This paper sets out to investigate the nature of the relationship between technological size and technological diversification historically, in a selected group of advanced countries. US patent data are used to construct indicators for the regression model which analyses the historical evolution of this relationship. The findings show that a big change has occurred over the last 30 years. In a period when multinational enterprises have sourced technology from all over the world, countries have concentrated their technological activities in traditional areas of strength.
This paper analyses the relationship between the degree of countries’ technological specialisation and their technological size from a historical perspective. It is well-known that larger countries have a tendency to spread their research activities across many technological fields while small countries tend to concentrate on narrow niches (Archibugi and Pianta, 1992a, 1992b; Dosi, Pavitt, Soete, 1990). Moreover, size matters as far as the cumulativeness and path-dependence of countries’ profiles of technological specialisation is concerned (Vertova, 1999). Differences due to countries’ technological size cause different technological strategies. Large countries can always draw from a greater pool of different kinds of resources (i.e. human, natural, and financial) which allows them to diversify their technological base more easily and in a variety of different directions. By contrast, small countries are constrained by the lack of the very same resources to concentrate their effort into few selected niches. At the beginning of the new century, a combination of abundant natural resources, a vast supply of qualified chemists, and the best education system of that time enabled Germany to diversify into all chemical-related technology, from dyestuff to pharmaceuticals, to fertilisers, to coal-based technology, to synthetic materials, etc (Haber, 1958). By contrast, the lack of the same important resources in Switzerland forced the country to concentrate only on certain branches of chemical technology, such as dyestuff and pharmaceuticals. Large countries can enjoy a greater domestic market, thus having the possibilities to exploit economies of scale and scope. In the case where the ‘demand-pull’ theory is applicable, inventions and innovations occurred in response to a growing domestic market with a growing demand (Chandler, 1990).
Alongside the recent literature about national systems of innovation (Edquist, 1997; Freeman, 1987; Lundvall, 1992; Nelson, 1993; Porter, 1990), particular attention has been given to small countries. Andersen and Lundvall (1988) stress the importance of user-producer interaction in the process of learning and searching, which is generally more important in smaller countries due to the presence of ‘family relations’ based on close personal contacts between all sides of the production system. Moreover, in small high-income countries, such as Switzerland and Sweden, a few dominant multinationals have been catalysing this interaction, therefore having a more interactive role than in larger countries. Walsh (1987, 1988) speaks about the technological strategies a small country can adopt, due to the limited availability of resources. All small countries have to face the shortage of resources for building an indigenous capacity to innovate, the emigration of scientists and engineers, the lack of a defence sector which can work as a stimulus for innovation, the international dependence on foreign large firms. Due to these intrinsic constraints, small countries adopt ad hoc technological strategies, by finding niches in the market, possessing one's ‘own’ multinational enterprise, and attempting to attract inward investment to increase internationalisation. All countries present very distinctive systems of innovation, but large and small countries can show particular features related to their size. They are therefore forced to pursue different technological strategies. Large countries are bound to technologically diversify just as small countries are to technologically specialise.
It is not an aim of the present paper just to confirm the well-known size-diversification relationship, because other works have already done that (Archibugi and Pianta, 1992a, 1992b). The main purpose here is to focus on the evolution over time of the size-diversification relationship. It is believed that technological diversification cannot be explained only in terms of size because a country does not diversify into technologically related fields purely because it has become technologically big. Technological diversification occurs also because technology becomes more complex and interrelated over time. As von Tunzelmann (1995) points out, while at the beginning of the century one technology was enough to produce one product, nowadays technological knowledge must be shared between different fields to produce a single product, thus increasing technological complexity and shortening the product life cycles. This increasing technological complexity is due to increasing technological inter-relatedness among previously distinct technological fields and the process of technological fusion (Kodama, 1992). Producing computers involves not only technological knowledge about semiconductors, but also knowledge about software, fibre optics, displays, networks, etc.. This increasing technological inter-relatedness has been found a significant explanation for firms’ technological diversification in recent decades (Cantwell and Fai, 1999).
The econometric work of this paper seeks to investigate the well-known size-diversification relationship historically and attention is given to its evolution over time. This paper is divided into 4 sections. The following section explains the statistical methodology adopted. Section 3 reports the results of the regressions and section 4 closes the paper with some general considerations.
The statistical and econometric analysis of this paper is based on the use of patent statistics as a proxy of countries’ profiles of technological specialisation. This section does not aim to argue the relative merits of patent statistics over other indicators of technological performance. Yet, since the use of patent statistics has always been controversial, many works have shown that patent statistics can be useful indicators of technological activities (Acs and Audretsch, 1989; Griliches, 1990; Pavitt, 1982; Scherer, 1983; Schmookler, 1966). Nevertheless, patent statistics can have some drawbacks. One main difficulty relates to the issue of weighting because some patents are extremely important and related to revolutionary breakthroughs while others are negligible. Furthermore, some inventions are never patented and not all the inventions, which are patented, become innovations (Archibugi, 1992; Basberg, 1987). Finally, there is a different propensity to patent across firms, industry and countries (Scherer, 1983; Mansfield, 1986). Despite all these disadvantages, patent statistics are unique because they provide the longest running historical record of technological activities and enable the creation of very long time series.
Since this paper deals with international comparisons, the problem of different national patent legislation is overcome by the use of foreign patenting in a common third country. Soete and Wyatt (1983) show that foreign patenting is a neutral indicator, finding a high correlation between domestic R&D activities and foreign patenting in different third countries. Yet, the US is commonly used as a third country for two main reasons. Historically, the US is the largest and most technologically advanced country of the world and it is reasonable to assume that inventions patented in the US have the greatest technological and economic impact. Moreover, the data collected by the US Patent Office are the most reliable due to the rigour and comprehensiveness of American patent legislation (Pavitt, 1985, 1988). However, the choice of the US as recipient country leads to the problem of comparing like with like, because domestic patents instead of foreign patents are used. Therefore, in the American case, the sectoral distribution of domestic patenting may differ from that one of foreign patenting. Nevertheless, this is not a serious problem here, because no sectoral analysis will be attempted.
Some of the weaknesses in using patent statistics are overcome through the use of the Revealed Technological Advantage (RTA) index, first used in his pioneering work by Soete (1987) and further developed by Cantwell (1989, 1991, 1992) and Patel and Pavitt (1987, 1989, 1991). The RTA index of a country in a particular technological sector is given by the national share of patenting in that sector divided by its national share of total patenting in all sectors. The RTA index is defined as follows:
RTAij = (Pij / å i Pij) / (å j Pij / å i å j Pij)
where Pij is the total number of patents of country i in sector j. For a given sector, the index varies around unity. A RTA above unity shows a relative technological advantage, while a RTA below unity shows a relative technological disadvantage. The international and intersectoral differences in the propensity to patent are eliminated by the RTA itself, which is normalised, thus taking into account differences in the propensity to patent across fields of activities (in the numerator), and across countries (in the denominator).
The patent records used here are based on the patent database compiled by one of us at the University of Reading. For the purposes of this paper, patents granted in the US between 1890 and 1990 to both individual inventors and companies (mainly large firms) and classified by host country are used. According to this classification, patents become indicators of the location of research and invention. Since each patent is classified according to the type of technological activity, the original patent classes identified by the American PTO (Patent and Trademark Office) can be grouped into 56 technological sectors, collecting together technologically related patent classes. Since this paper has a historical perspective, it is worth mentioning that the American PTO reclassifies all earlier patents in order to keep the classification historically consistent. The 100-year period under investigation is divided into four historical periods: the preWW1 period (1890-1914), the WW1/interwar period (1915-39), the WW2/postwar period (1940-64) and the most recent period (1965-90). Countries’ profiles of technological specialisation are, therefore, represented by four RTA distributions across 56 technological sectors in four historical periods. For each country in each historical period, the RTA index in a particular technological sector is calculated through the accumulation of patent stocks, from the beginning of the period to the end. Accumulated patent stocks are an appropriate measure of countries’ technological specialisation because they are consistent with the theoretical notion of technological change as a cumulative, path-dependent and incremental process.
In this paper, the size-diversification relationship is investigated with a regression model. Yet, before venturing into the explanation of the regression model, the variables need further elaboration. Each historical period is divided into two sub-periods, the preWW1 period (1890-1914) is divided into (1890-1901) and (1902-14); the WW1/interwar period (1915-39) into (1915-26) and (1927-39); the WW2/postwar period (1940-64) into (1940-51) and (1952-64); and finally the most recent period (1965-90) into (1965-77) and (1978-90), in order to run the regression within each historical period. Due to the historical perspective of this work, the RTA distributions can be heavily influenced by the small number of patents in some technological sectors, especially for the early years. To reduce this problem conditions are imposed. All technological sectors whose accumulated patent stock is less than 100 patents in each sub-period are omitted from the analysis. Yet, it has been noticed that even after the imposition of this restriction, the RTA distributions of some countries are found to be biased. Therefore, another cut-off point, this time at the country level, is introduced. All countries whose total number of patents is less than 1,000 in each historical period are not considered in the analysis. This problem occurs only for Italy and Japan and only in the preWW1 period. As demonstrated elsewhere, those restrictions are found to be sufficient conditions for the construction of RTA distributions which are approximately normal (Cantwell, 1991; Vertova, 1998).
Within each sub-period, the technological size of country i is measured by the total number of patents (TPi), which is taken in logarithmic form in order to have a linear regression. In this case, the logarithmic functional form generates a linear relationship because the size distribution across countries is approximately lognormal (Hart, 1995). Within each sub-period, new RTA distributions are calculated for each country, in the same way the others were calculated. The degree of technological specialisation of country i is measured by the Coefficient of Variation (CVi), expressed in percentage, and calculated on these new RTA distributions across 56 technological sectors as follows:
CVRTAi = (s RTAi / m RTAi) * 100
where s RTAi represents the standard deviation of the RTA distribution and m RTAi the mean of the same RTA distribution. This coefficient is a better measure of the simple standard deviation for two main reasons. Firstly, it takes into account possible changes in the mean over time, thus embodying a relative concept of dispersion. Furthermore, it is related to the Herfindahl index of concentration, which is more frequently used. In fact, H = (CV2 + 1) / n, where n is the number of sectors in the distribution (Hart, 1971). Keeping in mind that the CV is directly related to the standard deviation, which is a measure of dispersion around the mean, there are two possible outcomes:
Since the analysis has a time dimension, it must be kept in mind that a rise of the CV over time indicates increasing technological concentration and a fall indicates increasing technological diversification.
The regression model used to analyse the historical evolution of the size-diversification relationship is the following:
CVi = a + b LogTPi + e i
where CV is the Coefficient of Variation and LogTP is the logarithm of the total number of patents for country i. This model gives a negative sloped regression line because when a country increases its technological size (i.e. rise in the LogTP), it diversifies across many technological sectors (i.e. fall in the CV). The analysis of the size-diversification relationship with this model enable to distinguish two effects:
Graph 1 shows the four regression lines found within each historical period and Table 1 shows the separate effects of changes in the size-diversification relationship due to change in technological size and due to change in international specialisation. In Table 1 the effect of the rising technological size must is controlled and afterwards the international specialisation effect are measured, with the following calculations were made:
Column (7) and column (8) give the measure of the two different effects on the size-diversification relationship, the size effect and the international specialisation effects. The analysis of change in the size-diversification relationship is described by confronting the graphical results (Graph 1) with the statistical results (Table 1).
Graph 1:
Graph 1 shows that within the preWW1 period (line 1), the WW1/interwar period (line 2) and the WW2/postwar period (line 3), the regression lines remain almost unchanged, and the only movement is a minimum shift outward. Table 1 shows that the shift was due more to change in size than to change in international specialisation (i.e. the change in size is always higher than the change in international specialisation). Therefore, from the preWW1 period to the WW2/postwar period, there was no big change in the size-diversification relationship, the large countries kept on technologically diversifying and the smaller ones concentrating their technological specialisation. Moreover, Table 1 shows that these historical periods were those with the highest degree of technological diversification, on average (column 5). Different kinds of factors may have influenced the average tendency of countries to technologically diversify. The creation and growing of big business and the new method of carrying out research activity facilitate technological diversification. According to Freeman (1995), the greatest invention of the last decades of the nineteenth century was the method of invention itself within the industrial R&D laboratories. The presence of big companies with in-house laboratories favoured technological diversification because large firms tend naturally to diversify their technological activity in order to face increasing internationally competition (Cantwell, 1993; Patel and Pavitt, 1991). Historically, at the beginning of their existence, large firms tended to cluster in industries with similar characteristics and to be highly committed to research (Schmitz, 1993), which could have worked as a stimulus for technological diversification. The new relationship between science and technology could have been another stimulus. Management of large firms recruited university scientists and engineers for their laboratories, and inventions and innovations were carried out also in these industrial laboratories, not only within universities. More science went into the production system and, because this science came from different scientific disciplines, it could be used in different segment of the same industry, or in different industries altogether. The predominant technological paradigm of that period could have been another stimulus towards technological diversification. The Fordist technological paradigm symbolised a new era of mass-production and pointed the way for similar development in other industries. Mass-production took over everywhere, and the demand for the so-called ‘white goods’ increased substantially. These goods are characterised by a high level of technological inter-relatedness, therefore, firms were tempted to diversify their technological activity in order to cope with the high demand of those goods. The management literature suggests that at that time, technological diversification was a typical corporate strategy in order to grow and to become internationally competitive (Chandler, 1990). Yet, technological diversification was seen as means to accumulate new technological competencies, and not only to widen the market (Cantwell and Piscitello, 1999). Some changes occurred after the Second World War could have worked as incentive as well. At that time, governments believed that growth was the result of a high volume of investment to enhance technological progress and that the science-push model could work. Under the influence of war research, governments decided to directly intervene in the production of new technological knowledge. By 1950, the governments of most advanced countries had adopted policies aimed at fostering science and technology in order to increase the volume and variety of output and productivity. Moreover, this period was characterised by the highest level of spin-off from military and strategic research into commercial application. Many inventions and innovations were carried out to answer demanding military sectors, which needed newer innovations all the time because the Second World War had shown the importance of being at the technological cutting edge. The inventions of the transistor and the computer, developed for the American NASA laboratories in order to make difficult ballistic operations and calculations easier, were typical examples. All this technological spillover was a great stimulus to technological diversification, because many firms ventured into the development of new civilian application of military technology.
Graph 1 shows that the situation changed completely from the previous three historical periods (line 1, 2, 3) to the most recent period (line 4), with a big jump upward of the regression line 4. Its movement was caused by a negative size effect (-435.891) was summed up to a very high international specialisation effect (535.891). In the most recent period, the international specialisation effect, together with a negative size effect, led to an increased technological concentration of countries’ profiles of technological specialisation, on average (column 5). This international specialisation effect means that changes in the economic and institutional framework change the nature of the size-diversification relationship, and this result must be considered as an important historical shift. Since the mid-1960s, on average, countries increased their technological specialisation in their traditional areas of technological expertise. This result is consistent with Cantwell’s (1991) findings, showing a rise in the degree of technological specialisation in the period between the 1960s and the 1980s. Countries highly concentrated their profiles of specialisation in a period in which multinational enterprises played a major role in diffusing technological knowledge all over the world. In the last 30 years, an increased transfer of technology took place through an extensive use of international networks by multinational enterprises. In the most recent period, technological communities became transnational like never before. Initially, this trend was set by American firms, later European and Japanese firms joined in. Nowadays, most large firms have plants and research facilities in more than one country. Furthermore, globalisation and internationalisation of economy was also enhanced by the new prevailing technological paradigm. The rapid growth of information and communication technology enabled and supported the creation of these international networks. Large and small firms increased their co-operation in technology, quality control, training, investment planning and production planning thanks to the links made through computers. Also the increased availability of the public aspect of technological knowledge, commonly associated with information, worked as a stimulus for increasing globalisation and internationalisation of technology. Nowadays, engineers and applied scientists are taught pretty much the same things in schools in different countries. Technological information is therefore more open to the public than ever before. Furthermore, technological information is much the same in most countries. Fifty years ago there were major differences across industrialised countries in the technology employed and even in what the engineers knew. This is no longer true. Furthermore, firms face roughly the same market environment wherever their home base, due to the lowering of national barriers to trade as well as the recent convergence of living standards. Moreover, in the 1980s, international joint ventures on particular projects, especially large-scale research projects, began to appear in a number of industries. Transnational public programs of research and development (i.e. Eureka) increase the transfer of technological knowledge within firms of different countries.
The most recent period was therefore characterised by great diffusion of technological knowledge through multinational enterprises disseminated all over the world. This international environment should have led to a greater technological diversification than ever, because it is common knowledge that multinational enterprises naturally tend to diversify their technological base (Cantwell, 1993; Patel and Pavitt, 1991). Yet, in the most recent period countries tended to concentrate their technological activities as never before. Giving a reason for this trend is to answer the famous debate national versus global. Yet, this process of technological diversification by multinational enterprises does not work against national profiles of specialisation, but, on the contrary, enhances them. Multinational enterprises tend to reinforce national profiles of specialisation, which are historically rooted and quite stable, thus leading countries to concentrate their profiles of specialisation more than ever before on traditional areas of technological strength. Multinational enterprises find it easier to develop technological expertise and capability related to that one of the host country. In this way, they can count on a pool of technological knowledge already existing and organised in production form. Moreover, due to the local specificity of technological progress and research, multinational enterprises have an incentive to carry out research in countries that are centres of excellence for their fields. Therefore, multinational enterprises and their subsidiaries tend to complement locally specific profiles of specialisation and not to substitute them. As already found elsewhere (Cantwell, 1995), the general tendency has been for multinational enterprises to become more technologically diversified and for countries to become more technologically concentrate in their established areas of technological expertise.
This paper achieves two different results. Firstly, it confirms that the well-known size-diversification relationship exists also historically. Therefore, it is possible to state that, throughout the 100-year period, large countries have had a natural tendency to diversify their technological activities across many fields, while small countries have had a natural tendency to concentrate into specific technological niches. Moreover, the analysis of the historical evolution of the size-diversification relationship shows that a big change occurred in the nature of this relationship in the most recent period. Since the turn of the century until the mid-1960s, countries maintained an almost constant nature of the size-diversification relationship. Historically large countries were always technologically diversified, while small countries were always technologically concentrated. This is no longer true in the most recent period. Despite the increase in countries’ technological size (i.e. despite size effect), there was a common tendency of all countries to concentrate their technological activities. A structural shift in the nature of the size-diversification relationship has occurred. Change in international specialisation due to change in the economic and institutional environment led countries to concentrate their technological activities, despite their technological size. The interesting result is that, this process of technological concentration occurred in the period in which internationalisation and globalisation was enhanced by multinational enterprises. Contrary to what is sometimes presumed, the presence of multinational enterprises does not destroy national profiles of specialisation but reinforce them, with their desire to fit into local specificity.
This findings call for some public policy suggestions. Governments should understand that the role of the nation state has changed in the emerging global economy. The increasing globalisation of different economies and the internationalisation of technology has led to the debate about the possibility of still having ‘a nation’. Nobody denies that nations still matter, but instruments of public policy should be adapted to the new environment, in which a nation is constantly ‘invaded’ by multinational enterprises. National elements are still important due to their ability to influence the national structure of the economy at different levels. The macroeconomic environment can be influenced by national policies concerning the labour market, the financial system, and through different monetary, fiscal and trade policies. The education system is still national specified, and the universities and public laboratories have still a national dimension in their networks and interactions with other economic agents. The public infrastructures and laws, which determine the decision to either invest in the domestic country or abroad, are typically national. Finally, the government intervention especially in military programs, support and finance to R&D activities and educational support is still national. Nation exists also at the level of different forms of capitalism and organisation of production (Lazonik ????). The ‘American model’ was based on large firms, high expenditure in R&D, a significant role for the government intervention and a good research system. The ‘Japanese system’ had new features related to the greater role of the MITI (Minister for Trade and Industry) and the new emphasis on education and training.
Within the evolutionary tradition, the national aspect is even more important when tacit and uncodified knowledge, which is more difficult to transfer, is concerned. The new role of governments should be to support education and training, public research and universities, and to encourage firms to invest in research. Yet, in the light of this findings, the crucial question for a country is the extent to which a country should try to diversify its technological base of competence and expertise. Multinational enterprises fit into national profiles of specialisation, which are the result of accumulated technological competence and expertise of the host country, and in doing so support the process of technological concentration. It seems therefore more reasonable to adopt a public policy which enhances the traditional areas of technological expertise of a country, in order to attract the investment of those multinational enterprises which need that particular technological expertise. By contrast, if a country decides to develop technological areas that are remote from its technological core, it might never reach the level of competence required by multinational enterprises, which will then have no interest in coming and investing in the country. A country should not attempt to reach technological competence in every technological field, but should move within areas of accumulated technological expertise and gradually move from there to inter-related technological fields. This issue relates to the debate about the benefit of technological specialisation in high-technology products. The common argument is that if an economy does not have considerable strength in high-technology products, it will be disadvantaged. The logic behind this argument is that, advanced and technological supremacy in those fields provides the key opportunities for technical innovations in a wide range of downstream industries. It is believed that high-technology industries generate unusually large externalities, which flow to downstream industries, and which can benefit the whole economy. Yet, after a study about the national systems of innovation of 15 different countries Nelson (1992) concludes that "there does not seem to be strong empirical support for the proposition that national economies are broadly advantaged if their firms are especially strong in high technology, and disadvantaged if they are not." (Nelson 1992, p. 366). From this consideration, the previous suggestion for public policy can be reiterated. It appears that it is better for a country to maintain its profile of specialisation in traditional areas of technological competence, rather than trying to acquire technological expertise in new and unrelated fields.
Acs, Zoltan & David Audresch. 1989. Patents as a measure of innovative activity. Kyklos, 42(2): 171-180.
Andersen, Esben & Bengt-Åke Lundvall. 1988. Small national systems of innovation facing technological revolutions: an analytical framework. In C. Freeman & B-Å. Lundvall, editors, Small countries facing technological revolution. London: Pinter Publisher.
Archibugi, Daniele. 1992. Patenting as an indicator of technological innovation: a review. Science and Public Policy, 19(6): 357-368.
Archibugi, Daniele & Mario Pianta. 1992a. The technological specialization of advanced countries. Dordrecht: Kluwer Academic Publisher.
Archibugi, Daniele & Mario Pianta. 1992b. Specialisation and size of technological activities in industrial countries: the analysis of patent data. Research Policy, 21: 79-93.
Basberg, Bjørn. 1987. Patents and the measurement of technological change. Research Policy, 16: 131-141.
Cantwell, John. 1989. Technological innovation and multinational corporations. Oxford: Basil Blackwell.
Cantwell, John. 1991. Historical trends in international patterns of technological innovation. In J. Foreman-Peck, editor, New perspective on the late Victorian economy: essays in quantitative economic history, 1860-1914. Cambridge, U.K.: Cambridge University Press.
Cantwell, John. 1992. Japan's industrial competitiveness and the technological capabilities of the leading Japanese firms. In T.S. Arrison, C.F. Bergsten, E.M. Graham, M.C. Harris, editors, Japan's growing technological capabilities: implication for the US economy. Washington DC: National Academy Press.
Cantwell, John. 1993. Corporate technological specialisation in international industries. In M. Casson & J. Creedy, editors, Industrial concentration and economic inequality: essays in honour of Peter Hart. Aldershot: Edward Elgar.
Cantwell, John. 1995. The globalisation of technology: what remains of the product cycle model?. Cambridge Journal of Economics, 19(1): 155-174.
Cantwell, John & Felicia Fai. 1999. The changing nature of corporate technological diversification and the importance of organisational capability. In S.C. Dow and P.E. Earl, editors, Contingency, complexity and the theory of the firm. Essays in honour of Brian J. Loasby. Volume II. Aldershot: Edward Elgar.
Cantwell, John & Lucia Piscitello. 1999. Accumulating technological competence – its changing impact on corporate diversification and internationalisation. Industrial Corporate and Change, forthcoming.
Chandler, Alfred. 1990. Scale and scope: the dynamics of industrial capitalism. Cambridge, Mass.: Belknap Press.
Dosi, Giovanni & Keith Pavitt & Luc Soete. 1990. The economics of technical change and international trade. Hemel Hempstead: Harvester Wheatsheaf.
Edquist, Charles. 1997, editor. Systems of innovation. Technologies, institutions and organisations. London: Pinter Publisher.
Freeman, Chris. 1987. Technology policy and economic performance: lesson from Japan. London: Pinter Publisher.
Freeman, Chris. 1995. The ‘National System of Innovation’ in historical perspective. Cambridge Journal of Economics, 19(1): 5-24.
Griliches, Zvi. 1990. Patent statistics as economic indicators: a survey. Journal of Economic Literature, XXVIII(4): 1661-1707.
Haber, Ludwig. 1958. The chemical industry during the nineteenth century. Oxford: Clarendon Press.
Hart, Peter. 1971. Entropy and other measures of concentration. Journal of the Royal Statistic Society, 134: 73-84.
Hart, Peter. 1995. Galtonian regression across countries and the convergence of productivity. Oxford Bulletin of Economics and Statistics, 57(3): 287-293.
Kodama, Fumio. 1992. Technology fusion and the new R&D. Harvard Business Review, 70(4): 70-78.
Lazonik, William. 1992. Business organization and competitive advantage: capitalist transformations in the twentieth century. In G. Dosi, R. Giannetti, P. Toninelli, editors, Technology and enterprise in a historical perspective. Oxford: Oxford University Press.
Lundvall, Bengt-Åke. 1992, editor. National systems of innovation. Towards a theory of innovation and interactive learning. London: Pinter Publisher.
Mansfield, Edwin. 1986. Patents and innovation: an empirical study’. Management Science, 32(2): 173-181.
Nelson, Richard. 1992. National innovation systems: a retrospective on a study. Industrial and Corporate Change, 1: 347-374.
Nelson, Richard. 1993, editor. National systems of innovation. A comparative analysis. Oxford: Oxford University Press.
Patel, Pari & Keith Pavitt. 1987. Is Western Europe losing the technological race?, Research Policy, 122: 59-85.
Patel, Pari & Keith Pavitt. 1989. The technological activities of the UK: a fresh look. In A. Silbertstone, editor, Technology and economic progress. London: Macmillan Press.
Patel, Pari & Keith Pavitt. 1991. Large firms in the production of the world’s technology: an important case of ‘non-globalisation’. Journal of International Business Studies, 22: 1-21.
Pavitt, Keith. 1982. R&D, patenting and innovative activities. A statistical exploration. Research Policy, 11: 33-51.
Pavitt, Keith. 1985. Patent statistics as indicators of innovative activities: possibilities and problems. Scientometrics, 7(1-2): 77-99.
Pavitt, Keith. 1988. Uses and abuses of patent statistics. In A. van Raan, editor, Handbook of Quantitative Studies of Science Policy. Amsterdam: North Holland.
Porter, Michael. 1990. The competitive advantage of nations. London: Pinter Publisher.
Scherer, Frederic. 1983. The propensity to patent. International Journal of Industrial Organisation, 1(1): 107-128.
Schmitz, Christopher. 1993. The growth of big business in the United States and Western Europe, 1850-1939. London: Macmillan Press.
Schmookler, Jacob. 1966. Invention and economic growth. Cambridge, Mass.: Harvard University Press.
Soete, Luc. 1987. The impact of technological innovation on international trade patterns: the evidence reconsidered. Research Policy, 16: 101-130.
Soete, Luc & Sally Wyatt. 1983. The use of foreign patenting as an internationally comparable science and technology output indicator. Scientometrics, 5(1): 31-54.
Vertova, Giovanna. 1998. Historical evolution of national systems of innovation and national technological specialisation. Unpublished Ph.D. Dissertation. Department of Economics, University of Reading, Reading.
Vertova, Giovanna. 1999. Stability in national patterns of technological specialisation: some historical evidence from patent data. Economics of Innovation and New Technology, forthcoming.
von Tunzelmann, Nick. 1995. Technology and industrial progress: the foundations of economic growth. Aldershot: Edward Elgar.
Walsh, Vivien. 1987. Technology, competitiveness and the special problems of small sountries. STI Review, 1(?): 81-133.
Walsh, Vivien. 1988. Technology and the competitiveness of small countries: review’. In C. Freeman & B-Å Lundvall, editors, Small countries facing technological revolution. London: Pinter Publisher.