Changes in National Technological Competitiveness: 1990, 1993, 1996 and 1999

Georgia Tech’s Technology Policy and Assessment Center, with support from the US National Science Foundation, has been generating High-Tech Indicators (HTI)—measures of national technology-based export competitiveness since 1987. This paper reports the HTI results for 33 nations in 1999 in comparison with those of 1990, 1993 and 1996. HTI includes four ‘input indicators’ and a key ‘output indicator’—technological standing. We construct a new composite input indicator here and examine its predictive capability. Input indicators for 1990 and 1993 show intriguing relationships to 1999 technological standing. We compare the indicators for various groups—leading and emerging Western economies, rapidly developing Asian economies, former Eastern Bloc nations and lagging Latin American countries. The USA presently exhibits a dominant position, but signs strongly point toward increasingly broad-based competition in technology-based products. Introduction In modern society, national competitiveness is based primarily on technology. Science and technology constitute the decisive factors in the new productive forces. Developing nations that succeed rely heavily on technology for economic expansion. We believe that science and technology will constitute the core of competitiveness in the future. A variety of studies on technology-based competitiveness indicators have been conducted during the past decade to measure change in technical capabilities, predict trends in technology-aided development, assess the business environment and inform policy for enhancing national competitiveness. Generally speaking, the data used for measuring national competitiveness can be divided into two types. First, statistical data provide much information—e.g. gross domestic product (GDP), GDP per capita, research and development (R&D) expenditures, R&D personnel, exports of goods and services, educational infrastructure. These data can be collected from international, regional, or national statistical sources. Second, expert opinion data complement the statistics—e.g. compile the opinions of a set of knowledgeable persons. After data processing and mathematical calculations (statistical data, opinion data, or their combination), the results can be used to measure and compare national competitiveness. Several recent indicator Alan L. Porter is Director of the Technology Policy and Assessment Center (TPAC), and Professor of Industrial & Systems Engineering (ISyE) and of Public Policy, Georgia Tech, Atlanta, GA 30332-0205, USA. Tel: +1 404 894 2330; Fax: +1 404 894 2301, E-mail: [email protected]. J. David Roessner is Co-Director of TPAC and Professor of Public Policy Emeritus, Georgia Tech, and Associate Director, Science and Technology Policy Program, SRI International, E-mail: [email protected]. Nils Newman and Xiao-Yin Jin are Senior Researchers at TPAC, Georgia Tech, 781 Marietta Street, Atlanta, GA 30332-0525, USA, Newman’s E-mail: [email protected]; Jin’s E-mail: [email protected] ISSN 0953-7325 print; 1465-3990 online/01/040477-20 © 2001 Taylor & Francis Ltd DOI: 10.1080/0953732012009539 2 478 A. L. Porter et al. compilations provide interesting contrasts to our work; we plan to compare these in a forthcoming paper. Researchers at Georgia Tech’s Technology Policy and Assessment Center have been generating technology-based competitiveness indicators—‘High-Tech Indicators’ or HTI— since 1987. We initially demonstrated the feasibility of a set of country-level indicators of international competitiveness in high-technology industries, generating four ‘input’ and three ‘output’ indicators. Since 1990, we have compiled HTI every three years, with quite consistent component statistical and expert opinion components. Over time, and with the support and guidance of the Science Indicators Unit of the U.S. National Science Foundation, we have expanded the set of target countries to 33. We have examined and continue to assess the indicators’ reliability and validity. This paper describes the results of our latest survey and compares these to our 1990, 1993 and 1996 indicators. The paper is organized in three sections. In the Ž rst section, we summarize the 1999 web survey and statistical research background, including country coverage, expert opinion gathering and statistical measures. In the second section, we compare Ž ndings from 1990 to 1999. In the third section, we compare grouped ‘country sets’ of our 33 target countries (i.e. The big 3, Western Europe, English Heritage Nations, Eastern Europe, Asian Tigers, Asian Cubs and Latin America) analyzed in our study. The Conduct of HTI Country Coverage In 1987, we compiled data for 20 countries representing a range of regions and extent of industrialization. The second (1990) and third (1993) phases gathered data on an expanded set of 28 countries (29 countries in 1990—Germany subsumed West Germany and East Germany; also Russia replaced the USSR after the 1990 survey). The 1996 HTI added Poland, Venezuela and South Africa, but dropped Hong Kong because of its absorption into China in 1997, so the total number of countries totaled 30. (Regularization of statistics for China and Hong Kong remains problematic even for 1999, particularly in sorting out exports.) For the 1999 HTI we have added Ireland, Israel and the Czech Republic, yielding a total of 33 countries. Only limited ‘backŽ lling’ of statistical measures has been feasible. The countries are usually clustered in presentations as follows: · The ‘Big Three’—USA, Japan and Germany · Western Europe (UK, France, Netherlands, Italy, Switzerland, Sweden, Spain and Ireland) · English Heritage Nations + Israel (Canada, Australia, South Africa, New Zealand and Israel) · Eastern Europe (Russia, Poland, Hungary and Czech Republic) · Asian Tigers (Singapore, South Korea, and Taiwan) · Asian Cubs (Malaysia, China, Thailand, Indonesia, The Philippines and India) · Latin America (Mexico, Brazil, Argentina and Venezuela) Expert Opinion Collection In earlier years, opinion data from HTI experts were collected by mail and fax, then e-mail was added, and most recently, web-based questionnaires. For 1999, we published our questionnaire on the web (http://tpac.gatech.edu/hti99/ ) with background materials and followed up by e-mail to invite experts to participate in this HTI Panel. The webbased approach considerably facilitates collection internationally. This is vital as we seek National Technological Competitiveness 479 persons familiar with technology-based economic development in speciŽ c countries. One problem is that some experts, especially those from developing countries, are unable to access the Internet. This problem should lessen in the future. The expert opinion questionnaire has remained essentially the same from 1990 on. SigniŽ cant revision and country expansion from 1987 to 1990 discourages our inclusion of 1987 results in this comparative analysis. In our 1999 questionnaire, we made minor changes: · Adapted for web-based surveying · Added ‘software’ as one of nine sectors about which we inquired · Made minor format changes. The 1999 HTI questionnaire contains 16 questions (from A to P). Question A indicates the country being addressed. Question B requests self-assessment of one’s familiarity with technology-intensive development in that country. All the other questions from C to P relate to the seven competitiveness indicators. These consist of four ‘Input Indicators’: · NO: National Orientation to achieve technological competitiveness · SE: Socioeconomic Infrastructure to support a technology-based economy · TI: Technological Infrastructure to enable development, production, and marketing of technology-based goods · PC: Productive Capacity to e Ý ciently manufacture such goods and three ‘Output Indicators’: · TS: Technological Standing in manufacturing and export capabilities for hightechnology products · TE: Technological Emphasis in export mix · RTC: Rate of Technical Change. For instance, questions C (Strategy), D (Cultural values), E (In uential groups), and F (Entrepreneurial spirit) contribute to NO; questions G (Mobility of capital) and L (Foreign Ž rms encouraged) contribute to SE. Similarly, all other questions address a given aspect of one of these designed indicators that we feel is not adequately treated via statistical measures. The conceptual deŽ nitions of the seven indicators are the same as those used in the earlier studies. The Input Indicators (leading indicators) re ect national propensity for future technology-based competitiveness. The Output Indicators gauge current competitiveness. With the exception of TE, each indicator is comprised of both statistical and expert opinion data. See Appendix for details. The International Technology Indicators Panel includes both resident and foreign observers of given countries. Our criteria for inclusion include direct knowledge of the country and of the bases for technological competitiveness. Prototypical experts include embassy science attaches, faculty members and industry professionals. Attendees at international conferences and participants in journal advising and publishing relating to technology analysis, forecasting, management and so forth, are good candidates for the Panel. We seek balance among multiple perspectives. Over time turnover in membership is brisk—only 24% of the current respondents also participated in 1996. We invite various persons who appear to meet these criteria, but ultimately self-selection comes into play. The respondents indicate their familiarity on a self-report scale item. In our 1999 web survey, this self-rated expertise showed: 27.1% as expert, 44.0% as highly familiar, 26.2% as moderately familiar, and 2.7% as less familiar. Due to the general nature of our expert 480 A. L. Porter et al. inclusion criteria, we are cautious in interpreting sector-speciŽ c item responses on current and 15-year future prospects. The 1999 HTI expert opinion data were obtained from responses of the International Technology Indicators Panel during summer and fall, 1999. The resulting group of 303 experts (up from 207 in 1996) collectively provided 336 responses (up from 265 in 1996). The average number of responses per country was 10.2, ranging from six to 22. Only Ireland had six responses; six others had seven responses; eight or more experts assessed the other 26 countries. Most responses come from inside the country (79.2 % vs 20.8 % outside). Most experts (283/303, about 93%) responded for only one country. Twenty experts responded for two or more countries. Among the 303 experts, 72 of them responded both in 1996 and 1999. With respect to missing answers, we divide the responses into two groups. For questions A to O (15 questions), generally speaking, the HTI experts responded for all questions. The total answers for these 15 questions should be 15 3 336 5 5040, if experts answered every question; we received 5027 answers (99.7%). For question P (the last, multi-part question), we asked the respondents to characterize present and future (roughly 15-year time horizon) technology-intensive production in each of nine sectors for which they feel reasonably familiar. We received 5545 answers (91%). (Anonymous item-byitem response data for each country are available on request.) Statistical Data Collection The statistical measures re ect our explorations of a wide range of potential indicator components. Selection criteria include pertinence to our competitiveness model; availability of component measures for all, or nearly all, of the target countries; data quality; data accessibility for time-series construction; convergent validity (components show reasonable correlation with each other across countries); and divergent validity (components di Ú erentiate among countries). The following data sources exemplify those used to construct our indicators: · United Nations Statistical O Ý ce, Commodity Trade Statistics Section—1997 exports (most recent available, included in TS, TE and RTC) (United Nations, COMTRADE, 1999, New York) · Reed Electronics Research, Ltd, Yearbook of World Electronics Data 1999/2000, Surrey, UK) (collection of data on electronics production, data processing equipment purchases, sales, and exports, included in TI, PC, TS, TE and RTC) · The PRS Group’s Five-Year Investment Risk Assessment Index (for NO) (from The Political Risk Letter, for 1 July 1999) · Harbison–Myers Human Skills Index (for SE) (derived from the World Bank’s 1999 World Development Indicators, New York, Oxford University Press) · Numbers of scientists and engineers (for TI) (United Nations, Statistical Yearbook, 1998, New York) · International Monetary Fund, Direction of Trade Statistics Yearbook, 1998, Washington, D.C. (for TE). We seek the most recent data available for each HTI compilation, but readers should be wary of lags on some components. We are currently investigating new data sources and invite suggestions via e-mail or the Technology Policy and Assessment Center website. National Technological Competitiveness 481 Indicator Scaling We combine survey and statistical measures to compute the value for the seven indicators. Indicator formulae are presented as ‘S-scores’—each component’s raw data values are rescaled from 0 to 100. For survey data, 100 re ects the highest response for a question (a ‘5’ on a 1–5 questionnaire item response range becomes ‘100’). For statistical data, 100 represents the value attained by the country with the largest value among all 33 target countries. Other country scores are scaled proportionately between 100 and 0. Components are then averaged to generate indicators with approximately a 0 to 100 range. In that a given indicator combines several S-scores, typically no country will score 100 on the resulting indicators. For instance, ‘NO’ for 1990—the Ž rst indicator listed in Table 1—ranges from a low of 21.4 for Mexico to a high of 82.0 for Germany. Details of this process, including in-depth treatment of why we shifted from Z-scores (standardized scores) to S-scores, appear in earlier papers. HTI Findings from 1990 to 1999 A Composite Leading (Input) Indicator To enable comparisons, we present the 1999 HTI results together with those from 1990, 1993 and 1996. Table 1 consolidates the results of indicator calculation for the four input indicators and for TS, the most compelling output indicator. (Refer to the report on the website: http://tpac.gatech.edu for TE and RTC.) Figure 1 graphically depicts a composite input indicator (‘INPUT ’) for 1990, 1993, 1996 and 1999. This is simply the average of NO, SE, TI and PC for a given country and year. We explored alternative compositions using the 1999 data, including a multiplicative version. Conceptually, the case could be made for either. According to our model, a nation that aspires to enhance its technology-based export competitiveness ought to evidence strength in all four inputs—NO, SE, TI, and PC. The additive model implies that strength on one of these four dimensions could compensate for weakness on another. The multiplicative model implies that weakness on any one could seriously impede e Ú ective development. The composite indicator computed as the average (equivalently, the sum) correlates 0.90 with that computed as a product (NO 3 SE 3 TI 3 PC). Converting to S-scores does not alter this relationship. Rank scores for the average and multiplicative versions of INPUT correlate 0.995. A useful benchmark is the separate rating respondents provide on present and future (15 years ahead) technological competitiveness. Average 1999 INPUT correlates higher (0.89) than does the multiplicative input composite (0.79) with present ‘rated competitiveness’ (i.e. expert opinion item inquiring about technological competitiveness), and higher as well with future rated competitiveness (0.85 vs 0.69). Conversely, the 1999 multiplicative version correlates more highly with current TS— 0.88 vs 0.79, but this is not as suitable a gauge of an ‘input’ indicator. We thus choose the average of NO, SE, TI and PC as the composite input indicator, INPUT. We elect not to convert this average of S-scores into its own S-score to make comparisons of INPUT scores over time more e Ú ective. Table 2 shows how this composite, average INPUT indicator compares with its components in correlating with current TS, current rated competitiveness and future rated competitiveness. Among the individual input indicators, TI and PC appear more ‘short term oriented’. They correlate considerably more strongly with current TS and current rated technological competitiveness. In contrast, NO correlates more strongly with future rated competitiveness. The composite, INPUT, does well with both future 482 A. L. Porter et al. T ab le 1. In di ca to r va lu es fo r ea ch ye ar