number of distortions-currency speculation, political events such as wars or boycotts, and official currency intervention-that have little or nothing to do with changes in the relative prices of internationally traded goods.
For these reasons, an alternative currency conversion series-PPPs-has been developed (Ward 1985). PPPs take into account the cost differences across countries of buying a similar basket of goods and services in numerous expenditure categories, including nontradables. The PPP basket is therefore representative of total GDP across countries. When the PPP formula is applied to current R&D expenditures of other major performers-such as Japan and Germany-the result is a substantially lower estimate of total research spending than that given by MERs. (See figure 2-29.) For example, Japan's R&D in 1996 totaled $85 billion based on PPPs and $130 billion based on MERs; German R&D was $40 billion and $54 billion, respectively. (By comparison, U.S. R&D was $197 billion in 1996.)
PPPs are the preferred international standard for calculating cross-country R&D comparisons wherever possible and are used in all official OECD R&D tabulations. Unfortunately, they are not available for all countries and currencies. They are available for all OECD countries, however, and are therefore used in this report. Although there is considerable difference in what is included in GDP-based PPP items and R&D expenditure items, the major components of R&D costs-fixed assets and the wages of scientists, engineers, and support personnel-are more suitable to a domestic converter than to one based on foreign trade flows. Exchange rate movements bear little relationship to changes in the cost of domestically performed R&D. (See figure 2-29.) When annual changes in Japan's and Germany's R&D expenditures are converted to U.S. dollars with PPPs, they move in tandem with such funding denominated in their home currencies. Changes in dollar-denominated R&D expenditures converted with MERs exhibit wild fluctuations that are unrelated to the R&D purchasing power of those investments. MER calculations indicate that, between 1986 and 1996, German and Japanese R&D expenditures each increased in three separate years by 20 percent or more. In reality, nominal R&D growth never exceeded 12 percent in either country during this period. PPP conversions generally mirror the R&D changes denominated in these countries' home currencies.
The policy focus of many governments on economic competitiveness and commercialization of research results has shifted attention from nations' total R&D activities to nondefense R&D expenditures as indicators of scientific and technological strength. Indeed, conclusions about a country's relative standing may differ dramatically depending on whether total R&D expenditures are considered or defense-related expenditures are excluded from the totals. In absolute dollar terms, the U.S. international nondefense R&D position is still considerably more favorable than that of its foreign counterparts; the United States is not nearly as dominant, however, as when total R&D expenditures are compared. In 1996 (the latest year for which comparable international R&D data are available from most OECD countries), U.S. nondefense R&D was almost twice that of Japan's, but the non-U.S. G-7 countries' combined nondefense total was 17 percent more than nondefense R&D expenditures in the United States alone.
Between 1982 and 1990, growth in U.S. nondefense R&D spending was similar to nondefense R&D growth in other industrial countries (except Japan, where nondefense R&D expenditure growth was notably faster). As an equivalent percentage of the U.S. nondefense R&D total, comparable Japanese spending jumped from 45 percent in 1982 to 55 percent in 1990. (See appendix table 2-64.) During this period, Germany's annual spending equaled 26--29 percent of U.S. nondefense R&D spending. France's annual spending during this same period was equivalent to 17--18 percent of the U.S. total, and the United Kingdom's annual spending fluctuated narrowly between 14 and 16 percent of the U.S. total.
Since 1990, the worldwide slowing in R&D spending and the subsequent industrial rebound in the U.S. has narrowed the gap between U.S. nondefense R&D spending and that in the other G-7 countries. In 1996, the combined nondefense R&D spending in the six non-U.S. G-7 countries is estimated at $173 billion (in constant PPP dollars), compared with $148 billion (constant dollars) in the United States. Japanese and German nondefense spending relative to U.S. spending declined to 52 and 24 percent, respectively.
One of the most widely used indicators of a country's commitment to growth in scientific knowledge and technology development is the ratio of R&D spending to GDP. (See figure 2-30.) For most of the G-8 countries (that is, the G-7 countries plus the Russian Federation), the latest R&D/GDP ratio is no higher now than it was at the start of the 1990s, which ushered in a period of slow growth or decline in their overall R&D efforts. The ways in which different countries have reached their current ratios vary considerably, however. The United States and Japan each reached local peaks-at 2.7 and 2.8 percent, respectively-in 1990--91. As a result of reduced or level spending by industry and government in both countries, the R&D/GDP ratios declined several tenths of a percentage point, before rising again to 2.7 and 2.9 percent. Growth in industrial R&D accounted for most of the recovery in each of these countries. Electrical equipment, telecommunications, and computer services companies have accounted for some of the strongest R&D growth since 1995 in the United States. In Japan, spending increases were highest in the electronics, machinery, and automotive sectors and appear to be associated mainly with a wave of new digital technologies (IRI 1999). In addition, Japan's national government also has contributed to some of the renewed vigor in Japan's R&D spending. (See NSF 1997 for a summary of the Japanese government's intent to double Japan's R&D budget.)
By comparison-and with the notable exception of Canada, for which the R&D/GDP ratio has remained relatively level since the early 1990s-the other G-8 countries each report lower R&D shares now than at the beginning of the decade. The smallest share reductions occurred in Italy, the United Kingdom, and France (declining about two-tenths of a percentage point in each country, to current ratios of 1.0, 1.9, and 2.3 percent, respectively). In Germany, the R&D/GDP ratio fell from 2.9 percent at the end of the 1980s, before reunification, to its current level of 2.4 percent. The end of the Cold War and collapse of the Soviet Union had a drastic effect on Russia's R&D enterprise. R&D spending in Russia was estimated at 1.4 percent of GDP in 1991; that figure plummeted to 0.7 percent in 1992. Moreover, the severity of this R&D decline is masked somewhat in that while the R&D share was falling, it also was a declining share of a declining GDP. By 1997, R&D spending in Russia had inched back to about 1.0 percent of GDP.
Overall, the United States ranked sixth among OECD countries in terms of reported R&D/GDP ratios for the 1995--97 period. (See text table 2-14.) Sweden leads all countries with 3.9 percent of its GDP devoted to R&D-followed by Japan and South Korea (2.9 percent); Finland (2.8 percent); and Switzerland (2.7 percent). In general, southern and eastern European countries tend to have R&D/GDP ratios below 1.5 percent, whereas northern European nations and non-European OECD countries report R&D spending shares above 1.5 percent.
Compared with total R&D/GDP ratios, the relative position of the United States is slightly less favorable if only nondefense R&D is considered. Japan's nondefense R&D/GDP ratio (2.8 percent) exceeded that of the United States (2.1 percent) in 1996, as it has for years. (See figure 2-30 and appendix table 2-64.) The nondefense R&D ratio of Germany (2.2 percent) slightly exceeded that of the United States (again, in contrast to total R&D). The 1996 nondefense ratio for France (2.0 percent) was slightly below the U.S. ratio; those for the United Kingdom (1.7 percent), Canada (1.6 percent), and Italy (1.0 percent) were much lower. The most recent nondefense R&D/GDP ratio for Russia was a 0.6 percent share in 1994.
Consistent with overall R&D funding trends, however, the U.S. nondefense R&D/GDP ratio has been improving relative to each of the G-8 countries since 1994, when ratios reported for Japan and Germany exceeded that for the United States. France also reported devoting more of its economic output to nondefense R&D activities than did the United States, and the relative ratio of U.K. nondefense R&D spending to GDP was about equal to that in the United States. Led by industry's investments in research and predominantly development spending, the U.S. nondefense R&D/GDP ratio now matches or exceeds each of the world's other major R&D performing countries (except Japan).
Outside the European region, R&D spending has intensified considerably since the early 1990s. Several Asian countries-most notably South Korea and China-have been particularly aggressive in expanding their support for R&D and S&T-based development. In Latin America and the Pacific region, other non-OECD countries also have attempted to substantially increase R&D investments during the past several years (APEC/PECC 1997; RICYT 1998).
Even with recent gains, however, most non-European (non-OECD) countries invest a smaller share of their economic output on R&D than do OECD members (with the exception of Israel-whose reported 2.3 percent nondefense R&D/GDP ratio ranks eighth in the world). With the apparent exception of Costa Rica, all Latin American countries for which such data are available report R&D/GDP ratios below 1 percent. (See text table 2-14.) This distribution is consistent with broader indicators of economic growth and wealth. However, many of these countries also report additional S&T-related expenditures on human resources training and S&T infrastructure development that are not captured in R&D and R&D/GDP data (RICYT 1998).
As recently as 1990, R&D accounted for about 2 percent of the Soviet Union's GDP, with about 40 percent of that amount expended on defense-related activities (Gohkberg, Peck, and Gacs 1997). Indeed, the most advanced aspects of Soviet R&D efforts were undertaken in state-owned enterprises devoted to national security; much of the remaining R&D was performed in other large public industrial institutions in applied research fields that overlapped defense concerns. Most of the basic research was and continues to be in the physical sciences and engineering fields.
The introduction of a market economy to Russia brought about drastic economic restructuring, including a sharp decline in the dominance of state-owned enterprises and a 25 percent shrinkage in real GDP in just two years (IMF 1998). These trends, in turn, brought about major R&D downsizing; real R&D expenditures in 1992 collapsed to only 30 percent of the inflation-adjusted levels reported for 1990 (CSRS 1999). That is, real spending on R&D fell 70 percent with a resultant R&D/GDP ratio of about 0.7 percent. (See text table 2-15.) Reflecting the lack of core budgets, between 1990 and 1992 entire research institutes closed-including many well-equipped laboratories of the former military-industrial complex-and an estimated 19 percent of all researchers left their government R&D laboratories for the commercial sector or retirement or for other reasons, including emigration.
Between 1992 and 1995, Russian R&D spending continued to deteriorate, though at a slower pace, falling 25 percent in real terms (for a total decrease of 78 percent since the start of the decade) (CSRS 1999; OECD 1998b). The rate at which researchers left their labs accelerated, however; the number of researchers at government facilities declined 39 percent during the 1992--95 period, reflecting the effect of low and unpaid salaries, declining budgets for capital and research equipment, and generally inhospitable working conditions.
In terms of R&D spending, the situation in Russia has improved slightly since 1995. Fueled by government and industrial spending, growth in R&D exceeded inflation in 1996 and 1997. Similarly, funds from foreign sources (including funding from the European Union and the U.S. Civilian Research Foundation, among others) tripled between 1995 and 1997 and now account for 7 percent of domestic R&D spending in Russia (CSRS 1999). In spite of these recent gains, real R&D spending remains 13 percent below the levels reported for 1992 and 75 percent below the estimated levels at the beginning of the decade. Furthermore, the outflow of researchers from such activities is still an important concern, as is the belief that the younger generation is not choosing science and engineering careers to the same extent as previously. Between 1995 and 1997, an estimated 65,000 scientists and engineers left their R&D work, resulting in a researcher workforce level (455,000) that was less than half of the estimated 1990 level (993,000).