SUSTAINABLE GROWTH, EMPLOYMENT CREATION AND TECHNOLOGICAL INTEGRATION IN THE EUROPEAN KNOWLEDGE-BASED ECONOMY (SETI/SERD-2000-00126)

REVIEW OF THE THEORETICAL LITERATURE

by College of Europe (Bruges, Belgium), CeSPI (Rome, Italy), SPRU (University of Sussex, UK), Leverhulme (University of Nottingham, UK)

January 2002

1

Table of Contents 1. Introduction ............................................................................................................................................................................3 2. Technology and specialisation ............................................................................................................................................4 2.1 Technology and specialisation: theories and empirical evidence.....................................................................5 The consideration of technology like a key determinant of country and sectorial trade performance has generated an impressive development of theoretical and empirical proposals in the last years. Two themes in particular have received special attention in recent years:....................................................................................................................5 2.1.1 The impact of technology on trade competitiveness and specialisation........................................................5 2.1.2 Trends in specialisation patterns ..........................................................................................................................7 2.2 The impact of specialisation on international competitiveness and economic growth...............................9 Table 2.1: The impact of specialisation on economic performance.................................................................10 Methodology .............................................................................................................................................................................10 Dependent variable.............................................................................................................................................................10 Significant .................................................................................................................................................................................10 2.3 Specialisation and employment creation...............................................................................................................13 2.4 Open questions and directions for future research............................................................................................15 3. Services..................................................................................................................................................................................16 3.1 What are services?......................................................................................................................................................16 3.1.1 Services versus goods ..........................................................................................................................................16 3.1.2 The internationalisation of services ...................................................................................................................17 3.2 Trends in services........................................................................................................................................................19 3.2.1 Trade in services....................................................................................................................................................19 Table 3.1: World trade in goods and services .......................................................................................................20 Table 3.2: Trade in services by EU country and broad sector for 1999 ........................................................20 Table 3.3: Trade value by major economic region and selected industry (1990-1999).............................22 Table 3.4: Share in total services trade by major economic region and selected industry........................... (1990-1999)......................................................................................................................................................................23 3.2.2 FDI in services.......................................................................................................................................................23 Table 3.5: World foreign production in goods and services..............................................................................24 3.3 The regulatory environment....................................................................................................................................25 3.3.1 The free movement of services within the EU .................................................................................................25 Table 3.6: Determinants of Market Structure in Services (late 1980’s).........................................................26 3.3.2 The GATS and beyond ........................................................................................................................................27 3.4 Services, employment and economic growth .......................................................................................................28 3.4.1 Explaining the service economy .........................................................................................................................29 3.4.2 Technological change in services.......................................................................................................................30 3.4.3 Consequences on employment....................................................................................................................31 3.4.4 Consequences on growth: Direct and indirect effects.....................................................................................33 Table 3.7: Labour productivity by industry, 1987-1997....................................................................................33 3.4.5 The internationalisation of services and economic growth............................................................................35 3.5 Open issues and directions for future research...................................................................................................36 4. The Information and Communication Technology (ICT) revolution.........................................................................39 4.1 Theoretical frameworks............................................................................................................................................40 4.1.1 ICT as a new techno-economic paradigm.........................................................................................................40 4.1.2 ICT as a GPT..........................................................................................................................................................40 4.2 ICT and growth in the US and in Europe: A short review..............................................................................41 4.2.1 An overview of the US debate on ICT and growth.........................................................................................41 4.2.2 International comparisons of ICT and growth in the OECD area.................................................................43 4.3 Open Issues and Directions for Future Research...............................................................................................45 5. The internationalisation of technology ............................................................................................................................46 5.1 Theoretical framework ..............................................................................................................................................47 5.2 Recent empirical studies............................................................................................................................................50 5.2.1 Studies based on R&D and patent data.............................................................................................................51 5.2.2 Studies based on surveys .....................................................................................................................................52 5.2.3 Studies based on patent citations analysis ........................................................................................................53 5.3 Open issues and directions for future research...................................................................................................54 6. A review of the methodology............................................................................................................................................54 Bibliography .............................................................................................................................................................................59

2

1. Introduction The aim of the SETI project is to investigate the impact of technological globalisation, the development and diffusion of Information and Communication Technologies (ICT) and the rise of business services on national patterns of specialisation. Moreover, we are interested in assessing their impact on growth and employment potential in Europe. The interest for these issues arises from a number of stylised facts. First, Europe is losing competitiveness in sectors where the rates of growth of innovation and demand have been faster. Second, the process of internationalisation of R&D appears to have been more marked in Europe with respect to Japan and the US and this raises the issue of the impact of this trend on European national systems of innovation. Finally, the diffusion of ICT is producing an impact on the processes of internationalisation and the rise of business services. In this context we ask whether the poor performance of Europe in the new technologies has contributed to the slowdown in the catching-up process and the high rates of unemployment with respect to the US and what are the future perspectives. Different streams of literature are relevant for this project. First, there is a large literature on the impact of technology on international specialisation. This literature is the starting point for the development of the project. In this general framework we aim at assessing how the rise of the service sector, the internationalisation (globalisation) of technology, and the development and diffusion of ICT are shaping European specialisation patters. The phenomenon of internationalisation has been accompanied by the growth of particular activities within the service sector and both trends have been favoured by the development of ICT. Tertiarisation in a broad sense is not a new phenomenon, and the growth of the share of employment in the service sector follows a very long trend. Still the tertiarisation trend experienced over the last two decades has some distinctive features, of which the most marked one is the unprecedented growth of business services over the last two decades. The new service sector mix very diverse types of activities, requiring different levels of skill. The development of the highly qualified service activities in many new market niches is clearly linked with the diffusion of ICT. This opens differentiated opportunities to different countries. The business services sector constitutes a new phenomenon, characteristic of a new set of relations between firms and this raises some new questions: •

How will different economies adjust to the broad organisational challenge with which they are confronted? And to what extent new organisational paradigms, linked to the increasing role of services, will affect growth and employment performance at both the regional and national levels? Will, and to what extent, the process of internationalisation strengthen the diffusion of new technology?

The role of manufacturing in overall growth of the developed countries has declined over the last decades. Considering the increasing role of services as suppliers of high-quality inputs into the manufacturing process one might be tempted to ask whether business services can be the new engine of growth for the whole economy and how different economies will be favoured or hampered by this new phase of development. Because services are, by the strong path-dependency of their development, very much country-specific, the emergence of a new growth regime is marked by national specificities, although one might have expected the opposite in times of increased internationalisation. In this context we want to investigate whether there are interdependencies between national advantages in specific manufacturing sectors and the development of service activities. Anyway, the characters of internationalisation are different according to the types of sectors involved: this holds for 3

manufacturing and the service sector, as well as for the linkages between the two. Hence a relevant question is: •

How the internationalisation and tertiarisation processes will differ across sectors? And how existing specialisation patterns in Europe will affect future dynamics and constrain the outcome?

In order to address these questions we will define a framework growth model where the effects of specialisation, and innovation on growth and employment can be formally analysed and empirically tested along the lines already explored in Padoan (1998). Padoan (1998) presents an analytical description and an empirical estimation of a dynamic disequilibrium model where equilibrium growth is made to depend upon the structure of specialisation as well as on endogenous technological accumulation. A similar model can be developed in order to analyse the relationship between innovation and employment creation at a sectoral level. In this context it will be possible to formally analyse the relationship between growth and employment also within the service sector and the differences across countries. The purpose of this review is to present the state of the art in different theoretical fields that will be integrated within the project. These are: the relationship between technology and specialisation (Section 2); the process of tertiarisation (Section 3); the development and diffusion of ICT (Section 4) and the internationalisation of technology (Section 5). Particular attention will also be devoted to the consequences of these new trends on growth and employment opportunities. The review of the literature also aims at highlighting open issues that will be addressed in the development of the project. Finally, in the last section we review the methodology that we intend to use in the quantitative part of the project. Figure 1.1 summarises the various fields of literature relevant to the project and their interdependencies. Figure 1.1 A summary of the streams of literature relevant to the SETI project

Literature on the ICT revolution (Section 4)

Literature on the rise of business services and the internationalisation of services (Section 3)

Literature on technology & specialisation (Section 2)

Literature on the internationalisation of technology (section 5)

Impact on growth and employment creation (all Sections)

2. Technology and specialisation This section focuses on technology-based theories of trade and specialisation. Section 2.1 reviews the main theories and empirical evidence; Section 2.2 looks at the impact of specialisation on growth and international competitiveness; Section 2.3 examines the 4

relationship between specialisation and employment creation; finally, Section 2.4 emphasises the open questions and the directions for future research. 2.1 Technology and specialisation: theories and empirical evidence The consideration of technology like a key determinant of country and sectorial trade performance has generated an impressive development of theoretical and empirical proposals in the last years. Two themes in particular have received special attention in recent years: • The first relates to the way in which technology enters as an explanation of the direction of countries’ trade specialisation. • The second concerns to the characteristic traits of trade and technological specialisation patterns of advanced countries: degree of stability, degree of concentration and trends to convergence versus divergence over time. Next, we present a brief description of the theoretical and the empirical literature around these themes. It is important to emphasise that almost all empirical research is confined to the manufacturing sector. This is a serious limitation if we agree with Laursen (2000, p. 12) that "private service sector makes up the same proposition of the economy as does manufacturing in many developed countries" and that "service sector is not only a growing sector in development economies but also a more innovative sector in recent years". 2.1.1 The impact of technology on trade competitiveness and specialisation Under the label of ‘technology gap theory’ Posner (1961) introduced the idea that temporary monopoly profits can be appropriated, based on a technological lead, in an international trade context. Given the assumption that technology is not a free and universally available good, Posner argued that while technology might be important for trade in some sectors, and not in others, innovations made in one country (in technology intensive sectors) would benefit that country as long as the lead could be kept. That is, a country will have ample first-mover advantages in a given sector, until other countries have imitated the innovation. Hence, in the original formulation of Posner, once imitation has taken place, more traditional factors of adjustment and specialisation would take over and determine trade flows. However, as argued by Dosi and Soete (1988), there is not necessarily anything impermanent about the importance of technology in determining trade flows, since static and dynamic scale economies flowing from the initial break-through act to prolong the lead. Coupled with new product innovations, these scale economies might well secure a continuous trade flow. A formalised neoclassical treatment of aspects of the idea is found in Krugman (1985). Metcalfe and Soete (1984) also observe that trade can be due to the difference between national rates of diffusion of demand and capacity growth and to time lags in technology transfer with respect both to demand and production. While this type of trade should be transitory, it is possible that different diffusion patterns may result in different patterns of development within a technology, thus affecting countries’ long-run comparative advantages. Overall studies using the technology-gap approach to trade emphasise inter-country differences in technical change as the basis of international trade flows. In this framework it is variation across countries in innovation capabilities within each sector, rather than inter-industry differences in endowments, which matters in explaining the direction of trade. From an empirical point of view, the technology gap theory has gained support among others from Soete (1981) and Dosi et al. (1990). Based on cross-country regression analysis, for a single year, these two studies showed that among 40 sectors in about half of them 5

international competitiveness was influenced by technological advantages (measured as US patents) in the same sector. From a panel data perspective, in a dynamic setting - and in an aggregate country perspective - Amendola et al. (1993) found convincing support for the hypothesis as well. Also applying panel data - and from a sectoral as well as a country-wise perspective - Amable and Verspagen (1995) showed that competitiveness in trade was significantly influenced by technological capabilities (US patenting) in eleven out of the eighteen sectors in question, when using a dynamic specification of the model. Given that the principal sources of technological change (inducement mechanisms) differ between firms according to principal sector of activity, different explanations for trade should not be expected to be of equal importance across industrial sectors. Thus, if trade specialisation is determined to a large extent by technology, we should not expect the importance of ‘technology’ to appear along the same dimensions. Some studies have used the Pavitt taxonomy (1984) to identify the determinants of trade and specialisation distinguishing across groups of sectors.

Box 2.1: The Pavitt taxonomy Pavitt (1984) identifies differences in the importance of different sources of innovation according to which broad sector the individual firm belongs. The taxonomy of firms, according to principal activity, emerged out of a statistical analysis of more than 2000 post-war innovations in Britain and was explained by the sources of technology; the nature of users needs; and means of appropriation. Four types of firms were identified accordingly, namely supplier dominated firms, scale-intensive firms, specialised suppliers and science-based firms. Supplier dominated firms are typically small and found in manufacturing and nonmanufacturing sectors. Most technology comes from suppliers of equipment and material. Scale intensive firms are found in bulk materials and assembly. Their internal sources of technology are production engineering and R&D departments. External sources of technology include mainly interactive learning with specialised suppliers, but also inputs from science-based firms are of some importance. Specialised suppliers are small firms, which are producers of production equipment and control instrumentation. Their main internal sources are primarily design and development. External sources are users (science-based and scale-intensive firms). Science based firms are found in the chemical and electronic sectors. Their main internal sources of technology are internal R&D and production engineering. Important external sources of technology include universities, but also specialised suppliers. Even though the taxonomy was devised at the level of the firm, it has implications at the level of the industry, as we would expect the broad sectoral regularities of firms to be reflected in the aggregate behaviour of the sector. Thus, given the above description of the taxonomy, one would expect ‘own’ sector technology to be most important for gaining market shares in science based sectors, while downstream linkages should be expected to be more important in the case of specialised suppliers. For scale intensive sectors inter-sectoral linkages - but also to some extent R&D - should be of importance, while supplier dominated sectors should to some extent be expected to be determined by upstream linkages and by low unit labour costs. The Pavitt taxonomy has been criticised on a number of points, including a set of criticisms relating to the fact that the sectoral boundaries are not always straightforward. That is, firms (and sectors) cannot always easily be uniquely defined as one of the four Pavitt type firms. Some firms (and sectors) may have such attributes, so that they can be said to be affiliated to more than one of the Pavitt-type sectors .

Using the Pavitt taxonomy and including as explanatory variables for market shares investment, patents and unit labour costs, Amable and Verspagen (1995) find that there are a number of cases where the results seem to support the taxonomy. However there are also a number of cases where this is not true. Laursen and Meliciani (2000) also include some measures of forward and backward linkages and find support for the taxonomy. In particular they find that, unit labour costs appear to play the largest role in supplier dominated industries, while ‘own sector’ technology (proxied by patents) plays the largest role in science based industries. Upstream and downstream linkages are particularly important in determining market shares in those industries which have important linkages with the others (scale 6

intensive for upstream linkages and specialised suppliers for downstream linkages, respectively). Padoan (1998) also uses the Pavitt taxonomy to explain export and import shares in a dynamic model where knowledge accumulation affects trade performance and competitiveness but also trade enhance knowledge accumulation through imports. He finds that the results are consistent with theoretical expectations about the importance of price and non price determinants of trade performance in the four Pavitt’s macrosectors. The technology-gap approach, in the most recent literature, has been used more often to explain market shares rather than specialisation. Also the Pavitt taxonomy has been used in this context with the aim of pointing out that different “factors of competitiveness” might matter more or less in different industries. 2.1.2 Trends in specialisation patterns Taking into account the importance of technology like a key determinant of trade performance and its principal microeconomic features, the empirical literature has also addressed the following questions: a) What is the degree of concentration of trade and technological specialisation patterns? Do the countries tend to be more or less specialised over time? b) What is the degree of similarity among specialisation patterns of advanced economies? Do countries converge or diverge in terms of specialisation patterns over time? c) Are trade and technological specialisation patterns stable across countries and sectors? Empirical studies do not show convincing conclusions in relation to the issue of the degree of concentration of each national pattern of specialisation. The findings of Amendola et al. (1992, 1998) show that in each country trade specialisation appears to be more concentrated than technological specialisation, while the reverse is true in the work of Barcenilla (1999). Nevertheless, in the studies of Soete (1987), Archibugi and Pianta (1992a, 1992b), Amendola et al. (1992, 1998) and Barcenilla (1999) there is a general consensus in the following result: smaller countries display a level of specialisation higher than big countries like Germany or France. This result reflects the fact that smaller countries need to reach a minimal dimension in R&D and production to specialise in specific sectors. It is possible to notice a trend to despecialisation in time that is evident in the trade field (Barcenilla, 1999; Laursen , 2000) and more controversial in the technological field. Cantwell (1991) and Archibugi and Pianta (1992b) found an increase in technological specialisation. This is in contrast with the results of Barcenilla (1999) and Laursen (2000). The issue of convergence is a major theme in economic research. General findings on this issue show a clear economic or global convergence between advanced nations. The role of technology in this process is crucial, with innovations and product and process imitation leading to rapid diffusion of technical progress. Nevertheless, empirical studies on technological and trade specialisation have not reached clear conclusions. The papers of Beelen and Verspagen (1994), Soete and Verspagen (1994), Barcenilla (1999) and Laursen (2000) show a tendency of trade specialisation patterns to converge, in contrast with the findings of Wolff (1997) who recognises convergence in terms of factor endowments but not in terms of patterns of specialisation. The results of Barcenilla (1999) show that the process of convergence is more evident in the technological than in the trade field. The 7

opposite is true in the findings of Archibugi and Pianta (1998) or Amendola et al. (1998) who emphasise the existence of a paradox: a process of aggregate convergence is taking place with increasing specialisation in selected technological niches by industrialised countries. Finally, the hypothesis that national specialisation patterns evolve gradually and cumulatively has found great support in the empirical literature. It is especially true in the trade field. The results of Amendola et al. (1998), Barcenilla (1999) and Laursen (2000) confirm, in line with evolutionary theorising, that national export specialisation patterns are quite complex. The stability of technological specialisation in the short and medium term is also recognised by Pavitt (1988) and Cantwell (1989). However, the results of Barcenilla (1999) do not support the evolutionary hypothesis with respect to the technological field. Also Laursen (2000) specify that, in comparison, trade specialisation patterns are more stable than technological ones. On this matter, Amendola et al. (1992) find that over long periods, the persistence of technological specialisation tends to fade away because of the emergence of new technological paradigms and new industries. Recently, a study by Midelfart-Knarvik et al. (2000) has focussed on similar questions adopting a different theoretical framework that draws both from factor-endowment and newgeography theories of specialisation. One of the merits of this study is that it considers also some service industries. The idea is that closer European integration is likely to bring with it major changes in industrial location. Industries will move to exploit differences in countries' comparative advantages. Moreover integration may change the attractiveness of central areas relative to peripheral ones and may facilitate the clustering of activities that benefit from linkages with each other. The objectives of the study are to describe the changes in industrial location that have occurred in Europe in recent decades; to establish whether these are associated with countries’ economic structures becoming more or less similar, and industries becoming more or less spatially concentrated; to compare industrial location patterns in Europe and the US; and to identify the underlying forces that determine industrial location and assess the extent to which these have changed in recent years. The authors find that: a) Most European countries showed significant convergence of their industrial structure during the 1970s, but this trend was reversed in the early 1980s. There has been substantial divergence from the early 1980s onwards, as countries have become increasingly different from the average of the rest of the EU and, in bilateral comparisons, from most of their EU partners. b) The most dramatic changes in industry structure have been the expansion of relatively high technology and high skill industries in Ireland and in Finland. However, the specialisation process has occurred more generally, with nearly all countries showing increasing difference from the early 1980s onwards. c) Many, although not all, industries have experienced significant changes in their location. In particular a number of industries that were initially spatially dispersed have become more concentrated. These are mainly slow growing and unskilled labour intensive industries whose relative contraction has been accompanied by spatial concentration, usually in peripheral low wage economies. Amongst industries that were initially spatially concentrated, around half stayed concentrated. Significant dispersion has occurred in a number of medium and high technology industries and in relatively high growth sectors, with activity typically spreading out from the central European countries.

8

d) Services are in general more dispersed than manufacturing. Two trends – the general shift from manufacturing to services, and catch up by poorer countries with small initial services sectors – have reinforced this spatial dispersion of services. e) While the industrial structures of EU countries are diverging, those of US states are converging. However, in so far as it is possible to make any comparison of levels of industrial concentration between the EU and the US, they find that EU industries are still less concentrated than are those in the US.

2.2 The impact of specialisation on international competitiveness and economic growth Recently there has been a growing empirical literature on the relationship between specialisation and trade and growth performance. This literature is very heterogeneous both in the theoretical framework adopted and in the methodology. Moreover it mostly aims at testing some ideas of various theoretical frameworks but it is rarely directly linked to any specific model. In other words the estimated equations are not, in most of the cases, reduced forms of any specific theoretical model. One way of looking at the impact of the different composition of national activities across sectors on countries’ trade performance has been the use of constant market share analysis. This approach aims at detecting to what degree the changes in the market share of a country in the world market can be explained by the initial commodity composition of each country’s exports. In order to do so, it decomposes the overall change in market share into a structural and a competitiveness component; the first effect is the difference between the hypothetical share had the single commodity shares remained constant at the initial commodity shares, while the second effect is the difference between the actual share and the hypothetical share. Most recent studies using constant market share analysis include Guerrieri and Milana (1995), Fagerberg (1996) and Laursen (1999). Guerrieri and Milana find that structural effects contributed much to the remarkable gains in export shares achieved by Japan and the East Asian NICs over the period 1973-90: these countries were able to improve their comparative advantage in high-technology goods as a whole and in particular in the field of electronics. Fagerberg finds that, between 1965 and 1990, trade in commodities with high R&D expenditures grew faster than other trade, and these changes were more favourable for the large and medium-sized high-income countries of the OECD area. Laursen, applying constant market share analysis to technology (patent data), finds that the catching-up countries (with the exception of Japan) have been able to acquire market and technology shares despite their “wrong” specialisation pattern. Other studies, within different theoretical frameworks, have directly analysed the impact of the degree and the “quality” of specialisation (either technological or trade specialisation) on countries’ growth and trade performance. Table 2.1 summarises the main results emerging from these studies. In a study directed at investigating the technological specialisation of the most advanced countries, with no explicit theoretical claim, Archibugi and Pianta (1992b) made an attempt to capture the effects of specialisation on growth, estimating an equation where growth in industrial production is explained by specialisation in fast-growing patent classes, and an equation where it is related to the degree of specialisation (measured by the value of the Chisquared statistics, a measure of concentration). They found no evidence of a significant relationship between specialisation in the fast-growing patent classes and growth in industrial 9

production, while they found a strong link with the overall degree of specialisation. Their results suggest that there is a general advantage in being specialised that goes beyond the choice of the more dynamic sectors. In a similar study, but using only correlation analysis, Pianta and Meliciani (1996) found a significant correlation between the degree of technological specialisation (measured by the value of the Chi-squared statistics) and countries’ rates of growth. They also found that technological specialisation in electronics was not significantly correlated with countries’ rates of growth.

Table 2.1: The impact of specialisation on economic performance Study Archibugi and Pianta (1992b)

Methodology Regression analysis

Dependent variable Rate of growth of industrial production

Specialisation Specialisation in fast growing patent classes Degree of technological specialisation

Yes

Pianta and Meliciani (1996)

Correlation analysis

Rate of growth of GDP

Laursen (1999)

Regression analysis

Rate of growth of exports Structural technology effect Technology adaptation effect Structural market effect

No Yes Yes

Dalum, Laursen and Verspagen (1999)

Regression analysis

Sectoral growth of value added

Yes

Amable (1996)

Regression analysis

Rate of growth of GDP

Inter-industry trade specialisation Yes index Trade specialisation in electronics Yes

Pianta (1997)

Regression analysis

Rate of growth of sectoral technology Rate of growth of sectoral production Rate of growth of sectoral exports

Technological specialisation in the same sector Production specialisation in the same sector Trade specialisation in the same sector

No

Rate of growth of GDP

Specialisation in fast growing patent classes Specialisation in ICT

Yes

Meliciani and Simonetti (1998)

Regression Analysis

Technological specialisation in electronics

Significant No

Specialisation for each of the sector product groups

No

No No

Yes

Padoan (1998)

Regression analysis and simulation

Rate of growth of GDP, Sectoral approach export and import shares

Yes

Meliciani (2001)

Regression analysis

Rate of growth of GDP, Specialisation in fast-growing export and import shares patent classes

Yes

Laursen (1999) estimates a technology-gap model of international competitiveness where technological opportunity is assumed to affect countries’ ability to compete in international markets: assuming that technology is an important determinant of export growth, it can be expected that countries which have access to sectors (or activities) offering high (low) levels of technological opportunity will benefit from higher (lower) international demand. Within this framework, changes in trade performance are explained in terms of changes in unit labour 10

costs, investment, catching-up, a structural market effect (the advantage coming from being initially specialised in the classes that grew more in terms of exports), a structural technology effect (the advantage coming from being initially specialised in the classes that grew more in terms of patents) and a technology adaptation effect (the ability to move actively into the technological sectors with above-average rates of growth). The results give no support for the structural technology effect to have an impact on market shares, while a significant effect is found for the technology growth adaptation effect and the structural market effect. Laursen comments on the results by stating that ‘it appears to be more important for national systems of innovation to actively move into sectors offering above average technological opportunity, rather than being “fortunately” specialised initially’ (Laursen, 1999, p. 18). Dalum, Laursen and Verspagen (1999) explore the relationship between specialisation and growth, arguing that this can be justified on the grounds of both supply-side elements (specialisation leads to higher productivity growth in the form of learning, and to the extent to which different sectors or activities are characterised by different learning opportunities different specialisation patterns can offer different growth opportunities) and demand-side elements (Engel’s law, or the idea that as income rises demand for some goods grow more rapidly than for other goods). The supply-side elements are discussed in both the new growth theory and the evolutionary approach, while the demand-side arguments are characteristic of the post-keynesian and the evolutionary approaches. Dalum, Laursen and Verspagen do not commit to any of the specific modelling environments proposed by the theories; rather their approach is based on a combination of certain elements of all theoretical frameworks discussed above. In order to test the hypothesis that the structure of an economy matters for economic growth, they run a regression with the sectoral growth of value added as the dependent variable, and several variables (the growth rate of employment, the investment-output ratio, the number of patents granted in the US per employee, and the ratio of value added per employee relative to the maximum level for the sector, as a measure of the scope for catching-up) together with a measure of specialisation, as explanatory variables for a sample of OECD countries over the period 1965-88. The specialisation variable is computed as the revealed comparative advantage (RCA) for each of the sectors’ product groups; however owing to the large number of product groups within some sectors they include in the regression only up to three of the principal components. The results for the specialisation variable show that, for all sectors but textiles and basic metals, there is at least one, but often more, principal component for each time period (the time periods are 1965-1973, 1973-1979 and 1979-1988) that is significant. Over time the number of sectors where specialisation is significant appears to decrease. Looking at which activities have had a positive impact on growth, they find some unexpected results (for example that specialisation in semiconductors turns out to have had a negative impact on the rate of growth of the electrical goods sector in spite of high growth of trade in semiconductors). They conclude that a complex set of interacting supply and demand side factors can lead to unpredictable results as to which activities offer better opportunities for growth. Finally they also find that specialisation in some product groups has different impacts on growth over time (positive or negative), and therefore policies aimed at changing the specialisation patterns of the economy are risky and difficult to implement. Overall this study gives support to the view that specialisation can have an impact on growth, but fails to identify some characteristics within the different product groups that can be beneficial or detrimental to the growth process (it gives an ex post account of which product groups contributed positively to sectoral growth with no ex ante explanation).

11

Amable (1996) refers to the work of Grossman and Helpman (1991), Rivera Batiz and Romer (1991), Lucas (1988), Young (1991) and Matsuyama (1992) as theoretical models where trade affects economic growth. In the first two studies the exchange of technical information, the diffusion of knowledge, scale effects and the differentiation of innovations (the consequence of preventing duplication in research) act to enhance technical change, productivity and/or consumers’ utility. In these models therefore it is the opening up of trade with the resulting process of specialisation that can be beneficial for growth, independently from the sectors in which countries specialise. In the other studies not every effect of international trade is necessarily beneficial to growth. If international specialisation diverts a country from technology-intensive industries, the opening up of trade can have a negative effect on countries’ rates of growth. Referring to these streams of literature, Amable estimates an equation where growth depends on catching-up, education, investment and some measures of specialisation, in a sample of OECD as well as some Asian and Latin American countries over the period 1960-90 (panel data). He uses three different measures of trade specialisation: inter-industry specialisation, a trade dissimilarity indicator, and an indicator of comparative advantage in electronics. The inter-industry specialisation index (Michaely index) is high if a country’s trade structure is characterised by the coexistence of clearly defined export and import industries, irrespective of the aggregate trade surplus: the higher the index the more trade balances are dissimilar between industries, i.e. the higher is specialisation. The trade dissimilarity indicator measures the dissimilarity between a country’s trade pattern and international demand patterns: it is higher when a country exports commodities in an industry where international demand is relatively low. Its effect on growth is therefore expected to be negative. Considering that electronics has enjoyed fast rates of technological progress and has had pervasive effects throughout most industry and service activities, specialisation in electronics is expected to have a positive effect on economic growth. He finds a positive effect of inter-industry specialisation and also of specialisation in electronics, concluding that the positive influence of inter-industry specialisation does not mean that all patterns of specialisation are equivalent. On the contrary, countries specialised in electronics appear to have experienced above-average rates of growth, and this lends some support to the view that electronics is a ‘strategic’ branch of activity. Pianta (1997) investigates whether growth in technology, production and exports is higher in the sectors where countries show stronger specialisation (measured by comparative advantage). He finds for a pooled sample of the six largest OECD countries (US, Japan, Germany, France, UK, and Italy) and 19 sectors, between the 1980s and the 1990s, that national sectoral performances in either production, export or patenting are negatively related to existing specialisation patterns. He concludes by stating that a sort of ‘decreasing returns’ to specialisation appear to emerge that are at odds with the view of increasing benefits from comparative advantage: the presence of relative strengths within some sectors do not lead to higher rates of growth within the same sectors either in technology, or in production and exports. Meliciani and Simonetti (1998) look at the impact of specialisation in fast-growing patent fields and in ICT on economic growth, also controlling for the investment activity and the overall technological intensity of the countries. They find that specialisation in ICT has a positive and significant impact on GDP growth over the whole period (70-93), while specialisation is fast-growing patent fields becomes significant only from the 1980s. They explain this result considering that in the ‘70s many fast growing technology fields were still linked to the old technological paradigm, while the ICT paradigm becomes dominant starting from the 1980s.

12

Padoan (1998) estimates a dynamic disequilibrium model focused on trade specialisation and knowledge accumulation. The sectoral breakdown follows Pavitt taxonomy by grouping manufacturing goods into four macrosectors. The empirical estimation shows the importance of both price and non-price determinants of trade. Moreover simulation exercises confirm steady state results on the relevance of country differences in trade specialisation and in domestic knowledge accumulation in affecting growth and they also clarify the relationship between growth and changes in the structure of trade specialisation. Meliciani (2001; 2002) looks at the impact of specialisation in fast-growing patent field on economic growth through the balance of payments. She finds that countries specialised in fast-growing fields experience above average growth in export shares and below average growth in import shares, with an overall positive impact on GDP growth through the balance of payments. Overall the empirical evidence on the importance of specialisation for international competitiveness and growth is controversial. While most studies find a link between specialisation and growth and trade performance, there is no clear evidence whether specialisation matters per se, due to increasing returns, or if there are activities that offer better opportunities for growth. The rising importance of ICT, particularly over the 1980s, has to be taken into account in order to understand the performance of different countries in technology, trade and production. This can be an explanation for the limitations of specialisation per se to offer high growth opportunities in the more recent period. 2.3 Specialisation and employment creation The debate on the effect of technological change on unemployment focuses on the laboursaving impact of process innovations and on the capability of the system to compensate the initial displacement. Whereas the labour-saving effects of process innovations are only partly compensated elsewhere in the system, by most of the various compensation mechanisms, the job creation effects of product innovations can be very substantial and have a powerful compensating impact. To the extent therefore that the profits from labour saving process innovation are subsequently channelled, directly or indirectly, into the development of new products and services, this provides the best possibility to generate new employment. This has been particularly evident both in the US and in East Asia, but clearly Europe has not benefited so much from this mechanism (Vivarelli, 1995). The job-creating and job-replacing effects of technical change have to be investigated in order to understand whether the sectors where more technological accumulation is taking place are also the fields where production and employment grow. The debate over the microeconomic causes of unemployment (or of equilibrium unemployment) largely ignores product composition even though this could be quite relevant for determining equilibrium unemployment and its evolution over time. This point can be clarified by considering labour demand and supply effects. To the extent that labour demand is determined by product demand, the sectorial product composition affects demand for labour since different products or sectors display different rates of growth of demand. The references commonly found in the literature relate to the distinction between mature and new products and between industrial and service products. The demand for labour should grow faster in new sectors and in services. The available evidence confirms that the larger growth-elasticity of employment is found in services (in Europe as well as in the US) even though in the latter case the growth elasticity of employment is high even in the manufacturing sector. On the other hand, as shown in the literature on endogenous growth and on the service economy, a 13

well developed advanced service sector generates positive externalities for the industrial sector. In the second place, given that productivity growth in services is lower than in industry, the employment content of growth is enhanced by a higher share of services in output. A service orientation of employment has been under way for some time in the industrialised economies at approximately the same rate as that of the share of value added in industry and services (Baumol, Blackman, Wolff, 1989). Product composition matters in determining the position and the evolution over time of the demand for labour; however, the technological intensity for a given sectorial composition is also relevant. It is not obvious that sectors of high demand growth or high technological innovation are also those with the greatest growth of employment. As shown in Padoan (2000) the employment trend is worse in the so-called traditional sectors, and some of the high technology sectors display the best employment performance. The sectorial employment dynamic within the manufacturing sector is highly variable, while the sector's overall trend is negative. The main implication is that the production specialisation of a country and its evolution are relevant for structural unemployment. This is another element of diversification among European countries (and regions). The above considerations make it difficult to evaluate the role of the so-called technological component of unemployment, i.e., whether the introduction of technology intensive products which, per se, diminishes employment per unit - offsets the greater absolute employment created by larger output. Available empirical evidence seems to confirm that employment growth tends to be positive, or at least more sustained, in high-tech sectors. Moreover, there is evidence of a positive relation between innovation effort - as measured in terms of R&D expenditure as well as patenting - and employment. Of course, employment growth induced by technological innovation will be possible to the extent that an adequate, highly trained, labour force will be available i.e., to the extent that a recomposition of labour demand is matched by a recomposition of supply. The dynamics of world demand are frequently associated with traditional (or mature) sectors with low growth rates of demand, and innovative segments (often those with a high value added) with high growth rates (European Commission, 1998). Nevertheless, several studies offer a more complex picture. In a longer time-horizon we can discard cyclical components and concentrate the growth of international trade volumes (exports and imports of the main OECD countries) as reference. We note that some medium-to-high-added-value sectors have shown percentage growth rates lower than the industrial average in 1980-94 (and in 1970-94); this is the case of the petrochemical and steel segments, construction materials and food & beverages. On the other hand, trade volume have grown at a sustained pace in "traditional" segments such as furniture, footwear and clothing, in addition - as is to be expected - to chemistry (including pharmaceutics), machinery (including computers, consumer electronics and telecommunications) and vehicles. Thus, the equation "traditional equals static" is not always valid. This can help explain the relative employment creation performance of segments such as furniture, but it runs up against tangibly negative employment data in other traditional sectors (even in those with sustained trade growth). Several studies allow to classify the industrial sectors on the basis of a few parameters such as market concentration, the share of the factors of production in value added, returns to scale, intensity of R&D, and the number of patents (Davies and Lyons, 1996; OECD, 1996). A first group of sectors with high R&D and skilled-labour intensity is also associated with high rates of growth of world demand (computers, telecommunications, other chemicals and some segments of the transport equipment). Among the capital-intensive sectors, we must 14

distinguish between skilled labour intensive ones (food, motor vehicles) and those which employ mostly unskilled labour (textiles, steel, glass, rubber and plastics). Finally, we can identify sectors, respectively of high skilled-labour intensity (machinery, professional instruments, electrical equipment) and unskilled-labour intensity (clothing, furniture, leather, some metal products). In the high-tech and high added-value sectors, Europe as a whole displays a low "quality" of production specialisation, i.e., ability of generating value added and of rewarding factors of production with high returns. While, in fact, only the US and Japan are specialised, respectively, in computers and telecommunications equipment both are also specialised in vehicles (together with Germany and Spain), some EU countries are specialised in chemicals (including Germany, France, UK and Netherlands) and in industrial machinery (Germany, France, Italy, UK, Sweden and the Netherlands). Europe's limited presence in the high-tech sectors should be ascribed to a generally modest intensity of R&D in many European countries (excluding Scandinavia, as well as non-EU Switzerland), measured both by low input levels of R&D and low output levels of patents. 2.4 Open questions and directions for future research The literature reviewed so far is almost entirely confined to the manufacturing sector (one exception is the paper by Midelfart-Knarvik et al.). This is a serious limitation when we want to examine the impact of specialisation on growth and employment creation, especially in industrialised countries. In fact in the EU services as a whole account for approaching twothirds of output and employment. Moreover the development of new technologies have made the manufacturing and the service sector much more interdependent. In particular services (especially business services) are playing an increasing role as suppliers of high-quality inputs into the manufacturing process. Similarly the presence of delivery services can impact on the productivity of manufacturing to the extent that services supply intermediate inputs to the manufacturing process. Finally the close interaction between many service activities and other parts of the economy makes them potentially important conduits for the spread of new technologies, the diffusion of new knowledge and the promotion of organisational change. We can therefore ask the questions: how is the rise of services affecting countries specialisation patterns? And what are the links between advanced services and manufactures? The ability to develop or adopt the new technologies in one country - to the extent that it affects the development of high-quality business services - has not only a direct impact on national specialisation in the manufacturing sector. It also has an indirect effect through the interdependencies between the manufacturing and the service sector. This set of issues will be emphasised in Section 3. The literature on technology and specialisation has also mostly ignored the issue of the internationalisation (globalisation) of technology. The lack of interest in the location of technological activity by Multinational Enterprises (MNEs) can be attributed to the widespread belief that MNEs were concentrating innovation activity in their home country. While this could be the case until recently, in the past decade the significance of the internationalisation of R&D activities of MNEs has been growing. Moreover European firms, especially from small countries, have shown historically a higher tendency, with respect to American firms, to conduct both production and innovation activities abroad. The impact of these trends on national performance and, in particular, whether technological globalisation has weakened the links between a country’s knowledge base and its trade, growth and employment performance is a controversial issue. Section 5 will focus on theories and empirical evidence concerning the globalisation of technology. 15

The processes of transformation that we have emphasised (the rise of business services and the globalisation of technology) are strongly affected by the development and diffusion of ICT. Technological developments in telecommunications and information technology have boosted the potential of many producer services. Also the internationalisation of services has taken a major leap in the last two decades due to the revolution in ICT industries. Moreover ICT have a direct impact on countries’ specialisation patterns and growth and employment performance. Section 4 will deal with the most relevant issues raised by the diffusion of ICT. 3. Services In developed countries the services sector has become increasingly important over the last couple of decades. In the EU services as a whole account for approaching two-thirds of output and employment; for market services the share of output is close to 50%. Initially the growing importance of services was met with concern since the services sector exhibits relatively low productivity growth. The expansion of services could well attract labour away from relatively high growth manufacturing sectors into low-growth services industries. As a result, this trend would lead to a reduction in economic growth. By contrast, more recently many have come to the believe that there is something very positive about the growing importance of services or at least some groups of services. They put great emphasis on the role of producer services generally and information and communications technology (ICT) in particular. However to date there has been relatively little research, especially on the empirical side, that analyses the differential impact of individual services on economic growth. The aim of this section is to provide a review of the literature on the economics of services, and in particular the internationalisation of services, as a contribution to the wider literature review. Not only has the relative importance of services in domestic production increased, but also it has become increasingly clear that services are not as non-tradable as had been generally assumed. Even though this may never have been a very appropriate assumption, the internationalisation of services has taken a major leap in the last two decades due to the revolution in ICT industries and the trend towards deregulation in domestic markets. Throughout the section special attention is given where possible to three key services industries: distribution, management consulting and financial services. The section is structured as follows. In Section 3.1 some concepts and definitions related to the nature of services as well as to the internationalisation of services will be discussed. In Section 3.2 the recent developments in the internationalisation of services will be analysed. Section 3.3 reviews the regulatory environment for services both in the context of the EU and in the WTO. Section 3.4 examines how specialisation in services and the internationalisation of services affect economic growth and employment. Finally Section 3.5 discusses some open issues and directions for future research. 3.1 What are services? This section discusses first a number of concepts and issues related to the nature of services. Secondly, concepts related to the internationalisation of services will be examined. 3.1.1 Services versus goods Services have long been defined negatively as all economic activities whose output is neither agricultural nor a manufactured product. In a now classic study Fuchs (1968) tried to pin down the nature of services. Fuchs defines services on the basis of three key features. The first feature that distinguishes services from goods resides in the non-storable nature of 16

services. As a result services have to be produced and consumed in the same place and at the same time (generally referred to as the ‘physical proximity requirement’). Bhagwati (1984) and others have noted that due to the technological developments in telecommunications and information technology services the physical proximity requirement in the delivery of services may have been reduced. 1 However, the scope of separating receivers and providers geographically may be limited due to the intangible nature of services, the second feature emphasised by Fuchs. Factor services are not as in manufactures employed to produce physical outputs, but factors instead produce intangible assets. Consequently, problems of asymmetric information may arise as it is impossible to assess the quality of the service prior to purchase. Building a reputation is crucial to the strategy of services providers. Creating trust requires establishing long-term relationships which are importantly shaped by personal contacts. The intangibility of services explains therefore the path-dependence that characterises many service industries. Third, not only are services generally produced where they are consumed but they are also produced with the participation of the consumer, that is, services tend to be customised. Both intangibility and customisation tend to reinforce the ‘physical proximity requirement’. Hill (1977, p. 317-318) proposes a truly general definition distinguishing services from goods. “A good may be defined as a physical object which is appropriable and therefore transferable between economic units. A service – on the contrary – may be defined as a change in the condition of a person, or of a good belonging to some economic unit, which is brought about as a result of the activity of some other economic unit, with agreement of the former person or economic unit”. According to this definition services are no longer considered as immaterial goods, but are different in essence by being a change (Siniscalco, 1988). 2 More recently researchers seem to have abandoned the goal of defining services generally, instead they tend to adopt a more pragmatic approach acknowledging that the services sector comprises an extremely diverse set of economic activities. More attention is now directed towards developing sensible classifications that allow one to answer more specific questions. 3.1.2 The internationalisation of services Sampson and Snape (1985) argue that a definition based on geography as is used for goods, is not appropriate, because this would exclude, for example, tourism. Indeed, they define international transactions of services as ‘a transaction between the resident of one country and a resident of another’ (p. 172). However, this definition excludes for example foreign students. University students often spend one or more years abroad for study purposes and as such become residents of the country of study. A more appropriate definition should therefore be based on nationality. 3 Sapir and Winter (1994) implicitly acknowledge this as they include migration and consequently argue that migration policy can constitute a barrier to trade. The most common typology of international transactions in services is based on the analysis of movement of users and providers respectively that occurs for the transaction to take place 1

Bhagwati (1984) call this the ‘disembodiment effect’. For the measurement of services output Hill’s definition implies that one should focus on the change in the receiver, which corresponds more to value-added than gross services output (Siniscalco, 1988). Weiss argues in response to Siniscalco that the challenge is to construct sound price indices as the value of the change in condition to the receiver is reflected in its price. 3 Sampson and Snape’s definition however corresponds well to the existing balance of payments (BOP) statistics as BOP statistics register transactions between residents and non-residents. A definition based on nationality, that is one that distinguishes between ‘foreign’ and ‘national’ residents is theoretically more appropriate, but practically impossible to work with. However, analysis of the impact of GATS for example requires such data. 2

17

(Sampson and Snape, 1985; Sapir and Winter, 1994). This typology also provides the four ‘modes of supply’ employed in the GATS. Type 1 Neither the user nor the provider moves, but both reside in different countries. This is the case to which Bhagwati (1984) referred. Trade in services of this type is similar to trade in goods. For this reason it is also called commodity trade. This type of trade in services is likely to arise in financial services (and some professional services). Type 2 The user moves to the provider to receive the service. This the case in for example tourism and education, but is not likely to be important for the three industries of interest in this paper. This is also called commodity trade. Type 3 The provider moves to the user in order to perform the service. This type is referred to as factor trade (Ruane, 1990). This could be important for both financial and consulting services if physical interaction is required but not on a regular basis. Type 4 The provider establishes a foreign branch in order to obtain physical proximity to users abroad. This is probably the most common type of internationalisation of services. It is important for banking, consultancy and distribution. The first three types of transactions are generally referred to as trade, whereas type 4 involves the permanent movement of capital and possibly also labour. Theoretical models of trade in producer services are proposed by Markusen (1989), Melvin (1989), Francois (1990a and b) Jones and Ruane (1990), Van Marrewijk et el. (1997), De Vaal and Van de Berg (1999). Models of the latter type by Van Marrewijk et al. (1996), and Markusen, Rutherford, and Tarr (2000). Type 4 leads to the establishment of multinational service networks. Whereas international fragmentation of production in manufacturing is most often motivated by cost considerations and market access, for services the ‘physical proximity requirement’ seems to be the predominant consideration. Therefore, one may label such services providers as delivery services. Intuitively, it seems that service providers are more likely to establish several foreign affiliates within a single country than for manufacturers. For some services activities, those where economies of scale are particularly important and physical proximity may not be regularly required, it may be justified to concentrate them in one location. Such services could be called specialist or headquarter services. Specialist service activities are like to include management, marketing and product development. Given those two types of services the internationalisation of services may take two forms. First, vertical fragmentation refers to the international fragmentation of the integrated activities according to their different functions, that is, delivery and specialist activities tend to be separated geographically (Markusen, 1984). Second, horizontal fragmentation refers to the case where delivery and specialist services are located within the same country. The revolution in telecommunications especially is likely to promote the establishment of vertically fragmented networks as the intra-firm flow of information tends to be the more intensive in such networks. However, the extent of vertical fragmentation may be very limited as a result of diverging regulatory environment. Harmonisation in the EU may have affected the form of internationalisation, encouraging a further concentration of specialist services (form the national to the European level). Within the vertical form of international fragmentation it is possible to distinguish between vertically concentrated and vertically dispersed organisational structures. The former refers to

18

the case where specialist activities tend to be located within agglomerations and the latter where specialist activities are dispersed. Their location may be based on history. Similarly it may be useful to analyse the impact of the presence of delivery services on the productivity of manufacturing to the extent that services supply intermediate inputs to the manufacturing process. Further, do services follow manufacturers (Sapir, 1988) or are manufacturers attracted by advanced services infrastructure? 4 In addition to the commonly used typology based on international transactions, this section thus introduces a new functional typology which may help to analyse in more detail type 4 of internationalisation in services viz: vertical fragmentation (concentrated) vertical fragmentation (dispersed) horizontal fragmentation 3.2 Trends in services In this section the extent of the internationalisation of services is analysed. Ideally, one would consider trade in services (types 1 to 3), foreign direct investment in services, and migration related to the internationalisation of services (type 4). Data on migration with the aim of supplying or receiving a service is not available. 5 Only recently have statistical offices started collecting FDI data that distinguish between services and manufacturing. Disaggregated FDI data distinguishing different services sectors is available only for the United States. The review deals first with trade in services followed by FDI in services. 3.2.1 Trade in services Trade in services is measured according to the residential definition of international transactions. As a result international trade between ‘foreign’ and ‘national’ residents of the same country is omitted. Non-residents become residents if they reside in the country for one year or more. Trade in services in this section only approximates to types 1-3, thus underestimating the extent of internationalisation in services. Empirical research on trade in services is quite limited, not least because of the availability and quality of statistics of trade in services. Some have employed production and employment data in order to make inferences regarding the pattern of specialisation in a way to bypass data limitations on trade. However, it seems that such measures differ importantly from tradebased indices. Table 3.1 compares trade in goods with trade in services. The first point to be noted is that trade in services is much smaller than trade in goods. Nevertheless for a supposedly nontradable sector the share of services in world trade is rather about 15% to 20%. Moreover, the table shows that services have gained in relative importance from 15.2% of exports in 1980 (16.2% of imports) to 18.4% of total exports in 2000 (17.7% of total imports). These shares tend be somewhat higher in developed countries and slightly lower in developing countries.

4

Gross, Raff, and Ryan (2001) find evidence for the former in the early years of Japanese FDI in Europe, but observe that the picture is completely reversed in more recent years. 5 Although not very satisfactory Chang et al. (1999) use BOP category ‘compensation for employees’ which registers “the earnings of all natural persons established abroad for less than one year (p. 112)”. 19

Table 3.1: World trade in goods and services 1980

1985

1990

1995

1998

1999

2000

Merchandise (billion $) Exports 2035.19 1950.59 3441.68 5079.00 5446.93 5662.11 6364.37 Imports 2074.69 2009.43 3542.10 5217.86 5666.53 5898.81 6668.60 Services (billion $) Exports 365.1 382.1 782.7 1189.1 1333.6 1357.3 1435.4 Imports 400.9 400.2 818.3 1197.9 1328.7 1355.2 1436.9 Share of services in total trade (%) Exports 15.2 16.4 18.5 19.0 19.7 19.3 18.4 Imports 16.2 16.6 18.8 18.7 19.0 18.7 17.7

Source: WTO The relative importance of service sectors as well as the tradability of services differs greatly across individual sectors as well as countries. Table 3.2 represents trade in services for some broad sectors for all EU countries in 1999. Table 3.2: Trade in services by EU country and broad sector for 1999 Financial services

Austria Bel-lux Finland France Germany Greece Ireland Italy Netherlands Portugal Spain Sweden UK

Insurance Communi- Computer cation and services information services 5.2 3.6 0.6 8.3 3.8 1.8 5.7 13.1 8.5 6.6 4.9 9.4 -0.1 1.3 6.8

Other business services

Total services

3.6 2.6 7.1 6.3 0.6

5.2 4.8 6.8 6.3 1.1

Exports Imports Exports Imports Exports

3.0 4.7 18.1 25.6 0.0

Imports Exports Imports Exports Imports

0.0 0.1 7.8 3.6 16.0

0.1 -0.7 9.5 2.6 19.0

1.3 0.5 6.4 -11.9 23.4

4.7 0.6 4.8 -8.2 32.8

1.4 3.7 10.6 -7.7 24.8

1.3 3.2 11.1 -8.7 23.2

Exports Imports Exports Imports Exports

0.2 1.5 5.3 7.0 8.6

0.4 0.9 6.2 12.5 5.4

2.9 1.0 2.9 1.5 9.3

0.5 0.5 26.0 2.3 1.8

1.2 0.6 1.3 7.1 10.4

2.8 1.6 2.5 4.5 10.4

Imports Exports Imports Exports Imports

13.9 1.9 4.1 0.8 1.2

9.1 1.4 4.9 0.4 0.9

12.7 10.7 9.9 1.5 0.7

6.6 5.6 7.4 0.4 1.5

12.5 10.9 9.9 0.6 0.7

10.1 8.9 8.4 1.4 1.2

Exports Imports Exports Imports Exports

4.3 7.3 1.6 2.4 36.1

5.0 8.3 2.9 1.9 41.2

4.7 4.1 4.2 5.3 19.3

10.4 8.7 5.6 7.9 15.9

5.1 6.7 4.5 4.9 20.0

9.1 5.3 3.4 3.9 17.5

Imports

1.8

8.1

19.8

6.1

7.1

14.7

Source: OECD

20

Comparing imports and exports for the different countries, Germany and France have large trade deficits in services revealing that their comparative advantage is still in manufacturing. By contrast, Spain, Greece and the UK run a trade surplus on services revealing a comparative advantage in services. However, for individual services sectors the picture differs importantly. Spain and Greece derive their comparative advantage in services from consumer services, especially tourism, whereas the United Kingdom’s comparative advantage is largely based on producer services. Germany and France are large net importers of producer services. Elsewhere patterns of services activity are quite diverse. Financial services and insurance exhibit a similar pattern. The United Kingdom is the only country with a strong trade surplus in these services sectors, while all others EU countries run a deficit. This may reflect the role of US and Japanese financial services firms within the EU. 6 Belgium-Luxembourg, Greece, Ireland, and Portugal are net exporters of communications services, whereas France, Germany and Italy are net importers. This suggests an important role for country size in communications services. In computer services Ireland and the United Kingdom run notable trade surpluses. The category ‘other business services’ contains a diverse range of service activities. Broadly speaking, it consists of merchanting and trade related services and miscellaneous business, professional and technical services. The latter group (consulting in a broad sense) represents more than 90% ‘of other business services’ in the United Kingdom. Interestingly, Ireland appears to be a strong importer of other business services. Possibly, the dynamism of the Irish economy is reflected by the high demand for business services. Alternatively, it could reflect the use of home service providers by foreign affiliates in Ireland. Disaggregated data for the world as whole are not available. Chang et al. (1999) estimate them. Their results for 1996 reveal that transport and travel largely dominate global trade in services, together accounting for 58% of total services trade. Financial and insurance services account for about a fifth of the remainder. 7 Table 3.3 represents the value and table 3.4 the share of trade in total services and some selected services for Japan, the US and the EU. For Japan trade in total services peaked around 1996-1997 and fell back slightly afterwards. For both the US and the EU the value of trade in total services increased significantly over the period 1990-1999 and at a fairly constant rate. Overall, the US is strong net exporter of services, the EU is also a net exporter but by not nearly as much as the US, and Japan is a strong net importer. The only reversal in the pattern of trade relates to the EU. The EU moved from a net importer in insurance and other business services in the early 1990’s to a net exporter in the late 1990’s. In the late 1990’s the EU had become a net exporter of all three selected producer services.

6

For more information on the role of Japanese financial firms in the EU see Hawawini, G. and M. Schill (1994). The category ‘royalties and fees’ includes franchising, which under the GATS is a sub-sector of distribution services. 7

21

Table 3.3: Trade value by major economic region and selected industry (1990-1999) 1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

Other Business Services Japan

Exports

..

16 064

17 353

19 810

21 724

24 475

21 980

21 645

17 054

15 756

Imports

..

25 319

25 644

26 192

28 015

31 921

33 004

31 910

28 049

25 991

United States Exports Imports

17 223 10 608

20 643 12 131

22 799 12 540

23 092 13 823

27 530 16 035

28 895 18 067

33 148 20 377

38 600 22 894

40 432 25 453

41 612 28 579

European

Exports

44 814

52 378

92 188

87 959

93 518

108 240 127 991

128 580 136 463

137 919

Union

Imports Imports

47 828 ..

53 518 ..

92 052 ..

88 893 ..

91 829 158 908

107 348 118 986 183 962 202 796

117 847 131 118 203 485 217 614

139 561 228 272

Exports

..

89

144

233

188

313

2 837

1 849

1 609

2 033

Imports

..

1 491

1 112

1 147

565

468

2 979

2 675

2 134

2 712

United States Exports Imports

4 417 2 475

5 012 2 669

4 034 986

4 999 1 371

5 763 1 654

7 029 2 472

8 229 2 907

10 243 3 347

11 273 3 561

13 925 3 574

FINANCIAL SERVICES Japan

European

Exports

19 142

18 159

23 169

24 633

29 162

22 075

24 784

27 184

27 273

31 380

Union

Imports Imports

14 040 ..

14 373 ..

17 507 ..

18 462 ..

22 925 ..

13 839 19 541

14 383 23 419

14 882 24 234

14 684 24 235

17 992 ..

Exports Imports

.. ..

- 259 - 276

- 171 902

75 1 892

364 2 570

296 2 495

488 1 916

347 2 034

60 2 379

- 74 2 346

INSURANCE SERVICES Japan

United States Exports

230

491

682

1 020

1 676

1 296

2 168

2 473

2 189

2 295

Imports

1 910

2 467

1 324

3 095

4 034

5 360

3 885

5 873

9 080

4 078

Exports Imports

3 575 5 616

4 162 5 997

11 666 12 783

12 612 12 471

14 870 13 572

12 157 10 288

13 481 11 181

13 282 9 830

11 388 10 126

16 145 11 346

Imports

..

..

..

..

26 705

24 720

26 216

27 498

32 091

28 437

65 462

67 713

69 331

62 371

60 887

European Union

TOTAL SERVICES BY COUNTRY Japan

Exports

41 314

44 801

49 019

53 201

58 295

Imports

84 064

86 604

93 006

96 191

106 214

122 764 130 033

123 403 111 678

114 882

United States Exports Imports

146 433 162 591 117 045 118 039

175 134 184 046 116 476 122 281

198 977 131 878

217 360 238 100 141 447 150 850

255 289 260 691 166 487 182 687

269 578 191 289

European

Exports

381 532 388 608

469 570 440 966

473 528

523 871 554 736

560 052 575 141

590 655

Union

Imports

354 099 368 790

456 539 427 180

461 105

516 514 540 254

537 875 561 236

579 092

Source: OECD

22

Table 3.4: Share in total services trade by major economic region and selected industry (1990-1999) 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999

Other Business Services Japan United States European Union

Exports Imports Exports

… 35.9 35.4 37.2 37.3 37.4 32.5 31.2 27.3 29.2 27.6 27.2 26.4 26.0 25.4 25.9 25.1 11.8 12.7 13.0 12.5 13.8 13.3 13.9 15.1 15.5

25.9 22.6 15.4

Imports

9.1 10.3 10.8 11.3 12.2 12.8 13.5 13.8 13.9

14.9

Exports Imports

11.7 13.5 19.6 19.9 19.7 20.7 23.1 23.0 23.7 13.5 14.5 20.2 20.8 19.9 20.8 22.0 21.9 23.4

23.4 24.1

Financial Services Japan United States European Union

Exports Imports Exports Imports

3.0 2.1

0.2 1.7 3.1 2.3

0.3 1.2 2.3 0.8

0.4 1.2 2.7 1.1

0.3 0.5 2.9 1.3

0.5 0.4 3.2 1.7

4.2 2.3 3.5 1.9

2.7 2.2 4.0 2.0

2.6 1.9 4.3 1.9

3.3 2.4 5.2 1.9

Exports

5.0

4.7

4.9

5.6

6.2

4.2

4.5

4.9

4.7

5.3

Imports

4.0

3.9

3.8

4.3

5.0

2.7

2.7

2.8

2.6

3.1

Exports Imports Exports

0.2

-0.6 -0.3 0.3

-0.3 1.0 0.4

0.1 2.0 0.6

0.6 2.4 0.8

0.5 2.0 0.6

0.7 1.5 0.9

0.5 1.6 1.0

0.1 2.1 0.8

-0.1 2.0 0.9

Imports

1.6

2.1

1.1

2.5

3.1

3.8

2.6

3.5

5.0

2.1

Exports Imports

0.9 1.6

1.1 1.6

2.5 2.8

2.9 2.9

3.1 2.9

2.3 2.0

2.4 2.1

2.4 1.8

2.0 1.8

2.7 2.0

Insurance Services Japan United States European Union

Source: OECD, own calculations 3.2.2 FDI in services In order to assess the relative importance of type 4 transactions over time several proxies may be used. Chang et al. (1999) distinguish three statistical sources that could be used to proxy for commercial presence. These are: data on FDI, 8 affiliate sales, 9 and market size. Comparable data across countries distinguishing between goods and services are not available. Table 3.5 represents several indicators for world foreign production. Unfortunately, the data do not distinguish between goods and services. The share of services in foreign production has increased substantially over the last two decades. At present service activities account for over 50% of the stock of FDI (UNCTAD, 2001). 10 8

FDI statistics are defined on the basis of 10% ownership or more. However, for the present purposes the issue of control is probably more relevant, that is a majority of equity interest or in the board of directors. The GATS indeed uses a concept based on control instead of ownership. 9 Only the US has a tradition is administering affiliate sales (or Foreign Affiliate Turnover), although most OECD countries have begun to so recently (Chang et al., 1999). 10 For more information of FDI in services see UNCTAD and World Bank (1994). 23

Table 3.5: World foreign production in goods and services Value at current prices (Billions of dollars) 1982 1990 2000

Annualised growth rate (%) 1986-1990 1991-1995 1996-1999

1998

1999

2000

FDI inflows

57

202

1271

23

20.8

40.8

44.9

55.2

18.2

FDI outflows

37

235

1150

26.2

16.3

37

52.8

41.3

14.3

FDI inward stock FDI outward stock

719

1889

6314

16.2

9.3

18.4

19.8

22.3

21.5

568

1717

5976

20.5

10.8

16.4

20.9

19.5

19.4

151

1144

26.4

23.3

50

74.4

44.1

49.3

2465

5467

15680

15.6

10.5

10.4

18.2

17.2

18

565

1420

3167

16.4

7.2

11

3.2

27.3

16.5

1888

5744

21102

18.2

13.9

15.9

23.4

14.8

19.8

17454

23721

45587

5.7

5.3

7.8

16.8

5.3

12.7

2124

4381

7036

15.4

8.6

1.9

-1.5

3.9

Cross border M&As Sales of foreign affiliates Gross product of foreign affiliates Total assets of foreign affiliates Employment of foreign affiliates (thousands) Export of goods and nonfactor services

Source: UNCTAD

The first and most important thing to note is that growth in foreign production whether measured in terms of FDI flows or sales by foreign affiliates has been dramatic over the past 20 years. For comparison the development of trade is represented in the lower row of the table. Whereas trade has grown much more rapidly than world production during the 1980’s and early 1990’s - at an average rate of approximately 10% - it is dwarfed by the spectacular rise in foreign production. In the late 1980’s and early 1990’s FDI inflows increased at a rate slightly above 20% a year. More recently the growth rate even doubled to 40.8% for the period 1996-1999. In 1999 FDI inflows increased by 55.2%. For the year 2000, however, indications are that the growth rate of FDI inflows has fallen. However, it may be more appropriate to focus on the stock of FDI and sales by foreign affiliates. They also show a spectacular increase, although not so extreme and thus are more stable over time. The stock of FDI has been growing at a fairly constant annual rate of about 20% and sales by foreign affiliates between 10-20%. One could conclude therefore that foreign production has been growing dramatically over the last twenty years at a annual rate in the range of 10-20% which is significantly faster than the growth rate in trade or world production. In addition, it seems that FDI flows (especially the way they are registered) represent more than just the idea of expanding foreign production, that is, they may be driven by other factors as well (based on 10% ownership instead of control). Thus, as an indicator of commercial presence, data on the stock of FDI and affiliate sales are to be preferred. 24

Chang et al. (1999) analyse the different types of international transactions of services (Section 3.2) using disaggregated data for the US. Trade in services is based on the IMF’s BOP statistics, whereas commercial presence is proxied by affiliate sales. In the aggregate both types of transactions are fairly similar in importance. At a more disaggregated level their relative importance is fairly constant over time. For transport services, trade appears to be the most important mode of supply; for telecommunications, there is no significant difference; and for the remaining service industries, affiliate sales is the dominant form of internationalisation. For insurance and professional services, commercial presence is generally required as close co-operation between client and supplier may be necessary (unless reinsurance of risks is concerned). 3.3 The regulatory environment In this section a brief overview will be given of the regulatory environment of services both within the EU and in the global context. The General Agreement on Trade in Services (1995) is the first and so far only multilateral regulation which covers the internationalisation of services. Tradability of services has traditionally been considered to be very low for two reasons. First, trade in services may be difficult or even impossible as a result of the nature of services (see Section 3.1). Second, services tend to be heavily regulated. It is crucial to realise that regulation is justified whenever market failures are present. The most important market failures in service industries are imperfect information, imperfect competition and network externalities. Both the WTO and the EU do not intend to deregulate service, but instead they aim to liberalise international transactions in services for reasons that should become clear in Section 3.4. In spite of regulations being perfectly justified they can, and often do, still act as a barrier to trade. On the basis of the classification in Section 3.1.2 it is possible to distinguish the following barriers to the internationalisation of services: barriers to trade, barriers to the movement of users, barriers to the movement of providers, and barriers to FDI and migration. 11 3.3.1 The free movement of services within the EU The EU member states have committed themselves to establishing a common market allowing for the free movement of goods, services, capital and labour since the signing of the Treaty of Rome in 1957. From the discussion in Section 3.2 it follows that completely free movement of services can only be achieved if the free of factors of productions is also addressed. Although substantial progress had been made in the integration of goods markets, services market integration was not seriously dealt with before the Single European Market programme was initiated in 1986. That this was completely unjustified follows from data on the importance of services (Buigues and Sapir, 1993). In 1990 market services accounted for 48.2% of GDP and 42% of employment. The share of services in GDP had increased by approximately 10% since 1970. During the 1980’s 10.4 million jobs were created in market services, whilst 3.6 million jobs were lost in manufacturing. The SEM programme or “1992” initiative was meant to open up domestic services markets in order to improve competition and allow service providers to exploit economies of scale. In this context Sapir (1993) analysed the actual and potential degree of competition in a number 11

For an extensive analysis of barriers to trade in services see Hoekman and Primo Braga (1997). 25

of selected services industries on the basis of objective criteria. A characteristic that is almost generally shared amongst services industries is that service providers tend to compete on the basis of quality and not on price. 12 Table 3.6: Determinants of Market Structure in Services (late 1980’s) Sector

Insurance

Degree of Degree of product concentration differentiation High Medium

Sunk costs Intensity regulatory measures Medium High

Actual degree Potential of competition degree of competition Low Medium

Banking

High

Medium

Medium

High

Low

Medium

Distribution

High

Low

Low

Medium

Medium

High

Business services

High

Low

Medium

Low

Medium

Medium

Type of market structure Oligopolistic competition Oligopolistic competition Monopolistic competition Monopolistic competition

Source: Adapted from Sapir (1993)

However, the ‘new approach’ adopted in the Single European Act to tackle the remaining barriers to trade, although quite effective in goods markets, did not deliver the desired results in many services industries. Most of the problems encountered resulted from the densely regulated domestic services markets. The principle of mutual recognition did not prove as effective as in goods market. Therefore, the European Commission had to rely heavily on harmonisation, which was tedious given the degree of regulation. The European Commission tried to promote liberalisation of the services markets by adopting a sectoral approach.

Box 3.1: Banking in the Single European Market The Second banking Directive of 1 January 1993 introduced the principle of ‘single licence’ allowing banks and other credit institutions to set up branches and offer services that can be provided in all the Member States on the basis of such a licence. Home-country control implies that supervisory authorities of the country of origin are responsible for the behaviour of banks abroad. In addition, the Second Banking Directive entails harmonisation of essential requirements, that is, prudential regulation complemented by mutual recognition between national authorities of rules and regulations of the country of origin of foreign banks. In order the assess the impact of the SEM programme on banking one has take into account three developments (Pelkmans, 2001): the trend towards deregulation in most countries; the liberalisation of financial capital; and finally the Second banking Directive. Those developments generally increased competition, forcing banks to restructure their operations by a surge in M&A (in Netherlands, France, Belgium). However, M&A were largely domestic “and the share of domestic deals in total deals did not decline during the late 1980’s as was the case in other industries (Pelkmans, p. 124)”. Intra-EC trade in banking services surged in the second half of the 1980’s and early 1990’s as a result of deregulation and capital liberalisation (Gual and Neven; Pelkmans). Intra-EC FDI accelerated dramatically in the period 1984-1988 “but it only marginally affected actual penetration in terms of market share”, (Pelkmans, p. 124). Productivity increased more rapidly where deregulation was more pronounced.

Currently the European Commission is in the process of developing a new strategy. The key change is the desire to move from a sectoral to an all-inclusive generic approach putting more emphasis on the principle of mutual recognition and engaging in horizontal harmonisation, that is, designing rules and regulations that apply across different sectors. Service activities undertaken within manufacturing firms will also be addressed (likely to be important for 12

‘Distribution regulation is typically about land-zoning for shopping centres (with restricted entry) and shop opening hours, a sensitive issue for numerous family businesses everywhere (Pelkmans, p. 122)’. 26

distribution). The generic approach is inspired by the idea that traditional divisions between sectors become increasingly blurred. Fragmentation of production creates strong interdependencies between different sectors. Regulation or liberalisation in one sector should therefore be expected to have important knock-on effects in other industries.

Box 3.2: On-Line Trading in Distribution Distribution involves both wholesaling and retailing. Especially the potential of wholesale via the internet seems to offer great potential for on-line trading. Retail trade via the internet is limited to mainly non-food products. The latter are likely to remain largely in the domain of supermarkets. Hubertus (2001) gives an indication of the relative importance of individual product categories in on-lines sales. Financial services turn out to be the largest product category traded over the internet, but online sales are also important for computer products, books, music/video and travel.

3.3.2 The GATS and beyond The General Agreement on Trade in Services (1995-1997) is the first and so far only multilateral regulation which covers the internationalisation of services. The GATS applies to all four types of international transactions of services discussed in Section 3.1.2. The GATS consists of a general and a specific component. The general component consists of rules and regulations that apply across the board to government policies affecting international transactions in services. 13 The most important general obligation is the one of Most-Favoured-Nation (MFN) treatment. Article II of the GATS requires each Member to “accord immediately and unconditionally to services and service suppliers of any other Member treatment no less favourable than that it accords to like services and service suppliers of any other country”. The specific component includes the specific commitments of each WTO Member, its national ‘schedule’. National ‘schedules’ lists those services for which a WTO Member wishes to guarantee market access to foreign providers on a national treatment basis. Article XVI which relates to market access stipulates that a WTO Member cannot maintain or adopt, unless pre-specified in its national schedule, any restrictive measures. Restrictive measures include limitations on: a) the number of service suppliers; b) the total value of services transactions or assets; c) the total number of services operations or the total quantity of service output; d) the total number of natural persons who may be employed in a particular sector; e) specific types of legal entity through which a service can be supplied; and f) foreign equity participation (e.g. maximum equity participation). Article XVII:1 defines the concept of national treatment: “In the sectors inscribed in its Schedule, and subject to any conditions and qualifications set out therein, each Member shall accord to services and service suppliers of any other Member, in respect of all measures affecting the supply of services, treatment no less favourable than that it accords to its own like services and service suppliers”. Hence, the GATS does not allow discrimination between

13

Exceptions are services provided in the exercise of government authority, the air transport sector, air traffic rights and services related to exercise of traffic rights. 27

national and foreign service suppliers in those sectors where the WTO Member has committed itself to guaranteeing free market access. Hoekman and Primo Braga (1997) state that most countries scheduled only a part of their services sector. Developed countries made (limited) commitments for 47% of the total possible compared to 16% by developing countries. Mattoo (2001) argues that “the GATS has failed to deliver meaningful liberalisation. The Agreement is generally perceived, not as a scourge of protection, but as a rather stodgy reaper of liberalisation accomplished elsewhere (p. 3)”. Whether or not the GATS will prove effective in reducing barriers to trade in services will therefore depend on the commitments made by individual WTO Members in their national Schedules. The only study that attempts to assess the impact of GATS on trade in services is presented by Chang et al. (1999). They try to evaluate the impact of GATS systematically by services sector and mode of international transaction. Not surprisingly, many data problems arise and as such their study presents only a first step in research on liberalisation of trade in services and its consequences. The agenda for the next round of multilateral negotiation was set at the 4th Ministerial Conference of the WTO in Doha in 2001. Services will again feature prominently on the agenda, and market access and its liberalisation will move into a higher gear. “In services (…) the World Bank has estimated that welfare gains from a 50% cut in services sector protection would be five times larger than for non-services sector trade liberalisation”. In addition, the Doha Declaration sets for the first time ever the objective of establishing a multilateral set of rules and regulations with the objective of improving the potential of FDI globally and to establish a consistent competition policy. Furthermore, the right of the host country to regulate will be preserved. Finally, developing countries will obtain an opt-out clause.

Box 3.3: Liberalisation of Trade in Financial Services under the GATS Liberalisation of financial services is covered by the Financial Services Agreement (FSA) of GATS. In addition to the general rules set by the GATS, it contains the national schedules specifying rules and regulations with respect to market access in financial services. Prudential regulation need not be inscribed in the national Schedule. As restrictions to market access are clearly defined, new financial instruments may be left outside national Schedules and future negotiations may be impeded (Sorsa, 1997). “While the FSA goes far beyond previous regional trade agreements (…), in terms of actually dismantling trade barriers in the area of financial services, it has done little more than formalising the status quo (Cornelius, 2000, p.652)”.

Finally, Hoekman and Primo Braga (1997) draw attention to two interesting political economy aspects of the liberalisation of services. First, a host may be quite favourably disposed towards the liberalisation of services since a reduction in employment is not expected (commercial presence). Secondly, problems may arise where co-ordination between regulatory bodies with regard to foreign affiliates is required. 3.4 Services, employment and economic growth In this section the importance of services to economic growth and employment will be analysed. Before discussing how services may affect economic growth and employment it 28

will be useful to discuss the modern service economy in some more detail. Section 3.4.1 will therefore concentrate on the causes behind the move towards the service economy. Section 3.4.2 will discuss the nature of technological change in services. Section 3.4.3 will examine the link between technological change and employment in services. Section 3.4.4 in turn will deal with the consequence of the move towards the service economy for economic growth. More precisely, it will discuss the direct and indirect way in which this development may contribute or impede economic growth. Section 3.4.5 finally will discuss the importance of the internationalisation of services to economic growth. It will thus provide a motivation for the liberalisation of service markets in the context of the EU and the WTO. 3.4.1 Explaining the service economy Traditionally, the services sector has been associated with low productivity growth. Therefore, the rising importance of services in the economy has raised concern amongst academics and policy-makers alike. In the United Kingdom this was reflected in the late 1950’s by the introduction of an employment tax for services industries in order to discourage their growth. Baumol, Blackman and Wolff (1989) argue that the move towards a services economy constitutes a concern when services sector employment is associated with relatively low wages and low career opportunities, and when service growth withdraws workers from more productive manufacturing industries and thus impedes overall economic growth. Before turning to figures of services sector growth it is useful to gain some understanding of the underlying factors that contributed to this trend. Here, five factors will be discussed. First, there is strong evidence that the annual growth rate of labour productivity in manufacturing has been much higher than that in services in the OECD countries (see Klodt, 2000). Productivity grew at over 3 per cent a year in manufacturing compared with less than half this rate in services 14 . The superior performance of manufacturing reflects a much higher expenditure on innovation – Evangelista (2000) finds that innovation costs per employee in Italy in 1995 were three to four times higher in manufacturing than in services. Given that the improvement in labour productivity in services has been relatively small, while output trends have been fairly similar to those in manufacturing, the services sector absorbed a rising share of total employment in combination with rising relative prices of services. Second, economic development may have changed the pattern of tastes in favour of services. Clark (1957) argues that when income per capita increases demand will shift increasingly towards services while the expenditure share on manufacturers stabilises and will eventually fall. Baumol, Blackman and Wolff (1989) reject this explanation on the basis that the share of final expenditure on services is fairly similar in developed and developing countries. Their argument is supported by Klodt (2000) who found that in FR Germany between 1978 and 1990 the share of services in final demand was static and hence ‘non-homothetic preferences of consumers are not the decisive determinant of structural change on the demand side. The major driving force behind the transition to the service society is a rising service intensity all over the economy’ (p9). This reflects the increasing role of services as a provider of intermediate inputs in both manufacturing and services. Thus the finding by Rowthorn and Ramaswamy (1999) that demand is indeed one of the explanatory factors of deindustrialisation must be qualified in that it is demand for services as intermediate goods that is the key factor, not a faster growth of final demand for services versus goods.

14

Germany is unusual in having a rough parity of productivity growth in the two sectors, perhaps due to the limited presence of services activities with low productivity (ibid p12). 29

Third, and related to the previous point production processes are increasingly fragmented both within and across national borders. Across national borders fragmentation is mainly driven by relative factor price differences. Within countries economies of scale are likely to be the main force driving fragmentation of production. 15 It is a stylised fact that economies of scale are extremely important for production services (those result amongst others from the need to build up a reputation and establish a user network). Bhagwati (1984) refers to fragmentation of the latter type in the case of services as the ‘splintering process’. Bhagwati also argues that such services are typically technically progressive, as it is the presence of economies of scale that made them independent in the first place. 16 Actually the types of fragmentation are likely to reinforce each other. In order to split up production processes, producer services are needed to co-ordinate the activities between different locations. The technical revolution in ICT industries may have been needed to unleash the fragmentation process (Jones and Kierzkowski, 1990). Alternatively, the establishment of sophisticated international production networks may have increased the demand for specialised producer services. Even though the above does not suggest that economic development shifts tastes towards consumer services, economic development is associated with an increasingly complex production process which increases the demand for producer services. In addition, it seems that with economic development the drive towards specialisation becomes stronger. Fourth, trade with developing countries may also have contributed to the North specialising increasingly in services industries. Although most studies have found a limited impact of foreign competition on labour markets in developing countries, Feenstra and Hanson (1996, 1999) in their research on the international fragmentation of production find much stronger effects by focusing on the trade in intermediate inputs. Fifth, technological developments in telecommunications and information technology have boosted the potential of many producer services. Even though this development may not explain the rising importance of services in terms of employment, it is undoubtedly a key aspect of the modern service economy. 3.4.2 Technological change in services The multifaceted nature of innovation makes it difficult to make comparisons between services and manufacturing. Surveys suggest that expenditure on innovation is much higher per employee in manufacturing. However, this typically excludes expenditure for example on customisation and organisational change which may be important for innovation and which are found particularly in the services sector (Evangelista, 2000). Moreover, the close interaction between many service activities and other parts of the economy makes them potentially important conduits for the spread of new technologies, the diffusion of new knowledge and the promotion of organisational change (Miles 1993, 1996; Hauknes 1996). Survey evidence for Italy (Evangelista 2000) suggests that the pattern of innovation in services shares some common features with that in manufacturing – for example the dominant role of large firms. Similarly both sectors have a high rate of process innovation However, product innovation is reported to be much less common in services. Interestingly, the juxtaposition of production and consumption in services did not appear to hamper the ability

15

Sapir (1985) also mentions employment flexibility. The third argument does properly speaking not reflect an increase in the actual importance of services, but only an increase in the recorded importance of services (Sapir, 1985). 16

30

of the firms surveyed to distinguish between the two categories of innovation. Given the heterogeneous nature of the services sector, it would be surprising if there were not marked differences in the pattern of innovation between different categories of services. This is certainly borne out for Italy as Evangelista’s work demonstrates (see Box 3.4). Particularly interesting is the importance for some sectors, such as advertising, of the role of customer interaction and software development in the innovation process. The evident diversity suggests caution in generalising about technological change in the services sector as a whole. Box 3.4: Innovation patterns in Italian services (Evangelista 2000) Evangelista’s pioneering work in Italy suggests a classification of service activities according to their innovation behaviour. A primary distinction is in relation to the extent to which innovation expenditure is focused on R&D and design rather than investment. On this basis, R&D, Engineering and Computing are the core science and technology based service activities, with much above average expenditures on innovation, especially in the areas of R&D and design. These sectors have much in common with the ‘science-based’ firms and sectors identified by Pavitt (1984). Technical Consultancy is also in the high spend category but has in addition an important emphasis on close interaction with final customers, other consultancies and research institutes. All other service sectors are characterised by limited R&D and design activity and a reliance on investment for innovation (embodied technological change). Within this broad group there is a spectrum of behaviour. At one extreme, the ‘technology users’ reflect traditional user-producer links – the innovation pattern resembles Pavitt’s ‘supplier dominated’ category. The most characteristic activities in this group are Waste Management, and Transport (both Sea and Land). Security, Cleaning and Other Business Services also fall into this group. At the other extreme are service activities which have a high emphasis on close interaction with consumers or users and large investments in software – Evangelista’s ‘interactive and IT’ group. Innovation costs derive from the acquisition of know-how from consultancy firms, the purchasing and development of new software and other innovation sources such as training and marketing. This group includes especially Advertising, but also Banks, Insurance, Trade/Repair of Motors and Hotels. Most of the remaining service activities fall into an intermediate group, being technology users but with some emphasis on customer interaction and software – Legal Services, Other Financial Services, Tourist Services, and Retail.

Petit and Soete (1996) argue that technical change in services has its specificities. In the first place the development of goods and service markets are not submitted to the same type of constraints. The localisation of services and the interaction between customers and producers that occurs in these trades impose specific constraints to the development of new products and of new processes. Both process and product innovations in services will thus be more severely constrained by the willingness, abilities and original tastes and habits of the customers than they are for goods. Process innovations in manufacturing of goods are neutral for the product market and product innovations can be channelled by widespread advertising, marketing techniques or straightforward and rapidly diffusing demonstration effects. ICTs have also a specific impact in that respect as these technologies transform the basic context in which services can be perceived and delivered. Petit and Soete (1996) argue that ICTs change the tradeability of services and expand the potential of fields and forms of new markets. Conversely ICTs transform the markets of goods, with more customerisation and lasting relations, bringing thus characteristics of services to these markets. 3.4.3 Consequences on employment It is generally acknowledged that employment in our economies is increasingly dependent on services. As in other highly developed economies, the European countries are continuing their gradual move towards a service-based economy with today nearly 70% of the total labour force being employed in service activities (Petit and Soete, 1996). It is also generally acknowledged that services provide the key to future employment growth. Neither agriculture nor manufacturing have been able to generate sufficient output growth to offset, in the last two decades, the productivity growth following the diffusion of labour saving machinery and 31

the reorganisation of work and trades, impulsed by an increasing international competition. And while some high-tech manufacturing sectors have succeeded through the introduction of new and improved high income elastic consumer goods to generate new employment opportunities, their number has been falling steadily over time (Petit and Soete, 1996). Evangelista and Savona (2001) argue that technological change is likely to play an important role also with respect to the overall dynamics of employment in services. In particular the emergence and widespread diffusion of ICTs has an impact on employment through three main mechanisms, that is by: a) expanding final demand and/or shifting its composition from tangible goods to intangible, information and knowledge intensive services; b) changing the composition of intermediate demand both in services and manufacturing towards information and knowledge based inputs and processes; c) increasing labour productivity in some of the service activities traditionally affected by the so-called cost-disease or productivity-bias. The overall impact of technological change on employment in services is however very difficult to be empirically assessed because of the joint presence of positive and negative direct effects and the existence of a complex set of compensating mechanisms, operating both within the service sector, between the service and the manufacturing sectors, and increasingly at an international scale. In particular, mechanisms a) and b) above are likely to have a positive impact on employment, provided that they do not phase out pre-existing services, goods and intermediate inputs. On the other side, mechanism c) is likely to have a negative impact, although there might be positive compensating mechanisms on employment via the reduction of prices, income increases and the consequent growth of demand brought by technological change in services. As far as the relationship between technological change and employment in services is concerned, the lack of data and systematic analyses referring to the service sector is severe. Evangelista and Savona (2001) argue that in absence of robust statistical evidence, the literature has been taken over by optimistic scenarios on the employment perspectives linked to the emergence of the new technological paradigm. Most of the contributions have tended to emphasise the positive employment effects of ICTs, both at the firm level and in the service sector as a whole. The traditional view, which portrayed services as "sheltered" sectors, characterised by low productivity and poor technological performance, has been superseded by one emphasising the high technological performance of sectors such as ICT-related services, telecommunications, or high value added business services (the so called KIBS Knowledge Intensive Based Services). However, as already mentioned above, the fact that ICTs might also be used to cut down costs, rationalise production and delivery processes, save jobs and skills has often been overlooked. In other words, only the positive effects associated to the mechanisms a) and b) mentioned above have been taken into account. The debate on the skill–bias nature of technological change shows a similar bias. There is in fact much more emphasis on the jobs’ opportunities offered by the ICTs than on the jobs and skills made redundant by the adoption of new technologies. The contribution by Evangelista and Savona looks at the impact of innovation on employment in services using the empirical evidence gathered through the 1993-95 Italian innovation survey. The empirical evidence presented shows that the impact of innovation on employment varies greatly according to the type of innovation strategy pursued by firms, across industries and according to the level of qualification of the labour force. Innovation activities tend to substitute low skilled jobs with jobs with a higher level of qualification. Among small firms 32

and in more than half of the service sectors considered the net effect is positive, particularly in industries which have a strong scientific and technological base. The negative impact of innovation on employment is on the contrary concentrated among large firms, capital intensive industries and in all financial-related sectors (banking, insurance and other financial services). In these latter industries the labour-saving effect of innovation seems to be linked to the widespread use of ICTs. In the case of Italy, an overall negative impact of innovation on employment is found. It is argued that this result is affected by the specialisation of Italian economy in the most traditional service industries. 3.4.4 Consequences on growth: Direct and indirect effects In order to analyse the desirability of a services economy - that is, the contribution of increased specialisation in services to economic growth - one can distinguish between direct and indirect effects. From a theoretical point of view the direct effects can best be considered in a simple neoclassical growth model with exogenous technical progress. To the extent that the move towards the service economy reflects an improvement in the allocation of resources throughout the economy, deindustrialisation will change the level of steady- state output. In addition, if one assumes that exogenous technological progress may differ across sectors, deindustrialisation improves welfare if it moves resources into more technologically progressive industries. Thus, in the short-run deindustrialisation reflects a change in the allocation of resources reflected in a change in the level of output, whereas in the long-run it may change economic growth depending on relative productivity growth in the expanding industries compared with the declining industries. To appreciate the indirect effects of a shift towards services one needs an endogenous growth model. Deindustrialisation reflects an improvement if the expanding services sector exhibits a higher degree of sustained growth characteristics, such as positive externalities and knowledge creation, than the declining manufacturing industries. In other words the contribution of deindustrialisation resides in the degree of spill-over effects in services relative to manufacturing. Table 3.7: Labour productivity by industry, 1987-1997 Industry Total Private Sector Agriculture Mining Construction Manufacturing Durables Electronics Nondurables Services Transportation Communications Public Utilities Wholesale Trade Retail Trade Finance Other Services

Share (1992) Growth Rate (1987-1997) 100 1.1 2.1 2.2 1.7 4.4 4.3 -0.1 19.8 2.9 7.8 2.5 2.8 8.7 9.1 1.3 71.3 0.7 3.6 0.8 3 2.9 3.3 3.8 7.6 4.0 10.1 1.6 21.4 1.6 22.4 -0.7

Source: Triplett and Bosworth (2002) based on BEA data. 33

Table 3.7 reflects labour productivity by broad industry. As can be seen from Table 3.7, both services and manufacturing are extremely diverse. However, although manufacturing productivity growth indeed exceeds that of services, it is the electronics sector that is largely responsible for this differential. In services, wholesaling exhibits higher productivity growth than manufacturing. Finance and retailing experience lower but still above average productivity growth. Finance though, is inflated upwards due the stock-market boom. Consulting (not represented) experienced zero productivity growth. One should be careful in interpreting those numbers as there exist many measurement issues (see Text Box 3.5).

Box 3.5: Measurement Issues in Services The low productivity in producer services may be product of the mismeasurement of output (Triplett and Bosworth, 2002). Gullickson and Harper (1999) use input-output analysis to adjust for output that serves as an intermediate input to other industries. Triplett and Bosworth argue that there exists no universal method to the measurement of output in services industries. Triplett and Bosworth state that the share of business services has doubled over the last decade (also one of the fastest-growing export sectors). Approximately 80% is provided to other businesses. “Measuring the output of business services involves an effort to determine who should receive credit for the productivity gains recorded by the users of business services” (p. 9). Especially for consulting it is difficult “to define the firms’ activities in a way that leads to clear measures of their output”. Studies that have tried to collect such information on consulting are Nachum (1999) and Gordon (1999). Distribution suffers from the problem that wholesaling also occurs within firms or wholesaling may not be clearly distinguishable from retailing. Also output is often by measured by the value of sales, not the service activity itself (the margin).

A crucial aspect of a large and increasingly important share of services is that services serve as intermediate inputs to manufacturing or other service industries. This not only complicates the measurement of output as highlighted in Box 3.5, but it may also seriously underestimate the contribution of producer (intermediate) services to economic growth. The literature linking the use of producer services to productivity in manufacturing is just beginning to develop. However, it is felt that having a well-developed ‘service structure’ is a crucial ingredient to an economy’s performance. For some services industries the existence of positive externalities has been extensively emphasised. For more information on the role of a financial infrastructure in economic growth see Box 3.6. Another example is presented by the research and development industry which constitutes the key in the Schumpeterian type of sustained growth models. A general model that stresses that importance of specialised producer services is presented by Francois (1990a).

34

Box 3.6: Financial Development and Economic Growth The availability and quality of financial services is a crucial ingredient of economic development. For a review of the literature see Pagano (1993) and Levine (1997). The contribution of financial development to economic growth does not reside, as is often argued, in a positive relationship between the private savings rate and financial development – this link appears to b e (nearly) non-existent -, but in the positive relationship between financial development and intermediated savings. A positive effect of financial development on economic growth, with total private savings constant, can only be explained by either the lower cost of investment, i.e. a higher operational efficiency, or the better employment of capital, i.e. a higher allocative efficiency. Operational efficiency is primarily important in order to induce savings mobilisation, whereas allocative efficiency is most important for its impact on sustained economic growth. The provision of long-term finance (maturity transformation through liquidity insurance) is largely dependent on the ability of financial intermediaries to overcome incentive frictions resulting from information asymmetries when allocating capital. In order to solve for such frictions financial intermediaries engage actively in screening, monitoring and exerting corporate control.

3.4.5 The internationalisation of services and economic growth First of all, it should be noted that even in the absence of trade in services, services may still have an important international dimension. The availability of producer services or the ‘service infrastructure’ within a country may be an important determinant of a country’s comparative advantage.

Liberalisation of trade in services involves the free movement of services as well as free movement of factors of production. Liberalising trade in services affects economic growth in a number of ways. First, it will affect the allocation of resources. Traditionally, this has been considered as having only a static effect (changes in the level of output). However, Ventura (1997) argues in a recent paper that under certain conditions the law of diminishing returns applies only to world averages. The crucial point is that trade allows countries to specialise. By specialising towards more capital-intensive sectors one can circumvent diminishing returns. A similar argument should apply to the impact of foreign direct investment on capital accumulation. Secondly, liberalisation is likely to have a pro-competitive effect thus reducing prices and increasing efficiency. Services typically compete on a non-price basis. Instead increased competition is likely to encourage innovation. Third, trade in services will lead to external economies of scale in the form of a greater variety of inputs (Francois, 1990b). Also producer services may be available at a better quality or a lower cost. Fourth, the larger market size allows services providers to exploit economies of scale. It should be noted that this equally holds for research and development. The incentive to innovate will increase with the size of monopoly rents that can be reaped from successful innovations. Fifth, liberalisation will encourage international knowledge diffusion. This is especially important for the liberalisation of foreign direct investment. In a recent study Borensztein, De

35

Gregorio, and Lee (1998) show that foreign direct investment is associated with positive productivity spill- overs. Many empirical studies have investigated a positive relationship between trade liberalisation in goods and economic growth (see for example Edwards, 1993). Only very recently have studies begun to appear on liberalisation of services and economic growth. On the basis of the relative importance of services in production one might expect that trade liberalisation would yield even more important gains (Mattoo et al., 2001). Most empirical studies of liberalisation of services and economic growth are case studies. However, Kim and Kim (2001) conduct a general study on the link between the liberalisation of services in Korea in the 1990’s and productivity improvements in services and manufacturing. Markusen, Rutherford and Tarr (2000) examine the impact of the liberalisation of producer services on the economy as whole. The comparative static model predicts gains form liberalisation in the range of 3-15% of GDP. The source of these large gains resides in the indirect effect of services liberalisation on final goods sector productivity. “More service firms allow final goods producers to use more specialised expertise, in the same way that larger markets allow for specialised machine tools (p. 25)”. 17

Box 3.7: The Internationalisation of Financial Services and Economic Growth There exists an extensive empirical literature that tries to establish the causality from financial intermediation to economic growth. However, very little work has been done on the relationship between the internationalisation of financial services and economic growth. Francois and Schuknecht (1999) explore both theoretically and empirically the linkages between trade in financial services and growth. They find a strong positive correlation between liberalisation of trade in financial services and financial sector competition as well as between the financial sector and economic growth. They interpret their results as evidence for a strong causal relationship between trade in financial services and economic growth. Mattoo et al. (2001) analyse the impact of openness in telecommunications and financial services on sustained economic growth. For both sectors they find positive and statistically significant results, although the impact of financial service sector openness is much more pronounced than in telecommunications.

3.5 Open issues and directions for future research This section has reviewed the literature on a number of issues regarding the internationalisation of services. Services have become increasingly important in domestic production and play a significant role in globalisation. For policy-makers and researchers alike, services have moved to the top of the agenda. With regards to concepts and definitions the literature seems to be fairly well established. However, it may be useful to improve the understanding of the organisational structure of international service networks. In particular more attention should be directed towards an indepth analysis of the specific characteristics, the measurement, the internationalisation, and the impact on economic growth of distribution, financial, and consulting services. A particularly interesting avenue is to analyse the role of these three specific services in a country’s ‘service infrastructure’. How does the development of a country’s service 17

They note that although in the short-run this may harm wages of domestic skilled workers, the availability of cheaper imported services leads the service-intensive final good industries to expand, increasing the demand for skilled labour.

36

infrastructure affect productivity in the rest of the economy? Are distribution, financial, and consulting services equally important is this respect? What is the role of culture in establishing user networks? Does a service infrastructure attract manufacturing or do services travel with manufacturing? One possible direction of future research that will be followed in this project is the development of a taxonomy of the internationalisation of services. It is helpful to think of the international services firm as employing labour in two activities: delivery and specialist business services. The delivery activity will be responsible for the face-to-face contact with the customer and hence for tailoring the services provided to the needs of the individual consumer. The skill with which they are able to supply the service will depend on the input that they can get from the firm’s specialist business services activity which is a source of information, knowledge and expertise. These specialist activities will tend to be carried out in the firm’s head offices. In banking, for example, there will be certain financial activities such as futures market operations which are located in financial centres like London and Frankfurt. By contrast the delivery activity is likely to be more geographically dispersed so as to be close to customers. Services are generally distinguished from goods by their requirement for face to face contact between seller and buyer. This derives from the general non-storability of services combined with their characteristic intangibility. Although technological changes are eroding the distinction between services and goods, the face-to-face contact criterion remains a useful one as far as delivery activities are concerned. As communications have developed it is increasingly feasible for specialist and delivery activities to be geographically separated, not just within countries but also between them. International fragmentation of the production process has become increasingly important in manufacturing allowing individual parts of the production process to be relocated where costs are lowest. We use fragmentation below as an important criterion in distinguishing between forms of internationalisation in services. One might also like to consider the possibility that at least part of the delivery activity could take place outside the client’s country – this would allow the consumer to move to the producer of specialist services. Some foreign banks may, for example, have branch offices in London which enable them to access specialist financial services which they don’t themselves carry out in-house, and which they can then deliver back to their clients in their home country. We argue that the possible geographical separation of delivery and specialist business services within individual firms is a crucial aspect of internationalisation. However, the delivery activity is always located close to the consumer, enabling face-to-face contact, and the consumer is considered to be immobile. This might need to be reconsidered but conceptually the geographical separation of some delivery activities from consumers does not look interesting since they then share many of the features of the specialist services. Thinking of the services firm as employing labour in two activities – the specialist and delivery activities – suggests a basis for a typology of the internationalisation of business services. A fundamental distinction should first be made between vertical and horizontal forms of internationalisation. In vertical forms, a firm’s specialist business services activity would be located in the home country while the delivery activity is located in both the home country (for domestic consumers) and the foreign country (for foreign customers). When internationalisation takes a horizontal form, both specialist and delivery activities are in the same country; internationalisation occurs through the exchange of knowledge and expertise between specialist activities in different national subsidiaries of the same firm. In this view, the international firm delivers a branded, international product to a collection of national 37

markets. For example, in management consultancy KPMG aims to deliver a distinctive product to clients in different countries. It tries to secure an advantage over purely national firms from the ability to pool its expertise across countries 18 . In practice, systems are unlikely to be perfectly horizontal since the firm will usually have a national identity as well as a multinational presence. So it can be expected that there will be a headquarters country (home country) with at least some specialist services. But the extent to which these services are concentrated in one location clearly varies substantially across service sectors. Within the vertical form of internationalisation it is useful to make a further distinction between vertical concentrated and vertical dispersed arrangements. Vertical concentrated internationalisation occurs where the specialist activities of firms agglomerate in the same country. This may be driven by external economies (as e.g. in the City of London for banking). Hence only the home country firm will have (some) delivery and specialist activities in the same country. In internationally dispersed arrangements the country of location for the specialist activity differs between firms – there is no tendency for the activity to concentrate in one location. This dispersal could occur because there are disadvantages in agglomeration – e.g. higher wage costs – or at least no advantages. Further, there might be communication problems in the separation of specialist and delivery activities. Another possibility is that government intervention seeks to prevent the loss of specialist services – as in the airline business where governments have traditionally supported national carriers. The typology proposed is summarised in the table below which considers the simplest case of two countries each with a representative firm.

Specialist Delivery

Country 1 Country 2 Country 1 Country 2

vertical concentrated Firm 1 4 6 4 4

Firm 2 4 6 4 4

Vertical Dispersed Firm 1 4 6 4 4

horizontal Firm 2 6 4 4 4

Firm 1 4 4 4 4

Firm 2 4 4 4 4

In this typology, we have chosen to take a broad view of internationalisation as encompassing trade, FDI and international information flows. • Internationalisation in the services sector always requires FDI, either just for the delivery activity (vertical) or for both specialist and delivery activities (horizontal). • International flows of information within firms – from specialist to delivery activity in the vertical case, and between specialist activities in the horizontal case 19 - play an important role. • The vertical form of internationalisation of services generates international trade in services as conventionally understood. The trade flows are one-way in the case of vertical concentrated arrangements and two way where the form is dispersed. • Information flows may also require the temporary movement of personnel both from specialist activities and from delivery. 18

KPMG indeed has a ‘knowledge management system’ that enables knowledge sharing between KPMG’s 104,000 employees spread over 159 countries and 821 cities. For example KPMG England produced a document on eFinance that was used by KPMG activity in Germany (J. Ottway, Report on internship in KPMG, 2001). 19 In practice, trade might also take the form of flows of goods complementary to the information flows (e.g. a multinational retailing firm might ship own-label brands to subsidiaries from the home country).

38

For the three sectors identified by the SETI project for special attention the following table attempts to identify the nature of the activities undertaken: Retailing Banking Management consultancy

Specialist activity Product design, international procurement, design of delivery systems Specialised financial services Specialised management services, collating information, management activity training

Delivery activity Product delivery – retail stores Product delivery – branch banks Product delivery – branch consultancies

We have already stressed how the ICT revolution has played a major role in the process of fragmentation and internationalisation of services. In what follows we focus the attention on the theoretical and empirical contributions that have highlighted the major characteristics of the ICT revolution and its impact on the economy. 4. The Information and Communication Technology (ICT) revolution The all-purpose technological revolution based on Information and Communication Technologies (ICT) – born in the US - has by now spread in most of European countries. However, while the macroeconomic benefits of the ICT revolution are already apparent in some industrial countries, and especially in the United States, in several European countries they are still hardly quantifiable. As it is widely known, ICT contributes to labour productivity growth through both capital deepening and total factor productivity (TFP) growth. The rapid decline of prices of goods that embody the new technology stimulates unusual investment in these goods, thus resulting in significant capital deepening (ICT-related capital deepening). At the same time, technological change raises TFP growth in the innovating sector. Both phenomena can lead to an acceleration in overall productivity growth. Therefore, it is straightforward that the amount, diffusion and characteristics of ICT investment expenditure is a key issue of the ongoing technological revolution on the economic system. Two streams of literature have focussed on the role of ICT as a pervasive technology, affecting not only the productivity and employment level of the ICT sector, but of the whole economic system. These are the Schumpeterian literature on techno-economic paradigms and the neo-classical literature on General Purpose Technologies (GPT). The first theoretical framework emphasises the complementarity between technological and organisational change and the crucial role played by institutions in favouring or hampering the diffusion of the new paradigm in the different countries. The second approach is more formalised and focuses on the comparison between the social optimum and the outcome of a decentralised economy in the presence of GPTs which are characterised by innovational complementarities giving rise to increasing returns to scale. In this section, we first present these two frameworks (Section 4.1) and then provide a review of the empirical literature on ICT, productivity and growth with particular attention to comparisons between the US and Europe (Section 4.2).

39

4.1 Theoretical frameworks 4.1.1 ICT as a new techno-economic paradigm The concept of techno-economic paradigm has been used to refer to a set of guiding principles, which become managerial and engineering commonsense for each major phase of development (Perez, 1983, 1985, 1988; Freeman and Perez, 1988). A change in paradigm carries with it many clusters of radical and incremental innovations and has pervasive effects throughout the economy, spreading from the initial industries where it takes place to the whole economy. Such characteristics may be found in different waves of development in coal, steel, oil, and nowadays in microelectronics and telecommunications. Innovations in the field of microelectronics have led to a technological revolution that has directly affected some industries (computers, electronic components, telecommunication) and indirectly the whole economic system of the leading industrial countries. Moreover it has also involved organisational and social changes to the extent that we can properly talk of a new techno-economic paradigm. Freeman and Soete (1987) analyse the growing impact of the new techno-economic paradigm by considering its uneven diffusion from a few leading sectors to the economy as a whole. Considering in particular the service industries, they find that there is a group of service sectors based on information technology that are among the fastest growing activities but still account for a small proportion of service output and employment (software, databanks, computerised information services, design etc.). There is another group of service sectors that has also been affected by information technology with uneven increases in labour productivity among firms and countries. This includes the banking, insurance and retail services (this phenomenon is particularly important as it shows that gains in productivity are not necessarily higher within the manufacturing sector). The majority of service sectors have not been able to achieve major gains in labour productivity owing to the lack or very limited diffusion of information technologies. Overall the new paradigm linked to ICT has unevenly affected the different sectors of the economy; however its impact does appear to have had a pervasive nature. As a consequence countries’ ability to achieve high rates of growth and low rates of unemployment are likely to be linked to the extent to which they are able to produce and use the new technologies. 4.1.2 ICT as a GPT GPTs are characterised by the potential for pervasive use in a wide range of sectors and by their technological dynamism. As a GPT evolves and advances it spreads throughout the economy, bringing about and fostering generalised productivity gains. Most GPTs play the role of enabling technologies, opening up new opportunities rather than offering complete, final solutions. This phenomenon involves what we call innovation complementarities, that is, the productivity of R&D in a downstream sector increases as a consequence of innovation in the GPT technology. These complementarities magnify the effects of innovation in the GPT, and help propagate them throughout the economy. Like other increasing returns to scale phenomena, innovation complementarities create both opportunities and problems for economic growth through technical advance. In particular it can be shown that a decentralised

40

economy will have difficulty in fully exploiting the growth opportunities of GPTs resulting in “too little, too late” innovation (Bresnahan and Trajtenberg, 1995; Helpman, 1998). The literatures on techno-economic paradigms and GPTs share the view that in different time periods different technologies play a crucial role in their ability to impact on different sectors of the economy. They also share the view that such major innovations create imbalances in the system and that it takes time to exploit their full potential (this is one of the explanations of the productivity slow-down in the neo-classical literature). However the more formal treatment of the impact of these technologies in the GPTs literature comes at the expense of a less rich and articulated view on the relationship between technological and institutional change which is crucial in the literature on techno-economic paradigms. 4.2 ICT and growth in the US and in Europe: A short review 4.2.1 An overview of the US debate on ICT and growth Growth theory and growth empirics have undergone a revival from the mid-1980s. First, the dominant paradigm, that of the neoclassical growth model and its offsprings, has been revised in order to overcome its major limitations. Second, a new brand of growth models have been generated, the so-called “Endogenous growth models”. In the end, we have now available both new (“augmented”) neo-classical models and a “New growth theory” (see Barro and Sala-i-Martin, 1995; and Aghion and Howitt, 1998 for comprehensive reviews). Coming to growth empirics, the recent wave of research has first featured cross-country studies, then panel data econometrics and more recently time series and distribution dynamics (Durlauf and Quah, 1998). From the mid-1990s, that is almost a decade after the start of the new growth debate, the socalled “New economy” and its relations with growth have moved to the center of the stage. More and more researchers have started to study the conceptual links between the introduction of information and communication technologies (ICT) and economic growth, and evaluate their quantitative effects on national accounts. A reason why that happened was a mounting attention to the “Computer productivity paradox” of why productivity growth in the US was not strong in the 1980s and early 1990s despite the spread of ICT in the economy (Triplett, 1999). The paradox was due to many factors: first, there was (and possibly is) a measurement problem involved in the definition of the ICT sector itself, and then in the economic evaluation of the components of this sector (the ICT goods). The problem is now apparently solved as the OECD provides an official definition of ICT, and the ICT goods and services are evaluated taking into account their inner quality (in the US, with the tool of hedonic pricing: see OECD, 2001, for an assessment of methodological issues). Another problem was implicit in the definition of productivity itself and in the productivity dynamics assumed by neweconomy theorists. The distinction between the production and use of ICT is central in this case. According to a simple two-sector neoclassical framework, if we are to measure the growth contribution of technical progress in the sectors producing ICT, we have to compute total factor productivity (TFP) in the ICT-producing industries, as in this case technical progress is associated with an outward shift of the production function. If, on the other hand, we are to measure the impact of ICT utilization on the productivity of whole economy, we have to calculate the variation in average labour productivity (ALP) associated with the economy-wide rise in ICT investment (see Stiroh, 2001). This distinction is fundamental, because the overall impact of ICT on per capita output crucially depends on which is the main 41

channel of productivity improvement, and on the relative weight of the ICT sector vis-a-vis the rest of the economy. It is clear that the contribution of technical progress is the smaller, the lower the relative weight of the ICT-producing sector; on the other hand, in order for the effects of the ICT investment channel to become visible, more time is required for the new capital goods to fully generate a strong and permanent effect on labour productivity. A final question, related with the previous one, is whether productivity spillovers are associated with the ICT. In this case too, we can have spillovers due to the diffusion of technical progress from the ICT-producing sectors outside (but how relevant is this likely to be?), or productivity spillovers due to the use of ICT in the rest of the economy (network externalities; technical complementarities with other innovations generating in other sectors, such as the aircraft industry). In both cases, one has to distinguish between a neoclassical, constant-returns-to-scale (CRS) paradigm, and an alternative increasing-returns-to-scale (IRS) paradigm: no spillovers are admitted in the first one, but they are allowed in the second one. In general, empirical analysis is needed to discriminate across these models. Empirical studies on the contribution of ICT to growth have flourished in the US in recent years. They have mainly be inspired by the growth-accounting methodoloy, and they have provided slightly different answers to the above questions (also because of differences in the measurement approach). In the years of the so-called “Computer productivity paradox” (broadly speaking, the 1980s and the first half of the 1990s), the contribution of ICT services to growth was sluggish. According to the estimates obtained in growth-accounting studies, the contribution of ICT capital goods to the annual growth rate was of 0.35 percentage point in the period 1973-90, and of 0.40 percentage point in 1990-95 (Jorgenson and Stiroh, 2000). This is not a negligible contribution, but is possibly less than expected if one considers that investment in ICT amounted already to 15 percent of non-residential gross fixed capital formation of the US business sector in 1980, to rise to 22.5 percent in 1990 (OECD, 2001). Moreover, Moore’s Law of the doubling of the power of computer chips every eighteen months would have pointed out to a stronger impact of ICT investment on growth. As mentioned above, the paradox faded away in the second half of the 1990s, when the contribution of ICT to aggregate growth was estimated as substantially higher. Jorgenson and Stiroh (2000) report that the contribution of computer hardware alone moved from 0.19 percentage point per year in 1990-95 to 0.46 percentage point in 1995-98; adding computer software and communication equipment, the global contribution of ICT to growth is of 0.75 percentage point in the second period. Slightly different figures are found by Oliner and Sichel (2000), who estimate a contribution of computer hardware of 0.25 percentage point per year in 1990-95, and of 0.63 in 1995-99. Both studies agree that ICT capital accumulation has played an important role in the resurgence of US productivity growth in the late 1990s. The reason why ICT investment was so late in materialising in aggregate accounting data, is that the dimension of the ICT producing sector, on the one had, and of the stock of ICT capital goods in the whole economy, on the other hand, was still small at the beginning of the 1990s. This means that, although technical progress and TFP growth were strong in the ICT sector, only after a sustained boom in ICT investment during the 1980s and 1990s there were enough information technology inputs in the whole economy to have a significant effect on economywide productivity growth. This is not to deny the key role of technical progress in this neoclassical interpretation of the US productivity puzzle: quite the opposite, indeed, as the impressive decline in the relative prices of these high-tech industries, due to innovation and TFP growth, prompted massive investment in ICT in other sectors. In turn, this ICT investment may act either as a substitute or a supplement to other kinds of capital expenditure: however, in both cases one is likely to expect a positive contribution to growth through the 42

ALP channel, either because ICT investment replaces more traditional, and less productive capital goods; or through a capital deepening mechanism of higher capital-labour ratios. Notwithstanding the solution to the computer productivity paradox (the contribution of ICT to aggregate growth only becomes visible when information technology accounts for a large enough set of capital goods in the economy, or when the ICT producing sector accounts for a large part of GDP) a number of research questions are still open in the US debate. First, measurement debates are still under way, and especially the topic of if, and how, to adjust for the evolving quality of ICT goods is currently investigated. Second, ALP growth seems to remain relatively sluggish in ICT-intensive sectors such as finance, insurance and real estate services. The answers provided for this additional puzzles are that computers are still relatively new objects and that the full potential for labour productivity growth may take some more time to be exploited, or that computers are simply not so productive in some industries (Stiroh, 2001). 4.2.2 International comparisons of ICT and growth in the OECD area The academic and business debate on ICT and growth has eventually moved from a UScentered to an international dimension. In May 1999, the Economics Department of the OECD launched an ambitious two-year research project on “Sustainable growth and the New Economy”, which over time has provided a great deal of comparative studies on the nature and dynamics of innovation- and information-based growth. The starting point of comparative studies is twofold: first, there is evidence of a gap in the relevance of the ICT sector between continental Europe, on the one hand, and the US and the rest of the industrial countries, on the other hand, at the midst of the 1990s (Daveri, 2000). In fact great concern has been expressed about the position of most European countries, which have not been able to accumulate strong competencies in the fast-growing fields. Fagerberg, Guerrieri and Verspagen (1999) have argued that the problems that Europe faces in terms of low rates of growth and high rates of unemployment are partly linked to the unsatisfactory performance of European countries in science based industries and in particular in ICT. Over the 1980s European integration appears to have favoured natural resource-based and scaleintensive industries while Europe has experienced significant losses in export shares in R&D based industries. Second, the growth performances of continental Europe and Japan have been remarkably worse than those of the US in the 1990s, although the gap in ICT diffusion has been progressively closed during the decade (Schreyer, 2000). This evidence poses a number of intriguing research questions. In the first place, Is there still a gap in ICT adoption in Europe, or is it currently being filled? Second, What was (or, is) approximately the time-delay in ICT adoption between the US and other main industrial countries? and What is the time-delay in the contribution of ICT to growth? What are the key features of the US growth phenomenon of the 1990s, and what are the “systemic” requirements that seem essential in order for a “New Economy” to be established in OECD countries? Let us review some of the answers to this questions. International comparisons of the role of ICT in industrial countries have shown that two, or even three, groups of national patterns can be identified. According to Schreyer (2000, Table 1), at the end of 1996 the share of capital stock accounted for by ICT goods was about 2 percent in Italy and Japan, 3 percent in West Germany and France, 5 percent in Canada and the UK and over 7 percent in the US. Even within Europe, Daveri (2000) identifies laggards 43

(Italy, Spain and to a lesser extent Germany and France) and fast adopters (the UK, Netherlands, Sweden, Finland). However, once we have classified industrial economies according to their ICT endowments, the question becomes: Has the wedge between leaders and slow adopters been partially closed since the mid-1990s? While some authors offer a positive answer to that (see for instance Schreyer, 2000, p.5), recent evidence casts some doubts on this issue. According to the OECD (2001, Table 2), the annual rate of growth of constant price ICT investment over 1996-99, based on harmonised prices, has been 23 percent in the US, 22 percent in France but only 15 percent in Italy and Germany. Japan performed even worse, with an annual growth rate of 11 percent. Is is true that, as the national economies of continental Europe started behind, expenditure in ICT is now close to 6 percent of GDP in Western Europe with a strong catch-up effect vis-a-vis the US (Iammarino et al., 2001), the exceptional and sustained pace of economic growth in the US during the 1990s has magnified the denominator of this ratio and thus accomodated larger expenditure for ICT and larger national product. This brings us to the second theme outlined above. The US experienced an historically unprecedent period of uninterrupted growth during the 1990s, while the pace of economic growth has been sensibly reduced in continental Europe. Of course, many factors contributed to these outcomes, including fiscal consolidation in Euroland, accomodative monetary policy in the US, structural differences in labour, product and financial markets across the Atlantic, and also the higher weight of R&D in the US economy relative to Europe. In Japan, economic conditions severely worsened during the 1990s, and stagnation virtually prevailed over the decade. To what extent, and why, is the ICT sector also responsible for these developments? According to Daveri (2000), the growth contribution of ICT was substantial in the UK and Netherlands, and rapidly increasing in over the 1990s in Finland, Ireland and Denmark. On the other side, ICT contributed less to growth in France, Germany, Belgium and Sweden, and only marginally in Spain and Italy. Hence, with the exception of Sweden, the distinction between leading and slow adopters in Europe replicates the distinction between high and low contributions of ICT to aggregate growth. Moreover, the same line can be drawn between fast and slow growers. This interpretation is therefore consistent with the neo-classical tale described by Stiroh (2001), and the lag in European adoption of ICT (say, five to seven years) is the key for understanding the modest contribution to growth. If this is true, and following the US experience, one could expect a more relevant contribution of ICT to growth in Europe in the next future, and possibly several years of rising per capita output. Schreyer (2000) offers a different position, arguing that the gap in the growth contribution of ICT between the US and other G7 countries is not completely explained by the time-delay in the adoption of ICT technologies. In fact, the modest contribution in the first half of the 1990s in Europe and Japan occurred despite significant investment in ICT. Schreyer suggest that both a methodological bias (namely, different measurement techniques) and the lack of a strong ICT-producing sector in continental Europe could be part of the solution to the puzzle. We will come back on this in the next paragraph. Recent evidence on the growth contribution of ICT across industrial countries seem to reinforce some skepticism on possible “automatic” prospects for productivity growth in Europe. In a study adopting an harmonised price index and a new method for the treatment of software (considered as capital expenditure and not as intermediate consumption), the authors come to the conclusion that, while the role of ICT is relevant and increasing in the US, Canada, Australia and Finland, there is no evidence of an increasing contribution of ICT to growth in other industrial economies over 1995-99 (OECD, 2001). In order to better understand the main features of the US model of productivity growth, and the possible requirements for extending it to Europe, let us move onto the next section.

44

4.3 Open Issues and Directions for Future Research The main macroeconomic features of the US growth story in the 1990s can be summarised as follows. In the labour market, we observed a rise in employment rates and a fall in unemployment, while at the same time there was a sustained increase in labour productivity. These developments seemed to allow for a structural change in the behaviour of the labour market, shifting down the “natural” rate of unemployment or the NAIRU (non-accelerating inflation rate of unemployment), and thus leading to a low dynamics of aggregate prices in the presence of prolonged growth. Exceptional investment rates in historical perspective, and very low personal saving rates possibly due to the anticipation of future income growth (also reflected in high stock market values), explain the widening US current account deficit; at the same time, the strong value of the dollar is mainly interpreted as a side-effect of the US growth prospects. US productivity growth in the second half of the 1990s was strong both in ICT-producing and ICT-using sectors. However, the data show a clear distinction between the role of technical progress in the ICT-producing sectors, and the role of labour productivity growth in the ICTusing sectors (Jorgensen and Stiroh, 2000; Oliner and Sichel, 2000). Moreover, much of the acceleration in aggregate US productivity growth after 1995 can be traced to accelerations in the pace of technical progress in ICT-producing sectors - measured as faster relative price declines in these high-tech industries. No strong evidence of spillover effects has been found, hence the neoclassical constant-returns-to-scale paradigm looks appropriate as a heuristic framework to analyse the New Economy (Stiroh, 2001). Finally, software capital accumulation seems to have played a relevant role in US productivity dynamics in the second half of the 1990s (OECD, 2001). To what extent the US success story is (or was?) accounted for by the existence of a strong ICT-producing sector, which is lacking in several countries in Europe? In other words, is it there a question of comparative advantage in high-tech industries that is consistent with a windfall of technical progress in the US but not in continental Europe? Another issue concerns the role of structural factors in the recent comparative performance of the OECD economies, and in particular the impact of national factor markets in the different outcomes on the two sides of the Atlantic (and the Pacific as well). For instance, if we agree that the financing of innovation is clearly one of the key issues, as it deals with the Schumpeterian perspective on the role of financial markets that is more appropriate in a dynamic approach, then the contribution of capital markets to the development of the New Economy is crucial. We do not aim to yield exhaustive and final answers to such fundamental questions, but to further investigate this topic with particular attention to the performance of European countries. As a matter of fact, the central growth debate at the turn of the century is: Why and how the performance of the US economy has so radically changed in the last decade, bringing about a renewed American leadership in terms of productivity and income growth, and sweeping away many arguments in favour of an “automatic” convergence of the other industrial economies towards the US? How can Europe (and Japan) establish the right economic environment such that technology diffusion and market functioning allow for new a catching-up process towards the US? As far as the issue of comparative advantage is concerned, Roeger (2001) calibrates a twosector-two-skill growth model of the US and European economies, featuring both an ICTproducing and an ICT-using sector, and skilled and unskilled labour. The purpose of the calibrations is to evaluate whether structural differences between the US and Europe can 45

explain the growth rate gap of the 1990s: it is widely believed that factor markets are more flexible in the US, and this could account for the better macroeconomic performance. Roeger (2001) provides indeed some evidence showing that capital adjustment costs were probably lower in the US, as the variability of the cyclically adjusted investment to GDP ratio has been clearly higher in the US than in Europe. As far as the labour market is concerned, his evidence questions the idea that higher European adjustment costs may be mainly due to hiring and firing regulations, while more support is found to the hypothesis of wage inflexibility across skill groups in Europe. However, the main point is that even if one includes differential adjustment costs in Europe and the US in the model, the productivity growth gap is not explained by the features of factor markets. The main point of Roeger is that higher rates of TFP growth or technical progress, and not lower adjustment costs, are at the core of the productivity growth leap in the US; in turn, TFP growth is associated with the comparative advantage the US have in the production of high-tech goods, hence comparative (dis)advantage and not Eurosclerosis in general must be blamed for the inferior growth performance in Europe. If Roeger (2001) is right, namely there is no European comparative advantage in ICT production, then it becomes crucial for Europe to perform very well in ICT utilisation. Only a satisfactory performance in ICT utilisation can contribute to close the growth gap between Italy, and Europe in general, and the US. In this sense, it becomes relevant to investigate a number of issues: first, what was the record of aggregate ICT investment in the Nineties, and has the endowment gap with the US been filled so that we can now expect that the growth gap will be filled as well? Second, has ICT investment been spread economy-wide, or has it stayed confined in a few productive branches? Third, what are the main features associated with sectoral patterns of ICT investment, and is ICT intensity higher in industries where the potential for productivity growth seems larger? As we have already argued ICT have also played an important role in the process of internationalisation (globalisation) of technology. At the same time, the internationalisation of R&D has an impact on countries’ technological potential and competitiveness and thus affects also the way in which ICT technologies are developed and diffused in the economy. In what follows we focus the attention on the theoretical and empirical literature on the internationalisation of technology. 5. The internationalisation of technology Multinational companies (MNEs) play a dominant role in the innovation activities of their home country and control a vast proportion of world’s stock of advanced technologies. Their decisions in term of mode, location and exploitation of their R&D results greatly influence the home country’s technological potential and competitiveness (Patel and Pavitt, 1999). The growing significance of the internationalisation of R&D activities of MNEs over the past two decades has therefore been cause of some concerns among innovation policy makers. In Europe it has been advanced that the performance of R&D activities in foreign countries might result in a “hollowing out” of domestic capabilities, whenever national firms locate innovation activities in fast growing fields abroad. This is regarded as indicative of a weakening of the national innovation system and an erosion of the technological competitiveness (ETAN, 1998). In the United States the internationalisation of industrial R&D has brought with it worries about a possible impoverishment of the national technology base due to the increasing local R&D activities of foreign MNEs.

46

To be able to evaluate the potential impact on the home countries it is crucial to determine the extent and the nature of the R&D activity that has been relocated abroad. To this end the existing theoretical and empirical literature on the growing involvement of MNEs in foreign R&D will be reviewed. 5.1 Theoretical framework Although the global generation of technology by European MNEs dates back to the interwar period (Cantwell, 1995), little research on the topic was been carried out until the 1980s. We must attribute the lack of interest in the location of R&D activity by MNEs to the widespread belief that MNEs were concentrating innovation activity in their home country. Empirical evidence to support this assumption was provided using data on US MNEs. American companies were performing most of their R&D effort at home where high demand and advanced technological resources and capabilities were providing a constant stimulus to their innovation activities. European firms, especially from small countries, have instead shown historically a higher tendency to conduct both production and innovation activities abroad because of the lack of similar supply and demand conditions in their domestic market. Early studies of the process of internationalisation of R&D used as an analytical framework the product-cycle model of Vernon (1966, 1977). The original product-cycle model (1966) is based on the assumption that innovation is a demand-led process: it arises from a market stimulus. But according to this hypothesis, firms tend to be stimulated by the needs stemming from the nearest market, the home market. The home market plays a dual role in this model: it is considered the source of stimulus of innovation and at the same time it is the preferred location for performing R&D activities. Innovating firms will therefore concentrate their R&D effort at home where they can benefit from both the availability of scientists and engineers with the required skills, and the proximity to and, interaction with, potential customers. Economies of scale in R&D activity and agglomeration effects, as well as the need for the coordination and control of expensive and risky investments are also reasons for keeping R&D and the initial stage of production in a common location (Vernon, 1977). Therefore in this model internationalisation of production was limited to mature and standardised products and minor adaptive and development work would be the only R&D to accompany the foreign production. Nevertheless, this model, with its demand-led interpretation of the innovation process and the important role it gives to proximity to local costumers, has contributed significantly to our understanding of why overseas R&D is undertaken. As firms increasingly locate production closer to their customers and suppliers they need R&D laboratories to adapt the technologies and product developed at home to local conditions. The creation of such technical support laboratories, as defined by Hood and Young (1982), are then supposed to accompany the later stages of the production process abroad. Indeed they seem to follow a linear progression based on the age, growth and relative size of the international production of the MNE (Lall 1979). In this framework the technological advantages of the affiliates primarily reflect those of the home country (where the core of innovation activities continues to be concentrated) and foreign R&D units tend to enhance the existing parent-company technologies. This type of R&D site has been termed “home-base exploiting” (HBE) (Kuemmerle, 1997) or “assetexploiting” (Dunning and Narula, 1995). In general, the more embedded the foreign subsidiary, and the greater the intensity of the value-adding activity, the greater the amount of R&D activity. This leads to a duplication of its home base activities, since the host location is acting as a substitute for activities it may 47

have wished, ceteris paribus, to have undertaken at home (Zander, 1999), but now finds that it can undertake more efficiently elsewhere. On the other hand, evidence clearly suggests that this is intermediated by industry level effects (e.g., Lall, 1979; Patel, 1997). This type of foreign R&D operations has usually supported the expansion of international production, not in high-tech product groups, but rather in sectors where adapting the product to serve the local market is important. Over the last decade some of the factors encouraging centralisation of R&D activities have become less influential, i.e. the cost of technology transfer and the shortage of human capital and scientific infrastructures in the host countries. As multinationals have extended and diversified their global operations, they have set up global R&D networks. This strategy is based not only on the wish to rationalise R&D expenditures and to avoid duplication of R&D activities, but also indeed to absorb and acquire technological spillovers, either from the local knowledge base (be they agglomeration effects or from public infrastructure), or from specific firms. These R&D activities have been defined as “home-base augmenting” (HBA) (Kuemmerle, 1997) or “asset-seeking” R&D activity (Dunning and Narula, 1995). In such kinds of investments, firms aim either to improve their existing assets, or to acquire (and internalise) or create completely new technological assets by locating R&D facilities abroad. The assumption in such cases is that this provides access to location-specific advantages that are not as easily available in the home base and that might be associated with the presence of a lead market (Meyer-Kramer and Reger, 1999). Location decisions for this type of R&D facility are based not only on the technological infrastructure of the host country, but also on the presence of other firms and institutions, which may create externalities that investing firms could absorb. Another distinguishing feature of this type of R&D site is that it can be completely independent from the production process of the firm: it could set up a foreign R&D facility in a country even if it has no intention of producing there. There are several reasons why such HBA R&D activities would be hard to achieve from the home base. As suggested by Von Hippel (1994), when the knowledge relevant for innovative activities is located in a certain geographical area and it is very “sticky” 20 , the R&D activity should take place at that site, according to the principle of cost minimization. Foreign affiliates engaged in HBA activities are attracted to these technological clusters in order to benefit from the external economies and knowledge spillovers generated by the concentration of production and innovation activities. Among the reasons for such sticky knowledge, the argument for the tacit nature of knowledge often stands out. The tacit nature of technology implies that even where knowledge is available through markets, it still needs to be modified to be efficiently integrated within the acquiring firm’s portfolio of technologies. In addition, the tacit nature of knowledge associated with production and innovation activity in these sectors implies that “physical” or geographical proximity is important for transmitting it (Blanc and Sierra, 1999). While the marginal cost of transmitting codified knowledge across geographic space does not depend on distance, the marginal cost of transmitting tacit knowledge increases with distance. This leads to the clustering of innovation activities, in particular at the early stage of an industry life cycle, where tacit knowledge plays a particularly important role (Audretsch and Feldman, 1996). In other words, spillovers are stronger within a small geographical unit (see, e.g., Jaffe et al., 1993, Jaffe and Trajtenberg, 1996, Sjöholm, 1996, Maurseth and Verspagen, 2001).

20

Von Hippel (1994) defines stickiness as the incremental expenditure required to transfer that unit of knowledge. 48

However, merely establishing R&D activities abroad for the purposes of tapping into pools of scientific knowledge does not necessarily mean that firms will be successful in doing so. There are a wide variety of factors that determine the MNE’s efficiency in internalising spillovers. First, it is difficult to manage cross-border R&D activities. A dispersion of R&D activities across the globe requires extensive coordination between them – and particularly with the headquarters - if they are to function in an efficient manner with regards to the collection and dissemination of information. This acts as a centripetal force on R&D, and accounts for the tendency of firms to locate R&D (or at least the most strategically significant elements) closer to headquarters. Although these problems have been mitigated by developments in ICT which have facilitated the management and coordination of international research networks, geographical distance still represents a barrier for the transfer of knowledge. Such growing complex linkages, both within networks internal to the firm, and between external networks and internal networks, require complex coordination if they are to provide optimal benefits (see Zanfei, 2000, for a discussion). Such networks are not only difficult to manage, but also require considerable resources (both managerial and financial). It is no surprise, therefore, that external technology development is primarily the domain of larger firms with greater resources, and more experience in trans-national activity (Hagedoorn and Schakenraad, 1994). When firms engage in R&D in a foreign location to avail themselves of complementary assets that are location specific, they are aiming to explicitly internalise several aspects of the system of innovation of the host location. However, developing and maintaining strong linkages with external networks of local counterparts is expensive and time consuming, and is tempered by a high level of integration with the innovation system in the home location. Such linkages are both formal and informal, and will probably have taken years – if not decades – to create and sustain. Frequently, the most significant issues are ‘know-who’. Government funding institutions, suppliers, university professors, private research teams and informal networks of like-minded researchers take considerable effort to create, and once developed, have a low marginal cost of maintaining. Even where the host location is potentially superior to the home location - and where previous experience exists in terms of other value adding activities - the high costs of becoming familiar with, and integrating into a new location may be prohibitive. Keep in mind that firms are constrained by resource limitations, and that some minimum threshold size of R&D activities exist in every distinct location. As such, to maintain more than one facility with a threshold level of researchers must mean that the new (host) location must offer significantly superior spillover opportunities, or provide access to complementary resources that are simply not available anywhere else, and which cannot be acquired by less risky means more efficiently. This may also explain in part the growing use of M&A activity with regards to R&D. Buying an existing laboratory short-circuits the time taken to develop such linkages, as they come with ready-made networks and linkages (Narula, 1999). However, acquired laboratories are difficult to integrate them efficiently into the internal network of the firm, which may limit their efficiency, at least initially (Belderbos, 1999). The high costs associated with integrating into the host location’s systems of innovation – in contrast to the low marginal cost of maintaining its embedness in its home location’s innovation system – and the need for internal cohesion within the MNEs (Blanc and Sierra 1999, and Zanfei, 2000) explain why firms are reluctant to expand internationally. However, these costs must be tempered by supply-side considerations: the development of these technologies benefits from diversity and heterogeneity in the knowledge base, which might come from competitors, from interaction with customers and from other complementary 49

technologies. A single national innovation system is often unable to offer the required knowledge and skills in all the emergent areas with equal effectiveness (Patel and Pavitt, 2000). To overcome the size of their national public science system as well as the lack of advanced users, MNEs based in small countries, such as Switzerland and the Netherlands, have become highly internationalised early on, including in their R&D operations. More recently, MNEs in high-tech sectors have adopted a similar strategy to make up for the limitations of their national science base. The need to have multiple technological competencies is particularly compelling for firms operating in technology-intensive sectors (Granstand et al., 1997). Unlike mature technologies, where knowledge is codified and evolves slowly, new technologies are characterised by a high degree of tacit and uncodifiable knowledge, which does not follow a predetermined path. The reallocation of R&D activities in centres of excellence allows the MNEs to exploit other national technological specialisation. Thus, summarising, HBE activities are primarily associated with demand-based activities, with the internalisation of technological spillovers as a secondary issue. HBA activities, on the other hand are primarily undertaken with the intention to acquire and internalise technological spillovers that are host location-specific. HBE activity, broadly speaking, represents an extension of R&D work undertaken at home, while HBA activity represents a diversification into new scientific areas. While often reported to be a much smaller phenomenon in terms of international R&D expenditure (Patel and Vega, 1999; Gerybadze and Reger, 1999; Niosi, 1999), the number of HBA sites seems to be increasing quite fast, particularly in technology-intensive sectors, such as biotechnology, computers and telecommunications (Kuemmerle, 1999; Dalton and Serapio, 1999; and Patel and Vega, 1999). Despite the clear-cut distinction at the conceptual level between HBA and HBE activities, in practice it is very difficult to determine the specific role of a foreign R&D unit. An R&D facility might be involved in adopting innovations developed by the parent company and at the same time try to gain access to local technology sources (Zander, 1999). These two activities are complementary and they might also be consequential as shown by Kuemmerle (1999), who found that HBE sites usually get established before HBA R&D investment. 5.2 Recent empirical studies A number of recent studies have analysed the internationalisation of R&D activities by multinational companies using different methodologies and databases. The empirical analysis carried out to assess this phenomenon can be allocated in three groups. The first group focuses on the relative volume, scope and geographical pattern of foreign R&D using patent data (Patel and Vega, 1999; Cantwell and Janne, 1999; 2000) and R&D expenditure data by foreign affiliates (Dalton and Serapio, 1999). The second group of studies includes those based on surveys and case-studies whose main research question is to investigate the firms’ motivations in carrying out R&D activities abroad and the specific activities that are performed in foreign research facilities (Florida, 1997, Kuemmerle, 1999). The third group is a more recent stream of work which has examined the knowledge sources of foreign R&D units using patent citations analysis (Almeida, 1996 and Frost, 2001). Each type of methodology provides different insights into the process of internationalisation of R&D, although they all suffer from some shortcomings. Survey-based studies are able to offer a rich and detailed picture of a small number of cases, but they are mainly qualitative in nature and they can only provide a description of sector or location in a determined point in 50

time. Patent analyses, apart from being quantitative studies, are able to examine the internationalisation of R&D over a long time-span covering a larger sample of firms and sectors. The main disadvantages of using patent statistics are that they are not a satisfactory measure of the innovative activities in biotechnology and software and they are not able to account for the accumulation of un-codified knowledge in foreign locations. This is the case when foreign subsidiaries recruit talented scientists and technicians or when they acquire a host country laboratory. Patent citations analysis uses the citations recorded in patent documents as an indicator of knowledge spillovers. When an inventor cites another patent, this indicates that the knowledge contained in the cited patent has been useful in the development of the citing patent. This methodology can provide more insights in the process of technology sourcing of foreign R&D sites. Patent citations analysis presents the same disadvantages as patent data. In addition one should keep in mind that though suggested by the inventor, the final decision on which patents to cite in an application (which affects the claim to novelty of the invention, and thus the scope of the protection provided by the patent) lies ultimately with the patent examiners. They may therefore not always reflect the way the innovation process has evolved. 5.2.1 Studies based on R&D and patent data The study by Dalton and Serapio (1999) examines the magnitude, scope, sectoral distribution and country of origin of foreign R&D investment in the US as well the trend and destination of foreign direct investment in R&D by US MNEs. R&D expenditure performed by foreign affiliates in the United States tripled from 1987 to 1997, reaching a value of $19.7 billion, which corresponds to almost 15% of total company-funded R&D investment. 21 What emerges from this study it that the internationalisation of R&D is a phenomenon which occurs primarily between a small number of highly industrialised country and it affects mainly technology intensive sectors. Most of the R&D financed by foreign subsidiaries comes from firms whose parent is located in one of three countries: Germany, Japan, and United Kingdom. Foreign companies have concentrated their R&D efforts in the US in three industries (drugs and biotechnology, industrial chemicals, and electronic equipment) and in certain areas (such as Silicon Valley and greater Los Angeles, Detroit, and Princeton). Foreign companies have invested in the US for a variety of reasons which vary across industries. According to the results of the survey carried out by Dalton and Serapio, the most important function of foreign-owned laboratories in the biotechnology industry is “to take advantage of a more favourable environment for research”, “to cooperate with other US R&D laboratories” and “to engage in basic research”. These motives classify these R&D facilities as HBA sites. Major factors behind the R&D decisions of foreign firms in the automotive industry have been to meet American customers needs, to assist the parent company in satisfying US environmental regulations (a HBE strategy). In the electronic industry R&D investment motives are more mixed, both supply and demand considerations are considered important. However in the research activity related to high-tech products such as computers and semiconductors, technology acquisition emerges as the dominant motive. US MNEs’ investment in foreign R&D facilities has increased from $5.2 billion in 1987 to $14.1 billion in 1997 and overseas R&D funding accounts for 11% of total R&D performed in the US Most R&D expenditure is concentrated in five countries (Germany, United Kingdom, Canada, France and Japan) and in four industries (drugs, automotive, computers, and electronic components). Technology-oriented factors seem to explain the decision of US firms to locate 21

Part of the increase in foreign-funded R&D investment can be attributed to the wave of M&A in American pharmaceuticals and biotechnology, which has a significant R&D budget. 51

R&D abroad in the computers, hardware, software, industrial chemicals, and pharmaceuticals industries. Patenting activities of the largest American, Japanese, and European MNEs reported in the study by Cantwell and Janne (2000) show a similar pattern to the R&D data. In particular it emerges that European firms, mainly from relatively small countries, are the most internationalised in their R&D operations, while Americans and especially Japanese firms have retained a very centralized research structure. Patent statistics have been used by Patel and Vega (1999) to explore the relationship between the technological advantage of the multinational company at home and the technological strength of the host country. Using a sample of 220 firms with the highest volume of patenting outside the home country between 1990-96 the authors found that although foreign research efforts are concentrated in “high-technology fields” (such as pharmaceuticals, biotechnology, computers, telecommunications), they are mainly involved with developments of process and machinery technologies, i.e. they carry out HBE type of activities. Patent data suggest that firms tend to locate R&D abroad in their “core” fields, in the areas where they have a technological advantage (US MNEs in computers, German, UK and Swiss firms in organic chemicals and pharmaceuticals, and Japanese firms in computers, image and sound). For electronics and metals companies in almost half of the cases the technological advantage of the firm does not match the location’s technological strength, i.e. firms seems to go abroad to exploit technological knowledge accumulated at home. Chemicals, pharmaceuticals, mining, food and materials companies instead tend to locate their foreign R&D activities where they can find complementary technological assets. The study by Cantwell and Janne (1999) addresses a similar research question focusing on the international research strategy of European MNEs in European locations. In particular they found that European MNEs coming from the leading centre in their industries tend to adopt a more diversified spectrum of technological activities abroad so as to acquire complementary assets. Firms with headquarters in lower order centres in the same industry, when operating in foreign centres of excellence, seem instead to exploit their technological assets replicating their home country’s technological specialisation. 5.2.2 Studies based on surveys Early works based on surveys are the ones by Håkanson (1992) and Pearce and Singh (1992). Håkanson examines the impact of different host country characteristics in the location decisions of foreign R&D units. The location pattern of 20 chemical and engineering Swedish MNEs seems to show that ‘demand related’ factors are more important than the ‘supply related’ factors and that ‘political’ factors (such as trade barriers, the possibility of participating in government sponsored research programs) are also playing a role in determining the geographical location of foreign R&D operations. The results of the study by Pearce and Singh (1992) based on a comprehensive sample of MNEs operating in 30 industries seems to confirm that most overseas R&D is carrying out asset exploiting type of activities. However, more recent surveys seem to find substantial support for the increasing importance of ‘supply-side’ factors as motives for international decentralisation of R&D. The study by Florida (1997) surveys a sample of 207 R&D facilities in the US in four technology sectors (electronics, automotive, chemicals and materials, and biotechnology) with regard to the relative importance of their technology-oriented activities (HBA) and market orientedactivities (HBE). The findings of this study suggest that both types of activities are playing an important role in the overall activities of the sample laboratories. However technology52

oriented activities are relatively more significant, especially in R&D units operating in the biotechnology and pharmaceutical sectors, while R&D sites in the chemical and automotive sectors seem to concentrate on tasks related to the support of manufacturing activities and the adaptation of products to the local market conditions. The innovating performance of the laboratories in the sample confirms that these sites are not mere “listening posts” but are dedicated to the creation of new scientific and technological knowledge. The survey indicates also that one of most implemented strategies for gaining access to localised knowledge is the recruitment of high-quality scientists. Some of Florida’s results have been confirmed by a more recent survey by Kuemmerle (1999). This study analyses the activities of 238 foreign R&D facilities from 32 American, Japanese and European pharmaceuticals and electronics companies in different host countries over time and investigates the motives, location characteristics, and mode of entry for R&D facilities abroad. What emerges from this study is that technology sources have increasingly become a motivation for setting-up foreign R&D laboratories. Kuemmerle found that 38% of laboratories in the sample could be classified as HBA. The location of foreign R&D sites seems to match the distribution of knowledge sources they build upon. When the purpose of R&D is to try and gain access to localised knowledge, firms will establish centres in proximity to universities or national laboratories. When instead they are supporting manufacturing and marketing activities R&D sites they will be located near a lead market or in a cluster of competitors. Although in principle acquisition of a foreign laboratory could be a valid instrument for getting access to localised knowledge, Kuemmerle found that green field investment are the dominant form of entry both for the case of HBA and HBE sites. 5.2.3 Studies based on patent citations analysis Almeida’s (1996) study of the semiconductor industry tests the hypothesis of technology sourcing activities of foreign affiliates in the US using patent citations analysis. This methodology allows the author not only to assess whether or to what extent foreign subsidiaries draw from local sources of knowledge, but also to what degree they contribute to local knowledge. Following Jaffe et al. (1993) the author is able to identify the regional location of the innovation activity using the address of the inventor and to control for the geographical clustering of patenting activity by matching each citation listed in an affiliate’s patent with a control citation. The control citation is a patent with a similar technological and temporal profile to the cited patent but which itself is not cited by the subsidiary patent. It represents therefore the benchmark against which the localisation of the subsidiary citations is measured. The results seem to support the hypothesis that foreign subsidiaries build upon localised sources of knowledge: the patents cited by foreign affiliates are more likely to be originated in the US or in the same US States where affiliates operate. Using a similar procedure, Almeida finds that foreign affiliates contribute to the regional knowledge base: the patents granted to these foreign firms are cited more than expected by other patents originated in the same region. Frost (2001) builds upon and extends the work by Almeida investigating the geographic sources of foreign subsidiaries’ innovation process of across a much broader sample of US foreign subsidiaries in different industries and examining under what circumstances they draw from local sources of knowledge during their process of technological innovation. The results of this study show that both the characteristics of the subsidiary, such as the amount and the type of innovation activity carried out, and the technological specialisation of the home and host country are important in determining the geographic sources of innovation. Less innovative affiliates are more likely to build on the knowledge base of the parent company, 53

while more innovative subsidiaries, being more embedded in the local context, tend to draw upon local sources of knowledge. Frost found that foreign affiliates, that are devoting much of their R&D efforts in adapting technologies developed in the home country, are less likely to use technical ideas originating in the host country. 22 Similarly when the foreign affiliates are performing R&D activities on technical fields, in which the home country has a technological advantage and the host country presents a technological disadvantage, they seem more likely to cite home country’s patents. The opposite case arises when foreign subsidiaries are active in technological fields in which the host country has a greater technological advantaged with respect to the home country. 5.3 Open issues and directions for future research We can therefore briefly summarise the main trends that emerge from the studies that we surveyed: 1) the degree of R&D internationalisation has increased substantially in the last decade but it remains concentrated in a small number of industrialised countries; 2) MNEs in the biotechnology, pharmaceutical and electronic industries carry out R&D investment abroad to acquire and develop new technology, and in so doing they are learning from local sources of knowledge; 3) the home country technological advantages and the host country technological characteristics both play an important role in determining the type of R&D activity performed abroad. So far the existing empirical literature has analysed the learning behaviour and the technological sourcing of foreign subsidiaries and little interest has been devoted in analysing the impact that such innovation activities have on the home part of the multinational company and on other firms in the home country. From a policy perspective the existence of a knowledge transfer from the subsidiary to the home plant and possibly on to other firms in the home country is the most relevant issue. It might in fact be the case that the performance of R&D abroad - when asset-seeking in nature - might represent an opportunity for enhancing the technological expertise of other national firms. Assessing the existence and the scale of such phenomenon is extremely important for science and technology policy recommendations. Policy makers have a tendency to encourage domestic multinationals to maintain their R&D activity at home and to discourage its re-allocation to foreign countries, however it might be possible that the internationalisation of corporate R&D may lead to increases in the home country’s knowledge base and its competitive performance in general. One direction for further research is to examine this issue using patent data and in particular patent citation analysis to trace knowledge flows among subsidiaries, headquarters and other firms in the home country. The dataset compiled at SPRU, from information supplied by the US Patent Office, on the name of the company, the technical class, and country of origin of the inventor, (for each patent granted in the US from 1969 to 2000) is the ideal data source to carry out this type of analysis. 6. A review of the methodology In this section we review the methodology that we intend to adopt in the quantitative part of the SETI project. In particular we have the following requirements: a) to perform a consistent dynamic analysis whether or not the case under consideration is of steady-state from both a theoretical and empirical point of view and b) to maintain the structure studied at the 22

This results is in line with the definition of HBE sites. 54

theoretical level in the empirical analysis. Although other questions, basically concerning the dimension of the sample, might arise, either points are central to the present approach. The methodology employed is that of continuous time and poses its basis on the direct estimation of a simultaneous system of differential equations. Each equation is composed on the left hand side (l.h.s.) by a dependent variable, which is the time derivative of the variable of interest and, on the right hand side (r.h.s.), by other variables defining the difference between the actual value of the variable of interest and its desired value according to the behavioural rules of the agents involved in the model. This difference is then multiplied by a coefficient to evaluate the speed of adjustment of the convergence process. This formulation allows for every sort of theoretic formulation of general equilibrium whether linear or not, and with or without diffusion. Moreover the parameters can be readily estimated as they enter the objective function, and the estimation is independent from the simplifying hypothesis of steady-state. The techniques to perform the estimation are the same of that in use for system23 but, in general, full information methods (FIML or 3SLS) are preferred. Another advantage with respect to the ordinary methods is the sensitivity analysis that consists in the evaluation of the change in the dynamics of the system to variations in the underlying eigenvalues (Gandolfo, 1992; Wymer, 1997). This is possible because the method estimates the model as indicated by the theory so that changes in the eigenvalues are associated to changes in the parameters of interest. For this reason also simulations under different scenarios provides the analysis with useful insights. Moreover the choice of the explicit introduction of the adjustment speeds makes it possible to understand whether and how long it takes to reach the steady-state. Before going on it is important to stress that the term continuous is not referred to a frequency concept but to the fact that the reality under consideration is continuos. For this reason, contrarily to what it might have been thought, macro and international economics is the most appealing field for this approach. While, for further details, the reader is referred to the manual of Gandolfo (1981), in the remainder of this section there will be presented a model of continuous time of Padoan (1998) that evaluates the role of technology diffusion by means of international trade. The following model is addressed specifically to study the interaction between technology and trade and, even though it also considers the stock of R&D expenditure, cannot be compared to the models studied in the previous sections as in this case such a variable is exogenous. Moreover another aim of this model is also to investigate on sectoral effects according to the Pavitt (1984) taxonomy. Given these objectives the model adopts the simplifying assumption of not stating a functional form for output which is instead defined on the basis of the equilibrium condition for the balance of payments 24 . For this reason it needs some modifications to account more directly for growth issues. Apart from these adaptations, given its specificity, it constitutes a solid reference to how this methodology works in this field. The model will be first expounded in tabular form to be more easily commented below: Export Share. Traditional goods.  D log S a = α2 log S a∗ − log S a (6.1)  ∗  log S a = γ1 − β1 log P

(

)

23

Continuous time estimations, simulations and sensitivity analysis may be carried out by means of Wymer (1994) programs. 24 Specifically it refers to current account equilibrium condition: PX X = PM M, S M =M/Y, S X = X/W and then WPX S X = YPM S M . 55

Export Share. Scale intensive goods.  D log Sb = α3 log S b∗ − log S b  (6.2)  T ∗ log S b = γ 2 − β2 log P + β3 log T  w

(

)

Export Share. Specialised suppliers.  D log S c = α4 log Sc∗ − log S c  (6.3)  T ∗ log S c = γ 3 − β4 log P + β5 log T  w

(

)

Export Share. Science based goods.  D log S d = α5 log S ∗d − log S d  (6.4)  T ∗ log S d = γ 4 + β6 log T  w

(

)

High tech export share.  S  W  W  W  (6.5) S xh =  b  b  + S c  c  + S d  d   S x  W h   Wh   Wh  Aggregate Export Share.  S  W   S  W  D log S x = D log S a  a  a  + λa  a  a  +  S x  W   S x  W 

(6.6)

 S  W   S  W  + D log S b  b  b  + λb  b  b  +  S x  W   S x  W   S  W   S  W  + D log S c  c  c  + λc  c  c  +  S x  W   S x  W  S + D log S d  d  Sx

 Wd  S   + λd  d  W   Sx

Import share. Traditional goods.  D log S m a = α7 log S m∗ a − log S m a (6.7)  ∗  log S m a = γ 6 + β11 log P

(

 Wd    − λw  W 

)

Import share. High tech goods. DlogSm h=α8 logSm∗ h−logSm h  (6.8)  T ∗ log S = γ + β log P + β log +β14logSm m h 7 12 13  Tw 

(

)

Aggregate import share. S  S  (6.9) D log S m = D log S m a m a  + D log S m h m h   Sm   Sm 

56

Knowledge Accumulation. DlogT =α6 logT∗ − logT + β9 logTw + â10 logSmh (6.10)  ∗ logT =γ5 − β7 logF + β8 logSd

(

)

Output. DlogY =α1 logY∗ −logY (6.11)  ∗  logY = logW +logPx +logSx- logPm-logSm

(

)

VARIABLES Endogenous Sa : Export Share. Traditional goods Sb : Export Share. Scale intensive goods Sc: Export Share. Specialised suppliers Sd : Export Share. Science based goods Sxh: Export share. High tech export share Sx: Export share. Aggregate Export Share Sml: Import share. Traditional goods Smh: Import share. High tech goods Sm : Import share. Aggregate import share T: Stock of domestic knowledge Y: Output

Exogenous P: Relative price F: Stock of foreign R&D expenditure Tw: Stock of foreign knowledge W: Total foreign demand Wi: Sectoral foreign demand (i = A, B, C, D, H) Price of exports Pm : Price of imports D: d/dt

Where error terms are omitted for simplicity’s sake and there are three identities 25 (6.5), (6.6), (6.9) and eight behavioural equations. The variables pertaining to the latter are all expressed in naperian logs either to exploit the usual nice proprieties for coefficients and variables themselves in order to assign relevance to the rates of growth or to obtain a linearised system with a remainder not involving time. This, in fact, makes the conditions of the PoincarèLiapunov-Perron theorem automatically satisfied so that the linear approximation is uniformly good and the local stability analysis can be conducted correctly 26 . The system is very often non-linear either in variables or in the coefficients. Here the non-linearity pertains to the variables and is provoked by the identities as one can easily check. The system can be decomposed in three components: Trade, knowledge accumulation and output. The leading role is plaid by trade that refers to four sector according to Pavitt’s taxonomy. Each curly bracket refer to an adjustment equation where the starred variable represents the target according to a separate behavioural equation. Without considering the details of each equation as they involve problems out of the scope of this section, it is worth noticing that knowledge affects endogenously each trade sector but the one of traditional goods and conversely is affected by both high tech imports and exports. However this is the only endogenous link with the stock of knowledge. In particular there is not an endogenous domestic production process of knowledge. This aspect prevents the use of the model from growth analysis because it lacks of a direct link between output and knowledge. An evidence on this point has been reached in Maggi, Federici and Espa (2001) where the model has been 25

Identity (6.6) decomposes Dlog(S x W) and is obtained from Sx = (S a Wa + S b Wb + S c Wc + S d Wd )/ W and DlogWi 26 For further details the reader is referred to Gandolfo (1997) and Bellman (1953). 57

λi =

used for simulation purposes and the level of the stock of knowledge for the five involved countries (Germany, France, Italy, UK, Japan) penalised Japan with respect to the other countries. The parade was correctly restored with new simulations after the introduction of a new equation in line with Lucas (1988) concept of knowledge that outlives individuals. The new variable has a desired level that depends on income and the stock of knowledge with a feed-back effect on the latter. In this way it was possible to introduce an endogenous domestic effect on knowledge accumulation without altering the pre-existing nature of the model. However, notwithstanding this result a new estimation is necessary to better evaluate the effects of knowledge on output given that its desired level is, ultimately dealt with as a residual deriving from the current account equilibrium condition. Another important aspect is that this model has been estimated for any country while it might be interesting to study diffusion effects between countries with a single general model in order to exploit at most the sample size. A solution may be in principle the introduction of partial –instead of the simple- differential equations where the other independent variable that causes variations is the space. In this way it is also possible to use panel data for a more consistent estimate. This brief exposition is clearly not at all exhaustive of the issues pertinent to the continuous time methods but aims at give an intuition of the potential of this approach that is appositely shaped for the study of economic dynamic systems.

58

Bibliography Aghion, P.and P. Howitt (1998), Endogenous Growth Theory, MIT Press, Cambridge MA. Almeida, P. (1996), “Knowledge sourcing by foreign multinationals: patent citation analysis in the US semiconductor industry”, Strategic Management Journal, Vol. 17, Winter Special Issue, pp. 155-65. Amable, B. (1996), “The effects of foreign trade specialisation on growth”, paper prepared as part of the EU TSER project “Technology, Employment and European Cohesion”. Amable, B. and B. Verspagen (1995), “The role of technology in market shares dynamics”, Applied Economics Vol. 27, pp. 197-204. Amendola, G., G. Dosi and E. Papagni (1993), “The dynamics of international competitiveness”, Weltwirtschaftliches Archiv, Vol. 129, pp. 451-471. Amendola, G., P. Guerrieri and P. Padoan (1992), “International patterns of technological accumulation and trade”, Journal of International and Comparative Economics, Vol. 1, pp. 173-197. Amendola, G., P. Guerrieri and P. Padoan (1998), “International patterns of technological accumulation and trade”, in D. Archibugi and Michie, J. (eds), Trade, Growth and Technological Change, 141-167. Archibugi, D. and J. Michie (1995), “The globalisation of technology: a new taxonomy ”, Cambridge Journal of Economics, Vol. 19, pp. 121-140. Archibugi, D. and J. Michie (1997) (eds.), Technology, Globalisation and Economic Performance, Cambridge University Press, Cambridge. Archibugi, D. and M. Pianta (1992a), “Specialization and size of technological activities in industrial countries: The analysis of patent data”, Research Policy, Vol. 21,1, pp. 79-93. Archibugi, D. and M. Pianta (1992b), The Technological Specialization of Advanced Countries: A report to the EEC on International Science and Technology Activities, Kluwer Academic Publishers, Dodrecht. Archibugi, D. and M. Pianta (1994), “Aggregate convergence and sectoral specialization in innovation”, Journal of Evolutionary Economics, vol. 4, pp. 17-33. Archibugi, D. and M. Pianta (1998), “Aggregate convergence and sectoral specialisation in innovation: evidence for industrial countries”, in Archibugi, D. and Michie, J. (eds), Trade, Growth and Technical Change, Cambridge University Press, Cambridge. Audretsch, D. and M. Feldman (1996), “R&D spillovers and the geography of innovation and production,” American Economic Review, Vol. 86, pp. 253-273. Barcenilla, S. (1999), “Especialización tecnológica y especialización comercial. Evidencia empírica para los países de la Unión Europea”, Información Comercial Española, Vol. 781, pp. 85-102. Barro, R., and X. Sala-i-Martin (1995), Economic Growth, McGraw-Hill, New York. Baumol, W.J., S.A.B. Blackman and E.N. Wolff (1989), Productivity and American Leadership: The Long View, MIT Press, Cambridge, MA. Beelen, E. and B. Verspagen (1994), “The role of convergence in trade and sectoral growth”, in J. Fagerberg, B. Verspagen and N. Von Tunzelmann (eds.), The Dynamics of Technology, Trade and Growth, Edward Elgar, Aldershot. Belderbos, R. (1999), “Entry mode, organisational learning, and R&D in foreign affiliates: Evidence from Japanese firms”, NIBOR Research Memorandum RM/99/05. Belmann, R. (1953), Stability Theory of the Differential Equations, McGraw-Hill, New York. Bhagwati, J. N. (1984), “Splintering and disembodiment of services and developing nations”, The World Economy, Vol. 7, pp. 133-143. Blanc, H. and C. Sierra (1999), “The internationalisation of R&D by multinationals: a trade-off between external and internal proximity”, Cambridge Journal of Economics, Vol. 23, pp. 187-206. Borensztein, M., J. de Gregorio and J.W. Lee (1998), “How does foreign direct investment affect economic growth?”, Journal of International Economics, Vol. 45, pp. 115-135. Bresnahan, T.F. and M. Trajtenberg (1995), “General purpose technologies ‘Engines of growth’?”, Journal of Econometrics, Vol. 65, pp. 83-108. Buigues, P and A. Sapir (1993), “Market services and European integration, issues and challenges”, European Economy/Social Europe, 1993/3, reports and studies. Cantwell, J. (1989), Technological Innovation and Multinational Corporations, Bassil Blackwell, Oxford. Cantwell, J. (1991), “Historical Trends in international patterns of technological innovation”, in Foreman-Peck (ed.), New Perspectives on the late Victorian Age, Cambridge University Press, Cambridge. Cantwell, J. (1995), “The globalisation of technology: what remains of the product cycle model?”, Cambridge Journal of Economics, Vol. 19, pp. 155-74. Cantwell, J. and O. Janne (1999) “Technological globalisation and innovation centres: the role of corporate technological leadership and location hierarchy”, Research Policy, Vol. 28, pp. 119-44.

59

Cantwell, J. and O. Janne (2000), “The role of multinational corporations and national states in the globalization of innovatory capacity: the European perspective”, Technology Analysis and Strategic Management, Vol. 12, pp. 243-262. Chang, P., G. Karsenty, A. Mattoo and J. Richterung (1999), “GATS, the modes of supply and statistics on trade in services”, Journal of World Trade, Vol. 33, pp. 93-115. Clark, C. (1957), The Conditions of Economic Progress, London: Macmillan. Cornelius, P.K. (2000), “Trade in financial services, capital flows, and the value-at-risk of countries”, The World Economy, Vol. 23, pp. 649-672. Dalton, D. and M. Serapio (1999) “Globalization of industrial R&D: an examination of foreign direct investments in R&D in the United States”, Research Policy, Vol. 28, pp. 303-316. Dalum, B., K. Laursen and B. Verspagen (1999), “Does specialization matter for growth?”, Industrial and Corporate Change, Vol 8, pp. 267-288. Daveri, F. (2000), “Is growth in Europe an ICT story too?”, Università degli studi di Parma and IGIER, unpublished manuscript. Davies, S. and B. Lyons (1996), Industrial Organisation in the European Union, Oxford University Press, Oxford. Dosi, G. and L. Soete (1988), “Technical change and international trade”, in G. Dosi, C. Freeman, R. Nelson, G. Silverberg, and L. Soete (eds.), Technical Change and Economic Theory, Pinter, London. Dosi, G., C. Freeman, R. Nelson, G. Silverberg and L. Soete (1988), Technical Change and Economic Theory, Pinter, London. Dosi, G., K. Pavitt, and L. Soete (1990), The Economics of Technical Change and International Trade, Harvester Wheatsheaf., London. Dunning, J. and R. Narula (1995), “The R&D activities of foreign firms in the United States”, International Studies of Management and Organization, Vol. 25, pp. 39-73. Durlauf, S.N. and D.T. Quah (1998), “The new empirics of economic growth”, NBER Working paper n.6422. Edwards, S. (1993), “Openness, trade liberalization, and growth in developing countries”, Journal of Economic Literature, Vol. 31, pp. 1358-1393. ETAN-European Technology Assessment Network (1998), “Technology policy in the context of internationalisation of R&D and innovation. How to strengthen Europe's competitive advantage in technology”, Brussels, European Commission, Directorate-General Science Research and Development. European Commission (1998), European Competitiveness in the Triad, European Economy, Supplement A n.7, July, p.16. Evangelista, R. (2000), “Sectoral patterns of technological change in services”, Economics of Innovation and New Technology, Vol. 9, pp. 183-221. Evangelista, R. and M. Savona (2001), “The impact of innovation on employment and skills in services. Evidence from Italy”, paper presented at the Conference “The Future of Innovation Studies”, Eindhoven Centre for Innovetion Studies, Eindhoven, September. Fagerberg, J. (1996), Structural Changes in International Trade: Cause, Impact and Response, Working Paper, n. 547, NUPI (Norwegian Institute of International Affairs), Oslo. Fagerberg, J., P. Guerrieri and B. Verspagen (eds) (1999), The Economic Challenge for Europe: Adapting to Innovation-based Growth, Edward Elgar, Aldershot. Feenstra, R.C. and G.H. Hanson (1996), “Globalization, outsourcing, and wage inequality”, American Economic Review, Vol. 86, pp. 240-245. Feenstra, R.C. and G.H. Hanson (1999), “The Impact of outsourcing and high-technology capital on wages: Estimates for the United States, 1979-1990”, Quarterly Journal of Economics, Vol. 114, pp. 907-941. Florida, R. (1997), “The globalisation of R&D: results of a survey of foreign affiliated R&D laboratories in the USA”, Research policy, Vol. 26, pp. 85-103. Francois, J.F. (1990a), “Producer services, scale, and the division of labor”, Oxford Economic Papers, Vol. 42, pp. 715-729. Francois, J.F. (1990b), “Trade in producer services and returns due to specialization under monopolistic competition”, Canadian Journal of Economics, Vol. 23, pp. 109-124. Francois, J.F. and L. Schuknecht (1999), “Trade in financial services: Pro-competitive effects and growth performance”, CEPR Discussion Paper, No. 2144. Freeman, C. and C. Perez (1988), “Structural crises of adjustment, business cycles and investment behaviour, in Dosi, G. et al. (eds.). Freeman, C. and L. Soete (1987), Technical Change and Full Employment, Blackwell, Oxford. Frost, T. S. (2001), “The geographic sources of foreign subsidiaries’ innovation”, Strategic Management Journal, Vol. 22, pp. 101-123. Fuchs, V.R. (1968), The Service Economy, Columbia University Press. Gandolfo, G. (1981), Qualitative Analisys and Econometric Estimation of Continuous Time Dynamic Models, North Holland.

60

Gandolfo, G. (1992), “Sensitivity analysis in continuous time econometric models”, Computers Mathematics Application, Vol. 24, pp. 43-55. Gandolfo, G. (1997), Economic Dynamics, Springer-Verlag, Berlin. Gerybadze, A. and Reger, G. (1999), “Globalisation of R&D: recent changes in the management of transnational corporations”, Research Policy, Vol 28, pp. 251-274. Gordon, R. (1999), “Management consulting firms: Some approaches to output measurement”, Paper presented at the Brookings Workshop on Measuring the output of Business Services. Grandstand, O., P. Patel and K. Pavitt (1997), “Multi-technology corporations: why they have ‘distinctive core’ competencies”, California Management Review, Vol. 39, pp. 8-25. Gross, D.M., H. Raff, and M.J. Ryan (2001), “Intra- and inter-industry linkages in foreign direct investment: Evidence from Japanese investment in Europe”, unpublished manuscript. Grossman, G. and E. Helpman (1991), Innovation and Growth in the World Economy, MIT Press, Cambridge (Mass.). Gual, J. and D. Neven (1993) “Banking”, in European Economy/Social Europe, 1993/3, reports and studies Guerrieri, P. and C. Milana (1995), “Changes and trends in the world trade in high-technology products”, Cambridge Journal of Economics, Vol. 19, pp. 225-42. Gullickson, W. and M. Harper (1999), “Possible measurement bias in aggregate productivity growth”, Monthly Labor Review (February), pp. 47-67. Hagedoorn, J. and J. Schakenraad (1994), “The effect of strategic technology alliances on company performance”, Strategic Management Journal, Vol. 15, pp. 291-311. Håkanson, L. (1992), “Locational determinants of foreign R&D in Swedish multinationals” in Grandstand et al. (eds.), Technology Management and International Business: Internationalisation of R&D and Technology, Wiley, Chichester. Hauknes, J. (1996), “Innovation in the services economy”, STEP Report 7/96, Oslo. Hawawini, G. and M. Schill (1994), “The Japanese presence in the European financial services sector”, in M. Mason and D. Encarnation (eds.), Does Ownership Matter? Japanese Multinationals in Europe, Clarendon Press, Oxford. Helpman, E. (1998), General Purpose Technologies and Economic Growth, MIT Press, Cambridge. Hill, T.P. (1977), “On goods and services”, Review of Income and Wealth, Vol. 23, pp. 315-338. Hoekman, B. and C.A. Primo Braga (1997), “Protection and trade in services”, World Bank, Policy Research Working Paper, No. 1747. Hood, N. and S. Young (1982) “US multinational R&D: corporate strategies and policy implications for the UK”, Multinational Business, Vol. 2, pp. 10-23. Hubertus, J. (2001), “Distributive trades statistics: On-line retailing – a challenge for the future”, Statistics in Focus, Industry Services and Trade, Theme 4, No. 43. Iammarino, S., C. Jona Lasinio and S. Mantegazza (2001), “Sviluppo e diffusione dell’ICT: l’Italia negli anni ‘90”, Studi e Note di Economia, n.2/2001. Jaffe, A. B. and M. Trajtenberg (1996), “Flows of knowledge from universities and federal labs: modelling the flow of patent citations over time and across institutional and geographical boundaries”, NBER Working Paper 5712, Cambridge Massachusetts. Jaffe, A. B., M. Trajtenberg and R. Henderson (1993), “Geographic localization of knowledge spillovers as evidenced by patent citations”, Quarterly Journal of Economics, Vol. 108, pp. 577-98. Jones, R.W. and H. Kierzkowski (1990), “The role of services in production and international trade: A theoretical framework”, in R.W. Jones and A. Krueger (eds.), The Political Economy of International Trade, Basil Blackwell, Oxford. Jones, R.W. and F. Ruane (1990), “Appraising the options for international trade in services, Oxford Economic Papers, Vol. 42, pp. 672-687. Jorgenson, D.W. and K.J. Stiroh (2000), “Raising the speed limit: US economic growth in the information age”, Brookings Papers on Economic Activity, pp. 125-235. Kim, J. and J-D. Kim (2001), “Liberalization of trade in services and productivity growth in Korea”, in T. Ito and A.O. Krueger (eds.), Services Trade in the Asia-Pacific Region, East Asia Seminar on Economics Volume 11, forthcoming. Klodt, H. (2000), ‘Structural change towards services: the German experience’, University of Birmingham IGS Discussion Paper, 2000/07. Krugman, P.R. (1985) “A technology gap model of international trade”, in K. Jungenfelt and D. Hague (eds.), Structural Adjustment in Developed Open Economies, Macmillan, London. Kuemmerle, W. (1997), “Building effective R&D capabilities abroad”, Harvard Business Review, March-April, 1997, pp. 61-70. Kuemmerle, W. (1999), “Foreign direct investment in industrial research in the pharmaceutical and electronic industries: Results from a survey of multinational firms”, Research Policy, Vol. 28, pp. 179-93.

61

Lall, S. (1979), “The international allocation of research activity by US multinationals”, Oxford Bulletin of Economics and Statistics, Vol. 41, pp. 313-31. Laursen, K. (1999), “The impact of technological opportunity on the dynamics of trade performance”, Structural Change and Economic Dynamics, Vol. 10, pp. 341-357. Laursen, K. (2000), Trade Specialisation, Technology and Economic Growth, Edward Elgar, Cheltenham. Laursen, K. and V. Meliciani (2000), “The importance of technology based intersectoral linkages for market share dynamics”, Weltwirtschaftliches Archiv, Vol. 136, pp. 702-723. Levine, R. (1997), “Financial development and economic growth: Views and agenda”, Journal of Economic Literature, Vol. 25, pp. 688-726. Lucas, R.E. (1988), “On the mechanism of economic development”, Journal of Monetary Economics, Vol. 22, pp. 3-42. Maggi, B., D. Federici and G. Espa (2002), “Statistical economic approach to mix stock-flow dynamic models in macroeconomics”, Quaderni di Statistica e Matematica Applicata alle Scienze Economico-sociali, (forthcoming). Manzocchi, S., P.C. Padoan and C. Vicarelli (2000), “Specialisation and flexibility”, in Padoan, P.C. (ed.), Employment and growth in the Europe of monetary union, Edward Elgar, Cheltenam. Manzocchi, S., P.C. Padoan, P. Parascandolo and M. Tozzi (2000), “Employment and European regions”, in Padoan, P.C. (ed.), Employment and Growth in the Europe of Monetary Union, Edward Elgar, Cheltenam. Markusen, J.R. (1984), “Multinationals, multi-plant economies, and the gains from trade”, Journal of International Economics, Vol. 16, pp. 205-226. Markusen, J.R. (1989), “Trade in producer services and in other specialised intermediate inputs”, American Economic Review, Vol. 79, pp. 85-95. Markusen, J., T.F. Rutherford, and D. Tarr (2000), “Foreign direct investment in services and the domestic market for expertise”, NBER Working Paper, No. 7700. Marrewijk, C. van, J. Stibora, and A. de Vaal (1996), “Services tradability, trade liberalization and foreign direct investment”, Economica, Vol. 97, pp. 1180-1196. Marrewijk, C. van, J. Stibora, A. de Vaal, and J. Viaene (1997), “Producer services, comparative advantage, and international trade patterns”, Journal of International Economics, Vol. 42, pp. 195-220. Matsuyama, K. (1992), “Agricultural productivity, comparative advantage, and economic growth”, Journal of Economic Theory, Vol. 58, pp. 317-334. Mattoo, A. (2001), “Shaping future GATS rules for trade in services”, presented at the 11th NBER-Annual East Asian Seminar on Economics, Seoul, June 22-24. Mattoo, A., R. Rathindran and A. Subramanian (2001), “Measuring services trade liberalization and its impact on economic growth”, World Bank Working Paper, No. 2655. Maurseth, P.-B. and B. Verspagen (2001), “Knowledge spillovers in Europe. A patent citations analysis”, mimeo, Eindhoven University of Technology and Norwegian Institute of Foreign Affairs. Meliciani, V. (2001), Technology, Trade and Growth in OECD Countries: Does Specialisation Matter?, Routledge, Londra. Meliciani, V. (2002), “The impact of technological specialisation on economic growth in a balance-of-paymentsconstrained growth model”, Structural Change and Economic Dynamics, Vol. 13, pp. 101-118. Meliciani, V. and R. Simonetti (1998), “Specialization in areas of strong technological opportunity and economic growth”, in Eliasson, G., C. Green and C. McCann (eds.), The Microfoundations of Economic Growth, University of Michigan Press, Michigan, USA. Melvin, J.R. (1989), “Trade in producer services: A Heckscher-Ohlin approach”, Journal of Political Economy, Vol. 97, pp. 1180-1196. Metcalfe, J.S. and L. Soete (1984), “Notes on the evolution of technology and international competition”, in M. Gibbons, P. Gummett and B. Udgaonkar (eds.), Science and Technology Policy in the 1980s and Beyond, Longman, Harlow, Essex. Meyer-Krahmer, F. and G. Reger (1999), “New perspectives on the innovation strategies of multinational enterprises: Lessons for technology policy in Europe”, Research Policy, Vol. 28, pp. 751-76. Midelfart-Knarvik, K.H, H.G. Overman, S.J.Redding and A.J. Venables (2000), “The location of European industry”, Report prepared for the Directorate General for Economic and Financial Affairs, European Commission. Miles, I. (1993), ‘Services and the new economy’, Futures, Vol. 25, pp. 653-672. Miles, I. (1996) “Infrastructure and the delivery of new services”, in OECD Employment and growth in the knowledge based economy, OECD, Paris. Molero, J. (ed) (1995), Technological Innovation, Multinational Corporations and New International Competitiveness. The case of intermediate countries, Harwood Academic Publishers, Reading. Nachum, L. (1999), “Measurement of productivity in Swedish management consulting firms”, presented at Brookings Workshop on Measuring the Output of Business services, May 14. Narula, R. (1996), Multinational Investment and Economic Structure, Routledge, London.

62

Narula, R. (1999), “Explaining the growth of strategic R&D alliances by European firms”, Journal of Common Markets Studies, Vol. 37, pp. 711-23. Niosi, J. (1999), “The internationalisation of R&D: from technology transfer to the learning organisation”, Research Policy, Vol. 28. pp. 107-117. OECD (1996), Science, Technology and Industry Outlook, Paris. OECD (2001), The Impact of Information and Communication Technology on Output Growth: Issues and Preliminary Considerations, OECD, Directorate for Science, Technology and Industry, SWP 2001/11. Oliner, S., and D. Sichel (2000), “The resurgence of growth in the late 1990s: is information technology the story?”, Journal of Economic Perspectives, Vol. 14, pp. 3-22. Ottway, J. (2001), Report on internship in KPMG. Padoan, P.C. (ed) (2000), European Monetary Union and Employment, Edgar Elgar, Cheltenham. Padoan, P.C (1998), “Trade, knowledge accumulation and diffusion. A sectoral perspective”, Structural Change and Economic Dynamics, Vol. 9, pp. 349-72. Pagano, M. (1993), “Financial markets and growth. An overview”, European Economic Review, Vol. 37, pp. 613-622. Patel, P. (1995), “Localised production of technology for global markets”, Cambridge Journal of Economics, Vol. 19, pp. 141-153. Patel, P. (1997), “Localised production of technology for global markets”, in Archibugi, D. and J. Michie (eds), Technology, Globalisation and Economic Performance, Cambridge University Press, Cambridge. Patel, P. and K. Pavitt (1991), “Large firms in the production of the world’s technology: An important case of ‘non-globalisation’”, Journal of International Business Studies, Vol. 22, pp. 1-21. Patel, P. and K. Pavitt (1994), “Uneven (and divergent) technological accumulation among advanced countries: Evidence and a framework of explanation”, Industrial and Corporate Change, Vol. 3, pp. 759-787. Patel, P. and K. Pavitt (1999), “Global corporations & national systems of innovation: Who dominates whom?” in D. Archibugi, J. Howells and J. Michie (eds), Innovation Policy in a Global Economy?, Cambridge University Press, Cambridge. Patel, P. and K. Pavitt (2000), “National systems of innovation under strain: the internationalisation of corporate R & D”, in R. Barrell, G. Mason and M. O'Mahoney (eds), Productivity, Innovation and Economic Performance, Cambridge University Press, Cambridge. Patel, P. and M. Vega (1999), “Patterns of internationalisation of corporate technology: location vs. home country advantages”, Research Policy, Vol. 28, pp. 145-55. Pavitt, K. (1984), “Sectoral patterns of technical change: Towards a taxonomy and a theory”, Research Policy, Vol. 13, pp. 343-373. Pavitt, K. (1988), “International patterns of technological accumulation”, in Hood, N. and J.E. Vahlne (eds.), Strategies in Global Competition, Croom Helm, London. Pearce, R. and S. Singh (1992), Globalising Research and Development, London, Macmillan. Pelkmans, J. (2001), European Integration: Methods and Economic Analysis, Financial Times/Prentice Hall, Harlow, England. Perez, C. (1983), “Structural change and the assimilation of new technologies in the economic and social system”, Futures, Vol. 15, pp. 357-75. Perez, C. (1985), “Micro-electronics, long waves and world structural change: New perspectives in developing countries”, World Development, Vol. 13, pp. 441-63. Perez, C. (1988), “New technologies and development”, in C. Freeman and B.A. Lundvall (eds.), Small Countries Facing the Technological Revolution, Pinter, London. Petit, P. and L. Soete (1996), “Technical change and employment growth in services: analytical and policy challenges”, paper presented at the conference “Technology, Employment and Labour Markets”, Athens. Petit, P. and L. Soete (eds.) (2001), Technology and the Future of European Employment, Cheltenham, Elgar. Pianta, M. (1997), “Growth and specialisation in technology, production and trade: The limits of comparative advantage”, paper prepared as part of the EU TSER project “Technology, Employment and European Cohesion”. Pianta, M. and V. Meliciani (1996), “Technological specialization and economic performance in OECD countries”, Technology Analysis and Strategic Management, Vol. 8, pp. 157-174. Posner, M. V. (1961), “International trade and technical progress”, Oxford Economic Papers, Vol. 13, pp. 323341. Rauch, J.E. (1999), ‘Networks versus markets’, Journal of International Economics, Vol. 48, pp. 7-35. Rivera-Batiz, L. and P.M. Romer (1991), “Economic integration and endogenous growth”, Quarterly Journal of Economics, Vol. 106, pp. 531-555. Roeger W. (2001), The Contribution of Information and Communication Technologies to Growth in Europe and the US: A Macroeconomic Analysis, Economic Paper 147, EU Commission DG1. Rowthorn, R. and R. Ramaswamy (1999), “Growth, trade, and deindustrialisation”, IMF Staff Papers, Vol. 46, pp. 18-41.

63

Ruane, F. (1990), “Comment on papers by C.R. Neu and A. Sapir”, in R. Baldwin, C. Hamilton and A. Sapir (eds), Issues in US-EC Trade Relations, Chicago: University of Chicago Press, 269-276. Sampson, G.P. and R.H. Snape (1985), “Identifying the issues in trade in services”, The World Economy, Vol. 8, pp. 171-182. Sapir, A. (1985), “North-South issues in trade in services”, The World Economy, Vol. 8, pp. 27-42. Sapir, A. (1988), “International trade in telecommunications services”, in R. Baldwin, C. Hamilton, and A. Sapir (eds), Issues in US-EC Trade Relations, Chicago: University of Chicago Press, pp. 231-244. Sapir (1993), “Sectoral dimension”, European Economy/Social Europe, 1993/3, reports and studies. Sapir, A. and C. Winter (1994), “Services trade”, in D. Greenaway and L. Alan Winters (eds.), Surveys in International Trade, Blackwell, Oxford. Schreyer, P. (2000), “The contribution of information and communication technology to output growth: a study of the G7 countries”, OECD, Directorate for Science, Technology and Industry, DSTI/DOC (2000) 2, 23 March. Siniscalco, D. (1988), “Defining and measuring output and productivity in the services sector”, in H. Giersch (ed.), World Services in World Economic Growth, Symposium 1988, Institut für Weltwirstschaft an der Universität Kiel. Sjöholm, F. (1996). “International transfer of knowledge: the role of international trade and geographic proximity”, Weltwirtschaftliches Archiv, Vol. 132, pp. 97-115. Soete, L. (1981), “A general tests of the technological gap trade theory”, Weltwirtschaftliches Archiv., Vol. 117, pp. 638-666. Soete, L. (1987), “The impact of technological innovation on international trade patterns: The evidence reconsidered”, Research Policy, Vol. 16, pp. 101-130. Soete, L. and B. Verspagen (1994), “Competing for growth: The dynamics of technology gaps”, in L.L. Pasinetti and R. M. Solow (ed.), Economic Growth and the Structure of Long Term Development, London, Macmillan. Stiroh, K., (2001), “What drives productivity growth?”, Economic Policy Review, FED New York, Vol. 7, pp. 37-60. Triplett, J.E. (1999), “The Solow computer paradox: what do computers do to productivity?”, Canadian Journal of Economics, Vol. 32, pp. 309-334. Triplett, J.E. and B.P. Bosworth (2002), “Productivity in the services sector”, in R.E. Stern (ed.), Services in the International Economy, forthcoming. UNCTAD and World Bank (1994), Liberalising Trade in Services: A Handbook , United Nations, New York and Geneva. UNCTAD (2001), World Investment Report: Promoting Linkages, Geneva: United Nations. Vaal, A. de and Van den Berg (1999), “Producer services, economic geography and services tradability”, Journal of Regional Science, Vol. 39, pp. 539-572. Ventura, J. (1997), “Growth and interdependence”, Quarterly Journal of Economics, Vol. 112, pp. 57-84. Vernon, R. (1966) “International investment and international trade in the product cycle”, Quarterly Journal of Economics, Vol. 80, pp. 190-207. Vernon, R. (1977), Storm over Multinationals:The Real Issues, Macmillan, London. Verspagen, B. (1993), Uneven Growth Between Interdependent Economies, Aldershot, Avebury. Vivarelli, M. (1995), The Economics of Technology and Employment. Theory and Empirical Evidence, Edward Elgar, Aldershot. Von Hippel, E. (1994), “Sticky information and the locus of problem solving: implications for innovation”, Management Science, Vol. 40, pp. 429-439. Wymer, C. (1994), APREDIC and ESCONA programs. Wymer, C. (1997), “Structural non-linear continuous-time models in econometrics”, Macroeconomic Dynamics, Vol. 1, pp. 518-548. Wolff, E.N. (1997), “Productivity growth and shifting comparative advantage on industry level”, in J. Fagerberg, P. Hansson, L. Lundberg and A. Melchior (eds.), Technology and International Trade, Edward Elgar, Cheltenham. Young, A. (1991), “Learning by doing and the dynamic effects of international trade”, Quarterly Journal of Economics, Vol. 106, pp. 369-405. Zander, I. (1999), “How do you mean ‘global’? An empirical investigation of innovation networks in the multinational corporation”, Research Policy, Vol. 28, pp. 195-213. Zanfei, A. (2000), “Transnational firms and the changing organisation of innovative activities”, Cambridge Journal of Economics, Vol. 24, pp. 515-542.

64