continental, national and sub-national innovation systems ... · research policy 31 (2002)...

Research Policy 31 (2002) 191–211

Continental, national and sub-national innovationsystems—complementarity and economic growth�

Chris Freeman∗SPRU, University of Sussex, Falmer, Brighton BN1 9RF, UK

Abstract

The purpose of this paper is to discuss the relevance of innovation systems to economic growth rates over the last twocenturies. The focus is on complementarity (or lack of it) between sub-systems of society and on models of active learningin catching up economies. The paper discusses variations in rates of growth of economic regions and the extent to whichvariations may be attributed to “innovation systems”. The analysis is applied to Britain in the 18th century, the United Statesin the second half of the 19th century and the innovation systems ofcatching upcountries in the 20th century. © 2002 ElsevierScience B.V. All rights reserved.

Keywords:Innovation system; Economic growth; Institutional change; Economic history

1. Introduction

This paper discusses variations over time in ratesof growth of various economic regions and the extentto which these variations may be attributed to “inno-vation systems”. There has been a rapidly growingliterature on this topic during the 1990s (see, for ex-

� This paper was originally written for a research project onlocal systems of innovation in the Mercosur countries carriedout by a network of academic institutions in Brazil, Argentinaand Uruguay under the coordination of Jose E. Cassiolato andHelena M. M. Lastres of the Federal University of Rio de Janeiro,Brazil and sponsored by the Organisation of American States andthe Brazilian National Council for Scientific and TechnologicalDevelopment. A longer version was published in the book ‘Glob-alização e Inovação Localizada: Experiências de Sistemas Locaisno Mercosul’, Brasılia, IBICT, a collection of papers of the firstpart of the project (Cassiolato and Lastres, 1999). An extendeddiscussion of some of the themes in Sections 2 and 3 can nowbe found in Freeman and Louçã (2001).

∗ Tel.:+44-1273-67-8173.E-mail address:[email protected] (C. Freeman).

ample, Lundvall, 1992; Nelson, 1993; Mjøset, 1992;Villaschi, 1993; Humbert, 1993; Freeman, 1995;Reinert, 1997).

Much of this literature (see especially Hu, 1992;Porter, 1990; Patel, 1995) insists on the central impor-tance ofnationalsystems but a number of authors haveargued that “globalisation” has greatly diminished oreven eliminated the importance of the nation-state(notably Ohmae, 1990). Other critics have stressedalternatively (or in addition) that sub-national entities,such as provinces, industrial districts, cities or “Sili-con Valleys” are becoming, or have already become,more important than the nation-state (see, for exampleDeBresson, 1989; DeBresson and Amesse, 1991).

Unfortunately, at least in the English language, thesame word “Regional” is often used to describe twoentirely different phenomena, viz.

1. Geographical areas embracing several nation-statesand even entire sub-continents—the “Pacificregion”, “East Asia”, Eastern Europe, CentralAmerica, etc.

0048-7333/02/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved.PII: S0048-7333(01)00136-6

192 C. Freeman / Research Policy 31 (2002) 191–211

Table 1Comparative growth rates sub-continental regions 1965–1999a

1965–1980 1980–1989 1990–1999(estimate)

GDP % p.a.East Asia 7.5 7.9 7.2South Asia 3.9 5.1 5.5Africa (sub-Sahara) 4.0 2.1 2.7Latin America 5.8 1.6 3.1

GDP per capital % p.a.East Asia 5.0 6.3 5.7South Asia 1.5 2.9 3.4Africa (sub-Sahara) 1.1 −1.2 0.2Latin America 3.5 −0.5 1.2

a Source: World Bank Development Report (1991); own esti-mates in 1990s.

2. Geographical areas which are smaller sub-divisionsof nation-states, e.g. “states”, urban areas, counties,rural areas, etc.

This can be a source of confusion, so for this reason,this paper refers to the wider areas as “continental”or “sub-continental” and the smaller areas as “sub-national”.The inter-continental variations in growthrates are indeed very wide, as illustrated in Table 1,but the variations between countries have of coursebeen even wider. In particular, a group of countries,today referred to as “developed” or “industrialised”drew far ahead of the rest of the world (later knownas “under-developed”) during the last two centuries(Table 2, columns 5 and 6).

Table 2Estimates of trends in per capita GNP (1960 US$ and prices, 1750–1977)a

Year Developed countries Third World Gaps

(1) Total(billion ofUS$)

Per capita,US$

(3) Total(billion ofUS$)

(4) Percapita, US$

(5) = (2)/(4) (6) Ratio of the mostdeveloped to the leastdeveloped

1750 35 182 112 188 1.0 1.81800 47 198 137 188 1.1 1.81830 67 237 150 183 1.3 2.81860 118 324 159 174 1.9 4.51913 430 662 217 192 3.4 10.41950 889 1054 335 203 5.2 17.91960 1394 1453 514 250 5.8 20.01970 2386 2229 800 380 7.2 25.71977 2108 2737 1082 355 7.7 29.1

a Source: Bairoch and Levy-Leboyer (1981), pp. 7–8.

Abramovitz (1986) coined the expression “socialcapability” to describe that capacity to make insti-tutional changes which led to this divergence ingrowth rates. He was himself one of the pioneers of“growth accounting” but, as he cogently pointed out,the accumulation of capital and increase in the labourforce are not in themselves sufficient to explain thesevarying rates of economic growth. The huge diver-gence in growth rates which is so obvious a featureof long-term economic growth over the past two cen-turies must be attributed in large measure to the pres-ence or absence of social capability for institutionalchange, and especially for those types of institutionalchange which facilitate and stimulate a high rate oftechnical change, i.e. innovation systems. As we shallsee, attempts by Krugman (1994) and others to goback to the quantitative accumulation of capital andlabour as the main explanation of the “East AsianMiracle” are very unconvincing. Institutional changeswere essential for the accumulation of capital itself.

Many historians and economists had of coursealways emphasised the importance of technical andinstitutional change, as for example, Landes (1970) orSupple (1963). Indeed, going back to the early devel-opment of economic theory, List (1841) had stronglycriticised Adam Smith and other classical economistsfor what he perceived as their neglect of technologyand skills. In fact, Adam Smithdid recognise thegreat importance of science and technology but didnot consistently give it the prominence which Listthought that it merited.

C. Freeman / Research Policy 31 (2002) 191–211 193

The main concern of List was with the problem ofGermany catching up with England and for under-developed countries (as the German states then were inrelation to England), he advocated not only protectionof infant industries but a broad range of policiesdesigned to make possible or to accelerate industrial-isation and economic growth. Most of these policieswere concerned with learning about new technologyand applying it and many of them were applied incatching-up countries over the next century and a half(see Section 5).

After reviewing the changing ideas of economistsabout development in the years since the SecondWorld War, the World Bank (1991) concluded thatit is intangible investment in knowledge accumula-tion, which is decisive rather than physical capitalinvestment, as was at one time believed (pp. 33–35).The Report cited the “New Growth Theory” (Romer,1986; Grossman and Helpman, 1991) in support ofthis view but the so-called “New Growth Theory” hasin fact only belatedly incorporated into neo-classicalmodels the realistic assumptions which had becomecommonplace among economic historians and neo-Schumpeterian economists. Indeed, it could just aswell have cited List (1841), who in criticising apassage from Adam Smith said:

. . . Adam Smith has. . . forgotten that he himselfincludes (in his definition of capital) the intellec-tual and bodily abilities of the producers under thisterm. He wrongly maintains that the revenues of thenation are dependent only on the sum of its materialcapital (p. 183).

and further:

The present state of the nations is the result ofthe accumulation of all discoveries, inventions,improvements, perfections and exertions of allgenerations which have lived before us: they formthe intellectual capital of the present human race,and every separate nation is productive only inthe proportion in which it has known how to ap-propriate those attainments of former generationsand to increase them by its own acquirements(p. 113).

List’s clear recognition of the interdependence ofdomestic and imported technology and of tangibleand intangible investment has a decidedly modern

ring. He saw too that industry should be linked to theformal institutions of science and of education:

There scarcely exists a manufacturing businesswhich has not relation to physics, mechanics, chem-istry, mathematics or to the art of design, etc. Noprogress, no new discoveries and inventions can bemade in these sciences by which a hundred indus-tries and processes could not be improved or altered.In the manufacturing State, therefore, sciences andarts must necessarily become popular (p. 162).

The recent literature on “national systems of innova-tion” could be described as an attempt to come toterms more systematically with these problems ofsocial capability for technical change. List’s book on“The National System of Political Economy” mightjust as well have been entitled “The National Systemof Innovation” since he anticipated many of theconcerns of this contemporary literature. The mainpurpose of this paper is to discuss the relevance ofsystems of innovation to economic growth rate overthe last two centuries. A long-term historical ap-proach is essential for this purpose because of thevery nature of technical and institutional change. Theenormous gaps between different parts of the worldtook decades or even centuries to open up and theefforts to close them have also taken many decades.

The analysis starts with the case of Britain in the18th century because Britain was the first country toopen up a major gap in productivity, in technology andin per capitaincomes, compared with all other nationsand city states. The British case is discussed at somelength both because it was the first and also becauseit serves to introduce some basic problems in thetheory of innovation systems—notably the comple-mentarity (or lack of it) between various sub-systemsof society: science, technology, economy, politics andculture, and the complementarity between nationaland sub-national systems. The British slow-down and“falling behind” in the 20th century also illustratesthe relative rigidity of someorganisationalstructurescompared with informal institutions, a point empha-sised by Edqvist (1997a,b) in his thorough review ofnational systems theory.

Following this discussion Section 4 then takes upthe second major example of a national system forg-ing ahead of the rest of the world—the case of theUnited States in the second half of the 19th and first


half of the 20th century. The remainder of the paperthen discusses the innovation systems ofcatching upcountries, which have been described by Viotti (1997)in an excellent dissertation aslearning systems. Hemakes an interesting distinction betweenactive andpassivelearning systems and applies this distinctionto the example of South Korea and Brazil, an examplewhich is reviewed in Section 5.

Finally, Section 6 of this paper speculates about thepossible course of events in the next century, takingup the question of globalisation and convergence anddrawing some conclusions about the role of innovationsystems in future economic growth.

2. The concept of national systems

The capacity for technical and social innovationsdid of course strongly influence economic lifebeforenation-states became the dominant form of political or-ganisation. Although Adam Smith’s book was entitled“The Wealth ofNations” and his main concern was toexplain “the different progress of opulence in differentnations”, he nevertheless included a long discussion of“The rise and progress ofcitiesandtownssince the fallof the Roman Empire”. The contemporary discussionis therefore certainly not entirely new: changing formsof political organisation and territorial boundaries nec-essarily changed the nature of the debate. For AdamSmith, it was the widening gap in living standards andin manufacturing industry between Britain and otherpolitical units in Europe which most intrigued him.Some of these were powerful nation-states, such asFrance and Spain, others were still city states or smallprincipalities and still others were Empires.

Table 3National systems: “narrow” institutions (sources of innovations)

17th century Academies of Science, Royal Society 1662, “Proceedings” and Journals, Internationalism ofScience, Science Education

18th century “Industrial revolution” (factories), Technical Education, Nationalism of Technology,Consulting Engineers

19th century Growth of Universities, Ph.D. and Science Faculties, Technische Hochschulen, Institutes ofTechnology, Government Laboratories, Industrial R&D in-house, StandardsInstitutes

20th century Industrial in-house R&D in all industries, “Big Science and Technology”, Research Councils,NSF, etc., Ministries of Science and Technology, Service Industries R&D,Networks

Adam Smith’s discussion marked the transitionfrom policies which were mainly concerned with thepromotion and protection oftrade, with the financeof shipping, trading posts and cargoes, with the ship-building industry and naval power to policies whichwere mainly concerned withmanufacturing industry.

The city-state innovation systems of the Renais-sance had many remarkable achievements in craftindustries as well as in financial systems, shipping,the arts, medicine and science. We have nevertheless,started this account with 18th century Britain be-cause this was the time when Britain diverged fromits great trading competitors in Spain, Portugal andThe Netherlands and when the embryonic innovationsystems which had grown up in the period of theRenaissance developed into something new, associ-ated with the predominance of capitalistindustry.

A distinction has been made (Lundvall, 1992)between “narrow” definitions of national systems ofinnovation (Table 3) and a broad definition (illus-trated in Table 4). The narrow approach concentrateson those institutions which deliberately promote theacquisition and dissemination of knowledge and arethe main sources of innovation. The “broad” approachrecognises that these “narrow” institutions are embed-ded in a much wider socio-economic system in whichpolitical and cultural influences as well as economicpolicies help to determine the scale, direction and rel-ative success of all innovative activities. The decisivechanges which came about in 17th and 18th centuryBritain and later in the United States and other Euro-pean countries, were the elevation of science in thenational culture, the multiplication of links betweenscience and technology and the systematic widespreadembodiment of both in industrial processes in the


Table 4Some characteristics of British national system of Innovation during 18–19th century (broad definition)

Strong links between scientists and entrepreneursScience has become a national institution, encouraged by the state and popularised by local clubsStrong local investment bylandlords in transport infrastructure (canals and roads, later railways)Partnership form of organisation enables inventors to raise capital and collaborate with entrepreneurs (e.g. Arkwright/Strutt

or Watt/Boulton)Profits from trade and services available through national and local capital markets to invest infactory production and in infrastructureEconomic policy strongly influenced by classical economics and in the interests of industrialisationStrong efforts to protect national technology and delay catching up by competitorsBritish productivity per person about twice as high as European average by 1850Consulting engineers develop and diffuse best practice technology in waterwheels, canals, machine-making and railwaysPart-time training, night school, and apprenticeship training for new factory technicians and engineersGradual extension of primary, secondary and tertiary education

new workshops and factories (Table 4). The culturalchanges associated with the Renaissance were pushedeven further in the direction of secular instrumentalrationality and its application to industrial investment.

3. The British national system

The decisive differences between the city-state in-novation systems of the Mediterranean and the Britishnational system were in the role of science and therole of industry. The role of science is still disputedby historians. Some accounts argue that science inBritain in the 18th century lagged behind other Euro-pean countries, especially France and that it was notparticularly important for the success of the industrialrevolution. What these accounts tend to misconstrueis that it was not the location of a particular scien-tific discovery which mattered. These may have beenmore frequent outside Britain. What mattered for theindustrial revolution was the prevalence of a scientificculture. The treatment of Newton in Britain comparedwith the treatment of Galileo in Italy exemplifies thispoint. Newton was revered in Britain by both stateand church while the fate of Galileo was altogetherdifferent. Bacon (1605) had already proposed an in-tegrated policy for science, exploration, invention andtechnology at the beginning of the 17th century. Therewas an exceptionally fortunatecongruenceof science,culture and technology in Britain which made it pos-sible to use science, including Newtonian mechanics,on a significant scale in the invention and design ofa wide variety of new instruments, machines, en-gines, canals, bridges, water wheels and so forth. For

example, the British industrial revolution dependedon water power (not on steam power) for over halfa century. It was Joseph Smeaton in his papers anddrawings presented to the Royal Society in the 1770swhose experimental work made possible a scientificand technological breakthrough in the design of waterwheels more than doubling the productivity of waterpower. The use of iron rather than wood, first of allfor the gears and later for the entire water wheel,was made possible through Smeaton’s work as a con-sulting engineerfor the Carron Iron Works, by thenalready the largest iron foundry in Europe.

This is only one example, although a very importantone, of the positive interplay between science, technol-ogy, culture and entrepreneurship which characterisedthe British national system of innovation. Thecon-gruenceof these four sub-systems of society extendedalso to the political sub-system, which promoted allof these. According to many accounts (e.g. Needham,1954) it was the failure of the Chinese Empire to sus-tain congruence between these sub-systems which ledto the failure of China to sustain its world technologi-cal leadership. The conflict between church, state andscience (e.g. Galileo) hindered a more fruitful devel-opment of both science and technology in the Italiancity-states and elsewhere in Europe. The different roleof science in Britain and Italy has been especially welldocumented by Jacob (1988).

This was not the only factor which weakened thecity-state innovation systems. Even more decisivewere the scale economies made possible by factoryproduction, capital accumulation and specialiseddivision of labour. It was an Italian economist,Antonio Serra who first recognised the extraordinary


Table 5Labour productivity in cotton: operative hours to process 100 lbsof cottona

OHP

Indian Hand Spinners (18th century) 50000Crompton’s Mule (1780) 2000100-Spindle Mule (c. 1790) 1000Power-assisted Mules (c. 1795) 300Roberts’ automatic Mule (c. 1825) 135Most efficient machines today (1990) 40

a Source: Jenkins (1994), p. xix.

importance of increasing returns to scale but he diedin prison whereas Adam Smith was honoured by theBritish Prime Minister (“We are all your pupils now”).Enterprises and workshops were still very small in18th century Britain but the shift from cottage indus-try to factory production and the constant improve-ments in machinery were still enough to confer a hugeadvantage on British manufacturing firms. Nowherewas this more obvious than in the cotton industry(Table 5) where the combination of technical inven-tions, investment in machinery, factory organisation,and entrepreneurship in ever-wider markets (facili-tated still by naval power) opened an enormous pro-ductivity gap between Britain and all other producers.Some enterprises in the two leading industries, cottonand iron, already deployed hundreds of people by1800.

Investment inindustrycertainly owed a great dealto profits from trade but this could not have takenplace without a change in the culture and attitudes ofthe landlord and middle classes as well as changes inthe capital markets. The investment in the transportinfrastructure by the British landlord class was uniquein Europe and caused Marx to remark that Britainhad a bourgeois aristocracy. The bourgeoisie and thelandlords in Britain behaved differently followingthe victory of Parliament over the monarchy and thearistocracy in the English Civil War in the mid-17thcentury. This victory made irreversible political andsocial changes despite the restoration of the monar-chy in 1660. The investment in trade, transport andindustry became even more important than the owner-ship of land. The local political initiative of landlordsin promoting a wave of investment in canals in thelate 18th century was exceptionally important in theearly take-off of several key industrial districts whose

access to national and international markets hadhitherto been hindered by poor communications andtransport.

Schumpeter always maintained that the spread ofinnovations was necessarily uneven both with respectto timing and to space and this was certainly the casewith the spread of those innovations which comprisedthe British industrial revolution. They were not evenlyspread over all parts of the country, they affected onlya few industries at first and they diffused relativelyslowly to other European countries. The main centresof innovation, of urbanisation and of the rapid growthof new industries were not in the London region butin the North of England, especially in Lancashire andYorkshire, in the Midlands and in Scotland. Originally,the reasons for the success of the new “sub-national”industrial regions or districts had little or nothing todo with economic policy at a regional level. The mainadvantage of the North was probably the more rapidlyflowing rivers of the Pennine Hills which providedthe consistency and strength of flow for Smeaton’swater wheels. The iron industry was obviously alsoinfluenced by the location of wood for charcoal andlater, coal and iron ore deposits but much iron ore wasstill imported in the 18th century.

However, although geological factors such as riversfor navigation or water power and deposits of mineralsor the lack of them, played an important role in deter-mining the early growth of various industries, thesenatural advantages were soon overtaken bycreatedadvantages such as the transport infrastructure, thelocation of ports and access to skills and to markets.Lancashire enjoyed the advantage of the port of Liver-pool which was the centre of the North Atlantic tradewith North America. Many economists, and especiallyMarshall, pointed to the external economies whichresulted from many firms in the same industry locatedin the same industrial district where “the secrets of in-dustry are in the air” (Foray, 1991). To this day, theseexternal economies of agglomeration have continuedto be extremely important in industries as diverse assemi-conductors, toys, and machine tools. They arean essential part of the argument of Piore and Sabelfavouring small firm networks as against large massproduction firms. They are also one of the main rea-sons for some economists to propose that sub-nationalregional innovation systems have now become moreimportant than national systems themselves.


There is much in the experience of the Britishindustrial revolution which appears at first sight tofavour this view. Above all, the accumulated specia-lised skills in Lancashire were one of the major rea-sons for the extraordinary success of the Lancashirecotton industry, undoubtedly the leading sector of theBritish industrial revolution, accounting for 40% ofall British exports in 1850 and still for over a quarterof a much larger total in 1900.

In their explanation of the reasons for Britishdominance in cotton persisting throughout the 19thcentury, Mass and Lazonick (1990) attribute this“sustained competitive advantage” to a cumulativeprocess in which the development and utilisation ofseveral key productive factors reinforced each other.These affected labour costs, marketing costs and ad-ministrative costs. In all of these areas, industry scaleeconomies (Marshall’s external economies of scale)were important: In the case of labour:

During the nineteenth century the development andutilisation of labour resources provided the Britishcotton industry with its unique sources of compet-itive advantage. The major machine technologies. . . required complementary applications of ex-perienced human labour to keep them in motion.Experience gave workers not only specific cognitiveskills (of which a process such as mule spinningwas much more demanding. . . ) but also (and moreimportant over the long run) the general capabilityto work long hours at a steady pace without dam-aging the quality of the product, the materials orthe machines (Mass and Lazonick, 1990, p. 4).

Mass and Lazonick lay particular stress on the habi-tuation to factory work and cumulative skills of thelabour force but they also stress that the trade unionorganisation at that time and in that sector (surpris-ingly for some stereotyped ideas of British industrialrelations of later periods and other industries), wereparticularly congruent with incentives to sustain andincrease productivity. Great responsibility was givento the more skilled workers (who often had previousexperience in domestic craft work) for recruitment,training and supervision of the less skilled.

Besides the general habituation to factory work thatcame from growing up in factory communities andentering the mills at a young age, cotton workers

developed specialised skills in spinning particulartypes of yarn and weaving particular types of cloth(Mass and Lazonick, 1990, p. 5).

In common with other historians they point to theeconomies of agglomeration in relation to pools ofspecialised skilled labour in various Lancashire towns:Bolton (fine yarns), Oldham (coarse yarns), Blackburnand Burnley (coarse cloth), etc. Similar argumentsapply to the availability of skilled mechanics adeptat maintaining (and improving) the local machinery.The gains from increasing productivity were generallyshared with the skilled workers, whose union powerensured this.

By the 1870s cotton industries around the worldcould readily purchase British plant and equip-ment and even British engineering expertise. Butno other cotton industry in the world could readilyacquire Britain’s highly productive labour force;no other industry in the world had gone throughthe century-long developmental process that hadproduced the experienced, specialised and cooper-ative labour force that Britain possessed (Mass andLazonick, 1990, p. 8).

Similar arguments apply to the machine-buildingindustry and to mill and machine design. Whereas theearly mill-wrights came from the earlier tradition ofcorn-mills, wind-mills, etc., with the increased spe-cialisation and sophistication of machinery, specialand cumulative skills became increasingly importanthere too. All of this led to high levels of machineutilisation as well as lower initial costs of machinery.

Again in relation to material costs, the highly con-centrated Liverpool cotton exchange provided Lan-cashire with an exceptional advantage. Foreign buyersfound it cheaper to buy in Liverpool than anywhereelse in much the same way that the Amsterdam flowermarket re-exports to the entire world today. TheManchester ship canal and the railway from Liverpoolfrom 1830 onwards meant that transport costs forLancashire were extremely low. Lancashire spinnerscould avoid the heavy warehousing costs of moredistant competitors. It was not quite “just-in-time”but was well in that direction. The Liverpool marketgave Lancashire enormous flexibility in grades andtypes of cotton and spinners took advantage of pricechanges on a weekly basis. Furthermore, Lancashire


had a unique capability to work with inferior gradesof cotton for any market in the world and even tocope with the partial switch to Indian cotton at thetime of the American Civil War.

The world-wide marketing structure was yet ano-ther cumulative advantage of the Lancashire industry,which like all the other factors mentioned providedexternal economies for the firms in the industry.The structure of the Lancashire industry itself withthe very well-informed merchants, converters andfinishers meant that it had the capability to deliverwhatever product the customer demanded to any partof the world rapidly. Again, inventory, transport andcommunication costs could be kept low through thisindustry-wide advantage.

Similar economies of agglomeration applied toother industries such as pottery (Staffordshire), cut-lery (Sheffield), hosiery (Nottinghamshire) or wool(Yorkshire). There is no doubt that these sub-nationalsystems of innovation or industrial districts, asMarshall called them, made major contributions tothe success of the industrial revolution in Britain.However, it by no means follows that the nationalsystem was unimportant or that it was simply the sumof the sub-national systems. Each of the industrialdistricts could flourish not only because of the spe-cialised local advantages and institutions (pools ofskilled labour, exchange of experience, trade associ-ations, etc.) but also because ofnational advantagesconferred by British political, cultural, economic andtechnological institutions. Easy access to a large andrapidly growing domestic market as well as to for-eign markets, access to the capital market and a legalsystem which protected property and its accumula-tion, and access to a national pool of engineeringand scientific knowledge. In fact, the national andsub-national systems complemented each other.

It is hard to believe that the British industrial rev-olution would have been more successful if Britainhad been divided into 20 or 30 separate states, citiesand principalities as Germany and Italy then were.In fact, Friedrich List and most of those concernedto catch-up with Britain in the 19th century advo-cated a confederation of German states preceded bya Customs Union (Zollverein) and bound together bya national railway network and other national institu-tions because they perceived the many advantages ofa unitary nation-state.

The national advantages of Britain which comple-mented the specialised sub-national industrial districtswere admirably summed up by Supple (1963):

Britain’s economic, social and political experiencebefore the late 18th century explains with rela-tively little difficulty why she should have been anindustrial pioneer. For better than any of her con-temporaries Great Britain exemplified a combina-tion of potentially growth-inducing characteristics.The development of enterprise, her access to richsources of supply and large overseas markets withinthe framework of a dominant trading system, theaccumulation of capital, the core of industrial tech-niques, her geographical position and the relativeease of transportation in an island economy withabundant rivers, a scientific and pragmatic her-itage, a stable political and relatively flexible socialsystem, an ideology favourable to business andinnovation—all bore witness to the historical trendsof two hundred years and more, and provided mucheasier access to economic change in Britain than inany other European country (Supple, 1963, p. 14).

4. The United States national system of innovation

The economies of scale achieved for British firmsand the British industrial districts by the removal ofinternal trade barriers and by British trading and navalsuperiority were even more important in the rise ofthe United States economy. During the second half ofthe 19th century and the first half of the 20th centuryit was the United States economy which grew muchmore rapidly than any other (Table 6).

Not surprisingly, the country whose “national sys-tem of innovation” most closely resembled the Britishsystem in the 18th century were the former Britishcolonies of USA. However, in the first half of the 19th

Table 6Relative productivity levels (US GDP per hour= 100)a

1870 1913 1950

UK 104 78 57France 56 48 40Germany 50 50 30Fifteen countries 51 33 36

a Source: Abramovitz and David (1994).


century, despite a rich endowment of natural resourcesand many favourable institutions, growth was stillretarded by the lack of an appropriate transport infra-structure to take advantage of the natural endowmentand size of the country and its market. It was the ad-vent of railways and the new technologies of the late19th century which enabled American entrepreneursto forge far ahead of the rest of the world. At first,the United States imported much of this technologyfrom Europe. Many of the key inventions in cottonspinning and weaving were smuggled out of Britainand across the Atlantic by British craftsmen as it wasthen illegal to export this technology. Arkwright’swater frame was an example of a machine that wascarefully memorised and then re-built in America. Butfrom the very beginning American inventors modi-fied and reshaped these technologies to suit Americancircumstances. By the end of the century Americanengineers and scientists were themselves developingnew processes and products in most industries, whichwere more productive than those in Britain.

As we have seen, among those institutions mostfavourable to economic growth in Britain were thescientific spirit pervading the national culture and thesupport for technical invention. These features werereadily transferred to the United States and respectfor science and technology has been an enduring fea-ture of American civilisation from Benjamin Franklinonwards. As de Tocqueville observed in his classicon “Democracy in America” (1836):

In America the purely practical part of science isadmirably understood and careful attention is paidto the theoretical position which is immediately req-uisite to application. On this head the Americansalways display a free, original and inventive powerof mind (De Tocqueville, 1836, p. 315).

The early immigrants were obliged as a matter oflife and death, to learn by doing about agriculturaltechniques in the American Continent and agricul-tural research emerged early as an outstanding featurewith strong public support. Whereas in Europe, withthe partial exception of Britain, feudal institutionsoften retarded both agricultural and industrial deve-lopment, the United States never had any feudal in-stitutions either in agriculture or any other part of theeconomy. Moreover, the relative abundance of land,the westward-moving frontier, the destruction of the

native civilisations or their confinement to a relativelysmall part of the territory, all favoured a purely capi-talist form of economic development with a relativelyegalitarian distribution of income and wealth amongstthe white immigrants in the early period.

The big exception to these generalisations was ofcourse the slave economy of the South. It is difficultto assess the degree to which the economic growth ofthe South in particular and of the Union in general wasretarded by the prevalence of this slave economy butit was in the period which followed the victory of theNorth in the Civil War that the United States achievedrates of growth well above any previously achievedby Britain. This was a case of the sub-national sys-tem of the Southretardingnational economic growthuntil the victory of the Union and the major institu-tional changes which ensued, especially the abolitionof slavery. Even after its abolition, slavery left anenduring legacy of social and economic problems,some of which persist to this day but the maintenanceof the Union meant that the predominantly capitalistpath of development in the North and West prevailedin the whole country. In these circumstances an en-trepreneurial culture could flourish as nowhere else.

Historians such as Abramovitz and David (1994),who have examined American economic history afterthe Civil War point to several characteristics of theUnited States economy which were in combinationexceptionally favourable to a high rate of economicgrowth. These were (i) resource abundance in bothland, minerals and forests; (ii) an exceptionally largeand homogeneous domestic market facilitating pro-duction, marketing and financial economies of scale,especially in the extractive, processing and manu-facturing industries.

Abramovitz and David argue that the higher rela-tive price of labour in North America interacted withthese advantages to induce substitution of capital andnatural resource inputs for skilled labour. This stim-ulated already in the first half of the 19th century thedevelopment of a specific American labour-saving,capital-intensive technological trajectory of mecha-nisation and standardised production, which at thelower end of the quality range had enabled US manu-facturing to surpass British productivity levels alreadyby 1850. As the 19th century advanced, “the engi-neering techniques of large-scale production and highthroughput rates became more fully explored and


Table 7Comparative levels of capital/labour ratios 1870–1950 (USA= 100)a

Germany Italy UK Average of13 Europeancountries

Japan

1870 73 – 117 – –1880 73 26 106 68 121913 60 24 59 48 101938 42 32 43 39 131950 46 31 46 39 13

a Source: Abramovitz and David (1994, p. 8).

more widely diffused. American managers becameexperienced in the organisation, finance and operationof large enterprises geared to creating and exploitingmass markets (Abramovitz and David, 1994, p. 10).

The extent to which this specific American tra-jectory of tangible capital-using technology divergedfrom that of Europe (and Japan) can be clearly seenfrom Table 7. Until the 1880s the UK still had an over-all capital/labour ratio higher than that of the UnitedStates but by 1938, like all other countries, the ratiohad fallen to less than half that of the United States.The large cost reductions and productivity gainsassociated with this North American technologicaltrajectory could be illustrated from numerous indus-trial sectors. The extraordinary productivity gains inmining and mineral processing are emphasised in par-ticular by Abramovitz and David, whilst the produc-tivity gains in agriculture are very frequently cited byother historians. The examples ofsteelandoil are par-ticularly noteworthy because of the key role of thesecommodities in all kinds of tangible capital-usinginvestment projects, in capital goods themselves, intransport and in energy production and distribution.Cotton, iron, canals and water-power were the leadingsectors in the early British industrial revolution; oil,steel and electricity were the leading sectors in thehuge American spurt of growth from 1880 to 1913.

Viotti (1997) points out that these differences andthe case of the chemical industry in Germany meanthat catching up in the late 19th century and early20th century was rather different from what it hasbecome in the late 20th century. The United Statesand Germany, he suggests caught up byradicalinnovations in new industries, not by incrementalinnovations in cotton spinning and weaving. Today,

the late-comer countries may not have the option ofradical innovations in new industries and have no al-ternative but to pursue the path of imitation andlearn-ing. However, he makes a distinction betweenactivelearning systems andpassivelearning systems, takingthe examples of South Korea and Brazil to illustratehis point. We shall pursue this example in Section 5.

5. Late-comer catch-up in the 20th century

Very large economies of scale were characteristicof the forging ahead process in the United States,especially in steel, chemicals, oil, minerals and elec-tricity. Even after the Second World War, when theOECD (previously OEEC and ERP) organised manyEuropean missions to study the productivity gapsbetween European and American firms, they fre-quently stressed scale of plant and size of domesticmarket as two of the biggest comparative advantagesof US firms. This kind of thinking lay behind muchof the political impulse to establish first the EuropeanCoal and Steel Community and later the EuropeanCommon Market (Customs Union) and the EuropeanEconomic Community. Just as the German national-ists, following Friedrich List, believed that a GermanCustoms Union would greatly facilitate catch-up withBritain, so the European federalists believed that aEuropean Common Market would accelerate Euro-pean catch-up with the United States. This philosophystill influences the debate on EMU today.

Catch-up by Western Europe did indeed take placebetween 1950 and 1975 although it certainly cannotbe attributed uniquely to scale economies and mar-ket enlargement. As with the first British industrialrevolution, a general capability for institutional andtechnical change was essential and not merely scaleeconomies. European research and development activ-ities, technology transfer, education and training, andmanagement techniques were all greatly improved.Investment by the United States firms in Europe andby European firms in the USA also facilitated thetransfer of technology and management techniques.All these things were necessary even to achieve thescale economies themselves.

Not surprisingly, however, economists who wereinterested in catch-up by “late-comer” countries wereespecially impressed by the scale economies of North


American and European firms and when they cameto study catch-up phenomena, they stressed this pointin particular. Gerschenkron (1962, 1963) studied thecatch-up by 19th century German and later Russianfirms in the steel industry and argued that the new(late-comer) firms could acquire and use the latesttechnology, at much lower costs than those in thepioneering countries, by transfer agreements, inwardinvestment and the recruitment of skilled people.Even more important in his view was the fact thatthe pioneering firms and countries had already estab-lished a growing world market so that the catch-upfirms did not have to face all the uncertainties, costsand difficulties of opening up entirely new markets.Gerschenkron’s theory oflate-comer advantagesstressed that the pioneers could not possiblystartwithlarge plants, whereas the late-comers could move veryrapidly to large-scale production, while their maturecompetitors might be burdened with smaller plantsembodying now obsolete technology.

Jang-Sup-Shin (1995) in his study of theKoreansteel industry endorsed this Gerschenkronian explana-tion, pointing out that Posco, the largest Korean steelfirm was able to leapfrog European and Americanfirms with respect to plant size and technology andthus enter the world market as a low-cost producer.He extended this analysis to the case of the semi-conductor industry arguing that here too the plant-scaleadvantages of the Korean producers of memory chipsenabled them to leapfrog the European semi-conductorindustry and to compete with the most successfulJapanese and American firms. However, he acceptedthat the Gerschenkron late-comer scale economyadvantages had to be complemented by a “nationalsystem of innovation” explanation of successfulcatch-up since neither in the case of steel, nor in thecase of semi-conductors could catch-up have takenplace without many institutional changes, especiallyin education, training and R&D.

Bell and Pavitt (1993) pointed to another prob-lem with Gerschenkronian catch-up theory: a countrywhich simply instals large plants with foreign tech-nology and foreign assistance will not experiencethe build-up in technological capability over severaldecades, which has been characteristic of the lead-ing countries. Consequently, below capacity workingand low output capital ratios have often persisted indeveloping countries.Active learning policies of the

type prescribed by Bell and Pavitt, by Viotti and byAlice Amsden, will be essential to overcome thisdisadvantage of late-comers.

Perez and Soete (1988) provided a more generaltheory of the science and technology infrastructureneeded for effective catch-up. They showed that eventhe costs ofimitation could be rather high in theabsence of an infrastructure which is taken for grantedin mature industrialised countries. Even more impor-tant, they showed that these costs would vary system-atically at different stages of evolution of a productor a technology. Thus, while Gerschenkron could beregarded as the leading theorist of late-comer advan-tages, Perez and Soete reflected the experience innumerous developing countries of late-comerdisad-vantagesand difficulties in catching up the leaders intechnology. However, they did also point to “windowsof opportunity” in the acquisition and assimilationof technologies, provided the catch-up countries fol-lowed appropriate social, industrial and technologypolicies.

Gerschenkron himself argues that countries couldonly enjoy late-comer advantages if they could alsomake innovations in theirfinancial system so thatthe huge scale of investment needed for very largeplants could be accommodated. However, there isanother important point about late-comer advantageswhich Gerschenkron did not sufficiently explore:the plant scale economies of a particular histori-cal period were not necessarily characteristic ofallindustries or of other periods. As Perez and Soeteshowed, scale economies are industry-specific andtechnology-specific. In a number of industries, suchas aircraft or drugs, scale economies indesignanddevelopmentcosts were much more important thanplant-scale economies inproduction. In still otherindustries, scale economies inmarketing may bedecisive. In the steel industry itself as well as insemi-conductors, plant scale economies have changedwith technology.

Nevertheless, Gerschenkron’s theory of late-comeradvantages was an important complement to infantindustry arguments. The much higher growth ratesin some catch-up countries are obviously attributablein part to the fact that costs of technology acquisitionand implementation are lower and the risks and un-certainties are less in catch-up situations. Imitation isusually easier and less costly than innovation. A very


big gap in technology does provide apotential forfast catch-up.

Finally, as traditional international trade theory hasalways stressed, late-comer countries will usually en-joy labour cost advantages and these may be very largenow because of the wide disparities in world-wideper capitaincomes. These labour-cost advantages mayalso be reinforced by lower costs of particular materi-als and of energy, as well as climatic advantages. Allof these things as well as the narrow and the broadnational innovation systems affect the potential forcatch-up and its realisation.

It also seems to be the case that geographical andcultural proximity to nations which have led either inforging ahead or in catching up has a considerable ef-fect on rate of catch-up. It would be difficult otherwiseto explain the clear-cutinter-continentaldifferenceswhich are apparent. Britain was first caught up andovertaken by neighbouring European countries andby countries partly populated by British and otherEuropean immigrants (the United States, Canada,Australia, etc.). The most successful catch-up coun-tries in the late 20th century have been those whichare geographically (and in some respects culturally)close to the leading catch-up country of the 20thcentury—Japan.

Thus, it is not altogether surprising that East Asiancountries and some South-East Asian countries grewmuch faster than Latin American countries in the1980s and 1990s, despite the fact that the Asiancountries started from a much lower level of industri-alisation and productivity than most Latin Americancountries (Table 8). This would appear at first sight

Table 8Starting levels for industry, Latin America and Asia 1955a

Ratio of manufacturingto agricultural netproduct

US$ net value ofmanufacturing percapita

Argentina 1.32 145Brazil 0.72 50Mexico 1.00 60Venezuela 1.43 95Colombia 0.42 45South Korea 0.20 8Thailand 0.28 10India 0.30 7Indonesia 0.20 10

a Source: Maizels (1963).

to support a simple convergence theorem—the laterthe faster. But before jumping to any such conclu-sion it is essential to recognise that not all late-comercountries were catching up; some were falling furtherbehind and some were standing still. The countriesand sub-continents making the fastest progress haveactually varied enormously both in 19th and 20thcenturies. Uneven development is a much more ac-curate characterisation of growth than convergence.In a well-known paper, De Long (1988) showed thatBaumol’s (1986) attempt to establish the convergencetheorem using Maddison’s data was fundamentallyflawed:

. . . when properly interpreted Baumol’s finding isless informative than one might think. For Baumol’sregression uses anex postsample of countries thatare now rich and have successfully developed. ByMaddison’s choice, those nations that have notconverged are excluded from his sample becauseof their resulting persistent relative poverty.Con-vergence is thus all but guaranteed in Baumol’sregression, which tells us little about the strengthof the forces making for convergence among na-tions that in 1870 belonged to what Baumol callsthe “convergence club.”. . . Because [Maddison]focuses on nations which (a) have a rich data basefor the construction of historical national accountsand (b) have successfully developed, the nations inMaddison’s sixteen are among the richest nationsin the world today (De Long, 1988, pp. 1138–1139,emphasis added).

De Long showed that for another sample of coun-tries, drawnex ante, there was no correlation betweenstarting level of productivity and subsequent growthperformance.

Following up De Long’s critique of convergencetheory, Chabot (1995) showed that for the period1820–1870 the capacity for income growth varieddirectly not inverselywith countries’ 1820 level ofGDP per capita: “The result is directly antithetical towhat convergence theory predicts. The case is demon-strated remarkably well by two countries, the UnitedStates and the UK. Both were amongst the group ofthree to four most promising nations in 1820 and bothexperienced phenomenal subsequent growth rates.Clearly something more than the logic of convergenceand catch-up was at work.” (p. 60). Sections 3 and 4


Table 9Falling behind-rates of divergence or convergence towards the USlabor productivity level, 1870–1950a

Country Rate

Australia −0.95Austria −0.81Belgium −0.84Canada 0.10Denmark −0.46Finland −0.20France −0.37Germany −0.88Italy −0.37Japan −0.30The Netherlands −0.87Norway −0.15Sweden 0.04Switzerland −0.14UK −0.77

a Source: Maddison (1995), as quoted in Chabot (1995).

of this paper have attempted to show that the “some-thing more” was specific national innovation systems.

Chabot further showed that for the period 1870–1950, forging aheadby the United States in labourproductivity took place in relation to all other OECDcountries except the Nordic countries (Table 9). Onlyfor the period 1950–1992 did he find some limitedsupport for convergence and he concluded

in the whole of modern economic history, there isonly one period which offers any substantive sup-port for international convergence, and even that issubject to non-trivial qualification (p. 62).

Although Gerschenkron was undoubtedly right todetect some major advantages for late-comer countriesand to show thatin some circumstanceslate-comers,starting from a very low level of productivity couldenjoy growth rates much higher than the establishedleading countries, it certainly does not follow thatlate-comers willalwaystend to converge with the lead-ers. Whether they do so or not depends, as was sug-gested at the outset, on social capability for technicaland institutional change, i.e. on the national systemsof innovation and on the nature of the new waves oftechnology pervading the system. It also depends on afavourable conjuncture of international relationships.

Whilst it is true that most Asian countries were“later” late-comers than most Latin American coun-

tries in their industrialisation, their faster growth in the1980s and 1990s cannot be explained by this alone.Nor indeed can it be explained as Krugman (1994) at-tempted to do, primarily by fast growth of capital andlabour inputs, with technical change playing very littlepart. Neither is the World Bank Report (1993) on the“East Asian Miracle” entirely convincing in its expla-nation of the “miracle”. Both Krugman and the WorldBank Report greatly underestimate the role of techni-cal change and in particular they fail to recognise thedegree to which the “Tigers” since the 1970s concen-trated their investment, their R&D, their infrastructuraldevelopment, their training and their technology inputon the electronic and telecommunication industries.“Catch-up” is not a spontaneous process associatedwith late-coming, nor the inevitable outcome of marketforces in a liberal economy. By a combination of goodfortune and of good judgement, as well as proximityto Japan, the Tigers (and later some other Asian coun-tries) increased their commitment to the manufactureand export of electronic products at a time when theywere the most rapidly growing part of world exports.

Clearly, the huge achievements of the Tigers in ex-port performance in the 1980s and 1990s were greatlyfacilitated by the commodity composition of their ex-ports and imports, both of which were heavily relatedto the fastest growing sectors of the world economyand of world trade (Table 10). Singapore is the mostextreme example of this with electronic and telecom-munication products accounting for over a third oftotal industrial production and nearly two-thirds ofexports by the mid-1990s. Hong Kong, however,

Table 10Exports of ICT equipmenta to major OECD countries as a shareof total manufacturing exports to those countries from each of thefollowing countries

1970 1980 1992

Japan 21 17 29Germany 7 5 6USA 14 15 22France 6 5 6The Netherlands 6 7 8UK 6 7 13Taiwan 17 16 28South Korea 7 13 26Singapore 20 36 65

a Computers, office machines, telecommunication equipmentand other electronic equipment; source: OECD trade data base.


Table 11Indicators of national efforts in S&T-Brazil, South Korea, Japan and the United Statesa

Brazil South Korea Japan The United States

Scientists and engineers engaged inR&D (per million population)

235b (1993) 1990 (1992) 5677 (1992) 3873 (1988)

Expenditure for R&D as percentage of GNP 0.4c (1994) 2.1 (1992) 3.0 (1991) 2.9 (1988)Expenditure in R&D by source of funds (%) 1994c 1992 1992 1992

Government 81.9 17.2 19.4f 43.3f

Productive enterprise 18.1 82.4 71.0f 51.6f

Other – 0.4 9.6f 5.1f

Total patents granted in 1991 2479d 3741e 36100f 96514f

Patents granted to residents (%) 14d 69e 84f 53f

Number of tertiary students per100,000 inhabitants (1992)

1079 4.253 2340 5652

a Source: Viotti (1997). Numbers between parenthesis correspond to the year of the data.b Computed from MCT (1995).c CNPq (1995, p. 23).d MCT (1996, p. 37).e Lall (1992a, p. 197), the statistics of patents for South Korea are for the year 1986.f STA/NISTEP (1995).

transferred much of its electronics manufacturing toChina in the 1980s and 1990s. Elsewhere in Asianetworks of interdependent suppliers of electroniccomponents, sub-systems and final products were thetypical feature of rapidly growing manufacturing in-dustries. Both imports and exports of these productsgrew extremely rapidly. Not only did the World BankReport neglect the pattern of structural change in theEast Asian economies, it also neglected the changingpattern of research and development activities and oftechnology transfer.

The World Bank Reportdid recognise the im-portance of education. However, it differed fromthe earlier work of institutional economists such asAmsden (1989) and Wade (1990) in neglecting therole of government policy in many other areas, suchas the protection of certain industries, the promotionof exports, the subsidies to specific firms and thepromotion of R&D activities. In general, it ignoredindustrialpolicy and had little to say about training ortechnology transfer. In all these respects, the policiesof South Korea, Taiwan and Singapore have resembledthose prescribed long ago by Friedrich List. In point-ing this out, Freeman (1993) made some comparisonsand contrasts between Brazil and South Korea.

This comparison has been made much more thor-oughly since then by several Brazilian scholars(Villaschi, 1993; Albuquerque, 1997b; Viotti, 1997).

Albuquerque distinguishes several different categoriesof “national innovation systems”, while Viotti makesan interesting distinction between “Active” and“Passive” learning systems building on the earlierwork of Villaschi (1993) and Bell and Cassiolato(1993) on “learning to produce and learning toinnovate”. All of these accounts emphasise the roleof active policiesat the national and firm level in theimport, improvementand adaptation of technology ascharacteristic of successful catch-up.

Viotti (1997) presents impressive evidence of thecontrasts between Brazil and South Korea in thefields of education, and of research and development(Table 11). In presenting this evidence, he pointsout that the number of tertiary students per 100,000inhabitants in Brazil (1079) is approximately just afourth of that of South Korea (4253) and the concen-tration onengineeringis much greater in the Koreantertiary education system compared with both scienceand humanities. Amsden (1989) in her classic workon Korean catch-up, gave some vivid examples of therole of training in the steel and cement industries inthe early period of catch-up. This has become evenmore important in the electronic industry. For exam-ple, Samsung not only trains thousands of its ownemployees at every level but also regularly providesintensive training for its sub-contractors. Such inten-sive training efforts are essential for anactivepolicy


Table 12Emerging sources of technology in terms of ownership of US patentsa

Patents granted during

Year 1977–1992 Share 1983–1989 Share 1990–1996 Share

Taiwan 382 0.10 2292 0.41 11040 1.43South Korea 70 0.02 580 0.10 5970 0.77Israel 641 0.16 1507 0.27 2685 0.35Hong Kong 272 0.07 633 0.11 1416 0.18South Africa 491 0.12 699 0.12 787 0.10Mexico 245 0.06 289 0.05 314 0.04Brazil 144 0.04 212 0.04 413 0.05PR China 7 0.00 142 0.03 353 0.05Argentina 130 0.03 135 0.02 187 0.02Singapore 17 0.00 65 0.01 337 0.04Venezuela 51 0.01 122 0.02 192 0.02India 56 0.01 96 0.02 204 0.03East and Central Europe 3444 0.87 2417 0.43 1317 0.17

Sub-total 5950 1.51 9189 1.62 25215 3.26

Others 731 0.19 902 0.16 1494 0.19

Total 393629 100.00 565739 100.00 772927 100.00

a Source: Kumar (1997) based on data presented in US Patents and Trademarks Office (1997) TAF special report: all patents, alltypes—January 1977–December 1996, Washington, DC.

of continuous improvement of technology even morethan for simple imitation. Whereasstatic economiesof scale can be achieved simply by building a largerplant, the far greaterdynamic economies of scaledepend on such active learning policies and increas-ingly on engineering activities and in-house R&D atplant level, as shown so clearly by Bell and Cassio-lato (1993) and by Bell and Pavitt (1993). The rapidchange of product design and the associated changeof components characteristic of the electronic indus-try have further increased the importance of theseactive learning and training policies.

Viotti has further shown that Korean industry hasrelied far less on foreign investment than Brazil andhas relied more on imports of capital goods and onincreasingly active efforts to improve upon importedtechnology through thein-house R&D efforts ofKorean firms themselves. Hobday (1995) has pointedto the variety of strategies in the East Asian countries,all designed in different ways steadily to upgradelocal technological capability. The contrast betweenthe rapid rise in the performance of in-house R&Din firms in both South Korea and Taiwan, and its lowlevel, stagnation or non-existence in firms in most

developing countries is especially notable. Table 11illustrates the contrast with Brazil.

The rise of in-house R&D in the 1970s led to an ex-traordinary increase in numbers of patents (Table 12)and this is perhaps the most striking confirmation oftheactivelearning system in South Korea and Taiwan.Between 1977–1982 and 1990–1996, the number ofpatents taken out in the USA by Brazil, Argentina,Mexico and Venezuela nearly doubled from 570 to1106, but in the same period the numbers taken outby the four “Tigers” increased nearly 30 times over,from 671 to 18,763.

In the case of South Korea nearly half of the patentstaken out between 1969 and 1992 were in six tech-nical fields in the electronics area (semi-conductors,computers, imaging and sound, instruments, electri-cal devices and photocopying). In the case of Taiwanthe increase in total numbers of patents was evenmore remarkable but they were spread over a widertechnical area, including especially metal productsand machinery as well as electrical and electronic de-vices (Choung, 1995). Firms in both South Korea andTaiwan were so successful in their catch-up that theybegan to export technology themselves and to invest


overseas in older industrial countries like Britain aswell as in the less developed countries of South-EastAsia.

Although very useful for international comparativepurposes,United Statespatent statistics are of coursenot the only ones which are significant. Since theestablishment of the European patent, the EuropeanPatent Office (EPO) statistics are increasingly valu-able. Even more interesting for catch-up countries arethe measures ofdomesticpatenting in each individualcountry. In a pioneering paper, Albuquerque (1997a)analyses these statistics and more interestingly fromthe standpoint of this paper shows that the number ofBrazilian domestic patents have shown no clearly in-creasing trend in the last 15 years after reaching a peakin 1981–1982.

The extraordinary growth of East Asian patentingin the USA is not of course only an indication of ac-tive technology improvement. It is also influenced bythe commercial and export strategies of firms Hobday(1995) has particularly stressed the key role of learn-ing by exporting in East Asian catch-up. The successof East Asian firms in exports complements theirsuccess in obtainingUnited Statespatents. Moreover,insofar as they continue to make progress towards theworld technological frontier, they have encounteredincreasingly severe competition from the establishedtechnological leaders and from the pressures of theseleaders to liberalise all aspects of their economic,financial and trading systems.

The dominance of the United States firms in theworld software industries and in all the supporting ser-vices for the Internet and in world financial servicespresents some especially acute problems for thosecatch-up countries which are making good progressin closing the gap in manufacturing.

The liberalisation of capital movements which hasalready taken place exposes all countries to the in-stability and shocks which occur in any part of thesystem. Events in the late 1990s have shown that theseshocks can be propagated throughout the system andthat however good the narrow innovation system itis always still part of a broader global economic andpolitical system. Trends in the political, cultural andeconomic sub-systems are influenced very stronglyby institutions which are only tenuously related toscience and technology. These points are furtherdiscussed in Section 7 of this paper.

6. Conclusions

Section 3 of this paper suggested that it was fromthe time of the British industrial revolution in the18th century that fullnational systems of innovationemerged. Bairoch’s (1993) estimates of the wideninggap between developing countries and industrialisedcountries show this gap growing very rapidly fromthis time onwards and Sections 3 and 4 of the pa-per attempted to show that this huge gap between the“forging ahead” countries and the rest could best beexplained in terms of such concepts as the “nationalsystems of innovation”. Section 5 further showed thatthis concept is also necessary to explain varying ratesof late-comer catch-up and the failure of late-comercatch-up in some circumstances.

Whilst economic historians have had few difficul-ties in accepting the crucial importance of technicaland institutional change in the theory of economicgrowth, the growth modellers have found this a con-tinuing problem. Some early growth models put themain emphasis on accumulation of (tangible) capi-tal through investment and the growth of the labourforce and left all other influences to be subsumedin a “residual factor”. Although the so-called “NewGrowth Theory” broke away from this tradition andmoved “intangible investment” in education, researchand development to the centre of the stage, the oldapproach lingers on, as can be seen for example inKrugman’s (1994) attempt to debunk the “East AsiaMiracle” and to explain East Asian growth mainly interms of capital and labour accumulation.

The treatment of all the complexities of technicaland institutional change in early growth models camein for heavy criticism at the time both from historiansand from many economists, especially as most of themodels showed that the “Third Factor” apparently ac-counted for most of the growth. Balogh (1963) dubbedthe “Third Factor” the “Coefficient of Ignorance”while Supple (1963) concluded that “it must surely beclear that any discussion of the relationship betweencapital formation and economic growth necessarily en-tails the appraisal of a host of other issues. And these intheir turn lead to the conclusion that the accumulationof capital is in itself by no means the central aspect ofthe process of economic growth” (Supple, 1963, p. 22).

In response to this criticism various attempts weremade to disaggregate the residual factor in the aggre-


gate production function, notably by Denison (1962,1967), who used what Dosi (1988) described as an“entire Kama-Sutra of variables” in his efforts tomake systematic comparisons of growth rates. Yetnone of these efforts could survive the trenchant crit-icism of Nelson (1981) and others who pointed to thecomplementarityof all these variables. The contribu-tion of capital accumulation to growth depends notonly on its quantity but on itsquality, on the directionof investment, on the skills of entrepreneurs and thelabour force in the exploitation of new investment, onthe presence (or absence) of social overhead capitaland so forth.

A brave and highly original contribution to thegrowth modelling debate came from Adelman (1963).She recognised early on that the assumption of con-stant returns to scale in many models raised big prob-lems and in the so-called “New Growth Theory” thisassumption has been dropped in favour of Young’s(1928) increasing returns to scale (Romer, 1986;Grossman and Helpman, 1991).

These models usually also follow her in attemp-ting to assign a specific role to technical change (or asshe termed it “the stock of knowledge from appliedscience and technology”). In her model Irma Adelmanalso separated “Natural Resources” from other formsof capital in much the same way as the classicaleconomists separated land. This distinction is likelyto become increasingly important with the growingrecognition of the importance of ecological factorsand resource conservation in economic growth. Shealso separated technical change from other forms ofinstitutional change. Thus, she specified the produc-tion function as

Yt = f (KtNtLtStUt )

whereKt denotes the amount of the services of thecapital stock at timet; Nt stands for the rate of use ofnatural resources;Lt represents the employment of thelabour force;St represents “society’s fund of appliedknowledge”;Ut represents the “social-cultural milieuwithin which the economy operates” (Adelman, 1963,p. 9).

Adelman was also unusual in her frank recognitionof the immense difficulties in the production functionapproach and of the interdependence of her variables.For example,

. . . both the quality and the composition of thelabour force vary through time and are not indepen-dent of the rates of change of the other variables inthe system. Specifically, changes in the skills andhealth of the labour force are directly dependentupon changes in society’s applied fund of technicalknowledge (St ).

Like other modellers, she suggests that the con-ceptual problems “which arise from the heterogene-ity and incommensurability of the production factorsmay be reduced somewhat if we think of each inputas a multi-component vector rather than as a singlenumber”.

However, this is still not the greatest difficulty withthe production function approach. Again, as IrmaAdelman so clearly points out:

Even more difficult than the measurement problemsraised by these production factors are those posedby an attempt to quantify our last two variables.St

and Ut represent heuristic devices introduced pri-marily for conceptual purposes. . . At some time inthe future a method may be evolved for the ordinalevaluation ofSt andUt but such a method does notnow exist and accordingly neither variable can beused as an analytical tool (pp. 11–12).

This situation has scarcely changed since 1963 despitesome advances in the measurement of R&D and ofeducation and despite the somewhat greater realismabout technical and institutional change in the morerecent “new” growth models (Verspagen, 1992).

All of this does not mean that the modelling at-tempts and developments in growth theory of the pasthalf century have been a complete waste of effort.Adelman’s argument for the heuristic value of growthmodelling still stands and her own attempt to useher production function to illustrate the differencesand similarities in the growth theories respectivelyof Adam Smith, Ricardo, Marx, Schumpeter, Harrod,Kaldor and the neo-Keynesians is an excellent exam-ple of these heuristic advantages. But when all is saidand done the main conclusion of the whole debatehas been to vindicate the contention of many eco-nomic historians and neo-Schumpeterian economiststhat technical change and institutional change are thekey variables to study in the explanation of economicgrowth.


In his fairly sympathetic treatment of neo-classicalgrowth theory Gomulka (1990) concluded that

The cumulative effect of the theoretical and empir-ical work has been to highlight more sharply andwidely than ever before how really central is therole, in long-term economic growth, of the activi-ties producing qualitative change in the economy.Technological changes have assumed the primaryrole by virtue of their being typically the originalimpulses which tend to initiate other qualitativechanges. By the same token, the work has alsohelped to delineate the very limited usefulness ofthe (standard) growth theory based on the assump-tion that these qualitative changes are cost free andexogenously given (p. 19).

Sections 1–5 have attempted to justify the view thattechnical change and the institutions which pro-moted it played a central role both in the forgingahead process and in the catching up process. Thisis by no means to deny, however, the role of otherinfluences—economic, political and cultural—whichconstitute the “broad” national system of innovation.Although the work on national systems has made apromising start on the very complex task of “puttinghistory back into economics” it has certainly not doneenough to satisfy the critics, many of whom are un-easy about neo-classical growth models, but also lackconfidence in alternative “evolutionary” explanationsof economic growth.

Some of the most promising lines of future researchon national systems would appear to be in the studyof catch-up failure andfalling behind in economicgrowth. In cases such as Britain and Argentina, bothof which slowed down and fell behind in the 20thcentury, many of the explanations offered point to thelack of congruence between various sub-systems ofsociety, social institutions which have been favourableto economic growth in one period of technologicaldevelopment may not be so favourable when there arefundamental changes in technology.

Hughes (1982) has shown that Berlin and Chicagoforged ahead of London in the applications of electricpower in part because of the multiplicity of standardsin Britain and mis-management of the new infrastruc-ture. These kinds of points indicate that various “outof synch” phenomena had emerged towards the closeof the 19th century in Britain—between technol-

ogy and culture, technology and politics, technologyand the economy and even between technology andscience. In the mid-19th century, very few peopleforesaw this relative British decline. Even FriedrichList, the outstanding exponent of catch-up theory onthe Continent of Europe, died believing that Germanycould never overtake Britain. Much later on, in the1960s, the “Dependency” theorists were so impressedby the advantages of the United States and WesternEurope that they thought it impossible for countriesin Asia, Latin America or Africa ever to catch up.

The advantages of fore-runners may indeed appearoverwhelming at first to late-comers. Not only do theyapparently command an unassailable lead in tech-nology, but they also enjoy many static and dynamiceconomies of scale and privileged prestigious posi-tions in world markets. It is for this reason that suc-cessful catch-up is often referred to as a “miracle” (TheGerman and Japanese “miracles” of the 1950s, 1960sand 1970s; the Korean and Taiwanese “miracles” ofthe 1980s and 1990s). But if any process is to beregarded as a “miracle” it should be “forging ahead”rather than catching up.

At the present time (late 1990s) the United Statesappears to have enormous advantages compared withits principal competitors. The successful catch-up ofJapan and other East Asian countries was based ontheir intensive active learning inhardware design,development and manufacturing. Now, however, it isincreasinglysoftware design, development, produc-tion and marketing which is the key to commercialsuccess. Here the United States has some considerablefore-runner advantages. It has by far the strongestsoftware industry in the world with major advantagesin scale economies in business applications. This hasled in turn to English language domination in softwaregenerally and especially on the Internet—a global in-frastructure, dominated by US service providers andcontent providers. Finally, the United States firms alsolead in many other service industries and the victoryof the USA in the Cold War has established the USas the only “super-power” in a military sense. Thispower can be used to protect the interests of US firmsworld-wide including their intellectual property.

It is impossible to predict however how long theseadvantages can be retained despite the tightening ofintellectual property restrictions. Very many countrieshave rapidly growing young software firms including


Eastern Europe, as well as Eastern Asia, LatinAmerica and countries with strong English languagecapability, such as India. Moreover, political andsocial events may predominate over more narrowtechnological and economic factors.

Social scientists face a more complex problem thanbiologists because the “selection environment” con-fronting innovators is not simply the natural environ-ment but also several different sub-systems of humansocieties—scientific, technological, economic, polit-ical and cultural. Each of these has its own uniquecharacteristics and successful diffusion depends onthe establishment of some degree of congruencebetween them.

The growing environmental problems facing thewhole world may also impose a rather different pat-tern of economic and political development than thatwhich has prevailed in the 20th century. The devel-opment of environmentally friendly technologies andtheir universal diffusion may impose a more coop-erative civilisation and an entirely new pattern ofinstitutional change and of knowledge accumulation.The great variety of new possibilities in science,technology, economics, politics and culture meansthat despite the present-day predominance of theUnited States, permanent convergence based on UShegemony is a rather unlikely scenario. Viotti maybe underestimating the feasibility of new clusters ofradical innovations in catch-up countries.

The natural environment confronts all living crea-tures but the accumulation of scientific knowledge andof technological knowledge and artefacts are uniquelyhuman processes even though they may have origi-nated, as with other animals, in the search for food andshelter and the communication associated with thissearch. There are birds and mammals which also makeuse of “tools” in the sense of twigs, branches or stones,but the systematicdesignandimprovementof tools andother artefacts are uniquely purposeful human activi-ties, with their own partly autonomous selection envi-ronment. Economists often use a biological analogy toanalyse the competitive behaviour of firms in a capital-ist economy and the survival of the supposedly “fittest”firms. This is a case of the borrowing back of ananalogy which Darwinian theory originally took overfrom economics. But again the selection environmentwhich confronts firms in their competitive struggle isactually very different from the natural environment

confronting animals and plants and this economicenvironment is itself rapidly changing in ways whichare unique. Finally, the political system and the cul-tural milieu are again uniquely human and powerfullyinfluence the evolution of the economy, as they alsoreciprocally influence the evolution of science andtechnology. Evolutionary theories which dealonlywith the survival offirms(Alchian, 1951) oronlywiththe survival of artefacts or of nations are inadequatefor the study of economic growth (Freeman, 1994).

We have no alternative but to confront the uniquefeatures of human history, even though we may quitelegitimately search for patterns of recurrence and forexplanations of recurrence and of non-recurrence.One of the most obvious unique features is the rateof knowledge accumulation in human societies andthe varying modes of disseminating this knowledgebetween individuals and groups. These features areubiquitous and justify continuous attention by histo-rians of economic growth, searching both for regularpatterns as well as for the emergence of new features.

7. Speculations

In this search, economists and other social scien-tists will need to pay attention to changing systems ofinnovation at all levels—the global level, the continen-tal and sub-continental levels, the national level andthe sub-national level. This paper has concentrated ondevelopments at thenational level in the belief thatthe major phenomena of forging ahead, catch-up andfalling behind in 19th and the 20th centuries can mostplausibly be explained in terms of national systems,albeit in an international context and recognisinguneven development at the sub-national level.

All of this may change in the 21st century. In par-ticular, the capacity to use information and communi-cation technology will probably be a decisive factorin world competition and this in turn will lead tothe dominance of firms and networks with capabilityin serviceactivities. The models which economistshave used have largely been based onmanufacturingactivities, although of course agriculture and serviceshave always been important. Now manufacturing mayincreasingly be located outside Europe, Japan andthe United States. Major manufacturing firms such asGeneral Electric or Ericsson have for some time had


a deliberate strategy of increasing the share ofserviceactivities within their total turnover. Manufacturingemployment has already declined substantially infirms such as these and accounts now for much lessthan half the total, even though they are still oftenthought of as mainly manufacturing organisations.Financial, marketing, software, design and R&D ser-vices may often predominate now in the portfolio oflarge MNCs.

This does not mean that manufacturing will ceaseto be important. It will always be important, just asagriculture will always be important even though itoccupies a small proportion of the labour force. It maymore often be sub-contracted and competitive powerwill increasingly depend on the capability to manageinternational networks in production and marketing,with the core activities of research, design and devel-opment of software and hardware, based mainly in thenational “home” base as long as their base can providethe necessary supporting scientific, technological, ed-ucational, financial and communication infrastructure.

Powerin these networks will depend on a variety ofinformation services and knowledge-based activitiesbut not only on these. These networks are embeddedin social systems where increasing inequality is nowthe rule and to some degree exacerbated by these de-velopments. Both environmental and social problemsare likely to become more acute in these circum-stances. Political and cultural changes may then takeprecedence in the complex interactions between thevarious sub-systems of society at all levels of theglobal system. Breadth, enlightenment and socialsolidarity are essential in the end for any innovationsystem. Otherwise, as Gomulka has suggested, peo-ple may reject the constant turmoil and uncertaintyof contemporary competitive innovation systems andinsist on the primacy of quality-of-life objectives.

References

Abramovitz, M.A., 1986. Catching up, forging ahead and fallingbehind. Journal of Economic History 46, 385–406.

Abramovitz, M.A., David, P.A., 1994. Convergence and DeferredCatch-up: Productivity Leadership and the Waning of AmericanExceptionalism. CEPR Publication No 401. Stanford University,Stanford.

Adelman, I., 1963. Theories of Economic Growth and Develop-ment. Stanford University Press, Stanford.

Albuquerque, E., 1997a. Domestic Patents and DevelopingCountries, SPRU (Brighton) and UFRJ (Rio de Janeiro), Mimeo.

Albuquerque, E., 1997b. National Systems of Innovation andNon-OECD Countries: Notes About A Tentative Taxonomy,SPRU (Brighton) and UERJ (Rio de Janeiro), Mimeo.

Alchian, A., 1951. Uncertainty, evolution and economic theory.Journal of Political Economy 58, 211–222.

Amsden, A., 1989. Asia’s Next Giant: South Korea and LateIndustrialisation. Oxford University Press, Oxford and NewYork.

Bacon, F., 1605. The Advancement of Learning.Bairoch, P., 1993. Economics and World History: Myths and

Paradoxes. Chicago University Press, Chicago.Bairoch, P., Levy-Leboyer, 1981. Disparities in Economic

Development Since the Industrial Revolution. St. Martin’s Press,New York.

Balogh, T., 1963. Comments on the paper by Tinbergen and Bos.In: OECD, The Residual Factor in Economic Growth. OECD,Paris, pp. 180–187.

Baumol, W.J., 1986. Productivity growth, convergence and welfare:what the long run data show. American Economic Review 76,1072–1085.

Bell, M., Cassiolato, J., 1993. The Access of Developing Countriesto New Technology. Mimeo, FECAMP/UNICAMP, Campinas.

Bell, M., Pavitt, K.L.R., 1993. Technological accumulation andindustrial growth: contrasts between developed and developingcountries. Industrial and Corporate Change 2, 157–210.

Cassiolato, J., Lastres, H. (1999) (Eds.), Globalização e InovaçãoLocalizada: Experiências de Sistemas Locais no Mercosul, Ibict,Brasılia.

Chabot, C.N., 1995. Long-run productivity growth, technologicalchange and the international economy. M.Sc. Dissertation.SPRU, University of Sussex.

Choung, Jae-Yong, 1995. Technological Capabilities of Korea andTaiwan: An Analysis Using Patent Statistics. SPRU, Universityof Sussex, STEEP Discussion Paper, No. 26.

DeBresson, C., 1989. Breeding innovation clusters: a source ofdynamic development. World Development 17, 1–6.

DeBresson, C., Amesse, F., 1991. Networks of innovators: a reviewand introduction to the issue. Research Policy 205, 363–379.

De Long, B., 1988. Productivity growth, convergence and welfare:comment. American Economic Review 78, 5.

Denison, E.F., 1962. The Sources of Economic Growth in theUnited States. Committee for Economic Development, NewYork.

Denison, E.F., 1967. Why Growth Rates Differ: Post-WarExperience in Nine Western Countries. Brookings Institution,Washington, DC.

De Tocqueville, A., 1836. Democracy in America. OxfordUniversity Press, Oxford.

Dosi, G., 1988. Sources, procedures and microeconomic effectsof innovation. Journal of Economic Literature 36, 1126–1171.

Edqvist, C., 1997a. Institutions and organisations in the system ofinnovation: the state of the art. TEMA-T, Working Paper ISSN1101-1289. University of Linkøping.

Edqvist, C., 1997b. Systems of innovation approaches—theiremergence and characteristics, Chapter 10. In: Edqvist, C.(Ed.), Systems of Innovation, Technologies, Institutions andOrganisations. Pinter, London.


Foray, D., 1991. The secrets of industry are in the air: industrialcooperation and the organisational dynamics of the innovativefirm. Research Policy 20, 393–405.

Freeman, C., 1993. Technological revolutions and catching-up:ICT and the NICs, chapter. In: Fagerberg, J., Verspagen, B.,von Tunzelmann, N. (Eds.), The Dynamics of Technology,Trade and Growth. Edward Elgar Publishing Ltd., Aldershot,pp. 198–221.

Freeman, C., 1994. The economics of technical change: a cri-tical survey article. Cambridge Journal of Economics 18 (5),463–514.

Freeman, C., 1995. History, Co-Evolution and Economic Growth.Working Paper-95-76. IIASA, Austria.

Gerschenkron, A., 1962. Economic Backwardness in HistoricalPerspective. Harvard University Press, Cambridge, MA.

Gerschenkron, A., 1963. The early phases of industrialisation inRussia. In: Rostow, W.W. (Ed.), The Economics of Take-Offinto Sustained Growth. Macmillan, London.

Gomulka, S., 1990. The Theory of Technological Change andEconomic Growth. Routledge, London.

Grossman, G.M., Helpman, E., 1991. Innovation and Growth inthe Global Economy. MIT Press, Cambridge, MA.

Hobday, M., 1995. Innovation in East Asia: the Challenge to Japan.Elgar, Aldershot.

Hu, Y.S., 1992. Global or transnational corporations are nationalfirms with international operations. Californian ManagementReview 34, 107–126.

Hughes, T.P., 1982. Networks of Power Electrification in WesternSociety, 1800–1930. John Hopkins University Press, Baltimore.

Humbert, M. (Ed.), 1993. The Impact of Globalisation on Europe’sFirms and Industries. Pinter, London.

Jacob, M., 1988. The Cultural Meaning of the ScientificRevolution. McGraw-Hill, New York.

Jang-Sup Shin, 1995. Catching up, technology transfer andinstitutions: a Gerschenkronian study of late industrialisationfrom the experience of Germany, Japan and South Koreawith special reference to the iron and steel industry andthe semi-conductor industry. Ph.D. Dissertation. CambridgeUniversity, Darwin College.

Jenkins, D.T. (Ed.), 1994. The Textile Industries, Vol. 9. In: Church,R.A., Wrigley, E.A. (Eds.), The Industrial Revolution in Britain.Blackwell, Oxford.

Krugman, P., 1994. The myth of Asia’s Miracle. Foreign Affairs71, 62–78.

Kumar, N., 1997. Technology Generation and TechnologyTransfer in the World Economy: Recent Trends andImplications for Developing Countries, 9702. INTECH, UNU,MA.

Landes, M., 1970. The Unbound Prometheus: Technological andIndustrial Development in Western Europe from 1750 to thePresent. Cambridge University Press, Cambridge.

List, F., 1841. The National System of Political Economy, EnglishEdition. Longman, London, 1904.

Lundvall, B.-Å. (Ed.), 1992. National Systems of Innovation:Towards A Theory of Innovation and Interactive Learning.Pinter, London.

Maddison, A., 1995. Monitoring the World Economy, 1820–1992.Development Centre, OECD, Paris.

Maizels, A., 1963. Industrial Growth and World Trade. CambridgeUniversity Press and NIESR.

Mass, W., Lazonick, W., 1990. The British cotton industry andinternational competitive advantage: the state of the debates.Working Paper 90-06. Department of Economics, ColumbiaUniversity, New York.

Mjøset, L., 1992. The Irish Economy in a Comparative InstitutionalPerspective. National Economic and Social Council, Dublin.

Needham, J., 1954. Science and Civilisation in China. CambridgeUniversity Press, Cambridge.

Nelson, R.R., 1981. Research on productivity growth andproductivity differentials: dead ends and new departures. Journalof Economic Literature 19, 1029–1064.

Nelson, R.R. (Ed.), 1993. National Innovation Systems: AComparative Analysis. Oxford University Press, Oxford.

Ohmae, K., 1990. The Borderless World. Harper, New York.Patel, P., 1995. Localised production of technology for global

markets. Cambridge Journal of Economics 19, 141–153.Perez, C., Soete, L., 1988. Catching up in technology: entry barriers

and windows of opportunity. In: Dosi, G., et al. (Eds.), TechnicalChange and Economic Theory. Pinter, New York, pp. 458–479.

Porter, M., 1990. The Competitive Advantage of Nations. FreePress, Macmillan, New York.

. The Role of the State in Economic Growth. Working Paper,1997-5. S.U.M. University of Oslo.

Romer, P., 1986. Increasing returns and long-run growth. Journalof Political Economy 94, 1002–1037.

Supple, B. (Ed.), 1963. The Experience of Economic Growth.Random Publishing House, New York.

Verspagen, B., 1992. Uneven Growth between InterdependentEconomies: An Evolutionary View on Technology Gaps, Tradeand Growth. University of Limburg, Maastricht.

Viotti, E.B., 1997. Passive and Active National Learning Systems.PhD. Dissertation. New School for Social Research.

Villaschi, A.F., 1993. The Brazilian National System of Innovation:Opportunities and Constraints for Transforming TechnologicalDependency. D.Phil. Thesis. University of London, London.

Wade, R., 1990. Governing the Market. Princeton University Press,Princeton.

World Bank, 1991. World Development Report, 1991. OxfordUniversity Press, New York.

World Bank, 1993. East Asian Miracle. World Bank, Washington.Young, A.A., 1928. Increasing returns and economic progress.

Economic Journal 38, 527–542.

continental, national and sub-national innovation systems ... · research policy 31 (2002)...

Documents