trade and development report, 2006

UNITED NATIONS CONFERENCE ON TRADE AND DEVELOPMENTGENEVA

TRADE AND DEVELOPMENTREPORT, 2006

UNITED NATIONSNew York and Geneva, 2006

Chapter V

NATIONAL POLICIES IN SUPPORTOF PRODUCTIVE DYNAMISM

UNCTAD/TDR/2006

National Policies in Support of Productive Dynamism 149

The widening gap in relative income levelsbetween rich and poor countries has been a majortrend in the world economy over the past 250 years.On one estimate, the difference in per capita in-come between the richest and the poorest countryin the world was about 5:1 before the IndustrialRevolution; today this difference has increased to400:1 (Landes, 1998). While the exactitude of thesenumbers is debatable, there can be little doubt thatthe world economy has been on a long-term pathof substantial and growing divergence in relativeproductivity levels and living standards, both be-tween developed and developing countries andamong developing countries themselves.

The slow per capita income growth in de-veloping countries has left millions of people inpoverty. Nevertheless, recently the growth per-formance of many developing countries has im-proved, especially since the beginning of the currentcommodity price boom in 2002. Rapid growth ina few highly populated developing countries, es-pecially China and India, has helped lift a sub-stantial number of people out of poverty, in thesecountries themselves as well as in other develop-ing countries that have benefited from spillovers

of fast growth in Asia. But improved growth per-formance in the developing world will need to bemore broad-based and sustained over a long pe-riod of time if there is to be more substantial progresstowards achieving the Millennium DevelopmentGoals (MDGs) and eradicating poverty.

Around the long-term trend towards diver-gence in relative productivity and per capita incomelevels, a number of initially backward countrieshave succeeded, at different times, in catching upto the productivity and income levels prevailingin the frontier countries. It is well known that thecurrent global technology leader – the UnitedStates – was itself once on a catch-up path withrespect to the then economically and technologi-cally leading country, the United Kingdom. Aus-tralia, Canada, New Zealand, some Latin Americanand many Western European and Scandinaviancountries began catching up about 50 years priorto the First World War. Japan is a prominent ex-ample of catch-up during the decades before andafter the Second World War, as are the East Asiannewly industrializing economies (NIEs) since the1960s (TDR 1997, Part Two, chap. II), and Chinaand India more recently (TDR 2005). Fast growth

Chapter V

NATIONAL POLICIES IN SUPPORTOF PRODUCTIVE DYNAMISM

A. Introduction

Trade and Development Report, 2006150

in these Asian developing countries, particularlyin China, has even succeeded in pulling alongsome of the lagging economies in Latin Americaand Africa.

Explaining this diversity in the pattern ofdevelopment and determining what governmentpolicy can do to help achieve economic catch-upis among the oldest and most controversial issuesin economics. There is no clear-cut policy pre-scription for success, but investment, technologyadoption and structural change have traditionallybeen considered among the main critical factorsfor sustained economic catch-up. Giving supportto earlier findings (Levine and Renelt, 1992), re-cent empirical analyses underline the importanceof investment in physical capital as a “very strong-ly” robust growth determinant (Sala-i-Martin, Dop-pelhofer and Miller, 2004; Tsangarides, 2005). Butfor sustainability of economic growth, it is im-portant that output expansionbe based not merely on capi-tal accumulation, but also ona continuous rise in labour pro-ductivity and the maintenanceof productive dynamism overtime, as is obtained throughthe acquisition of technologi-cal mastery over a broad rangeof activities, especially inmanufacturing. The develop-ment of a strong manufacturing sector has been atthe core of all successful catch-up experiencesover the past 250 years, which suggests thatachieving a lasting productivity-based increase inmanufacturing is indispensable for a sustained risein income levels and, ultimately, the eradicationof poverty.1

Industrialization strategies have varied widelyacross developing countries over the past 50 years.Especially during the 1960s and 1970s, much ofLatin America, Africa and parts of South Asiaemployed import-substituting industrializationstrategies oriented towards the domestic marketand based on a plethora of protective measuresand other government interventions. Many ofthese countries subsequently abandoned thosestrategies for a variety of reasons, including theirfailure to promote development and because ofthe policy conditionality of multilateral lendinginstitutions. Consequently, they began to view

unfettered market forces and deep integration intothe world economy as the most promising meansto economic development during the 1980s and1990s. There is some dispute as to the merits of theimport-substituting industrialization strategy as aparadigm (Bruton, 1998), while the outcome ofthe liberalization strategy is generally judged dis-appointing (TDR 2003; World Bank, 2005). In anycase, the annual rate of real economic growth av-eraged about 2.0–2.5 per cent in Africa and LatinAmerica during the 1980s and 1990s, which isonly about half that of these countries’ growth per-formance during the 1960s and 1970s.

By contrast, the East Asian NIEs recordedan average annual rate of real economic growthof almost 9 per cent during the 1960s and 1970sand more than 7 per cent during the 1980s and1990s. Their successful economic catch-up andindustrialization, in particular until the mid-1990s,

have been associated with out-ward-oriented industrializationstrategies and strategic integra-tion into the world economy.Proactive trade and industrialpolicies2 played a key role inthe pace and direction of struc-tural change and economicgrowth particularly in the Re-public of Korea and TaiwanProvince of China (TDR 1996).

Similarly, the recent star performers among de-veloping countries in terms of economic growth,particularly China, India and Viet Nam, have notfollowed orthodox policy prescriptions of relyingon unfettered market forces, broad-based liberali-zation and deep integration into the world economy.

Given that economic policies relying on un-fettered market forces have failed to deliver theexpected development results over the past twodecades, many developing countries that hadclosely followed the prescriptions of the Wash-ington Consensus have begun to reconsider theuse of proactive trade and industrial policies intheir development strategies. Arguing that “it isfair to say that nobody really believes in the Wash-ington Consensus anymore” (Rodrik, 2006: 2)appears to be an exaggeration. Nonetheless, thereasoning put forward by Rodrik (2004), alongwith the establishment of a task force on Indus-trial Policies and Development within the Initia-

Exclusive concentration onallocative efficiency impliesthat too little attention ispaid to stimulating thedynamic forces of markets.


tive for Policy Dialogue directed by Nobel laure-ate Joseph Stiglitz at Columbia University, andthe publication of a recent study by the WorldBank (2005: xiii) whose “central message ... isthat there is no unique universal set of rules ...[and that we] need to get awayfrom formulae and the searchfor elusive ‘best practices’”,have given new vigour to theindustrial policy debate.

In spite of the revitaliza-tion of the debate, much con-troversy remains in develop-ment policy discourse concern-ing the rationale for proactivetrade and industrial policies andthe feasibility for developing-country governments to adopt them. Some havequestioned the efficacy of such policies, tendingto associate them with failed inward-looking,import-substituting strategies, a comprehensiverange of open-ended interventions and a strongbias towards protectionism (Krueger, 1990). Oth-ers have argued that sectoral policies conferredonly modest benefits in the economic catch-up ex-periences in East Asia after the Second World Warand that, due to the associated high risk of pro-tracted rent-seeking and other potentially adverseeffects, developing countries “should be excep-tionally cautious before embarking on such poli-cies” (Pack, 2000: 64). By contrast, most of therecent development literature argues that indus-trial policies were indeed an important supportivefactor for East Asia’s economic catch-up. For ex-ample, according to the WorldBank (2005: 83) “the role ofactivist industrial policies isstill controversial but is likelyto have been important.”

The rationale for proactivetrade and industrial policies hasbeen questioned also becauseof their possible adverse ef-fects on efficient resource al-location (Bora, Lloyd and Pangestu, 2000; Packand Saggi, 2006). But a major theme in historicalanalyses of economic catch-up in mature and lateindustrializers (Amsden, 2001; Chang, 2002;Rodrik, 2006) is that exclusive concentration onallocative efficiency implies that too little atten-

tion is paid to stimulating the dynamic forces ofmarkets that underlie structural change and eco-nomic growth. As widely argued in the recent lit-erature (Akyüz, 2005; Chang, 2005; Gomory andBaumol, 2000; Hausmann and Rodrik, 2003;

Rodrik, 2004; Singh, 2005;Stiglitz, 2005; World Bank,2005), industrialization andeconomic catch-up are notgenerally the result of a natu-ral process simply based onan efficient allocation of re-sources. As recently stated byRodrik (2006: 5), “market orgovernment failures that affectaccumulation or productivitychanges are much more costly,and hence are more deserving

of policy attention, than distortions that simplyaffect static resource allocation.” A recent studyby the World Bank (2005: 10) also argues that“growth entails more than the efficient use of re-sources”. This is particularly true for developingcountries where economic growth entails dynamicchanges in the structure and technology-contentof production.

From this perspective, successful industriali-zation and economic catch-up need to be inter-preted as a process of cumulative causation. Sup-portive national economic policies advocated bythis interpretation focus on strengthening the dy-namic forces of markets related to informationexternalities in the context of innovative invest-ment, coordination externalities associated with

complementarities in invest-ment, production and consump-tion, and dynamic economiesof scale resulting from cumu-lative production experience.Strategic integration into theworld economy helps to maxi-mize the benefits of these ex-ternalities at the level of thenational economy. But all ofthese externalities are inti-

mately linked to departures from the competitiveequilibrium ideal of conventional economictheory; if unsupported by proactive national eco-nomic policies, such externalities cause a subop-timal provision in the volume and industry com-position of investment.

Many believe that the newinternational trading rulesreduce the degree offreedom available todeveloping countrypolicymakers ...

... causing them torelinquish policies thatfavour industrialdevelopment.


Another aspect of the argument that questionsthe wisdom of adopting proactive trade and indus-trial policies as an integral element of developmentstrategies relates to the feasibility for developing-country governments to implement such policies.There is indeed a widespread belief that the newinternational trading rules and regulations, whichhave emerged following both the conclusion ofthe Uruguay Round of multilateral trade negotia-tions and the rising number of regional and bilat-eral trade arrangements, reduce the degree offreedom available to developing-country policy-makers to the point that they are required to relin-quish policies that favour domestic industrialdevelopment and associated trade-related strate-gies (see, for example, Das, 2003; Gallagher, 2005;Chang, 2006).3

This raises the following questions: What arethe principles on which developing-countrypolicymakers should base their formulation of in-

dustrialization and technological upgrading strat-egies? Which principles would they need to heedin formulating the set of policy instruments appro-priate to the specific conditions of their economies?And what degree of freedom remains for policyimplementation, given the increased importanceof international rules and commitments stemmingfrom international trade agreements? Sections Band C address these questions. Section D discussesthe main features of industrial development andtechnological upgrading in developing countriesover the past decade with a view to examiningwhether and to what extent commitments frominternational trade agreements have required de-veloping countries to abandon the use of policyinstruments on which they had previously relied,what kinds of policies they have adopted instead,and with what effect on industrial development.The final section draws conclusions on options forpolicy innovation at the national and internationallevel.

B. Stimulating the dynamic forces of markets

1. Maintaining productive dynamism

Countries at the earliest stages of economicdevelopment can increase per capita income andliving standards simply by capital accumulationthat allows a fuller use of underutilized labour andnatural resources. This is the case in particular forcountries seeking to diversify away from the pro-duction of primary commodities. But sustainedeconomic success to enable countries to go be-yond these early stages depends on continuousimprovements in productivity. The basic policyquestions facing these developing countries arehow to maintain productive dynamism and tech-nological upgrading as the key to successful eco-

nomic development and structural change, andhow best to promote trade and investment to thatend. From this perspective, when formulating eco-nomic policies related to industrialization and struc-tural change, developing-country policymakersneed to take into account the interrelationshipbetween income growth, productivity gains andchanges in production structure in an open economy.

Four arguments4 in favour of a proactive na-tional economic policy designed to support pro-ductive dynamism and technological upgradinghave received the most attention:5 (i) the presenceof dynamic scale economies that gives rise to in-creasing returns of scale at the firm level; (ii) com-plementarities in investment, production and con-


sumption that, if unchecked, result in coordina-tion failures; (iii) information externalities asso-ciated with investment in goods or modes of pro-duction that are new for the respective economy;and (iv) strategic trade integration, which de-scribes the open-economy pattern of public sup-port policies that is motivated by the above threearguments. This form of integration represents amix of import substitution through temporary pro-tection and export promotion using temporary sub-sidies, and embeds industrial policy in a wider,outward-oriented industrialization strategy.

(a) Strengthening the creative functions ofmarkets

Dynamic scale economies are important inthat substantial productivity growth can resultfrom an increase in investment. This is becauseinvestment in physical capital includes the compo-nents of technological change that are embeddedin machinery and equipment and investment inhuman capital allows for the efficient use of in-creasingly sophisticated technologies. Moreover,learning by doing, resultingfrom cumulative experiencewith a production process, fa-cilitates incremental improve-ments in product design andorganizational practices thatsupport productivity growth.These productivity gains inturn, combined with produc-tivity growth resulting from other factors – for ex-ample, increased specialization of productionwithin firms and the introduction of more special-ized capital equipment – drive the profitability ofa firm’s activities and its further investment.

Demand is an important incentive for invest-ment by entrepreneurs. Only when they expect tohave a large enough market for their new activitiesto be profitable will they engage in new investment.Developing countries with a large population andsome minimum level of domestic purchasingpower may be able to generate domestically a sub-stantial part of the demand needed to support in-dustrialization.6 But most developing countrieswill need to generate exports as a vent for output,because a small economy is hardly able to main-

tain the circular causal links between productiv-ity growth and large-scale output and investmentsimply by trying to meet domestic demand. It isthus the interaction of supply and demand factorsin the investment process that translates produc-tivity growth into further investment and main-tains productive dynamism and technologicalupgrading.

Changes in the production structure influencethe intensity of the interrelationship between in-come and productivity growth. Traditionally, thebelief has been that industrialization, particularlythe development of manufacturing activities, of-fers the greatest scope for productivity growth.This is because manufacturing provides a largepotential for the division of labour as well as vir-tually unlimited scope for technological change.Moreover, the income and price elasticity of de-mand of most manufactures exceeds that for otherproducts.

Turning to the second argument in favour ofindustrial policies, the presence of complemen-tarities in investment, production and consump-tion is also generally considered to be greater in

manufacturing than in othersectors because manufacturingactivities give rise to more andstronger forward and back-ward linkages. For example,investment and profits of onemanufacturing firm dependnot only on its own output andfactor inputs, but also on the

output and factor inputs of other manufacturingfirms that provide intermediate production inputsor use the firm’s output as production inputs intheir own production.

This type of interdependence among differ-ent individual firms, which increases the profitsof both of them, has been referred to as “pecuniaryexternal economies” in the economic literature.When investment creates pecuniary externaleconomies, its economy-wide impact exceeds itsprivate profitability. Hence, profitable investmentcan fail to develop unless investment in upstreamand downstream activities occurs at the sametime.7 In this sense, a key problem of entrepre-neurs, acting as independent agents and only intheir self-interest, is how to coordinate investment

Substantial productivitygrowth can result from anincrease in investment.


so as to exploit external economies. As noted byScitovsky (1954: 150), market prices are not ca-pable of providing a signalling device to transmitinformation about present in-vestment plans and future pro-duction conditions when thereare reciprocal pecuniary exter-nal economies.

The above two argumentsin favour of industrial policyhave generally been inter-preted as providing a ration-ale for temporary protection. By contrast, the thirdargument supporting the adoption of industrialpolicy stresses that protection of the manufactur-ing sector per se does not provide an incentive foran entrepreneur to undertake innovative invest-ment and create new production capacity at an in-ternationally competitive level of productivity.8

Following Meade (1955) and Baldwin (1969),Rodrik (2004) argues that large uncertainties re-lated to the profitability of investment associatedwith the manufacture of products or the adoptionof modes of production which are new for the re-spective economy, as well as to the speed of entryof imitative entrepreneurs, give rise to informa-tion failures.

This information problem results from thefact that each potential investor that creates pro-duction facilities for new products or introducesnew modes of production faces fixed start-upcosts. But the investor usuallydoes not know the cost func-tion of these new activities,and therefore whether theywill be profitable and whetherthe sunk costs can be recov-ered, because the productioncosts of modern, non-tradi-tional activities can be deter-mined only after the initial in-vestment has been made.9 Inaddition, an entrepreneur who discovers the bestway to produce a particular product incurs the riskof potential imitators entering the market tooquickly to allow the realization of sufficient prof-its to cover the initial sunk cost.10 Hence, imita-tive entry reduces the private return that the inno-vative investor can realize, but at the same time itincreases its social return because of the spillovers

from imitation that allow the newly discoveredcost structure of the economy to be exploited by awide range of entrepreneurs. Moreover, if the in-

vestment fails, the innovativeentrepreneur will bear the fullcost of the mistake. Thus, thepotential innovative entrepre-neur’s initial lack of knowl-edge of the cost structure ofnew products or the use of newproduction processes causesan information failure. If un-checked, this information fail-

ure results in a suboptimal provision in the vol-ume and industry composition of investment.

According to this argument, the main task ofpublic support policies is to address the informa-tion externalities entailed in discovering the coststructure of an economy (Meade, 1955: 256–257;Baldwin, 1969; Rodrik, 2004). Acquiring knowl-edge of the underlying cost structure of an economythat determines the evolution of production pat-terns over time is a discovery process (Zeira, 1997;Hausmann and Rodrik, 2003). In this sense, in-dustrial policy needs to be defined not in terms ofan expected outcome (i.e. an altered sectoral struc-ture of production), but in terms of a process –“one where firms and the government learn aboutunderlying costs and opportunities and engage instrategic coordination ... with the aim of uncover-ing where the most significant obstacles to restruc-turing lie and what type of interventions are most

likely to remove them” (Rodrik,2004: 3).

This reasoning on the roleof policy implies a shift in em-phasis concerning policy in-struments. In terms of support-ing innovative investment, pro-tection is a rather blunt instru-ment. But, as noted by Meade(1955) and Rodrik (2004), and

succinctly stated by Baldwin (1969: 298), “[w]hatis needed, of course, is a subsidy to the initial en-trants into the industry for discovering better pro-duction techniques”.

From this perspective, economic develop-ment, technological upgrading and structural changeamount to a cumulative process of emerging new

A key problem ofentrepreneurs is how tocoordinate investment toexploit external economies.

Large uncertainties relatedto the profitability ofinvestment in new activitiesgive rise to informationfailures.


and more dynamic economic sectors, with tradi-tional activities being phased out or performed inmore productive ways. This process goes throughseveral stages, distinguished by the changing rela-tive importance of economic sectors and activi-ties. Empirical evidence shows that in the courseof economic development the sectoral allocationof employment changes. In a study covering awide cross-section of countries, Imbs and Wacziarg(2003) note that during the earlier stages of eco-nomic development countries diversify (i.e. theyspread economic activity more equally across sec-tors), but relatively late in the development proc-ess, at around a per capita income of $9,000, theystart specializing again. They suggest that an in-crease in a country’s productivity level, relativeto the rest of the world, drives the tendency to-wards diversification, while adecline in trading costs – stem-ming from a decline in transportcosts or tariffs, or from an in-crease in agglomeration econo-mies resulting from forwardand backward linkages – leadsto a shrinking range of goodsproduced domestically, thusfostering specialization (Imbsand Wacziarg, 2003: 82–83).They also emphasize that “in-creased sectoral specializa-tion, although a significant development, appliesonly to high-income economies. Countries diver-sify over most of their development path” (Imbsand Wacziarg, 2003: 64). Thus economic develop-ment appears to be closely related to the acquisi-tion of technological mastery over an increasingrange of products.

Rodrik (2004) provides a powerful restate-ment of the need to address information problemsrelated to the key importance of innovative invest-ment for diversification and technological upgradingin developing countries. This restatement comple-ments the other two arguments, mentioned earlier,that provide a rationale for industrial policy: theneed to support dynamic economies of scale andovercome coordination failures.

A frequent argument made against industrialpolicy is that if there are industries with a poten-tial comparative advantage, but domestic privateinvestors fail to develop the necessary activities

because of insufficient financing possibilities,government policy should be directed at the fulldevelopment of domestic financial markets, ratherthan providing industrial policy support. In otherwords, policy should address the financial marketimperfections with a view to moving the economytowards fully developed factor markets and com-petitive equilibrium. However, this reasoning isvalid only when investment is actually financedfrom sources external to the firm. It has beenshown that much of the investment that drovesuccessful industrialization in East Asia relied onprofits as a source for investment (TDR 1996).

Moreover, the full development of domesticfinancial markets takes time. As a result, the aboveargument has led to the suggestion that the task of

determining whether the pros-pects for the domestic infantindustry are profitable be leftto foreign investors whose de-cisions are based on productionexperience elsewhere. Someauthors (e.g., Pack and Saggi,2006) argue that the promotionof foreign direct investment(FDI) should play a key role inindustrial development and na-tional economic policies shouldbe limited to the creation of

locational advantages – such as the provision ofappropriate physical infrastructure and assuringappropriate education and health services for thelabour force – with a focus on the provision ofincentives to attract FDI. This view considers FDInot only as eliminating information constraintsregarding the profitability of innovative invest-ment, but also as delivering a bundle of assets thatincludes additional capital investment, productivity-enhancing technology and best corporate normsand practices. Moreover, it is assumed that theknowledge initially transferred to an enterprisethrough FDI will spill over to other firms in thesame industries.

Such optimism about the economic growth,technology transfer and productivity consequencesof FDI has led many countries to adopt invest-ment regimes that offer special financial incentivesto foreign enterprises. However, empirical evi-dence points to considerable variation in thebenefits that host countries actually reap from FDI

In the cumulative processof technological upgradingand structural change therelative importance ofeconomic sectors andactivities changes.


inflows, and much depends on the establishmentof an appropriate regulatory and fiscal framework.

Kumar (2005: 179–186) cites a number ofrecent empirical studies showing that knowledgespillovers may not take place, especially in de-veloping countries, and domestic enterprises mayactually be affected adversely. In some cases FDImay be immiserizing by crowding out domesticinvestment. Moreover, the interests of a trans-national corporation (TNC) may diverge from thehost country’s developmental objectives, due tothe TNC’s strategy of pursuing global profitmaximization. Thus its decisions to source pro-duction inputs locally or from international sup-pliers may not be taken on the basis of efficiencyconsiderations alone. Also, TNC affiliates indeveloping countries tend tobuy the bulk of their inputsfrom their parent companies orother associated suppliers, andhence generate few domesticlinkages.

One explanation for thisvariation in FDI-related devel-opmental benefits is that theeffects of FDI on domestic in-vestment and growth partlydepend on the mode of entry(UNCTAD, 2000). For example, FDI in new plantequipment (i.e. “greenfield investment”) adds tothe existing capital stock and may indicate alonger-term commitment of the foreign investorto producing in the host country. However, in thepoorer countries that are most in need of externalfinancing of investment, much of the greenfieldinvestment has occurred in fuel and mineral in-dustries, thus making little contribution to thediversification and development of competitivemanufacturing activities. In other cases, FDI hasoften taken the form of mergers and acquisitions,thus making no addition to the host country’s pro-ductive assets.11

Apart from the nature of FDI inflows them-selves, national policies also determine the extentto which FDI contributes to technology transferand linkage creation. Restrictions on entry in theform of contractual obligations on technologytransfer, ownership ceilings, the provision ofincentives only for entry in specific targeted eco-

nomic sectors, and performance requirements re-lated to purchases of intermediate inputs fromlocal suppliers can play an important role in in-creasing the developmental impact of FDI. Thesemeasures try to establish positive, complementaryinteractions between foreign and domestic invest-ment so that they can have a favourable impacton the host country’s productivity performance.To what extent such attempts are successful oftendepends on the leverage of host countries overforeign firms. It is clear that the larger a host coun-try’s domestic market and the more developed itsindustrial production structure, the better it willbe able to offer auxiliary activities that foster theprofitability of TNC activities. It will thereforebe in a good position to demand concessions interms of technology transfer and input sourcing

from domestic suppliers in ex-change for access to a largedomestic market and a largedomestic network of inputsuppliers.

National policies that aimto create locational advantagesbased on cost differentials, forexample through favourabletax treatment or relatively lowunit labour costs, rapidly riskbecoming ineffective as a re-

sult of small cost changes or the emergence ofalternative host countries. By contrast, supportpolicies designed to create a dense network of in-termediate input suppliers can be an importantmeans to attract or retain TNC activities, developdomestic supply capacity and foster technologi-cal upgrading. Local availability of high-qualityintermediate inputs at world market prices pro-vides pecuniary externalities for TNC activities.The profit incentive for the TNC to produce inthe host country will be higher, the more inten-sively it uses intermediate inputs, and the higherwill be the savings stemming from lower importsand the associated lower trade costs of such pro-duction inputs. As explained in the literaturerelated to new economic geography models (e.g.Puga and Venables, 1996, 1999), a similar mecha-nism applies when the host country provides alarge market for the output of the TNC: TNC prof-its will increase because of the scale effects ofadditional demand.12 Thus, support policies de-signed to provide a network of competitive input

Developing-countrypolicymakers need apragmatic and strategicperspective on how FDIcan fit into their widerdevelopment agenda.


suppliers or output users can be a key factor in ahost country’s locational advantage.

Thus there is a need for investment policiesto mobilize domestic resources as well as FDI.Such policies should be combined with an appro-priate regulatory and fiscal framework to ensurethat the expected development gains will be ob-tained. An excessive focus ofnational industrialization poli-cies on attracting FDI wouldbias the national economicpolicies towards “external in-tegration” at the expense of“internal integration”, in thesense of a denser set of linksbetween consumer, intermedi-ate and capital goods indus-tries (Wade, 2003a). This isthe case, in particular, when ahigh import content of exports, including thoseassociated with the activities of TNCs, mislead-ingly suggests successful industrialization andtechnological upgrading of domestic production,when, in reality, the domestic value added ofthese activities is small. Consequently, developing-country policymakers need a pragmatic and stra-tegic perspective on how FDI can fit into theirwider development agenda in ways that bringabout structural and technological change.

(b) Embedding investment promotion in awider industrialization strategy

The above arguments indicate that develop-ing countries may be well advised to adopt – inthe context of a private-sector-led, market-basedeconomy – a broader industrial strategy, whichcombines temporary protection with public sup-port that nurtures the innovative capabilities ofdomestic private enterprises and increases theirrate of investment and technological upgrading.Indeed, the fourth argument in favour of proactivenational economic policies relates to the patternof support; it stresses the importance of combin-ing industrial policy with a strategy that takes intoaccount the relationship between trade and growth.

As mentioned earlier, one linkage betweenexports and growth is through market size. Ex-

porting enlarges the market for domestic produc-tion, and thus offers scale economies at the firmlevel made possible by mass production tech-niques. It also provides a range of externalities atthe industry level, including economies of spe-cialization and agglomeration. Moreover, thenexus between the availability of foreign exchangeearnings from expanded exports and the need for

foreign exchange to financeimports of capital goods andintermediate inputs – requiredto build up industrial capacityand competitive strength – re-duces technological constraintsthat would otherwise impedethe development process.

But these linkages be-tween trade and growth do notnecessarily imply the rapid

opening up of markets. While the experience ofsuccessfully industrializing economies points tothe importance of strong export performance (TDR2003), cross-country regressions attempting toestablish a causal link between import liberaliza-tion and growth have failed to deliver robust find-ings (Rodriguez and Rodrik, 2000). This is partlydue to processes of cumulative causation (i.e. thefact that the levels of past and present activitiesin a sector are a determinant of current patternsof comparative advantage). As argued by Burgessand Venables (2004: 118), in these instances, broad-based “liberalization actually impedes growth byinhibiting infant industries and local accumula-tion of knowledge.”

The difficulty in establishing a robust em-pirical causal relationship between openness totrade and higher growth is also due to the fact thatsuccessfully integrating developing countries haveadopted a wide variety of trade policy approaches.These range from partial liberalization through theestablishment of export processing zones (as inChina and Mauritius) and opening up different sec-tors at different speeds (as in India), or ambitiousbroad-based unilateral trade liberalization (as inChile), to a combination of unilateral trade reformsand an aggressive pursuit of regional and bilateraltrade deals (as in Mexico). Moreover, these differ-ent trade policies have been combined with variouscomplementary policies. As a result, econometricstudies encounter severe methodological problems

It is important to combineindustrial policy with astrategy that takes intoaccount the relationshipbetween trade and growth.


related to measurement, reverse causation andomitted variable bias.

While recognizing that there may be severaldifferent means of trade integration that can suc-cessfully support economic growth, in most his-torical patterns of successful industrialization, in-dustrial policy has been part of a wider strategyin which the structure of imports and exportsprogresses through a number of stages. During theearliest stages of economic development, produc-tion and exports consist largely of primary com-modities, while imports comprise mainly manu-factures. Subsequent stages of industrializationgenerally involve, first, increased production (andreduced imports) of manufactured producer goods,accompanied by increased imports of machineryand equipment. This is followed by a stage of netexports of consumer goods and a reduction in im-ports of capital goods. Finally,a stage of mature industriali-zation is reached in which mostcapital goods are producedat home and basic consumergoods are imported (Kaldor,1966; Akyüz, 2005). From thisperspective, acquiring the abil-ity to competitively producegoods that were previously im-ported is inherent in rapid eco-nomic change, and implement-ing some temporary protectiondoes not imply adopting an “anti-trade” strategy,because import replacement needs to go hand inhand with policy-supported export development.

While this process follows a clear trajectoryof progress towards the efficient production ofmore technology- and knowledge-intensive prod-ucts, it does not converge to a predefined point.Rather, choice is involved across a whole rangeof industries and products in each stage of devel-opment, influenced by geography, size, relativefactor endowments, the decision of entrepreneursand policy. The trade policies used to animate thiscomplex process can be characterized as “strate-gic trade integration” – a more measured approachto liberalization combined with proactive indus-trial policies and outward orientation (TDR 1996).

The precise policy mix will depend on thestage of industrial development and the particular

requirements of different manufacturing indus-tries. Thus the specific product category candi-dates for public support policies in a country willdepend on many factors, and are likely to changeduring the course of economic development astheir skill and technology content gradually in-crease. During the initial phase of industrial ex-pansion that emphasizes resource-based and la-bour-intensive manufactures, price signals result-ing from traditional comparative advantage andreflecting an economy’s relative abundance ofnatural resource and low-skilled labour endow-ments can provide strong investment incentivesfor entrepreneurs. Since these sectors tend not tobe very demanding in terms of technological mas-tery, the start-up costs of investment designed todiscover their cost structure in the domesticeconomy are likely to be small. As a result, sup-port measures can be of relatively small size, and

can be phased out after a shortperiod of learning and expan-sion in world markets.

As these basic industrialsectors mature and becomeinternationally competitive,they are likely, over time, toencounter difficulties in com-peting on international mar-kets as domestic wages rise,low-cost competitors emerge,and sector-specific limits of

learning and productivity growth approach. Hence,more dynamic and skill- and technology-intensiveindustries need to be promoted, and any existingprotection and support to the traditional industriesneed to be phased out. Industries in the medium-technology-intensive range typically include elec-trical machinery, basic chemicals, automobiles,consumer electronics and semiconductors – sec-tors that, historically, have played a key role insuccessful late-industrializing countries. Furtherindustrial upgrading will allow some industrial-izing economies to develop production and exportactivities in high-technology-intensive manufac-tures, such as aerospace industries and biotech-nology.

In all successful industrialization experiencesof the twentieth century the gradual process oftechnological upgrading has followed this generalpattern. But it is clear that the sequential devel-

Public support policieshave to shift to otherproduct categories in thecourse of economicdevelopment as their skilland technology contentgradually increase.


opment of individual industrial sectors will differacross countries, depending on initial conditions,such as geography, economic size and relativefactor abundance, as well as on policy choices.

These considerations suggest a structure ofsupport policies as described in a stylized mannerin figure 5.1. In this schema, policy support for aspecific product category is introduced once thetechnological barriers to entry are no longer outof reach for domestic manufacturers. It is with-drawn once domestic manufacturers have attainedtechnological mastery, once the increase in domes-tic wages makes production no longer profitableat an internationally competitive level, and whenbenefits from economies of scale and learning bydoing approach exhaustion. Thus the key featureof the stylized representation in figure 5.1 is thatsupport policies follow a non-linear path; that is,any specific product category is a candidate forpublic support policies only for a limited periodof time.

It is difficult to establish meaningful bench-marks for the size of support policies; these willdepend on product-specific and, in particular,country-specific characteristics. Thus the exactpositions and slopes of the lines in figure 5.1should be seen merely as a schematic illustrationof a general pattern. In this stylized representa-tion, the relative strength of support policies foreach product category depends on the incremen-tal increase in skill and technology content as pro-duction moves from one product category to thenext, and on a country’s cumulative manufactur-ing experience. Previous manufacturing experi-ence will have allowed a country to benefit fromdynamic external economies (e.g. through thecreation of technological capacities) and to estab-lish a network of auxiliary manufacturing activi-ties. As a result, it is likely that the level of supportrequired to move from medium- to high-technology-intensive products will be lower than that requiredto move from traditional industries to medium-technology-intensive products. Moreover, during

Figure 5.1

STYLIZED REPRESENTATION OF SUPPORT POLICIESFOR DIFFERENT CATEGORIES OF MANUFACTURES

Source: Adapted from Akyüz, 2005: 22.Note: This is a stylized representation and should not be viewed as a precise mapping of relative levels of support measures

required for specific product categories in individual countries. For the allocation of individual products to the fourcategories, see TDR 2002, annex to chapter III.RL: Resource-based and labour-intensive manufactures.LT: Low-technology-intensive manufactures.MT: Medium-technology-intensive manufactures.HT: High-technology-intensive manufactures.


the earlier stages of industrialization, which con-centrate on resource- and labour-intensive prod-ucts, there would be no protection against importsof other manufactures13 since they will be an im-portant source for satisfying domestic demand,given that there is not yet any significant domesticproduction of these goods. By the time technologyupgrading towards medium-technology-intensiveproducts sets in, support for resource- and labour-intensive products is assumed to have been fullyphased out.

The recent economic literature also suggeststhat support policies may be required for skill- andtechnology-intensive manufactures to enter glo-bal markets. For example, as stated by Gomoryand Baumol (2000: xiii), comparative advantages“based on natural resources still exist ... but moredominant today are advantagesthat can be acquired. These canbe advantages conferred by be-ing established in an industryand gaining thereby either spe-cialized knowledge or econo-mies of scale or scope” (empha-sis in original). Acquired, as op-posed to natural, comparativeadvantage plays a key role inmedium- and high-technology-intensive manufactures for which economiesof scale or scope and high start-up costs are a keycharacteristic. To quote again Gomory and Baumol(2000: 6), “much of modern technology requiresactivities to be carried out on a very large scale inorder to be economical and competitive. Conse-quently entry into one of these industries, againstan entrenched competitor, is slow, expensive, andvery much an uphill battle if left to free-marketforces.”

Thus, while the process of industrial upgrad-ing and strategic trade integration implies outwardorientation, it is not a process that is driven byunfettered market forces. Rather, beyond the ear-liest stages of industrialization, the provision oftemporary support to promote increasingly higherskill- and technology-intensive industries helps theeconomy to progress through a series of overlap-ping industries, as well as to continuously increaseproductivity and acquire technological masteryover a wider range of products. Thus, policychoices are crucial to countries’ trading patterns,

because international trade does not lead to one,uniquely determined, best economic outcomebased on natural national advantages. On the con-trary, “there are many possible outcomes that dependon what countries actually choose to do, what ca-pabilities, natural or human-made, they actuallydevelop” (Gomory and Baumol, 2000: 5, empha-sis in original).

Two additional observations regarding thestylized representation in figure 5.1 are worthmentioning. First, the scheme can be interpretedas referring not only to industrial sectors but alsoto specific activities within the same sector. Tech-nological progress and reduced trade costs havegiven rise to international production-sharing,whereby activities with different factor intensityare carried out at different locations. Thus it may

be possible for a country tostart producing in an industrialsector by carrying out labour-intensive functions and under-taking gradual technologicalupgrading, leading eventuallyto its being able to carry outthe most technologically in-tensive activities in that par-ticular sector. Second, one rea-son to combine temporary pro-

tection and temporary subsidies is that the main-tenance of dynamic scale economies requires bothsuccessive innovative investments and learningprocesses. Temporary subsidies facilitate innova-tive investments, while temporary protection al-lows learning processes to unfold. But the poten-tial for learning in each specific activity dimin-ishes with growing experience in that activity, sothat learning and innovative investment dependon each other: new innovative investment opensnew possibilities for further learning, which in turnprovides the basis for the productive use of anew round of innovative investments, and so on(Mayer, 1996).

The question often arises as to whetherdeveloping-country governments have the admin-istrative and institutional capability to design andimplement well-conceived support policies. Forexample, Pack and Saggi (2006: 28) point to thealleged severe information constraints of indus-trial policy: “The range and depth of knowledgethat policy makers would have to master to im-

Industrial upgrading impliesoutward orientation, but it isnot a process driven byunfettered market forces.


plement a successful policy is extraordinary. Theywould have to understand the relevance of, andbe accurately informed about, a huge range ofcomplex questions and have the ability to accu-rately evaluate very subtle differences.” However,the same governments are ex-pected to carry out other com-plex tasks, such as implement-ing trade and financial liber-alization and privatization. In-deed, many of them are sup-ported in these efforts throughcapacity-building assistance,for example, for the implemen-tation of WTO agreements.

These implementation-related arguments arebased on the view that industrial policy tends torely on excessively complex methods and targetsfor a wide range of industrial sectors at the sametime. Indeed, import substitution strategies insome countries in Latin America and South Asiarelied on a wide-ranging and often overlappinguse of import quotas, exchange controls and do-mestic content rules in addition to industrial tar-iffs. This often made it difficult to determine howmuch support an administrative regulation actu-ally provided, which government institution wasultimately responsible for that support, and whobenefited, how much and from which measure.

However, as shown by Amsden (1989, 2001)and Wade (1990), compared to the broad-basedsupport in much of Latin America during the1960s and 1970s, the scope ofsectoral support policies andprogrammes in the post-SecondWorld War cases of successfuleconomic catch-up was rathermodest, yet, in aggregate, muchmore effective. Implementingthese kinds of support policiesdoes not require sophisticatedcalculations, and the associ-ated information requirementsdo not go beyond those neededfor routine decision-making by managers in TNCs,such as estimating demand elasticities and thetechnology requirements for internationally com-petitive domestic production of particular goods.Most importantly, a key objective of the institu-tional mechanisms that are intimately linked with

these kinds of support policies is to promote in-formation exchange between the private sector andthe government with the aim of identifying themost serious obstacles to diversification and struc-tural change, and determining the kinds of poli-

cies most likely to remove theseobstacles.

Another argument againstproactive trade and industrialpolicies, which also addressesimplementation issues, is that itrisks giving rise to rent-seeking;that is, State intervention cre-ates additional “wastes” thatmay more than offset the ben-

efits it produces, because private agents divertresources to unproductive activities in order tocapture rents generated by government interven-tion. But as in the case of information constraints,institutional mechanisms exist that can substan-tially reduce the risk of rent-seeking.

Thus, while it is correct to say that success-ful sectoral support policies require a certain im-plementation capability on the part of govern-ments, there are various institutional arrangementsto address implementation issues, and, more gen-erally, to generate, distribute and revoke economicrents and coordinate investment in a way thatmeets wider development goals. These institu-tional arrangements are discussed in chapter VIbelow. What is important at this point is to recog-nize that the acceleration of industrial develop-

ment and technological catch-up is not just a question ofpolicy reform; it also requiresthe strengthening of adminis-trative capacity and institu-tions.

In sum, the adoption ofproactive trade and industrialpolicies can be anchoredfirmly in economic theory ifthe recent revival of arguments

supporting the creative functions of markets,which played a key role in the theoretical debateamong early development economists (Meade,1955; and the authors discussed in Toner, 1999,and Ros, 2000), and the more recent theoreticalcontributions regarding the creation of new areas

The maintenance ofdynamic scale economiesrequires both successiveinnovative investments andlearning processes. ...

... Temporary subsidiesfacilitate innovativeinvestments, whiletemporary protection allowslearning processes tounfold.


of comparative advantages (Gomory and Baumol,2000; Puga and Venables, 1996, 1999) are takeninto account. From this perspective, proactivetrade and industrial policies can enhance the in-formation and coordination mechanisms of marketsand help economies achieve technological mas-tery and international competitiveness in a rangeof increasingly technologically sophisticated prod-ucts. The following section addresses key issuesrelating to how the rationale for proactive tradeand industrial policies can be made operationaland translated into concrete national economicpolicies.

2. Principles and types of policies forstimulating the dynamic forces ofmarkets

The formulation of any prescription for de-velopment policy must recognize the large differ-ences between developing countries and the needto respect their unique characteristics. Neverthe-less, there are some common features that permitconsideration of general policy principles for de-veloping countries as a whole, while bearing inmind that such principles need to be translated intospecific types of policies adjusted to the particu-lar circumstances of individual countries.

For the implementation of proactive trade andindustrial policies there is a range of choices regard-ing general principles and specific types of policies.General principles refer to the characteristics ofeconomic policy, for example, the balance betweenprivate initiative and public policy support, or theextent to which policies apply horizontally acrossmultiple sectors, or selectively at the sectoral oreven subsectoral level.14 Specific types of policiesrefer to specific measures that operationalize suchgeneral principles.

(a) General policy principles

One general principle concerns the balancebetween private initiatives and public policy sup-port. Latin America’s import-substitution strategyof the 1960s and 1970s has sometimes been char-

acterized as “State-led industrialization” as op-posed to the so-called “market-led industrializa-tion” strategy pursued in East Asia since the 1960s(World Bank, 1993). However, it has become gen-erally accepted that this characterization is a misin-terpretation of the historical facts (see, for example,World Bank, 2005). The main difference betweenthe strategies pursued in these developing regionsis that industrial policy has not been as concertedand coherent in Latin America as in East Asia.For example, Bruton (1998: 912) notes that in-dustrial policies under import-substitution regimesoften levied tariffs on an ad hoc basis, with theconsequence that “a great hodgepodge of rates ap-peared, with virtually no evidence of any consid-eration of costs or efficiency.” Moreover, theimpact of specific trade and industrial policymeasures cannot be expected to be proportionalto their intensity. The intensity of intervention inLatin America was higher than in East Asia. How-ever, whereas support policies in East Asia werestrategically designed and implemented, in LatinAmerica, governments often adopted unsystematicand overlapping measures, making it difficult toestablish a clear link between policy measures andtargets.

Modern support policies give the lead role toprivate enterprises, supporting their innovative in-vestments as well as their efforts to get importedtechnologies to work well under local conditions.This support is complemented by trade policy sup-port designed to achieve international competitive-ness in increasingly more technologically sophis-ticated products.

A second general principle of proactive tradeand industrial policies that aim at strengtheningthe creative forces of markets is that in order tofoster diversification and technological upgrad-ing subsidies should be given only to investmentthat is undertaken to discover the cost function ofnew goods or new modes of production in the re-spective economy.15 This implies that such policiesshould not be employed as defence mechanismsto support industries where production and employ-ment are threatened the most by foreign competitorsthat have successfully upgraded their production.For example, this general principle does not sup-port selective trade protection or other selectivesupport measures that many developed countriesare still applying in agriculture or in labour-


intensive manufacturing sectors such as the cloth-ing industry.

Neither does this general principle support alarge number of contemporary industrial policymeasures which focus on at-tracting FDI and related export-oriented activities.16 Rodrik(2004: 28) emphasizes the cur-rent importance of such meas-ures, stating that “industrialpolicies [privileging exportsand foreign investment] haverun rampant during the lasttwo decades.” This widespreadsupport is based on the expec-tation that FDI inflows willfacilitate industrialization anddevelopment in host countries by enabling themto benefit from foreign investors’ production tech-nology, organizational and managerial skills, andmarketing know-how, as well as by entering theirmarketing networks. Moreover, host countriesexpect to benefit from knowledge spillovers andother favourable externalities of FDI. However,as already mentioned, empirical studies havefound very little systematic evidence of techno-logical and other externalities obtained in thismanner. Perhaps most importantly, FDI inflowsand export promotion may reinforce host coun-tries’ existing comparative advantage based on therelative abundance of natural resources or cheaplabour, ignoring the importance of productivitygains and structural transformation that is at theheart of the rationale for proactive trade and indus-trial policies.

A third general principle is that policy sup-port should not be open-ended. Instead, it shouldbe given only on the basis ofclearly established operationaland achievable goals, observ-able criteria for monitoring itand specific time horizons.Regarding the latter, a keyquestion is how one can ensurethat the policy support lastslong enough to motivate entrepreneurs to invest,but short enough to force investors to keep im-proving productivity. Historical experience showsthat where supporting policies were of a blanketand virtually permanent character, they failed to

curb inertia and rent-seeking. In East Asia, by con-trast, attaining the policy objectives has beenachieved by the establishment of strict perform-ance criteria related to productivity, as verifiedthrough performance in international markets.17

The rapid growth of exportsprovides the key demand stimu-lus to ignite a cumulative proc-ess of high investment, highprofits, high savings and highgrowth. Moreover, export per-formance provides a clear, neu-tral standard to evaluate theperformance of firms receiv-ing public policy support. Ex-port data have the additionaladvantage that they cannot befaked, and that they provide a

relatively clean measure of the relative competi-tiveness of domestic producers. The establishmentof clearly defined and quantifiable performancecriteria also makes it easier to identify failures andwithdraw any associated support. Finally, by im-posing performance standards on investors, thegovernment subjects itself to evaluation by ob-jective criteria. Indeed, the aim of performancerequirements is not for the government to pickwinners, but to know when there is a loser.

A fourth general principle is to base the de-termination of policy measures on an intense dia-logue between ministries, industry associationsand research institutions; that is, on a delibera-tion process, rather than on autonomous decisionsof government entities. The aim of this process isto exchange information on the government’s vi-sion regarding structural change and developmentstrategies, on the views of industry associationsregarding business opportunities and investment

constraints, and on research in-stitutions’ assessments of na-tional and international tech-nology developments. The lead-ership and decision-makingpower of the individuals whoparticipate in the deliberationprocess play a fundamental

role in its success. Participants from business as-sociations need to be representative, with suffi-cient economic and political weight. As for gov-ernment officials, Wade (2006) points out that theauthority for carrying out support policies must

The lead role of privateenterprises needs to becomplemented by policysupport to achievecompetitiveness intechnologically increasinglysophisticated products.

Policy support should notbe open-ended.


vest in agencies with demonstrated competence,while the implementing agencies must be moni-tored closely by and be accountable to a principalwith a clear stake in the outcomes and with politi-cal authority at the highest level, and they mustmaintain the capacity to reinvent and refashionthemselves as economic con-ditions change. Institutionalissues are addressed in moredetail in chapter VI below.

In sum, modern publicpolicies combine private ini-tiatives with public support.This should be embedded inprocesses that rely on reciprocal control mecha-nisms and on information and coordination com-mitments from both sides. An important objectiveof such processes is knowledge generation. Whilethese processes are inevitably characterized bytrial and error, it is important to minimize the eco-nomic costs of mistakes. Yet, attempting to pre-vent any mistakes risks leading to no innovativeinvestment at all.

(b) Specific types of policies

As already mentioned, how these generalprinciples are translated into specific types of poli-cies depends on a country’s particular initialconditions and stage of economic development.Also, they should be the outcome of a delibera-tion process to determine which public supportpolicies are likely to have the greatest impact.Nonetheless, a brief discussion of a number ofpossible types of policies may be useful for illus-trative purposes.18

Fiscal incentives, direct public credit andsubsidies are measures that lower the cost of in-novative investment. They can thus reduce uncer-tainty among potential investors as to the profit-ability of innovative investment that can be eas-ily emulated. Fiscal incentives may take the formof tax deductions and tax credits for particulartypes of innovative activities or the acquisition ofnational or imported, embodied or disembodied,technology related to innovation. Direct publiccredit may take the form of loans by developmentbanks for innovative investment and the acquisi-

tion of technology, and be granted with preferen-tial interest rates and favourable repayment sched-ules. Subsidies may be allocated to entrepreneursby competition according to their projects’ poten-tial to bring about diversification and technologi-cal upgrading as well as knowledge spillovers or

the creation of forward andbackward linkages.

Venture capital organiza-tions can play an importantrole in providing risk capital,since obtaining loan finance isparticularly difficult for inno-vative investment, given that

the profitability of the innovation and its poten-tial market are not yet known. But in addition touncertainties and asymmetric information regard-ing the profitability of a project and the potentialopportunistic behaviour of entrepreneurs, venturecapital organizations themselves often face financ-ing constraints. The resulting desire for zerodefault may lead to underfunding. In these circum-stances, development banks and other publicactors that are motivated by social returns andexternalities, rather than private profit, can play acrucial role. In particular, when domestic sourcesof investment finance are constrained, credits fromdevelopment banks can also be an alternative toFDI for financing investment.

Undertaking research and development (R&D)activities in public research institutes constitutesa third specific type of public policy. A major prob-lem, however, is that the resources devoted by thegovernment to R&D may be substantial in fiscalterms, because of the proportion of the budget theyabsorb, yet insufficient to cover a broad scienceand technology infrastructure and to provide ameaningful level of subsidies for R&D. Thesebudgetary constraints are exacerbated where theprovision of funds is more horizontal and less tar-geted. One solution could be to deploy a schemeallowing for the partial recovery of public R&Doutlays through royalty payments by the privateusers of public research output commensurate withtheir profits. Another possibility could be to in-troduce a system of allocating subsidies for R&Dthrough competitions designed in conformity withthe general innovation promotion strategy. Giventhe current income boom from natural-resourceexports in many developing countries, a further

The aim is not to pickwinners, but to know whenthere is a loser.


possibility could be to earmark income from Statecompanies, and royalties or tax income for R&Din research institutions designed to generate in-novations either for product diversification, re-search into new technologies, or capacity buildingamong suppliers.

Technology development can also be sup-ported by the creation of science and technologyparks that provide incentives (for example, in theform of tax breaks, subsidized credits, or permis-sion to exceed normal debt-equity ratios) for theestablishment of firms that identify, transfer, diffuseand absorb foreign industrial technologies andsubsequently undertake innovation. Such incentivesmay be complemented by offering attractive sala-ries in order, for example, to encourage the returnmigration of skilled nationals. Developing-countrygovernments could also consider paying or subsi-dizing royalty payments, and support the applicationof technology by negotiating on behalf of domes-tic firms that are able to apply the technology.

In addition to the use of these traditional in-struments, the potential of strategic standard set-ting for technology development has recentlygained attention. In the economics of technologi-cal change, formal standardization closely inter-acts with international property rights protection(UNIDO, 2005). The timing and scope of techni-cal standard setting plays an important role in thediffusion of patents and the related new productsand technologies. Thus it contributes to channel-ling collective efforts towards technologicalprogress. On the other hand, standard setting canbe a mode of selection for the use of product andprocess innovations that are protected by intellec-tual property rights and thus favour one set of in-novative firms over another. Consequently,strategic setting of compatibility standards can bea means of stimulating domestic technology re-search efforts and the creation of non-proprietarytechnology. Successful pursuance of these goalscan increase the ability to exercise leverage innegotiations with overseas patent holders and, asa long-term objective, help to develop domesticproprietary technology. These issues apply mostlyto high-technology-related product markets, suchas, recently, in the information and communica-tions technology industries, which have seen arapid succession of patented new technologies inan environment of multiple standards.19

The vast majority of developing-countrymarkets are likely to be too small, with too littlepurchasing power to impose technology standardsthat favour the production of domestic firms. Thisis because foreign firms producing in conformitywith existing internationally applied technologystandards serve much larger user bases and canrealize economies of scale and learning effects.Inappropriate standard setting by developingcountries may therefore stifle technology transferto their economies. However, standards are oftenshaped by market needs and users’ preferences,rather than simply by technology requirements.If a developing country, or a group of developingcountries, can provide a large enough user basewith promising market potential, it may rival anexisting technology that enjoys property rightsprotection. Given its large domestic market, itslarge pool of educated researchers and experiencedreturnees from overseas, and its substantial ex-penditure in high-tech research, China (either onits own or in concert with other Asian economies)appears to have acquired such a position (Ernst,2004; Linden, 2004), but this is an unusual caseamong developing countries.

Governments need to have a clear vision fortheir economy’s future technology developmentif they are to benefit from the support to technol-ogy upgrading and the development of proprietaryknowledge that strategic standard setting can pro-vide. Only when promising new technologies areidentified at their very early stages can standardi-zation influence basic research activities andsubsequent pilot production. Moreover, standardsetting should specify the performance of compo-nents, rather than their design, in order to avoidconflicts with patents protecting those compo-nents. Governments should also offer attractivelicensing schemes to provide incentives for inno-vative, R&D-intensive companies to participatein standardization processes (UNIDO, 2005).

Specific policy measures related to strategicintegration include selective liberalization throughdifferentiated intervention, granting duty drawbacksand establishing temporary admission regimes forselected imports (e.g. capital equipment and in-termediate inputs), and the creation of exportprocessing zones that offer preferential tax andcustoms treatment. Such measures have been suc-cessfully employed for industrial development in


a wide range of developing countries, and haveoften been complemented by selective treatmentof FDI inflows through, for example, restrictionson FDI entry, ownership ceilings, barriers to hos-tile takeovers, or the imposition of performancerequirements.

As already mentioned, the above discussionof specific policy measures is intended to illus-

trate some of the options available to developing-country policymakers; it is not intended to providean exhaustive list. In addition, it is doubtful whetherany of these policy instruments can be used suc-cessfully in isolation. Rather, policymakers willneed to have a vision for the economic developmentof their country in order to tailor these instrumentsto local conditions and link them with other poli-cies in support of development.

Unlike monetary and financial multilateralarrangements, discussed in chapter IV, the multi-lateral trading regime is organized around a set ofnegotiated, binding and enforceable rules andcommitments. Negotiated under the aegis of theGATT/WTO, these rules provide the basis forregulating international trade. The core principlesof this regime are reciprocity and non-discrimi-nation, as reflected in the most-favoured nation(MFN) rule and the commitment to national treat-ment (i.e. equal treatment of domestic and foreigngoods and enterprises in domestic markets). Leav-ing aside a number of general exceptions,20 as wellas exemptions that specifically apply to develop-ing countries (see below), the multilateral tradingregime is thus intended to provide what is oftencalled a “level playing field”, by extending thesame legal rights and obligations to all memberStates of the WTO.

Since the mid-1980s, rapid and broad-basedtrade liberalization has been a central conditionattached to loans from multilateral lending organi-zations, as well as to aid flows and debt relief frommajor developed-country donors. But the current

wide scope of multilateral governance in the areaof trade is associated with the Uruguay RoundAgreements (URAs) and the establishment of theWTO in 1995. The Uruguay Round (UR) broughtabout industrial tariff reductions, negotiated on arequest-and-offer basis, rather than through the useof a formula approach based on a percentage re-duction in average tariffs, as well as through “zero-for-zero” reductions for some product groups,including under the Information TechnologyAgreement.21 Moreover, the UR resulted in a newset of agreements on trade in goods – an exten-sion of the General Agreement on Tariffs andTrade (GATT) which the WTO absorbed – as wellas additional agreements on so-called “trade-re-lated” activities. These include the Agreement onTrade-related Aspects of Intellectual PropertyRights (TRIPS), the General Agreement on Tradein Services (GATS) and the Agreement on Trade-related Investment Measures (TRIMs), as well asthe Agreement on Subsidies and CountervailingMeasures (SCM). It also established a unified andbinding dispute settlement mechanism. The Agree-ments were adopted as a so-called “single under-taking” – countries had to accept the package of

C. Restrictions imposed by international agreementson policy autonomy: an inventory


Agreements in its entirety. The resulting expan-sion of the scope of the multilateral trading regimemeans that key aspects of countries’ regulatoryregimes that affect how na-tional economies operate havebecome subject to multilateraldisciplines.

The multilateral trade re-gime has accorded exemptionsto developing countries. In ne-gotiations, they are allowed togrant less-than-full reciprocityunder Article XXVIII bis of theGATT, adopted in 1958. More-over, the so-called “enabling clause”, adopted in1979 and generally known as special and differ-ential treatment (SDT), accords developing coun-tries exemptions to the MFN rule, by allowingthem to benefit from more favourable market ac-cess conditions. However, the UR brought achange in perspective on SDT. Prior to the Round,exemptions from the MFN rule and the principleof reciprocity were seen as a recognition by theinternational community that in order to providesome kind of parity between developed and de-veloping countries, developed countries needed togive developing countries access to their marketswithout requiring them to open up their own mar-kets on a reciprocal basis. These exemptions alsogave developing countries some possibilities topursue legally their own nationally determined de-velopment policies. Following the UR, SDT hasbasically come to mean that developing countries,and especially the least developed among them,are accorded longer transitionperiods for full implementa-tion of all rules and commit-ments in the WTO.

Especially since the early1990s, many developing coun-tries have increasingly com-plemented multilateral tradenegotiations in the WTO withregional or bilateral agree-ments, including with devel-oped countries and regions, inparticular the United Statesand the EU. Regional or bilateral agreements withlarge developed countries offer substantial ben-efits to developing-country members as they usu-

ally provide greater market access than multilat-eral agreements, and often include a wider rangeof products than traditional trade preference

schemes such as the General-ized System of Preferences(GSP). Moreover, their adop-tion is generally expected tolead to additional FDI. On theother hand, greater integrationoften involves additional stepstowards regulatory disciplines,and thus further constrains thede jure ability of developingcountries to adopt appropriatenational regulatory and devel-

opment policies, particularly with regard to FDIand intellectual property rights.

The constellation of these rules and commit-ments, as well as the associated exceptions andexemptions, constitute a complex legal structurethat offers scope for different interpretations andpractices. Against this background, this sectionconcentrates on the often voiced concern that,since the rules and commitments of the interna-tional trading regime restrict the de jure ability ofdeveloping nations to adopt national developmentpolicy, they limit the possibilities for governmentsto deploy policies in support of further produc-tive and technological development. More specifi-cally, there is concern that these rules and com-mitments could deny the use of the very policymeasures that were instrumental in the develop-ment of today’s mature and late industrializers.To the extent that this is the case they thus reduce

the flexibility of national gov-ernments to pursue their de-velopment objectives. Anotherconcern is that these rules andcommitments, which in legalterms are equally binding forall countries, in economic termsmight impose more bindingconstraints on developing, com-pared to developed, countriesbecause of the differences intheir respective structural fea-tures and levels of industrialdevelopment. The discussion

in this section concentrates on rules and commit-ments associated with the TRIMs, SCM andTRIPS Agreements and tariff regulations.

The rules and commitmentsof the international tradingregime restrict the de jureability of developing nationsto adopt nationaldevelopment policy.

Rules and commitments,which in legal terms areequally binding for allcountries, in economicterms might impose morebinding constraints ondeveloping countries.


1. The Agreement on Trade-relatedInvestment Measures (TRIMs)

One important area that the URAs havebrought under multilateral jurisdiction involves arange of investment measures that have been usedby many developing and some developed coun-tries as part of broad strategies aimed at nurturingdomestic industry and achieving technology trans-fer. The TRIMs Agreement is designed to clarifythe relationship between a country’s investmentpolicy and the core rules of the multilateral trad-ing regime by identifying measures consideredincompatible with national treatment and forbid-ding the application of quantitative restrictionsthat link imports to export per-formance (e.g. trade or foreignexchange restrictions) or ex-port restrictions based on do-mestic sales.22

The Agreement does notdefine “trade-related invest-ment measures”, nor does itprovide objective criteria foridentifying them. The appen-dix to the Agreement gives guidance to govern-ments to decide which of their measures violatethe Agreement by providing an “illustrative list”of prohibited regulations. Countries are requiredto notify the WTO of such measures and elimi-nate them following the termination of transitionperiods.23

The imposition of performance requirementson foreign investors is a key regulatory measurethat has been affected by the TRIMs Agreement.Many governments, in developing and developedcountries alike, have used performance require-ments, which generally aim to increase the linkagesbetween foreign investors and local manufac-turers. One commonly employed performancerequirement concerns local content regulations,which are intended to increase domestic valueadded, thereby generating additional national in-come and employment, as well as encouraging thetransfer of technology. Other frequently usedperformance requirements relate to export per-formance or trade balancing, which require firmsto match their use of imported inputs in their ex-port products with an equal share of domestically

produced inputs in order to integrate the affiliatesin the host countries into their global/regional pro-duction networks. Foreign exchange balancingrules, which require foreign investors to meet for-eign exchange needs for imports through exports,rather than by converting local earnings into for-eign exchange, have also often been used.

Chang (2002), for example, shows that to-day’s developed countries extensively employedperformance requirements to maximize domesticvalue added. A number of developed countriescontinued the use of performance requirements inthe early post-Second World War period (WTOand UNCTAD, 2002). Local content requirementswere also a widely used instrument that strength-

ened backward integration andincreased domestic value added,in particular in the automobileindustry.

Developed countries haveincreasingly replaced explicitperformance requirements withtrade policy measures thatachieve essentially the sameobjectives as performance re-

quirements but are consistent with WTO rules(Kumar, 2005: 185). One example is screwdriverregulations (i.e. regulations governing imports bytrading partners of parts and components), whichhave been used by the EU (Safarian, 2003).

While developed countries extensively em-ployed performance requirements in one form oranother at earlier stages of their industrial devel-opment, developing countries have only recentlystarted to use these policy tools to foster their in-dustrialization and technological upgrading. Thisis closely related to the increasing importance ofinternational production networks, where devel-oping-country exports often include a high importcontent of technology-intensive parts and compo-nents, while domestic value added mostly consistsof wages paid for simple assembly activities. Inthis context, domestic content requirements havebeen used to increase technology transfer to de-veloping-country producers and to foster the useof domestically produced parts and components.Empirical evidence on the effectiveness of suchmeasures suggests that well-conceived perform-ance requirements “that have clear objectives and

TRIMs have affected theimposition of performancerequirements on foreigninvestors.


are effectively enforced are not only able to meettheir objectives, but may also bring significantfavourable externalities to the host countries”(Kumar, 2005: 193). However, developed coun-tries have brought a number of cases against de-veloping countries before the WTO dispute set-tlement mechanism, especially in the automotivesector, invoking the rules and commitments of theTRIMs Agreement.24

The TRIMs Agreement does not restrict theprovision of incentives to attract FDI, even thoughthe economic effect of such incentives may besimilar to the provision of subsidies, and eventhough such incentives may affect internationalinvestment and trade flows as much as domesticcontent requirements. This is the case, in particular,for activities in international production networkswhere TNCs are known to practice trade-restrictingpolicies with respect to their foreign affiliates(Kumar, 2005: 194).

Regional and bilateral investment agreementscan be considerably more restrictive than TRIMsbecause they address all measures regulating FDI,and not only those that are considered “trade re-lated”. Moreover, many such agreements allowfirms, rather than just governments, to bring casesto arbitration. Thus they go much further towardsregulatory harmonization. By contrast, developingcountries’ bilateral and regional trade agreementswith developed countries play a peculiar role inthe area of TRIMs, as they weaken rather thanreinforce multilateral commitments. This is be-cause, through the rules of origin, local contentrequirements have also, by definition, been in-cluded in preferential trade agreements betweendeveloping countries and large importing coun-tries/regions, such as the United States and the EU.Given that developed-country parties to such tradeagreements can tailor local content requirementsto their needs, these measures have not beenbrought before the dispute settlement mechanismof the WTO. Di Caprio and Amsden (2004: 23)therefore argue that preferential trade agreementspresent developing-country WTO “members withan escape hatch from limitations on that particu-lar aspect of TRIMs.”

It also needs to be recognized that FDI-regulating measures that do not violate nationaltreatment or impose quantitative restrictions con-

tinue to be consistent with WTO rules. For exam-ple, governments can impose technology transferrequirements which specify that a foreign com-pany conduct a certain proportion or type of itsresearch and development activities locally andtransfer or license a specified technology to do-mestic firms. Or a licence could be granted forthe establishment of an assembly plant only if theforeign investor simultaneously establishes a plantthat produces required intermediate inputs. Gov-ernments can also require that domestic investorsretain a proportion of a firm’s equity or that a spe-cific percentage of their technology personnel berecruited domestically (Shadlen, 2005a: 759).25 Inreality, however, only countries with substantialleverage over foreign investors are able to use suchmeasures.

2. The Agreement on Subsidies andCountervailing Measures (SCM)

The SCM represents another outcome ofthe UR that impinges directly on national rule-making authority. It addresses multilateral disci-plines for regulating the provision of subsidies,as well as the use of countervailing measures tooffset injury to an industry in the importing coun-try caused by imports that are subsidized in thecountry of origin. The SCM covers mainly the in-dustrial sector; special rules apply to agriculture,and the General Agreement on Trade in Services(GATS) has no rules on subsidies (although thecurrent WTO-negotiations are addressing this is-sue).

The SCM defines a subsidy as a financialcontribution made by a government or any publicbody within the territory of a WTO member thatconfers a benefit. Such benefits can result fromdirect payments, foregone revenues and rights,government guarantees and equity participation,the provision of goods and services below marketvalue, or from differential application of certainrules to different sectors and activities, such asbank credits directed to specific sectors and ac-tivities with preferential conditions.

The Agreement represents a significant tight-ening of disciplines compared with the pre-UR


regime, which did not include comprehensive rulesand regulations on the use of subsidies, and al-lowed developing countries greater leeway touse subsidies for export pro-motion and import substitu-tion. It broadens the scope ofregulations relating to subsi-dies as it binds WTO mem-bers26 (except for the poorestamong them, as discussed be-low) and extends to measuresof subnational governments,State-owned enterprises and private entities thatcarry out functions that would normally be vestedin the government.

As the Agreement only applies to specificsubsidies, that is, those targeted at an enterprise,industry, or group of enterprises or industries, itaffects the selective function of policy. Non-specificsubsidies are not affected because they are pre-sumed not to distort the allocation of domesticresources; these include subsidies for the provi-sion of physical and social infrastructure, or sub-sidies resulting from low energy taxes that ben-efit all enterprises, as well as subsidies earmarkedfor specific enterprises according to their size orsimilar criteria. The Agreement prohibits subsidiesthat are conditional on export performance or onthe use of domestically produced goods (but coun-tries with a per capita income below $1,000 areexempted from this prohibition) and makes spe-cific subsidies “actionable”, which means that theyare subject to challenge through multilateral dis-pute settlement or countervailing action. Whilethis distinction between specific and non-specificsubsidies is straightforward in legal terms, in prac-tice it is not always easy to dif-ferentiate (Anderson, 2002).This may leave some room fordeveloping countries to designsubsidies that help import-competing or exporting firmswithout contravening WTOdisciplines.

Article 8 of the originalSCM provision defined certainspecific subsides as non-ac-tionable. Subsidies extended to research fell in thiscategory, as did subsidies in the pursuit of regionalor environmental objectives.27 The permitted sub-

sidies for R&D included the financing of venturecapital funds and the provision to the private sec-tor of technologies and innovations developed in

government research laborato-ries. Also included in this cat-egory was public procurementpolicy in support of the pro-liferation of domestically de-fined standards for particulartechnologies. Moreover, in or-der to support a shift in eco-nomic activity to new products

or the use of new technologies, activities couldbe subsidized as long as they were in the pre-com-petitive phase (i.e. before they resulted in the pro-duction of goods that were exported or subject tosignificant import competition).

It is, however, important to note that theprovision that classified these subsidies as non-actionable came up for review in 2000, when noagreement over its extension could be reached.Thus these subsidies have now become actionable.

The Doha Declaration revisited this issuealong with the proposal of some countries to al-low certain subsidies for development. More spe-cifically, it stated that the Ministerial Conference“takes note of the proposal to treat measures im-plemented by developing countries with a view toachieving legitimate development goals, such asregional growth, technology research and devel-opment funding, production diversification anddevelopment and implementation of environ-mentally sound methods of production as non-actionable subsidies, and agrees that this issuebe addressed ... [as an outstanding implementa-

tion issue]. During the courseof the negotiations, Membersare urged to exercise due re-straint with respect to challeng-ing such measures” (WTO,2001: 6). Meanwhile, however,the issue of Article 8 subsidiesseems to have been eclipsed bynegotiations on other issues.

According to AguayoAyala and Gallagher (2005),

this call for restraint has been respected, and de-veloped and developing countries alike continueto use such subsidies under a tacit agreement not

Subsidies impose a cost onpublic budgets, whichdeveloped countries canafford more easily thandeveloping countries.

The SCM Agreementaffects the selectivefunction of policy.


to challenge them under the dispute settlementmechanism. To the extent that this is the case, theSCM agreement is a good illustration of how WTOrules and commitments that are equally binding,legally, impose more binding constraints on de-veloping countries economically. Firstly, subsidiesimpose a cost on public budgets, which developedcountries can afford more easily than developingcountries. For example, Aguayo Ayala and Gallagher(2005: 19) estimate that in 2003 the EU-15 spenta total of about 50 billion euros on Article-8-typesubsidies, mainly consisting of State aid and Struc-tural Fund payments. This corresponds to about25 per cent of developing countries’ total annualgross domestic expenditure on R&D (UNESCO,2005).

Secondly, Article-8-type subsidies are of con-cern primarily to developed countries in their questto develop high-tech capabilities and technologi-cal innovations. They differ from subsidies con-ditional on export performance or on the use ofdomestically produced goods, which were fre-quently used by the late industrializers to fosterindustrialization and technological catch-up. In-deed, Article-8-type subsidies can be a key de-vice for developed countries in their shift awayfrom the provision of basic funding for scientificR&D towards a strategic approach that establishesand targets research priorities in frontier sectorssuch as information and communications technol-ogy, biotechnology and nanotechnology, alongsidenew challenges arising in more traditional sectors,such as health care, national defence and the en-vironment. These are areas that many developed-country policymakers have come to consider ascrucial for economic growth and national pros-perity (see section D below).

Probably the most serious drawback of theSCM Agreement for development is that it pro-hibits making subsidies conditional on export per-formance. This has been an important instrumentin East Asia’s reciprocal control mechanisms,which have often been identified as key to thegreater success of industrial policy in that regioncompared to Latin America (Evans, 1995). Thusthe SCM Agreement withdraws a major monitor-ing standard that outward-oriented sectoral strat-egies in East Asia used successfully to ensure thatsupport was given only to those enterprises thatwere able to compete in international markets. It

is possible to establish other performance stand-ards under a reciprocal control mechanism (suchas the percentage of technology personnel em-ployed, the percentage of sales contributed by newproducts and the allocation of retained earnings).But none of these alternatives enable a perform-ance-based incentive policy that ensures interna-tional competitiveness and minimizes the risk ofabuse and rent-seeking.

These effective asymmetries cast some doubton arguments, such as made by Amsden (1999),that the bark of WTO law is worse than its bite.According to this argument the SCM Agreementformally leaves open the possibility of support-ing industrial upgrading, as developing countriesmaintain the ability to provide “boundless” sub-sidies for science and technology and the devel-opment of human capital. The main problem, theargument goes, is that developing countries havefailed to take advantage of the major types of non-actionable subsidies. It is probably true that sub-sidies is an area where, in principle, the main chal-lenge for many developing countries is to use theexisting flexibilities of the multilateral regulationsthrough innovative policy measures. However, inpractice, budgetary constraints may prevent somedeveloping countries from using subsidies as partof their industrial policies.

3. The Agreement on Trade-relatedAspects of Intellectual PropertyRights (TRIPS)

The TRIPS agreement establishes globalmandatory minimum standards for the grantingand protection of intellectual property rights inseveral areas, particularly copyrights and patents.28

It also provides a dispute resolution and enforce-ment mechanism. Countries are free to decide howto implement these provisions in accordance withtheir own legal and institutional systems. Appli-cation of TRIPS in developing countries (exceptthe LDCs29) has been mandatory since 2000. Ac-cording to Article 7, protection and enforcementof these rights must contribute to the promotionof technological innovation and the transfer anddiffusion of technological knowledge in order toimprove social and economic welfare. They must


also ensure a balance between the rights and obli-gations of the parties.

Many believe TRIPS to be the most contro-versial of URAs because of its potential to restrictaccess of developing countries to technology,knowledge and medicines. Theimportance of the Agreementfor industrial development liesin the fact that the procure-ment of proprietary knowl-edge has been among the keydeterminants of both early andlate industrialization. The his-tory of intellectual propertyrights protection shows thatcountries with low levels oftechnological capacity havegenerally used weak standards until they reacheda level of development at which their industriescould benefit from intellectual property rights pro-tection. Chang (2002) points out that many of thenow developed countries did not adopt intellec-tual property rights legislation or strict intellec-tual property rights standards when they were inthe process of economic catch-up.

Prior to their implementation of the TRIPSagreement, developing countries’ patent regimestypically included instruments to restrict the pri-vate rights of (largely foreign) patent holders(Amsden, 2001). Such instruments aimed to cre-ate more opportunities for local firms to accessforeign innovations, thereby encouraging learningand technological progress via imitation. This ena-bled these countries to movebeyond a critical thresholdlevel for domestic technologi-cal skills and promote nationalfirms that were eventually ableto engage in export activities.Knowledge procurement oc-curred in different ways, butreverse engineering from im-ported goods played an impor-tant role. This was facilitatedby relatively weak enforcement of intellectualproperty protection, particularly of patents.

The TRIPS Agreement severely restricts re-verse engineering and other forms of imitative in-novation since it upholds the private rights of pat-

ent holders. As a result, it tends to limit access ofdeveloping countries to proprietary knowledge.More precisely, TRIPS has introduced a numberof limitations on developing countries in design-ing their patent regimes. It broadens the scope ofpatents by requiring countries to extend patent pro-

tection to all fields of technol-ogy, while previously, coun-tries could deny patents to cer-tain types of goods or inven-tions in order to encourage re-verse engineering; it extendsthe duration of patent protec-tion uniformly to 20 years,while previously, countriescould offer patents of short du-ration; it reduces the scope ofexceptions, which are limited

to very specific cases; and it limits governments’ability to regulate patent holders, while previously,countries could make the granting of patents thatprovided monopoly benefits conditional upon lo-cal production or licensing and on the transfer oftechnology to local users (Shadlen, 2005a).30

The kinds of limitations introduced by TRIPSimplies an asymmetry that favours the producersand holders of protected intellectual property –mainly in developed countries – at the expense ofthose trying to gain access to protected intellec-tual content, mainly in developing countries.Moreover, the Agreement requires developing coun-tries to expand and enhance their intellectual prop-erty regimes, while providing very little to effec-tively facilitate and promote their access to tech-

nology. Indeed, the provisionsin the Agreement are specific,binding and actionable withregard to the protection of in-tellectual property, and non-compliance with these provi-sions can be challenged underthe WTO’s dispute settlementmechanism. By contrast, pro-visions regarding technologytransfer and technical coopera-

tion, which are of importance mainly for devel-oping countries are of a “best endeavour” natureand vaguely worded, making them difficult to en-force. As a result, non-compliance with these pro-visions is difficult to prove and, on a practicallevel, subject to no penalty.

... at the expense of thosetrying to gain access toprotected intellectualcontent, mainly indeveloping countries.

TRIPS implies anasymmetry that favours theproducers and holders ofprotected intellectualproperty, mainly indeveloped countries ...


Another expression of this asymmetry of fa-vouring incentives for the creation of patentableknowledge at the expense of the dissemination anduse of such knowledge is the implied additionalcost – in the form of royalties – to developingcountries of acquiring useful technology. The po-tential economic costs of the TRIPS Agreementfor developing countries to acquire patentableknowledge may be illustrated by the fact that in2001 only five developed countries (France, Ger-many, Japan, the United Kingdom and the UnitedStates) accounted for 83.6 per cent of the total pat-ent applications filed in the EU, Japan and theUnited States. And 82 per cent of scientific arti-cles worldwide were published in the OECD area,nearly two thirds of which were from G-7 coun-tries (OECD, 2005: 9 and 40). The surplus in theOECD-countries’ technology balance increasedfrom $9.6 billion in 1993 to $30.4 billion in 2003(OECD, 2005: 203). More-over, technology transfer is in-creasingly taking place withinmultinational firms, which re-duces the importance of con-tractual and non-equity modesof technology transfer andmakes it increasingly difficultfor developing countries to ob-tain useful technology on acommercial basis as envisagedby TRIPS.

While acknowledging that TRIPS wouldcause a significant revenue transfer from devel-oping to developed countries, it has sometimesbeen argued that its application would bring abouthigher returns to knowledge generation, which inturn will spur knowledge diffusion to developingcountries, including through increased flows ofFDI. However, there is no persuasive evidence forthis (Correa, 2000). A further economic handicapis that patentable research is increasingly carriedout in private entities, with the result that mostresearch activities are driven by their expectedeconomic pay-off. Given the limited financial re-sources of most developing-country firms, thereis a bias in the research agenda against those ar-eas that are primarily of importance to develop-ing countries.

Even though TRIPS has placed significantconstraints on countries’ autonomy in intellectual

property matters, it has left room for variationacross countries. For example, developing coun-tries can impose stringent rules on patent disclo-sure (i.e. disclosure of the intervention that is suf-ficiently clear and comprehensive for a skilledperson in the related activity to reproduce the in-ventive step), and subsequently grant narrow pat-ents, i.e. patents that protect a very limited rangeof variations and thus offer no – or little – protec-tion for variations that are not explicitly claimed.Or they can liberally grant improvement patentsto local actors and protect their “minor” innova-tions, which often refer to incremental innovationsthat build on more fundamental discoveries andare thus crucial for tailoring imported technolo-gies to local conditions. Such flexibilities allowlocal actors to “invent around” patents withoutgovernments risking litigation for infringement.Kumar (2003) argues that the patent regime in

place in Japan after the Sec-ond World War until the 1980sprovided for the granting ofnarrow patents, and that thisregime served as a model forthe late industrializers in Asia.Another example is flexibleuse of compulsory licencesthat allow a government to au-thorize itself or third parties touse a patent without the per-mission of the patent holder.Compulsory licences histori-

cally have been an important component of coun-tries’ patent regimes, and they are granted in awide range of situations (UNCTAD and ICTSD,2005). The TRIPS Agreement continues to leavecountries with a significant degree of autonomyin this regard, as it grants countries “considerableleeway to impose non-voluntary licensing of pat-ented interventions for any legitimate purpose andwithout undue constraints” (Reichman and Hasen-zahl, 2003: 2).

However, many developing countries haveengaged in regional and bilateral trade agreementswith developed countries that often foreclose partof the autonomy left open to developing countriesby TRIPS. For example, the United States – andto some extent the EU (Shadlen, 2005b) – usesregional arrangements to introduce legislation andpractices that go beyond the levels of intellectualproperty protection under TRIPS (USTR, 2004).31

Regional and bilateral tradeagreements with developedcountries often foreclosepart of the autonomy leftopen to developingcountries by TRIPS.


One of the greater obligations imposed by manyregional and bilateral trade agreements concernsthe reduced ability of governments to use com-pulsory licensing as a policy instrument (see, forexample, Maskus, 1997). In general, regional andbilateral trade agreements do not allow develop-ing governments to issue compulsory licencesexcept during declared states of national emer-gency, and even then they require increased lev-els of prior negotiations with the patent holder;moreover, where such licences are granted, theagreements substantially restrict the rights of thelicensee (Shadlen, 2005a).32

International harmonization of substantiveand enforcement rules on intellectual propertyrights has been further pursued at the World Intel-lectual Property Organization (WIPO), especiallyin the ongoing negotiations on a Substantive Pat-ent Law Treaty (SPLT). Discussions on the workprogramme of the SPLT so far have not led to anagreement as to whether aspects such as prior art,grace period, novelty and inventive step shouldbe included, as suggested by developed countries.Developing countries fear that adoption of thedeveloped countries’ proposal would eventuallyresult in the further harmonization of nationalpatent laws in areas of patent law that have so farbeen left to the discretion of national legislation.This would risk further reducing developing coun-tries’ flexibilities to decide on the stringency ofrequirements for disclosure and the standards forgranting patents, because it would eliminate coun-tries’ ability to determine what an invention is andhow the patentability standards are set (see, forexample, Correa, 2005).

4. Industrial tariffs

The use of industrial tariffs is in many re-spects not the best tool to promote diversificationand technological upgrading. For a number ofdeveloping countries, domestic markets are toosmall to sustain the scale needed for productionto be internationally competitive. Hence, tariffsmay end up protecting infant industries that areunable to come anywhere near world market priceand quality combinations. Industrial productionneeds to have an export component to reach an

efficient scale, and protection alone may well dis-courage efforts to export. Also, as noted earlier,protection can easily be abused, in the sense ofbeing unrelated to efforts to improve productiv-ity: once granted, firms will lobby vigorously tomaintain the protection. Therefore, industrial tar-iffs need to be used with great caution.

In spite of the numerous drawbacks of tariffuse, developing-country policymakers may behesitant to abandon industrial tariffs, mainly forthree reasons. First, tariffs remain an importantsource of fiscal revenue for many developingcountries. According to Kowalski (2005), shouldtariffs be completely abolished, many low-incomecountries would need to replace, on average,around 18 per cent (and in some cases over 50 percent) of their tax revenues with sources other thanimport duty (see also Laird, Vanzetti and Fernandezde Cordoba, 2006: 7). While the importance oftrade taxes in total revenue collection generallydeclines with economic development, in upper-middle-income countries import duties accounted,on average, for about 12 per cent of total revenuein the late 1990s. Improved tax collection andbroadening of the tax base can reduce the revenueshortfall resulting from declining taxes. However,many developing countries have already substan-tially lowered the share of import duties in theirtotal revenues over the past two decades,33 whilelow-income countries in particular have been un-able to recover the revenues lost from trade liber-alization (Baunsgaard and Keen, 2005). As a result,they are likely to experience difficulty in findingsupplementary sources of revenue that further tar-iff reductions would necessitate. Moreover, the de-cline in government revenue resulting from areduction in import duties may lead to an increasein public deficits or a decline in public investment.But public investment has a crucial impact on eco-nomic development because it seeks to improveeducation, health and other social indicators. Andpublic investment is often complementary to privateinvestment, so that a decline in public investmentbelow a critical level can seriously compromisean economy’s development prospects (TDR 2003).On the other hand, tariff cuts could lead to a sub-stantial increase in imports, with lower tariff rateslevied on a higher volume of imports; in princi-ple, this could maintain the value of import lev-ies, but this is unlikely to occur because ofbalance-of-payments constraints.


Second, the provision of subsidies, ratherthan broad-based protection, could provide the in-centives required for innovative investment, asdiscussed above. However, as already mentioned,the SCM Agreement has highly circumscribed theuse of subsidies in areas where, formerly, both themature and late industrializers of today activelyused them during their eco-nomic catch-up. Moreover, asmentioned earlier, tight budget-ary constraints limit developingcountries’ ability to use subsi-dies.

Third, and perhaps mostimportantly, the economic im-pact of changes in industrialtariffs is often assessed in terms of welfare gainsor losses resulting from the reallocation of exist-ing resources. From this perspective, a trade policyaimed at low and uniform tariffs across industrialsectors with full binding coverage will maximizea country’s welfare benefits.34 But such an assess-ment pays little attention to the implications oftariff cuts and harmonization for capital accumu-lation, technological change and productivitygrowth that underlie industrialization and eco-nomic development.

It may be useful to recall that industrial tariffswere the main element of protection that today’sdeveloped countries used during their industrialdevelopment. As illustrated in table 5.1 and analysedin some detail by Bairoch (1993), the United Statesmaintained average industrial tariffs at around 40 percent, and never below 25 percent except for short periods,throughout most of the periodbetween 1820 and 1945. Re-garding the United Kingdom,Bairoch (1993: 46) notes thatprior to its substantial move to-wards free trade with the repealof the Corn Laws in 1846, Brit-ain had achieved its techno-logical lead “behind high and long-lasting tariffbarriers”. He also notes that the country had ac-tively used infant industry protection, export sub-sidies, import tariff rebates on inputs used inmanufacturing for export, and export quality con-trol. Table 5.1 also shows that at the beginning ofeconomic catch-up in West European countries

following the Second World War, the level of tar-iffs on manufactured products was also fairly high(see also, Chang, 2002).

Comparing tariff levels at similar levels ofper capita income (measured at purchasing powerparity) shows that average tariffs in today’s de-

veloped countries were muchhigher when they had similarper capita income levels as to-day’s developing countries (seealso Akyüz, 2005). In this sense,tariff policy in today’s devel-oping countries appears to berelatively liberal. Towards theend of the nineteenth century,when the United States had a

per capita income similar to today’s weighted aver-age level in developing countries (i.e. about $3,700in 1990 dollars measured in purchasing powerparity), the level of its weighted average appliedtariffs on manufactured goods was close to 50 percent, compared to 6.5 per cent in developing coun-tries today (tables 5.1 and 5.2). In 1950, when theUnited States had evolved as the world’s techno-logical leader with a per capita income more thandouble the average level in today’s developingcountries, the level of its weighted applied tariffson manufactured products still exceeded the cur-rent level in today’s developing countries. Whenthe United States had the same level of per capitaincome as the Republic of Korea today, itsweighted applied tariffs were higher (7.0 per centcompared to 4.5 per cent), and when it had thesame per capita income level as Brazil, China or

India today, its tariffs were sev-eral times higher. This is alsotrue, to varying degrees, for theEuropean countries in table 5.1(i.e. Germany, France and theUnited Kingdom).

These comparisons of therelative levels of tariff protec-tion between the developed

countries during their catch-up phases and today’sdeveloping countries do not tell the whole story.Developed countries also benefited from the ad-ditional protection of natural trade barriers in theform of transportation and information costs,which were higher in the past than they are today.More importantly, the productivity gap between

A flexible tariff policyconsists of maintainingbound tariffs at a higherlevel and ...

... modulating applied tariffson particular industrialsectors around a loweraverage level.


developed and catching up countries, which is themain justification for tariff protection in catch-upperiods, is much greater now than it was in the past.Thus in order to obtain the same degree of actualprotection, today’s developing countries wouldneed to impose relatively higher tariffs than thosethat were used by the now developed countriesduring their catch-up periods (Chang, 2002: 67).

Chang (2002) shows that the great importanceof tariffs in promoting economic developmentuntil the 1920s was associated with the underde-velopment of other instruments of public policy.Governments’ limited abilities to raise tax rev-enues circumscribed their use of subsidies. More-over, non-tariff measures such as quotas, voluntaryexport restraints and anti-dumping were developedonly after the Second World War, before evolvinginto standard instruments in support of industrialdevelopment. Therefore, in a sense, the limited

range of trade policy instruments available to de-veloped countries until the 1920s resembles thesituation faced by developing countries today,given that WTO rules and commitments curtailthe use for economic catch-up of instruments suchas export-related subsidies, performance require-ments for foreign investors, and reverse engineer-ing and imitating of foreign technology.

As proposed by Akyüz (2005), in such cir-cumstances, it would be important for developingcountries to be able to modulate applied indus-trial tariffs in order to pursue a pattern of publicsupport policies such as that illustrated in figure5.1 above. That is, the variation of applied tariffslevied on particular product categories, in accord-ance with their path of technological upgrading,could be a key instrument of sectoral policy. Tobe sure, this kind of tariff policy does not implyeither the imposition of high applied tariffs for all

Table 5.1

TARIFFS ON MANUFACTURED PRODUCTS AND PER CAPITA INCOME IN SELECTEDDEVELOPED COUNTRIES, 1820–1980

Country 1820a 1875b 1913 1950 1980

Tariffs, weighted averages

(Per cent)

United States 35–45 40–50 44.0 14.0 7.0

United Kingdom 45–55 0 0.0 23.0 8.3

Germany 8–12c 4–6 13.0 26.0 8.3

France .. d 12–15 20.0 18.0 8.3

Per capita income

(1990 international dollars)

United States 1 257 2 445 5 301 9 561 18 577

United Kingdom 1 707 3 191 4 921 6 907 12 928

Germany 1 058 1 821 3 648 3 881 14 113

France 1 230 1 876 3 485 5 270 15 103

Source: Tariff data from Bairoch, 1993: 40; income data from Maddison, 2001: 264, 276–279.a Very approximate rates. Range of averages, not extremes.b Per capita income data for 1870.c Prussia.d Numerous and large restrictions on imports of manufactured products render calculations of average tariff rates insignificant.


sectors at any one time or the imposition of highaverage applied tariffs. On the contrary, it is likelyto result in lower average applied tariffs than wouldbe the case if tariff policy werelooked at from a tariff line bytariff line perspective.

This kind of flexible tar-iff policy would be best accom-modated by a strategy of main-taining bound tariffs at a rela-tively higher level (or main-taining a large part of industrialtariffs unbound) and modulat-ing applied tariffs on particu-lar industrial sectors around a relatively loweraverage level. This would be possible if industrialtariff reduction obligations from internationalagreements extended only to average tariffs, andnot to individual tariff lines,35 which has indeedbeen the case in all multilateral trade agreementsconcluded so far.

Apart from supporting diversification andtechnological upgrading, this tariff policy patternprovides several additional advantages (see alsoAkyüz, 2005: 26). First, it would balance multi-lateral disciplines with national policy flexibility,because it would encourage countries to chooseapplied levels of their industrial tariff lines withinthe overall limit of an average bound tariff, ratherthan seeking revenue maximization or accommo-dation of wide-ranging demands from lobbygroups. Second, it would encourage policymakersto view tariff protection for specific industries atthe lower rung of the technology ladder as a tem-porary measure, to be phasedout and replaced by tariff pro-tection for industries at higherrungs of the ladder until theyare able to compete in worldmarkets. Third, as a conse-quence, it would encouragepolicymakers to take a longer-term view of their economy’stechnological developmentand multilateral commitments.

A number of developing countries have main-tained a tariff regime that allows them to modulateapplied tariffs on manufactured goods. Table 5.2shows that for developing countries as a group and

for all individual developing countries in the ta-ble, except China, bound tariffs on manufacturessignificantly exceed applied tariffs,36 thus leaving

room to adjust tariffs in supportof domestic producers. More-over, many developing coun-tries have less than full bindingcoverage or deploy significant-ly different levels of both boundand effectively applied tariffsacross manufactured goods, asshown by relatively high valuesof the coefficient of variationthat reflects intersectoral dis-persion. Among the countries

in the table, India maintained the greatest degreesof freedom, as its tariff regime combined relativelyhigh levels of bound and applied tariffs, as well assizeable intersectoral dispersion and a relativelylow binding coverage.37 This tariff profile left In-dia significant space for tariff modulation. By con-trast, Chile has a relatively low level of tariffs,very little intersectoral dispersion and full bind-ing coverage. China and Mexico have also con-served relatively little flexibility in their tariffprofile. The other countries in the table occupyintermediate positions as they either conserve rela-tively high tariff levels but little intersectoral dis-persion and (close to) full binding coverage (Ar-gentina, Brazil and Egypt), or relatively low tar-iff levels but with some intersectoral dispersionand less than full binding coverage (the Republicof Korea).

The current multilateral negotiations on non-agricultural market access (NAMA) are set to re-

duce this flexibility in tariffsetting and binding that devel-oping countries have main-tained. The framework adoptedfor modalities of industrial tar-iff reductions, as contained inAnnex B of the so-called JulyPackage (WTO, 2004) stipu-lates a reduction in tariffs ac-cording to a non-linear Swissformula, and an increase inbinding coverage. While at the

time of writing (June 2006) the definition of fullmodalities remained to be negotiated, the overallobjective of the adopted approach is to bind andreduce all industrial tariffs with a view to harmo-

The current multilateralnegotiations on NAMA areset to reduce the flexibilityin setting and bindingtariffs.

Since the Uruguay Roundreduced the freedom to useother policy instruments,the relative importance ofindustrial tariffs hasincreased.


Table 5.2

TARIFFS ON MANUFACTURED PRODUCTS AND PER CAPITA INCOME IN SELECTEDDEVELOPING COUNTRIES AND COUNTRY GROUPS, 1985–2005

Applied tariffs Bound tariffs

Simple Weighted Simple Weighted Bindingaverage average average average coverage Memo item:

Coefficient Coefficient Per capitaCountry Year (Per cent) of variation (Per cent) (Per cent) of variation (Per cent) (Per cent) incomea

Argentina 1985 28.3 .. 26.2 .. .. .. .. 10 008

1990b 14.8 0.5 13.6 31.7 0.2 32.5 100.0 10 755

1995 13.1 0.6 9.0 31.6 0.2 32.0 100.0 11 254

2000 15.3 0.5 11.2 31.6 0.2 32.2 100.0 12 174

2005 10.9 0.7 9.4 31.6 0.2 32.7 100.0 12 222 c

Brazil 1985 60.7 .. 53.5 .. .. .. .. 6 640

1990 34.4 0.5 28.1 30.7 0.2 28.7 100.0 6 497

1995 13.7 0.6 12.7 30.7 0.2 30.0 100.0 6 940

2000 17.0 0.4 14.8 30.6 0.3 29.6 100.0 7 301

2005 12.6 0.6 9.1 30.6 0.3 28.6 100.0 7 531 c

Chile 1985 19.8 .. 18.5 .. .. .. .. 4 969

1990b 11.0 0.1 10.9 25.0 0.0 25.0 100.0 6 764

1995 10.7 0.2 10.6 25.0 0.0 24.9 100.0 7 999

2000 9.0 0.0 9.0 25.0 0.0 25.0 100.0 9 115

2005 5.0 0.5 4.4 25.0 0.0 25.0 100.0 9 993 c

Mexico 1985 12.1 .. 11.1 .. .. .. .. 7 870

1990d 14.4 0.3 13.0 35.0 0.1 34.9 100.0 7 758

1995 12.5 0.5 7.7 35.0 0.1 34.8 100.0 7 619

2000 18.0 0.4 14.6 35.0 0.1 65.4 100.0 9 046

2005 8.5 1.0 2.8 35.0 0.1 35.5 100.0 9 010 c

Chinae 1985 41.9 .. 33.2 .. .. .. .. 1 181

1990b 40.0 0.8 35.6 9.5 0.7 8.9 100.0 1 944

1995f 21.3 0.7 18.2 9.5 0.7 7.9 100.0 2 971

2000 15.8 0.6 13.5 9.5 0.7 6.0 100.0 3 928

2005c 9.5 0.7 5.8 9.5 0.7 5.3 100.0 5 419

India 1985 101.9 .. 99.4 .. .. .. .. 1 385

1990 79.9 0.5 70.8 35.5 0.5 29.3 71.5 1 701

1995g 28.9 0.3 21.1 35.4 0.5 31.2 71.3 2 154

2000h 30.7 0.3 28.3 35.4 0.5 30.0 71.3 2 480

2005 17.7 0.4 12.6 35.5 0.5 28.0 71.5 2 885c

Rep. of Korea 1985 23.4 .. 22.5 .. .. .. .. 6 649

1990 12.8 0.2 11.4 11.0 2.1 7.3 95.4 9 792

1995 7.8 0.2 7.3 11.2 2.2 7.2 95.4 13 597

2000i 8.0 0.3 6.2 11.3 1.8 6.1 94.8 15 143

2005c 7.2 2.1 4.5 11.3 1.7 6.4 94.8 18 840

Egypt 1985 37.5 .. 30.8 .. .. .. .. 2 845

1990 .. .. .. .. .. .. .. 2 896

1995 24.0 0.9 22.2 28.5 0.7 26.4 99.3 3 025

2000j 19.2 0.7 17.5 28.5 0.7 25.1 99.3 3 326

2005k 19.0 0.9 16.9 28.4 0.7 24.9 99.2 3 729

/...


nizing them, both across countries and across in-dividual tariff lines.

In sum, a developing country’s tariff policyneeds to be part of a long-term industrializationstrategy. Selective trade liberalization should be inline with a country’s ability to achieve technologi-cal upgrading. In addition, temporary protectionshould be combined with export promotion asso-ciated with quantitative targets that are easy to

monitor and allow governments to withdraw sup-port from firms that do not achieve upgradingtargets. Given the numerous drawbacks and risksassociated with their use, tariffs need to be imple-mented with considerable caution. But since theURAs reduced the degrees of freedom for devel-oping countries to use other policy instrumentsdesigned to support diversification and technologi-cal upgrading, the relative importance of industrialtariffs has increased.

Table 5.2 (concluded)

TARIFFS ON MANUFACTURED PRODUCTS AND PER CAPITA INCOME IN SELECTEDDEVELOPING COUNTRIES AND COUNTRY GROUPS, 1985–2005

Applied tariffs Bound tariffs

Simple Weighted Simple Weighted Bindingaverage average average average coverage Memo item:

Coefficient Coefficient Per capitaCountry Year (Per cent) of variation (Per cent) (Per cent) of variation (Per cent) (Per cent) incomea

Memo item:Developing countries

1985 .. .. .. .. .. .. .. 2 946

1990 26.1 0.9 20.5 26.7 0.7 17.7 68.6 2 875

1995 14.8 0.9 9.5 32.6 0.5 19.8 73.2 3 225

2000 14.0 0.8 11.3 36.5 0.5 19.1 70.7 3 612

2005c 10.5 1.1 6.5 30.2 0.5 15.6 67.9 3 915

Source: Tariff data for 1985 from UNCTAD, 1994. All other tariff data from UNCTAD, TRAINS Database at WITS. Income datafrom World Bank, World Development Indicators (WDI) Database.

Note: The data in the table refer to manufactures, and thus exclude tariffs applied in a number of sectors (such as extractiveindustries) that are included in industrial tariff data (i.e. the subject of NAMA negotiations). Data for developing coun-tries are only indicative, because averages are based on less than full country coverage for some years.

a PPP (constant 2000 international dollars), data for developing countries as a group refer to the median.b 1992. c 2004. d 1991.e Data for applied tariffs are based on a more recent list, that includes a larger number of tariff lines than the list used for

data on bound tariffs. Thus, prior to China’s accession to the WTO in 2001, the numbers in the table for applied tariffscan exceed those for bound tariffs, even with full binding coverage.

f 1996. g 1997. h 2001.i 1999. j 1998. k 2002.


The objective of this section is, first, to pro-vide empirical evidence of industrialization andtechnological upgrading in developed and devel-oping countries over the past 25 years and, sec-ond, to present a few selected case studies on theassociated trade and industrial policies. It attemptsto set the general context in which countries haveundertaken trade and industrial policies and ex-amines whether, how and to what extent multilat-eral trade rules and commitments have affectedcountries’ autonomy in policy-making and imple-mentation.

1. Industrial dynamism:recent empirical evidence

The degree of expansion of their manufac-tured exports and improvement of their share inworld trade, particularly in high-tech products, isoften taken as a measure of the pace of industri-alization and technological upgrading in develop-ing countries. However, the higher import contentof domestic production brought about by trade lib-eralization, together with the greater participationof developing countries in import-dependent, la-bour-intensive, low-value-added processes in in-ternational production networks, implies thatincreases in their manufactured exports may of-ten have taken place without commensurate in-creases in income and value added, as discussedin TDR 2002.

Table 5.3 presents data on the shares of de-veloped and developing economies in world manu-

facturing trade and production over the past25 years. Comprehensive data on manufacturingvalue added (MVA) are available only up to 2003,so that they do not reflect the more recent impactof trade and industrial policies. The data show thatsuccess in exporting manufactures is not an ap-propriate indicator of a country’s industrial de-velopment. They reveal a pattern comprising thefollowing features:

• The shares of developing economies both inworld manufactured exports and world MVAshowed a sharp increase during the period1980–2003, but growth in exports was muchstronger than in value added. This contrastswith the experience of developed countries,whose share in world manufacturing exportsfell between 1980 and 2003, while their sharein world MVA rose significantly.

• There has been wide variation in industrialperformance across developing regions, lead-ing to a rise in the concentration of industrialactivities. South and East Asia are the mostindustrialized in the developing world; theircombined share in total world MVA has in-creased the most, more than doubling since1990 to exceed 17 per cent in 2003. The LatinAmerica and Caribbean region has experi-enced the strongest decline in its share ofworld MVA, the sharpest fall being in the1980s and early 2000s.

• China succeeded in more than tripling itsshare in both world MVA and world manu-factured exports between 1990 and 2003. Itsexperience closely resembles that of the Re-

D. Industrial dynamism and national policies:recent experiences


public of Korea between 1980 and 2000,which (together with the Taiwan Province ofChina) is often portrayed as exemplifyingsuccessful economic catch-up among the lateindustrializers.38 This outcome strongly con-

trasts with that of Mexico, whose share inworld manufactured exports increased morethan fivefold during the 1990s, while its sharein world MVA only about doubled during thesame period. Moreover, both these shares de-

Table 5.3

SHARE OF SELECTED DEVELOPING ECONOMIES AND REGIONAL GROUPS IN WORLDMANUFACTURING VALUE ADDED AND MANUFACTURED EXPORTS, 1980–2003

(Percentage share)

Share in world Share in worldmanufacturing value added exports of manufacturesa

Region/economy 1980 1990 2000 2003 1980 1990 2000 2003

Developed countriesb 64.5 74.1 74.9 73.3 74.1 77.9 67.3 65.4

Developing countries 16.6 17.0 22.8 23.7 18.9 18.3 28.9 29.7

Latin America and the Caribbean 7.1 5.6 5.4 4.4 4.3 2.4 4.7 4.1

Argentina 0.9 0.8 0.8 0.5 0.2 0.3 0.3 0.3Brazil 2.9 2.2 1.1 0.9 0.8 0.8 0.8 0.8Chile 0.2 0.1 0.2 0.2 0.2 0.2 0.2 0.2Mexico 1.9 1.1 2.0 1.7 0.8 0.5 2.7 2.2

South and East Asia 7.4 8.7 15.2 17.2 7.6 13.6 21.7 22.7

China, Taiwan Province of 0.6 1.1 1.3 1.1 1.3 2.3 2.7 2.3Republic of Korea 0.7 1.4 2.2 2.3 1.1 2.2 3.1 3.0

ASEAN-4 1.2 1.5 2.4 2.8 1.0 2.0 4.2 3.7

Indonesia 0.4 0.5 0.9 1.1 0.2 0.4 0.8 0.6Malaysia 0.2 0.2 0.5 0.5 0.4 0.7 1.6 1.5Philippines 0.3 0.2 0.3 0.3 0.2 0.2 0.7 0.5Thailand 0.3 0.5 0.7 0.8 0.2 0.6 1.1 1.1

China 3.3 2.6 6.6 8.5 1.0 1.7 4.3 6.5

India 1.1 1.1 1.2 1.4 0.3 0.5 0.7 0.9

Africa 0.9 0.9 0.8 0.8 5.4 2.6 1.8 2.0

Source: UNCTAD secretariat calculations, based on UNIDO, Handbook of Industrial Statistics 1996; UNIDO, InternationalYearbook of Industrial Statistics, 2006; World Bank, World Development Indicators online; Taiwan Province of China,Monthly Bulletin of Statistics online; UN COMTRADE and UNCTAD estimates.

Note: Calculations in current dollars.a To ensure data comparability, the definition of this product category follows industrial statistics. It therefore includes

processed primary products in addition to manufactures, as defined in trade statistics. For further discussion of thisstatistical issue, see Wood and Mayer, 1998. Using the definition of manufactures in trade statistics (i.e. SITC 5 through8 less 68) has a negligible effect on the shares of the individual countries listed in the table. By contrast, it results in anumber of sizeable changes for country groups. Most of these changes are confined to 1980 when, based on thedefinition of manufactures in trade statistics, the share in world exports of manufactures was 15.5 per cent for developingcountries and 78.1 per cent for developed countries (using the UNIDO country classification). The remaining discrepancywith respect to TDR 2002, table 3.5 for 1980 trade data is due to data (re-)estimation, in particular for China.

b To ensure data comparability, the definition of this group is that used by UNIDO prior to 2006. Hence, contrary to thecurrent standard definition of the United Nations, it does not include the Czech Republic, Estonia, Hungary, Latvia,Lithuania, Poland and Slovakia.


clined during the early 2000s. Brazil alsoshows an interesting development, with itsshare in world manufactured exports remain-ing largely unchanged over the past 25 yearsand its share in world MVA falling signifi-cantly, during the 1980s, and even more soduring the 1990s.

These different experiences in industrial de-velopment across individual developing countriesare closely related to changes in the compositionof the respective countries’ industrial activities,as shown in figure 5.2. The figure concentrateson Brazil, Mexico and the Republic of Korea, be-cause comprehensive data for China are not avail-able. The Republic of Korea shows the classicpicture of successful industrial structural changeand technological upgrading. While the shares ofresource-intensive and labour-intensive productsin that country’s total industrial activities fell dur-ing the period 1980–2003, the share of technology-intensive products grew continuously, to reach72 per cent in 2003.

Brazil and Mexico show an entirely differ-ent picture. Mexico experienced a slight increasein the share of technology-intensive products inits total industrial activities between 1995 and2000, which is likely to have been associated withthe growing activities in the automobile sectorfollowing the entry into force of the North Ameri-can Free Trade Agreement (NAFTA) in 1994. Never-theless, resource-intensive manufactures havemaintained the highest share in the country’s in-dustrial activities. Brazil experienced little changein the relative importance of the three categoriesof industrial production over the period 1980–2003.

2. National policies for industrialdevelopment: some recent experience

(a) Latin America

Most countries in Latin America adoptedcomprehensive economic reform programmesduring the 1980s and early 1990s, which empha-sized more stringent monetary and fiscal policies,liberalization, privatization and deregulation (asnoted in chapter II). These were accompanied by

Figure 5.2

SHARE IN TOTAL MANUFACTURING VALUE ADDEDOF MAJOR PRODUCT CATEGORIES IN THE

REPUBLIC OF KOREA, BRAZIL ANDMEXICO, 1980–2003

(Per cent)

Source: UNCTAD secretariat calculations, based on data forMexico and the Republic of Korea from OECD, Struc-tural Analysis (STAN) Database; and on data for Bra-zil from ECLAC, Program for the Analysis of IndustrialDynamics (PADI) Database.

Note: The shares are based on values in 1985 dollars. Fora detailed explanation of the product categorization,see the notes to this chapter.


the discontinuation of inward-oriented industri-alization strategies and, in most cases, the disman-tling of institutional structures, such as develop-ment banks, that had been of major importance tothose strategies. However, in the mid-1990s, therewas a growing belief that, even though the reformshad been successful in bring-ing about macroeconomic fi-nancial stability, they were notachieving the promised resultsin terms of economic growthand sustainable improvementin the countries’ balance of pay-ments. Moreover, the industrialrestructuring process that ac-companied economic reformhad led to the undesired out-come of premature deindustri-alization (i.e. a decline in theshare of industry in aggregate output and employ-ment accompanied by the growing importance ofthe primary sector, rather than by that of the serv-ices sector as has generally been the case in thebenign process of deindustrialization in developedcountries). As discussed, for example, in TDR 2003and Cimoli et al. (2006), the pattern of industrialproduction and trade, which had resulted frompolicies that relied on unfettered market forces,was shifting the composition of output and exportstowards natural-resource-based products at theexpense of those sectors that have the greatestpotential for productivity growth and technologi-cal upgrading (i.e. manufactures, and particularlythe high-technology-intensive ones).

In the mid-1990s, a sig-nificant number of countries inLatin America – in particularBrazil, Colombia and Mexico– adopted medium- or long-term plans for the (re-) devel-opment of their industrial sec-tor, and a number of othercountries began to implementpolicies with the same objec-tive, though through a less for-malized strategy. The general approach of theseplans and strategies has been to: (i) maintain theirmacroeconomic orientation, emphasizing financialstability and broad-based trade liberalization, of-ten accompanied by financial liberalization, witha view to achieving an efficient allocation of re-

sources in response to signals from world marketprices; and (ii) complement this macroeconomicpolicy with microeconomic policies designed tomake their domestic enterprises internationallycompetitive and facilitate their international inte-gration. These microeconomic policies have of-

ten been embedded in nationalcompetitiveness strategies,39

sometimes specifically tar-geted at small- and medium-sized enterprises (SMEs), withthe general aim of fostering la-bour productivity and techno-logical innovation in existingindustries. Such business pro-motion policies have beencombined with measures de-signed to attract FDI throughimprovements in the host coun-

tries’ locational advantages.40 FDI has also beensought with the objective of developing produc-tion, particularly for export activities in new in-dustrial sectors for the respective host country(Melo, 2001; Peres, 2006).41

The specific policy measures employed toachieve the central objective of enabling domes-tic enterprises to gain a competitive edge over theirforeign competitors have differed across the coun-tries of the region. In general, they have beenaimed at promoting exports, output growth andinvestment, and higher productivity and competi-tiveness. Tax and credit incentives have been themost important types of policy instruments for thepromotion of exports, particularly of non-traditional

exports. Such fiscal incentiveshave been characterized by a re-duced use of subsidies and anemphasis on tax refunds on do-mestic inputs or duties paid onimported inputs, and the crea-tion of export processing zones.

The main objective ofcredit policies to promote ex-ports has been to provide ac-

cess to working capital or initial investment fi-nancing for new export activities. These loans aregenerally offered at domestic market rates, whichare usually higher than those of international fi-nancial markets. Brazil’s programme for the fi-nancing of exports (PROEX) is a major excep-

In the mid-1990s, asignificant number ofcountries in Latin Americaadopted plans for the (re-)development of theirindustrial sector ...

... attempting to enabledomestic enterprises togain a competitive edgeover their foreigncompetitors.


tion, as it allows Brazilian exporters of certaingoods and services, or their foreign importers, toobtain trade finance on international market con-ditions (i.e. granting mainly interest equalizationpayments).42 Embraer, the Brazilian manufacturerof regional aeroplanes, has been one of the firmsto use this programme. How-ever, in 1996–1999, Canada,home to Embraer’s main com-petitor, Bombardier, chal-lenged the compatibility ofPROEX with WTO rules andcommitments. The WTO dis-pute settlement panel ruledthat the subsidies granted un-der PROEX were inconsistentwith the SCM Agreement’sprovision prohibiting the use of subsidies contin-gent upon export performance. While Brazil ap-pealed against this ruling,43 it also changed the pro-visions of PROEX. Subsequently, ruling on a fur-ther challenge from Canada, the WTO panel re-port of July 2001 established that the revisedPROEX falls under the exceptions provided inAnnex I of the SCM Agreement, and thus is notagainst WTO rules and commitments.44 While thisillustrates that WTO rules and commitments canreduce the degrees of freedom of national policy-making, it also highlights the fact that (i) the WTOprovides a transparent legal structure to deal withdisputes, and thus minimizes the risk of “tradewars”, (ii) while much leeway in policy-makingmay have been lost under multilateral rules, coun-tries can nevertheless retain a degree of flexibil-ity through creative policy-making, and (iii) sig-nificant administrative and ne-gotiating capacities are re-quired to fully benefit from theWTO regime.

Policies to promote out-put growth and investmenthave also emphasized creditand tax incentives. Generally,these incentives have been provided to all eco-nomic sectors, but some have a sector-specificdimension. Contrary to much of the policiesadopted in connection with the inward-orientedindustrialization strategies of the past, which fo-cused on support to manufacturing activities, re-cent fiscal incentives and loans by developmentbanks have emphasized extraction industries (such

as the oil, mineral or forestry sectors), tourism, ora variety of services sectors (ranging from infra-structure to the film industry). In some countries(e.g. Brazil and Mexico), a variety of sectors ben-efit from credit and tax incentives, but most coun-tries have focused support on a narrow range of

sectors (Melo, 2001: table 3).45

However, Mexico’s recent plan(contained in the National Planfor Development, 2001–2006)to adopt sector-specific policiesto stimulate investment witha view to generating greaterdomestic value added andstrengthening the linkagesamong local production chainshas been hampered by insuf-

ficient budgetary funds and by long delays in im-plementation (Moreno-Brid, Rivas Valdivia andSantamaria, 2005: 14).

The promotion of scientific and technologi-cal upgrading has been an important element inpolicies designed to improve enterprise produc-tivity and international competitiveness. As withexport promotion and support to output growthand investment, the provision of credit and fiscalincentives have been the main types of policy in-struments used to promote technological upgrad-ing. Other instruments include government fund-ing of R&D projects and strengthening coopera-tion between public research institutions and pri-vate enterprises. However only limited budgetaryresources seem to have been made available fortechnology development and innovation. For ex-

ample, in Brazil, public fund-ing covered only 10 per centof private sector R&D activi-ties during the period 1998–2000; much of the rest wassourced from company profits,given that the high domesticinterest rate provided little in-centives for financing such ac-

tivities through loans (De Negri, 2006). Anothertype of policy used to promote scientific and tech-nological development has been support to SMEs,which are often considered highly important ininnovation. Particularly in countries of the AndeanCommunity (Bolivia, Colombia, Ecuador, Peru,and Venezuela) or in Central America (Peres,2006), this has taken the form of support to SME

Incentives have beenprovided to all economicsectors, but some havea sector-specificdimension ...

... emphasizing extractionindustries, tourism, orservices.


clusters (i.e. an agglomeration of SMEs in thesame or related lines of business located in a givengeographical area).46

While the bulk of policy support mentionedso far has been applied to all economic sectors,many countries in Latin America have also usedsector-specific measures. For example, credit andfiscal incentives have been directed mainly to at-tract FDI in high-technology-intensive sectors,such as the information and communications tech-nology sector, and in the automotive industry.However, the shift away from relatively central-ized tax policies supportive of economic devel-opment and their increasing devolution to regionaland local governments has often led to “fiscalwars”, whereby different re-gions and municipalities offerincreasingly generous incen-tive packages to attract newTNCs or trigger their reloca-tion away from existing TNCproduction sites within thesame country. In Brazil, forexample, in order to make theautomotive sector more attrac-tive to FDI, individual incen-tive packages amounted to asmuch as $300,000 per job,leading some observers to con-clude that subsidies of this sizeare likely to exceed the gains from reallocatingplants within Brazil (Christiansen, Oman andCharlton, 2003).

Such incentives were offered to TNCs in thehope that they would provide technological andknowledge spillovers to domestic producers, aswell as facilitate the integration of such produc-ers into international markets, such as throughtheir participation in TNC-managed internationalproduction networks (Lugones, 2006). Mexico,among the countries in Latin America, has gonefurthest in this regard. In December 1993 (i.e. justbefore NAFTA took effect), Mexico enacted a newlaw on FDI that simplified administrative proce-dures and eliminated virtually all restrictions onFDI in manufacturing. The law also provided forthe progressive removal of all performance re-quirements on FDI in the automotive sector.Moreover, imported inputs for re-exportation wereallowed to enter the country tax-free. As a result

of such tax benefits “manufacturing firms that relyon foreign inputs entering as temporary importspay approximately 30% lower input costs thansimilar firms which use locally produced inputs”(Moreno-Brid, Rivas Valdivia and Santamaria,2005: 22). This is probably why, during the pe-riod 1990–2000, producers of auto parts that en-joyed such tax benefits through the maquiladoraregime increased output and employment muchmore than those that did not benefit from such aregime, even though the latter experienced morerapid productivity growth (Mortimore and Barron,2005: 20). Indeed, it is doubtful whether the pro-vision of such benefits has supported the entry ofdomestic enterprises into international productionnetworks and contributed to domestic industrial

development. Rather, it mayultimately have benefited main-ly the automobile manufac-turers in the United States byincreasing their competitive-ness vis-à-vis their Asian com-petitors in the United Statesmarket (Mortimore and Barron,2005).

Partly in reaction to this,in the early 2000s, the Mexi-can Government attempted totransform the export platforms,which had mainly carried out

assembly activities of imported production inputsfor re-export to the United States. It sought to con-vert them into manufacturing centres that wouldproduce auto parts in addition to assembling theminto vehicles, and aimed at other large marketsfor automobiles in addition to the United States,such as the EU and Japan. Free trade agreementswere used as a major instrument to that effect,because it was expected that the rules of originassociated with these agreements would bringabout increased levels of local content. However,according to Mortimore and Barron (2005: 25–26),this strategy has largely failed, mainly becausevery few enterprises located in Mexico were ableto provide parts and components that would meetinternational price and quality standards. Thisdemonstrates the difficulties of the Mexican au-tomotive industry in re-establishing local produc-tion linkages and furthering technological upgrad-ing, which may have been due to the adverse ef-fects of previously overgenerous treatment of FDI.

A major shortcoming ofindustrialization strategiesin Latin America has beenthe lack of coherencebetween the measurestaken at the micro leveland the macroeconomicenvironment.


If comprehensive tax incentive packages to attractFDI had not been offered, there may have beensufficient additional budgetary revenues to signifi-cantly increase the Government’s ability to pro-vide greater funding support to R&D and innova-tive activities.

In sum, the microeconomic policies that gov-ernments in many Latin American countries haveadopted since the mid-1990s may have been suc-cessful in supporting the international integrationof domestic enterprises and in attracting more FDIto the region. On the other hand, these policies donot appear to have significantly moved produc-tion and export patterns towards products of highertechnology intensity. According to Peres (2006),it is not easy to assess the success of the micro-economic policies in these terms because in manycases they have been adopted in the absence ofquantitative criteria that would have enabled anobjective assessment of their effects.

Nevertheless, a significant shortcoming of thecurrent industrialization strategies in many LatinAmerican countries relates to the lack of coherencebetween the adopted microeconomic measures andthe prevailing macroeconomic environment. Theorthodox macroeconomic policy stance of mostcountries in the region helped to bring inflationunder control and establish a reasonable degreeof monetary and fiscal discipline. However, partlydue to a loss of macroeconomic policy autonomyresulting from rapid liberalization and closer in-tegration into the global economy, financialstabilization has often been accompanied by sig-nificant instability in key prices, such as realwages, exchange rates, interest rates and assets,that exert a strong influence on resource alloca-tion and investment decisions. High interest andexchange rates have exacerbated this situation andimpeded investment and technological change(see, TDR 2003, chap. VI). Thus the macroeco-nomic framework has not been conducive to thecreation and expansion of productive capacity andthe improvement of productivity and internationalcompetitiveness, which were the main objectivesthat the microeconomic measures sought toachieve.

In addition, the emphasis on export promo-tion through measures that apply to all economicsectors has tended to boost international com-

petitiveness of domestic enterprises within a pro-duction and export pattern based on existingcomparative advantage related to the abundantendowment in most countries of the region of natu-ral resources and/or low-skilled labour. However,these sectors are generally not capable of gener-ating sizeable growth in productivity and valueadded. This policy stance has been reinforced bythe fact that tight budgetary constraints of manyLatin American countries have prevented themfrom allocating sufficient financial resources tosupport R&D and innovation activities and thefurther development of technological capabilities.Innovation policies have been too broad-based,which, although substantial in overall fiscal terms,47

has meant that they have been insufficient formaking a meaningful contribution to R&D andinnovation activities in individual sectors. More-over, budgetary constraints were exacerbated bythe generous tax incentives to FDI. More recently,the current commodity price boom and the asso-ciated higher budgetary revenues have given manyLatin American governments greater flexibility inthis respect. It is also worth noting that techno-logical and knowledge spillovers from FDI havefallen short of expectations, and have been insuf-ficient to improve domestic technology andproductivity to create a competitive domestic ex-port sector in high-value-added manufactures. Thedisappointing impact of FDI on industrial devel-opment may have been partly due to the overlygenerous incentive packages offered in competingfor that FDI. In some instances, this shortcomingmay also have been related to the fact that rulesand commitments in international trade agree-ments prohibit the use of performance criteria forFDI of the kind applied in the East Asian NIEsduring their economic catch-up.

(b) China

The structural transformation of China’seconomy is of particular interest for several rea-sons. First, it has been accompanied by very rapideconomic growth, which led China to become thefourth largest economy in the world by 2005. Sec-ond, China’s economic development over the past30 years has been based on various developmentstrategies, including central planning, inward-oriented import-substituting industrialization along


with a strong export orientation, and an open-doorpolicy regarding FDI. These strategies were usedat different times, but in some cases also simulta-neously in different parts of the economy. Third,the various types of proactive economic policiesand wide range of instruments that have shapedChina’s economic development have continuouslybeen adapted to changes in the underlying devel-opment strategy, as well as to changing circum-stances in the domestic and international environ-ment. This has been the case, in particular, in therun-up to, and in the aftermath of, China’s acces-sion to the WTO. Thus the process of China’s eco-nomic transformation over the past 30 years maybe characterized as “experimental gradualism”,with the use of heterodox poli-cies in a creative and often in-novative manner.

Distinct from economictransformation in most otherdeveloping countries, China’sprocess of industrialization andstructural change has been partof a general economic transi-tion from a centrally plannedeconomy towards a marketeconomy. Thus, much policysupport has consisted of a gradual and selectiveadoption of regulations that have governed thepace and pattern of the transition towards a mar-ket economy. Key elements in this transition havebeen the reduction of the role of State-owned en-terprises (SOEs), and the gradual and selectiveintroduction of market incentives through theregulatory reform of price systems and of the re-gimes governing domestic labour mobility, exter-nal trade and FDI. On the other hand, governmentpolicy has played an important role in directingboth domestic and foreign investment towardsspecific sectors.

Investment promotion has mainly taken theform of sizeable public investment in physicalinfrastructure, direct government financing, theprovision of credit at preferential interest rates,and tax rebates. In the absence of an efficient do-mestic securities market, bank loans were themajor source of corporate finance. Fixed invest-ment expanded faster between the late 1980s andthe late 1990s, directed at targeted industries andsectors that benefited from preferential credit,

mainly from State-owned banks. Moreover, manu-facturing industries and industries based on non-agricultural raw materials enjoyed tax rates up to80 per cent lower than those imposed on otherindustries (Lu, 2001: 342 and 348).

FDI has played a pivotal role in China’s chang-ing industrial structure. It was attracted throughthe creation in 1980 of four Special EconomicZones, where imported inputs and exports wereexempted from duties and new enterprises wereoffered extended periods of tax exemption. Since1992, inward FDI has increased, as firms in Tai-wan Province of China and Hong Kong (China),under increasing pressure from rising wages in

their labour-intensive indus-tries, were driven to find newlow-wage production locationselsewhere to maintain their in-ternational competitiveness.This contributed to accelerat-ing industrial restructuring inChina from heavy to light in-dustry. Later, in the 1990s, agrowing number of interna-tional production networks, inwhich TNCs organize severalsuppliers in different loca-

tions, supported industrial restructuring from lightindustry to capital- and technology-intensive in-dustries. Chinese locations have come to play acentral part in such networks, in particular in elec-tronics. Since 1999, on average, about 15 per centof total FDI flows to China have been invested inthe electronics industry (MOFCOM, 2006a).

In addition to the provision of fiscal incen-tives, the government has influenced the sectoraldistribution of FDI by screening potential FDIinflows. Guidelines and regulations were issuedexplicitly identifying “prohibited”, “permitted”,or “encouraged” types of FDI. The latter offeredincentives to FDI in high-tech industries throughextensive preferential treatment such as tax rebatesand/or exemptions, duty free imports of capitalequipment and better access to public infrastruc-ture and utilities such as gas and electricity(MOFCOM, 2006a). In order to attract FDI in-corporating more sophisticated technologies forexport-oriented industrial production, the Govern-ment streamlined administrative procedures andoffered incentive packages in “free trade and high-

In China, much policysupport has consisted of agradual and selectiveadoption of regulationsregarding the pace andpattern of the transitiontowards a market economy.


technology development zones”, including theprovision of heavily subsidized land and energy.

China’s foreign trade policy has been an in-tegral part of its strategy for industrial develop-ment. It was very industry-selective, with exten-sive import restrictions before China’s accessionto WTO. Until then, foreign trade had supportedindustrial restructuring in two ways. First, gradualand phased trade liberalization enabled importsof technology which China would not otherwisehave access to, and which were essential for struc-tural changes in Chinese industry (see TDR 2002).Second, foreign trade allowed the export of sur-plus production, without which these structuralchanges, based on a combination of a large sur-plus of labour and rapidly increasing investment– both domestic and foreign –would not have been sustain-able given the narrowness ofChina’s domestic market, inparticular before 1990.

Exporting firms benefitedfrom various pricing, tax andloan privileges, as well as sup-port for technological upgrad-ing, to maintain and increasetheir exports. With the deepening of China’s eco-nomic reforms and, in particular, the decentrali-zation of foreign trade, which led to a massiveentrance of private enterprises, many of the in-centives have been phased out, non-tariff barriersgradually dismantled and tariff barriers loweredsignificantly. Indirect instruments such as tax re-bates have become increasingly important to boosttrade in the Government’s “encouraged” industrialsectors.

Over time, technological upgrading has be-come one of the greatest challenges in Chineseeconomic development. China’s exports continueto have a relatively high import content, particu-larly of technology-intensive parts and compo-nents, as indicated by the fact that in 2005, 55 percent of the exports fell into the category of process-ing trade, and in the same year 88 per cent of Chi-nese high-tech exports came from foreign-fundedenterprises (FFEs) and 66 per cent from whollyforeign-owned enterprises (MOFCOM, 2006b). Asrecognized in China’s five-year plan for 2006-2010, making scientific and technological ad-

vancement and domestic innovation a drivingforce for structural change and social developmentis of particular policy relevance at China’s cur-rent stage of industrialization. So far, only a smallnumber of firms have been reinvesting a signifi-cant share of their profits in R&D. The relianceon FDI for technological upgrading appears tohave weakened domestic innovation and applica-tion of knowledge: since the 1990s, FFEs in Chinafiled by far the largest proportion of patent appli-cations, with local firms filing less than 20 percent (Cao, 2004: 8). This may further inhibit tech-nological progress in Chinese industry. Anotherproblem lies in the application of patents: only10 per cent of the domestically owned patentswere applied in production annually between 1985and 2003, in stark contrast to an annual average

of 60 to 80 per cent in devel-oped countries.48 The lack ofinitiative on the part of domes-tic enterprises to innovate andupgrade technology may alsobe due to the wide technologygap between Chinese firmsand competitors in advancedcountries, which makes catch-ing up very costly and highlyrisky for individual enter-

prises. Moreover, within international productionchains Chinese firms have tended to be locked intolabour-intensive activities, a tendency that hasbeen reinforced by China’s tax rebate system. Thesystem was successful inasmuch as it contributedsignificantly to China’s trade expansion, but it hasdone so by encouraging processing and assemblingactivities.

As a result, advanced technology in China’sexport industry is still highly concentrated in af-filiates of TNCs. This is evidenced by the highand growing share of FFEs in China’s high-techexports compared to their share in the country’stotal exports. In 2005, the share of FFEs in high-tech exports was 88 per cent, compared to 45 percent in 1995, while the share of FFEs in China’stotal exports rose from 31 per cent to 58 per centin the same period (MOFCOM, 2006b). At the be-ginning of China’s opening up, policies towardsFDI included measures aimed at coercing tech-nology transfer and enhancing backward linkages.The FDI approval process frequently included ex-plicit provisions for technology transfer in the

Technological upgradinghas become one of thegreatest challenges inChinese economicdevelopment.


form of local content requirements and produc-tion export quotas. In addition, joint ventureprojects served to obtain technology through col-laboration in production, research or training. WithChina’s accession to the WTO, in particular itscommitment to abide by the TRIMs Agreement,these practices had to be abandoned. Besides, asChina’s FDI regime has been liberalized and ad-ministrative power increasingly decentralized,there has been growing competition among localgovernments to attract FDI. As a result, incentivesto foreign investors have tended to become moregenerous, so that they frequently benefit frombetter than national treatment with little pressurefor technology transfer, and there is risk of a raceto the bottom in bidding for FDI. Although TNCsin China appear to have considerably increasedtheir investment in R&D, this has been drivenmainly by the abundant human resources avail-able there, and their aim to create R&D centresclose to the potentially rapidly growing Chinesemarket in order to adapt advanced technology tospecific demands of local consumers (UNCTAD,2005: 110–111). Overall, however, the level ofdiffusion of competitive technology of TNCs inChina is still low.

With the end of the WTO transition period,most of the elements of China’s earlier industrialpolicy have been phased out, in particular infantindustry trade protection measures, preferentialinterest and tax rates, as well as some forms ofdirect financial assistance to industries. This hasbrought new challenges for the design and imple-mentation of industrial policy. For instance, whenthe clauses covering technology requirement andexport content in the Law of Wholly ForeignOwned Enterprises were repealed upon China’saccession to the WTO, many FFEs separated fromtheir local joint-venture partners to becomewholly-owned foreign enterprises, making tech-nological and other spillovers from FDI moredifficult to obtain. Moreover, the possibilities forthe Government to support domestic industrieshave diminished considerably. Recently, Canada,the EU and the United States requested consul-tations with China concerning its regulatingof imports of automotive parts and componentsin order to support the development of the Chi-nese automobile industry, which they consideredto be inconsistent with some WTO/GATT agree-ments.

Although direct intervention favouring do-mestic industries has declined with the growingimportance of the private sector and China’s ac-cession to the WTO, the Government is stillretaining a guiding role via indirect instrumentssuch as taxation, the provision of guidelines forscience and technology development49 and certainforms of public financial support for related ex-penditures. At present, a reform of the system ofvalue-added tax rebates is under discussion.50 Inthe past, these rebates, which favoured raw mate-rials and parts and components used as inputs formanufactured exports, had a significant impact onthe structure of China’s trade and industrial de-velopment in favour of processing trade. Now,increasingly it is believed that relying too muchon the comparative advantage of labour-intensiveproducts may discourage the upgrading of Chi-na’s position to higher value-added production inthe international supply chain. Another ongoingdebate is about the possible elimination of the dualcorporate tax system applied to foreign investorsand domestic companies.51 So far, foreign inves-tors benefit from lower tax rates of between 15 and24 per cent compared to domestic companies thathave to pay 33 per cent. In addition, foreign in-vestors are entitled to tax holidays if they investin “encouraged sectors” or poor regions in China.Changes in these government policies are boundto have an impact on China’s future industrialstructure.

(c) Recent industrial policy in France

France has often been characterized as theEuropean representative of State-led develop-mentalism. Particularly prior to economic liber-alization in the 1980s, France pursued a develop-mentalist industrial policy supported by subsidies,credit controls, indicative planning, and directintervention in State-owned enterprises. While lib-eralization eliminated government control over theallocation of credit by banks and other financialinstitutions, which previously had been the prin-cipal tool of industrial policy, subsidies survived,although they were used to a lesser extent. Theimportance of indicative planning was also dras-tically reduced, and, in any case, had already be-come less development orientated, given that inthe aftermath of the economic slowdown of the


1970s industrial policy was directed more at avoid-ing bankruptcies and unemployment rather thanat espousing an enlightened vision of economicdevelopment.

French industrial policy has traditionally fo-cused on the development of sectors designatedas being of national interest, such as steel and com-puter technology in the 1960s,nuclear power and telecom-munications in the 1970s,electronics in the 1980s, andhigh-technology sectors morerecently. The institutional struc-ture in which industrial policyis embedded and the instru-ments used have been evolv-ing over time in response to changes in the worldeconomy, in particular the change in the interna-tional monetary system after 1973.52 There wasalso concern that the strategy of nurturing nationalchampions might cause industrial policy to be“captured” by the economic sectors it was de-signed to serve and develop, with the result thatthe instruments used would serve to create sectoralrents rather than accelerating national economicmodernization.

France’s recent industrial policy reforms marka further evolution in the choice of institutionalframework and policy instruments. Reflecting thegeneral trend towards decentralization in thatcountry, the new approach gives substantiallymore weight to local and regional governmententities, even though the central Government re-tains a strong role. The new initiative also marksa shift away from a State-led approach based onthe nurturing of a few large national enterprisesto a private-sector-led approach. It relies on part-nerships between firms (both large enterprises andSMEs, most of which collaborate with foreign,particularly German, enterprises), educationalcentres and research institutions in which the Statemainly plays a merely facilitating role.

France’s new industrial policy started, in2002, a process aimed at defining a strategy forthe promotion of clusters of competitiveness(“pôles de compétitivité”).53 This resulted in theidentification of 74 such clusters in 200554 and thelaunch of six clusters in March 2006. This newindustrial policy is based on the observation that

having successfully narrowed the technology gapwith the world leader, the United States, duringthe 1950s, 1960s and 1970s, in the past few yearsFrance’s industrial development has not kept pacewith technological progress and economic growthin some other developed countries, particularly theUnited States. This is reflected in (i) a decline inthe contribution of manufacturing to total value

added relative to that in otherdeveloped countries (in addi-tion to the common trend indeveloped countries towards agrowing services sector); (ii) aspecialization pattern biased infavour of traditional, relativelylow-technology-intensive sec-tors (such as agro-industry and

transport equipment) at the expense of dynamichigh-tech sectors; and (iii) growing internationalcompetition from developing countries and Cen-tral European economies in France’s traditionalindustrial sectors, which has been accompaniedby the relocation of some activities in these sec-tors from France to relatively low-wage countries,as well as by an erosion of the international com-petitiveness of France’s existing high-tech sectors(Beffa, 2005).

Some observers have identified insufficientresearch and innovation activities as the root causeof the concentration of France’s industrial speciali-zation in relatively low-technology-intensive sectorsand of its difficulties in developing high-tech sec-tors (Beffa, 2005; Jacquet and Darmon, 2005).Thus the main objective of France’s new indus-trial policy is to promote research and innovationand improve industrial efficiency. It is expectedthat this will help increase the country’s growthpotential and social cohesion, change its patternof industrial specialization by according greaterimportance to high-tech sectors, and enable theachievement of the highest level of technologicalcompetencies (Jacquet and Darmon, 2005: 72).

The clusters of competitiveness bring to-gether enterprises, educational centres and privateand public research institutions to work in part-nership on common projects with a view toattaining a critical mass of economic activitywithin a geographical area. They aim to achievetechnological innovation that will improve thecompetitiveness of French enterprises on interna-

France’s new industrialpolicy promotes clusters ofcompetitiveness ...


tional markets of substantial size or growthpotential. There are two types of clusters: (i) a rela-tively small number of clusters, the researchactivities of which are organized around a spe-cific area of technology with innovative activitiesaimed at applications at the technological fron-tiers (e.g. in biotechnology, nanotechnology andspace industries); and (ii) a larger number of clus-ters that conduct more applied research closelytargeting existing industrial sectors and marketswith a significant growth potential at the globallevel (CIADT, 2004; Jacquet and Darmon, 2005:63–74). Both types of clusters emphasize activepartnerships for innovation.

The clusters were selected on the basis of acompetitive process with the objective of identi-fying projects capable of making a significantcontribution to the development of enterprises forwhich innovation is central to their competitive-ness. Under the general oversight of the Inter-min-isterial Committee for Regional Planning and De-velopment (CIADT), which was chaired by thePrime Minister and which also had the final deci-sion-making authority, the selection process wasconducted by independent experts from the busi-ness, research and educational sectors, as well asby government experts at bothlocal and ministerial levels. Theassessment criteria included thecluster’s potential for creatingvalue added through innova-tion, for playing a leading roleon international markets, forits reliance on a partnershipwith different actors, and thecoherence of its economic de-velopment strategy with that of the geographicalregion in which it is located (CIADT, 2004).

The State’s financial contribution to the clus-ters amounts to 1.7 billion euros for the period2005-2007. The direct financial benefits take theform of subsidies, tax exemptions and reducedsocial contributions, as well as specific financialsupport and guarantees. These are supplementedby priority treatment in terms of the provision ofIT-equipment and speedy administrative proce-dures, staffing of public research institutions,appraisal and exchange of technological knowl-edge, and a range of other measures (Jacquet andDarmon, 2005: 70–71 and 83; OECD, 2006: 77).

The funds are granted on the condition that thesupported activities are not relocated (Jacquet andDarmon, 2005: 70).

While increasing the local ownership ofprojects, the decentralization, combined with thehigh fragmentation of local and regional Stateentities into multiple levels, has increased the ad-ministrative complexities of decision-making andposes an additional challenge to maintaining co-herence in government actions. In addition, theselection process includes a sizeable element ofsubjectivity, as no strictly defined quantitativecriteria are used. Moreover, it is not clear how theclusters’ performance will be evaluated. Perhaps,most importantly, the call for project submissionvastly exceeded expectations, resulting in the crea-tion of more than four times the initially expectednumber of recognized clusters, while the fundsallocated for their support were only doubled. Thisraises the question as to whether the allocatedfunds are sufficient to obtain the objective ofreaching a critical mass (OECD, 2006: 16).

The new industrial policy in France has beendesigned under the general umbrella of the so-called “Lisbon Strategy”. This Strategy, adopted

at the European Council Sum-mit in 200055 and updated in2005, aims at increasing R&Dintensity in the EU. The actionplan of 2005, which reformu-lated the Strategy’s priorities,provides an integrated ap-proach to improving the con-ditions for business invest-ment in R&D and innovation

in order to meet the goal of increasing overall in-vestment in research in the EU from 1.9 per centof GDP to 3 per cent by 2010 (Commission of theEuropean Communities, 2005).56

Within this EU-wide framework, the Frenchapproach is distinguished by its greater emphasison a required increase in the contribution of pub-lic funds to industrial research. Beffa (2005), forexample, notes that in the United States this shareis between 12 per cent and 21 per cent, dependingon the industrial sector, while in France, it is only12 per cent on average for all industrial sectors.Moreover, France allocates much fewer of thesefunds than the United States to non-defence-

... with the main objectiveto step up research andinnovation and improveindustrial efficiency.


related research in industrial frontier technologies.The funding provided to the competitiveness clus-ters attempts to narrow both these gaps.

This policy of broad-based sponsorship ofpartnerships between government, business, andeducational and research institutions to advanceindustrial R&D and innovation has been a gen-eral tendency in developedcountries over the past fewyears. It reflects a move awayfrom public support to the de-velopment of productive ca-pacity towards fostering inno-vation for the development ofknowledge-based industries.Within this new strategy, sup-port measures appear to bemainly of a general nature, butin practice they imply according priority to par-ticular industrial sectors that have been identified,in one way or another, as offering considerablepotential for innovation. As noted by Weiss (2005:732), developed-country governments have imple-mented an extensive range of programmes to pro-mote high-tech firms. These include support forpre-competitive R&D, facilitating access to ven-ture capital, and the expansion and upgrading ofa sophisticated infrastructure for the promotionand protection of intellectual property, informa-tion and telecommunications, and the appraisaland exchange of technological knowledge viapubic-private collaborative projects.

One reason for this shift towards the promo-tion of R&D and innovation activities has beenthe perception that outsourcing activities or therelocation of entire production units to cheaperlocations no longer involves only labour-intensiveassembly stages; it is also increasingly affectingmore skilled activities. This has caused concernbecause it is generally believed that, contrary

to the outsourcing of labour-intensive activities to lower-wage regions, a process thatactually may improve an out-sourcing firm’s internationalcompetitiveness, outsourcingof high-tech activities deprivesan economy of part of its dy-namic development potential.Anxiety over the outsourcingof IT-based services to India

has perhaps been the most vivid expression of thisconcern.

Multilateral trading rules provide sufficientlatitude for developed countries to implement thisstrategy. As discussed in the previous section, theprovision in Article 8 of the SCM Agreement al-lows subsidies for R&D and regional and envi-ronmental development activities, although theyare now actionable. The fact that the developedcountries have the budgetary capacity to providemassive public support to such activities highlightsthe asymmetry involved in the use of Article-8-typesubsidies.

Developed countries havethe budgetary capacity toprovide massive publicsupport.


Experience with reforms over the past 15 years,as well as recent developments in economic theoryconcerning the creation of new areas of compara-tive advantage, provide a strong rationale for theadoption of proactive trade and industrial policies.

However, specific policy measures that suc-cessful countries have adopted cannot easily beemulated by other countries. Nevertheless, thereare some common general principles underlyingtheir success, and governments, through creativepolicy-making, could choose specific types of pub-lic support policies adapted totheir country’s particular localconditions, including its stageof economic and institutionaldevelopment.

An assessment of the ex-tent to which various interna-tional trade arrangements haverestricted the degrees of free-dom of developing countriesto pursue proactive trade and industrial policiesgives a mixed picture. On the one hand, WTO rulesand commitments have made it far more difficultfor developing countries to combine outward ori-entation with the unorthodox policy instrumentsthat the mature and late industrializers employedto promote economic diversification and techno-logical upgrading. The rules and commitmentslimit policy space in three areas. First, they se-verely restrict the use of subsidies to develop lo-cal production of new products or new modes ofproduction; probably the greatest obstacle to sen-sible industrial policies in this context is the pro-

hibition under the SCM Agreement to provide sub-sidies contingent on export performance.57 Sec-ond, they prohibit the imposition on foreign in-vestors of performance requirements that favourtechnology transfer and the use of domesticallyproduced components. And third, they make itdifficult or costly for domestic producers to un-dertake reverse engineering and imitation throughaccess to technology that is covered by patent orcopyright protection. Given these constraints, theURAs, by implication, lead to an increase in therelative importance of temporary protection in the

form of industrial tariffs. De-veloping countries thus mayfind that tariffs are one of thefew policy options left, and inthis respect it may be of inter-est to modulate applied tariffson particular industrial sectorsaround a stable average levelof industrial tariffs, in line witha country’s pace and pattern oftechnological upgrading. How-

ever, even this option is likely to be limited by thecurrent WTO negotiations, as well as by RTAs.

On the other hand, under the current set-upof multilateral trade rules, countries still have thepossibility to pursue policies that will be able tohelp generate new productive capacity and newareas of comparative advantage. Such types ofpolicies largely concern the provision of publicfunds in support of R&D and innovation activi-ties. Countries in a position to use the WTO rulesand commitments to this effect can continue tosupport their own industries, target national cham-

E. Conclusions: options for policy innovation

International tradearrangements have limitedpolicy space in severalareas.


pions, and generally promote national efforts to-wards technological advancement.

The case studies in section D of this chapter,which attempt to shed some light on the kinds ofpolicies that have been used to support industrialdevelopment and technological upgrading over thepast few years, indicate that:

• Coherence between macro- and microeco-nomic policies is crucial. A macroeconomicpolicy stance that leads to high domestic in-terest rates and an overvalued exchange rateis not conducive to investment that can bringabout productivity growth and improve theinternational competitiveness of domestic en-terprises, even when microeconomic andstructural policies provide incentives for suchinvestment.

• A coherent policy strategy that supportsindustrial development and technological up-grading also requires a pragmatic and strategicapproach aimed at making FDI fit into thedevelopment agenda in a way that would helpbring about not only faster and more sustainedgrowth, but also structural and technologicalchange. However, in the current internationaleconomic environment, where many coun-tries compete with each other for FDI, onlycountries with a skilled labour force and alarge enough domestic market and purchasingpower potential are likely to have sufficientleverage over TNCs to secure technologytransfer and productivity spillovers. On theother hand, even those countries will find itdifficult to exert such leverage if there is in-tensive domestic competition for increasedFDI through generous incentive packages atthe provincial or municipal levels – a lessonlearned from the experiences of some LatinAmerican countries and China, as notedabove.

• The shift in emphasis from trade and indus-trial policies based on protectionist measuresfor a limited number of industrial activitiestowards the provision of public funds to sup-port all economic sectors has a significantfiscal impact. Subsidies imply a cost to pub-lic budgets, in the form of a loss of fiscalrevenues (e.g. through tax exemptions) or an

increase in fiscal expenditure (e.g. throughthe provision of subsidies), while protection-ist measures in the form of tariffs providefiscal revenues. Thus, as emphasized by Wade(2006), by pursuing public support policiessolely on the basis of the provision of publicfunds, developing countries risk encounteringserious budgetary and financial constraints,which will allow substantially lower levelsof support compared to those provided in thepast and those that developed countries con-tinue to be able to provide.

• It also needs to be borne in mind that WTOrules and commitments carry the threat ofsanctions, but the eventual imposition bytrading partners of retaliatory tariffs or othermeasures depends on the actual damage.58

Consequently, as long as the damage causedby a trading partner’s infringement of rulesis small, a WTO member State is unlikely toinvoke the dispute settlement mechanism andinitiate the imposition of sanctions. It wouldappear that this confers additional degrees offreedom on countries whose importance inworld trade is relatively small.59

Thus, developing-country governments maywish to take advantage of the degrees of freedomin national policy-making that have remained un-touched by the URAs. Indeed, the observation thatmultilateral rules still allow countries a certain de-gree of freedom to adopt open-economy industrialpolicies and that infringements are liable to chal-lenge only when the dispute settlement mechanismis invoked, has led to the hypothesis that “[w]hatconstrains sensible industrial policy today islargely the willingness to adopt it, not the abilityto do so” (Rodrik, 2004: 32).

However, the asymmetries in the URAsshould not be underestimated. They result fromthe fact that while the negotiated agreementsextend to all WTO members in the same way interms of legal obligations, they are much moreburdensome for developing countries in economicterms. This implies that it is crucially importantto look at the “level playing field” metaphor notin terms of legal constraints, but in terms of eco-nomic constraints, considering countries’ differentstructural features and levels of industrial devel-opment.


Moreover, what is left of the degree of free-dom for developing-country policymakers afterthe URAs has been reduced through regional andbilateral free trade agreements with developedcountries. These agreements typically extend therange of disciplines beyondthose found in the URAs re-garding investment regulationand intellectual property rightsprotection.

Current trade negotiationsthreaten to further curtail thedegree of freedom for devel-oping country-policymakers. At the multilaterallevel, the threats are probably greatest in the in-tellectual property rights negotiations being con-ducted under the aegis of WIPO – where devel-oped countries are pushing for further harmoni-zation of national patent regulations – and in thecurrent multilateral trade negotiations on indus-trial tariffs.60 Concerning the negotiations on in-dustrial tariffs, employing a harmonizing formula(across products and/or countries), cutting tariffsline-by-line rather than just on average, would re-duce the degree of freedom for developing coun-tries to protect some industrial sectors while lib-eralizing others. Most importantly, it will reducethe flexibility to increase tariffs that had previ-ously been cut, as shown by Laird, Vanzetti andFernández de Córdoba (2006).61 Maintaining ex-isting degrees of freedom for national policy-making would imply extending the notion of flex-ibility to the right to exempt some sectors fromtariff-binding and tariff-cutting commitments, andto the right to maintain the av-erage level of tariffs at negoti-ated levels while being able toflexibly raise and lower tariffsin specific sectors, as deemednecessary for industrial up-grading and development.

It needs to be recognizedthat developing-country policy-makers willingly signed on tomany of the commitments ininternational trade agreements, which implied areduction in their de jure policy autonomy. Thiswas motivated by the expectation that the result-ing benefits would far outweigh the costs of suchcommitments. Another possible reason for doing

so may have been the fear of adverse reactions bymultilateral lending agencies, international finan-cial markets and foreign investors. As Finger andNogues (2002) note, at the end of the UR, devel-oping countries were faced with the choice of ac-

cepting what was proposed orrisk being marginalized in theinternational trade regime.62

Regarding bilateral trade agree-ments, it appears that the em-phasis on export promotion asa development strategy ledmany developing countries tobelieve that securing and in-

creasing access to developed-country markets bysigning free trade agreements is almost an end initself.

More recently, however, developing countrieshave been making concerted efforts to prevent afurther reduction of their policy autonomy and torecover some of their lost autonomy (Gallagher,2005: 12). This implies that, in light of their ex-perience with adherence to the existing multilat-eral rules and disciplines, many governments to-day believe that too much policy autonomy wasconceded during the UR, without gaining muchin return. According to this view, some of the con-cessions developing countries made in the URAs,such as in TRIPS and TRIMs, were on the under-standing that these were in exchange for devel-oped countries’ providing improved market ac-cess. However, as discussed in chapter III above,developed countries have largely failed to followthrough on their side of the deal.

The Doha Work Pro-gramme has yet to deliver onthe development promise ofthe Doha Declaration. Theeventual outcome may wellfurther reduce flexibility inpolicy-making by developingcountries, particularly in thearea of industrial tariffs. Onthe other hand, a failure of theongoing multilateral negotia-

tions could result in greater importance beinggiven to regional or bilateral free trade arrange-ments as the legal mechanisms that define rulesand disciplines in international trade. While thesearrangements may improve developing countries’

Developing countries mayfind that tariffs are one ofthe few policy options left.

Some degrees of freedomin national policy-makingremained untouched by theURAs, but the asymmetriesin the URAs should not beunderestimated.


access to developed-country markets, they mayentail a reduction in the degree of freedom in na-tional policy-making that could be greater thanthat emerging from a Doha Round Agreement.

This could make it even more difficult to developthe supply capacity needed to take advantage ofimproved export opportunities.

Notes

1 As discussed in TDR 2005, India’s reliance on theservices sector is the major exception to this pat-tern but to what extent India can sustain rapid eco-nomic growth without rapidly expanding its manu-facturing sector is an open question.

2 There is no generally accepted definition of indus-trial policy. Chang (1996: 60) defines it “as a policyaimed at particular industries (and firms as theircomponents) to achieve the outcomes that are per-ceived by the state to be efficient for the economyas a whole” (emphasis in original). In a similar vein,Pack and Saggi (2006: 2) argue that “industrialpolicy is basically any type of selective interven-tion or government policy that attempts to alter thesectoral structure of production toward sectors thatare expected to offer better prospects for economicgrowth than would occur in the absence of such in-tervention, i.e., in the market equilibrium”. In thissense, the desired outcome of industrial policy canbe considered to be the creation of new productioncapacity and new comparative advantage, enablingan economy to progress upwards in the internationaldivision of labour. By contrast, Rodrik (2004: 3)stresses that “the analysis of industrial policy needsto focus not on the policy outcomes – which areinherently unknown ex ante – but on getting thepolicy process right” (emphasis in original).

3 It has also been questioned whether developing-country governments have the administrative andinstitutional capability to design and implementactive trade and industrial policies. This issue is ad-dressed in chapter VI.

4 A further argument, associated with the literatureon “strategic trade theory” initiated by Brander andSpencer (1985), relates to international rent shift-ing on the basis of strategic interdependence amonga small number of firms. However, the policy out-

come of this argument is very sensitive to even smallchanges in the underlying model assumptions. More-over, the model is often based on the interdependenceof a duopolist structure (i.e. a market structure thatis most relevant for high-tech enterprises in devel-oped countries). Thus, apart from isolated cases, suchas that relating to the aircraft firms Embraer andBombardier, there appears to be little in strategictrade theory of relevance for developing countries.

5 These arguments stem largely from the concept ofcircular and cumulative causation that posits a cir-cular relationship between growth in productivityand growth in total output. The main proponents ofthis concept – including Young, Rosenstein-Rodan,Hirschman, Myrdal, Kaldor and, more recently,Cornwall, McCombie and Thirlwall – employ simi-lar or related analytical tools as those of recent em-pirical studies of late industrialization (e.g. Amsden,1989, 2001; Wade, 1990, 2003a; TDR 1996, 2003).They argue that there are pervasive and significantincreasing returns and externalities; complemen-tarities in investment, production and consumption;endogenous technical change and factor creation;imperfect information, and a dependence of the capi-tal-labour ratio on the size of the market, which,taken together, contradict the conditions for gen-eral equilibrium (see Toner, 1999, and Ros, 2000,for detailed accounts).

6 This also largely depends on the extent to whichproductivity growth translates into an increase inaggregate demand (i.e. issues related to income dis-tribution).

7 With international trade in intermediate goods, do-mestic producers may import their production inputs.However, such imports are likely to pose problemsof technology adaptation similar to those related tothe purchase of foreign machinery and equipment.


8 To be sure, innovation in developing countries doesnot generally mean pushing out the technology fron-tier. Rather, it means developing products or proc-esses at home that are new for the specific economybut that may already be well established in worldmarkets.

9 Technical evaluations may provide some indicationof the cost structure, but undertaking such an evalu-ation itself represents an initial investment.

10 Noland and Pack (2005: 4) argue that the existingtechnological barriers will cause the entry of imita-tors to be slow, so that the innovative enterprise willhave a considerable time period during which it willnot face much competition from local firms. How-ever, if the speed of imitative entry depends on thedegree of technological innovation, it is likely thatthe size of the start-up investment will also dependon the size of existing technological barriers, therebyextending the time period required to recover thestart-up investment.

11 In the discussion on the volatility of different typesof capital inflows (TDR 2003), it has been notedthat in the evolving international environment,where finance has assumed a more prominent rolein shaping international economic relations, liber-alization of financial flows and related innovationsin financial market instruments allow for hedgingof FDI flows. This tends to blur the distinction be-tween FDI and other types of capital flows by mak-ing FDI much more footloose and less stable thanthe kind often proposed in the argumentation in sup-port of FDI as the key driver of industrialization indeveloping countries.

12 In new economic geography models, the structureof production in individual countries is determined,as in traditional trade theory, by the interaction be-tween country characteristics and industry character-istics. But while traditional trade theory focuses onrelative factor endowments of countries and factorintensities of goods, the mechanisms of new economicgeography models also take into account market sizeand countries’ geographical distance from the mar-kets of the main developed countries, as well as thetransport intensity of the industrial sectors, includ-ing the level of transport costs and the dependenceon intermediate inputs. Mayer (2004) examines therelevance of these models for developing countries.

13 In the past, many countries maintained restrictionson imports of luxury consumer goods, but this wasmotivated not so much by industrial policy as byforeign-exchange management considerations.

14 Moreover, as a fundamental rule, it is clear that, tobe successful, any kind of trade and industrial policyrequires a stable macroeconomic environment con-ducive to investment.

15 Policy support for product or process innovation willbe more successful if it can be directed at those ac-

tivities with the highest potential to crowd in com-plementary investment and create technologicalspillovers. But the creation of linkages and inter-firm spillovers very much depends on, among otherthings, the prevailing industry structure (i.e. whetherall activities in an industrial sector are combined inlarge firms, such as in the chaebols of the Republicof Korea, or whether there is a dense network ofsmaller firms with forward and backward linkages).It is probably easier for large enterprises to exploitscale economies before potential imitative competi-tors enter the market, as well as to benefit fromspillovers. On the other hand, this reduces the casefor supportive policies designed to reduce the costof innovative investment. An alternative may be tocombine more horizontal support, targeted at newactivities and processes more generally, with moreselective measures aimed at fostering diversifica-tion and structural change.

16 Lall (2004), for example, builds a classification ofdifferent types of industrial policy around the atti-tude towards FDI based on a “competitiveness strat-egy”, which seeks to identify the kind of public sup-port required to attract FDI while laying the groundfor knowledge spillovers.

17 Experience suggests that performance criteriashould be related to productivity growth and struc-tural change, rather than to a multiplicity of objec-tives such as rent transfer to particular groups on anethnic, family, gender or interest group basis.

18 In addition to using formal policy tools, govern-ments can also seek to exercise influence throughinformal administrative guidance, coercing recalci-trant firms if necessary. Wade (2003a: xxi–xxii), forexample, describes how “nudging” foreign firms toswitch supplies from imports to domestic produc-ers, or nudging established industries quickly toprovide markets for firms in innovative industries,was used in Taiwan Province of China. This kind ofpersuasion involved a mix of methods, such as prom-ises of goodwill for future ventures, or delaying thegranting of permission to import (that had earlierbeen approved quickly and automatically).

19 Examples include the incompatibility of standardsbetween IBM personal computers and Apple Mac-intosh, computer chips made by Intel and otherfirms, or competing standards for third-generationtelephone handsets, optical disk storage or high-defi-nition televisions. By contrast, open source softwareis an example of global compatibility because itmakes the source code of an application availablevia the Internet.

20 Departing from the MFN rule, there are provisionsthat allow free trade agreements and customs un-ions among WTO members under certain conditions.Moreover, the so-called “escape clause” allows aWTO member to suspend its obligations as a tem-


porary emergency measure, accompanied by theadoption of adjustment policies.

21 The agreements also included liberalization com-mitments relating to non-tariff barriers, as well ascommitments in the areas of agriculture and serv-ices, but these are not considered here.

22 Investment-related disciplines of the multilateraltrading regime are also contained in the GeneralAgreement on Trade in Services (GATS) as part ofmode 3 (i.e. supply through commercial presence).For a detailed discussion of this area, see Wade,2003b; and Cho and Dubash, 2005.

23 These transition periods were five years for devel-oping countries (i.e. until the end of 2000), and sevenyears for the least developed countries (LDCs), withsome further extensions granted to countries expe-riencing implementation difficulties for develop-ment, finance or trade reasons.

24 This concerns disputes Nos. 51, 52, 65 and 81 againstBrazil; Nos. 146 and 175 against India; Nos. 54, 55,59 and 64 against Indonesia; No. 195 against thePhilippines, and Nos. 339, 340 and 342 against China.For details, see www.wto.org/english/tratop_e/dispu_e/dispu_status_e.htm.

25 Since TRIMs applies only to trade in goods, gov-ernments can impose local procurement require-ments with respect to services such as banking, in-surance and transport, as long as such measures re-main possible under GATS disciplines.

26 This is in contrast to the Subsidies Code of the To-kyo Round, which was voluntary and extended onlyto national governments.

27 Subsidies for research had to be for activities con-ducted by firms or by higher education or researchestablishments on a contract basis with firms, onthe condition that the assistance covered not morethan 75 per cent of the cost of industrial research or50 per cent of the cost of pre-competitive develop-ment activity. Eligible regions were defined as thosewhose per capita income did not exceed 85 per centof the country’s average or those whose unemploy-ment quota had been at least 110 per cent of thecountry’s average over a three-year period. Regard-ing environmental objectives, subsidies were per-mitted for the “promotion of adapting existing fa-cilities to new environmental regulations”.

28 The Agreement refers to seven areas of intellectualproperty: copyrights, trademarks, geographical in-dications, utility models, patents, integrated circuitsand undisclosed information.

29 The LDCs have been granted a general transitionperiod until 1 July 2013, and an additional exten-sion until 1 January 2016 with regard to patents andundisclosed information relating to pharmaceuticalproducts.

30 While the focus of this discussion is on the impactof the TRIPS Agreement on industrial development,

it should be noted that access to medicines in de-veloping countries has gained considerable atten-tion. This has mainly a humanitarian dimension, butit can also be of considerable importance for phar-maceutical industries in developing countries. TheDoha Declaration clarified the need to interpretTRIPS from a public health perspective, and thusimproved the Agreement’s developmental aspectsin terms of access to medicines. It explicitly recog-nizes the flexibility within TRIPS to grant compul-sory licences and the right of countries to determinethe grounds on which these are granted. The WTOdecision of 30 August 2003 waives the limitationon exports of generic products if they go to coun-tries having insufficient manufacturing capacity.

31 For specific examples, see Morin, 2003; and Shadlen,2005b.

32 The advent of the avian influenza made compulsorylicensing a global issue because of the widespreadperception that patent protection of the apparentlyonly efficient drug in this area is a barrier to prepara-tions for combating a potential pandemic. For a de-tailed discussion of the impact of developing coun-tries’ bilateral free trade agreements with the EUand the United States on intellectual property rightsthat affect access to medicines, see Correa, 2006.

33 Kowalski (2005: 11) points out that in upper-mid-dle-income countries, the share of import duties intotal government revenue fell from about 20 per centin the late 1980s to about 7 per cent in the early2000s; the respective shares for lower-middle-in-come countries were 25 per cent and 16 per cent,and for low-income countries 27 per cent and 19 percent respectively.

34 Full binding coverage and uniform tariffs also con-tribute to greater predictability of trade policy andmarket access, and thus foster the stability of theinternational trading system (Francois and Martin,2002). But this is true only if the resulting loss offlexibility in tariff policy is not replaced by a greaterapplication of non-tariff measures, which are gen-erally less transparent than tariffs. Moreover, it isexchange-rate instability that often constitutes themost serious threat to the predictability and stabil-ity of international trade flows and the internationaltrading system (TDR 2004).

35 Moreover, developing countries would need to beable to raise tariffs in particular industrial sectorswithout much cost. GATT Article XVIII: A andXVIII: C allows countries to remove tariff conces-sions or use quotas for infant industry protection,but in order to do so they have to “negotiate” and“compensate”. While these obligations maintaintransparency and help to avoid abuse, this proce-dure can be cumbersome and involve costly com-pensation. The so-called “escape clause” under Ar-ticle XIX allows a WTO member State to suspend


its obligations under certain conditions in order tosafeguard its industry. However, these safeguardscan be invoked only as temporary emergency meas-ures and must be accompanied by adjustment; thusthey do not provide an instrument for promotingcompetitive industrial production.

36 This discrepancy between bound and applied tar-iffs is partly due to unilateral trade liberalizationthat many developing countries have undertakeneither voluntarily or as a result of conditionalitiesimposed by multilateral lending institutions.

37 This statement relates to a comparison of India withthe other countries in the table, but, as noted above,not with respect to the now developed countrieswhen they were at India’s current level of per capitaincome.

38 However, industrial upgrading in the Republic ofKorea has largely relied on national enterprises,while it appears that much of the high-value-addedindustrial activity in China occurs in wholly foreign-owned enterprises, as discussed below.

39 This emphasis on international competitiveness iswell illustrated by the fact that one of the main policyactions included in Brazil’s Multi-Annual Plan forthe period 1996–1999 was a reduction of the so-called “Brazil-cost”, that is, “the extra labor and fis-cal costs producers (both foreign and domestic) haveto bear when producing in Brazil as opposed to pro-ducing in foreign countries” (Melo, 2001: 10).

40 Peres (2006) also notes that in bilateral or multilat-eral free trade negotiations, representatives fromLatin American countries sometimes attempt toimprove export opportunities for new industrial ac-tivities in order to promote industrial development.

41 This has been the case, in particular, for countriesclosely linked to the United States markets, eitherthrough geographical proximity or formal trade ar-rangements, such as Mexico and the smaller Cen-tral American and Caribbean countries.

42 For details on PROEX, see www.bb.com.br/appbb/portal/gov/ep/srv/fed/AdmRecPROEX.jsp. In asense, it could be argued that, in terms of exportfinance, PROEX simply seeks to bring Brazilianexporters on an equal footing with their competi-tors in countries that have sustained macroeconomicstability and strong financial markets.

43 Moreover, in 2001–2003, Brazil challenged the com-patibility with WTO rules and commitments of thelow-interest financing provided by the CanadianGovernment to a foreign importer of Bombardieraircraft. In February 2002, the WTO dispute settle-ment panel ruled that this aid constituted an illegalsubsidy.

44 As pointed out by Goldstein (2002: 112), to pre-vent abuse of the programme, financing must be atmarket rates plus a risk premium; loans must be forno longer than 10 years, and they must cover no

more than 85 per cent of the purchase in question.For a legal assessment of the WTO dispute settle-ment panel ruling, see Doh, 2003: 14–15.

45 For details on the criteria used for the selection ofspecific sectors, see Peres, 2006.

46 Employment creation and regional developmenthave been additional objectives of support to SMEclusters.

47 Gross domestic expenditure on R&D accounts, onaverage, for only 0.6 per cent of GDP in LatinAmerica and the Caribbean, compared to 1.2 percent in China, 1.8 per cent in the EU and 2.8 percent in the United States (UNESCO, 2005).

48 See China Daily, Speed application of high-tech ad-vance, 14 February 2006.

49 National Guideline on Medium and Long-term Pro-gramme for Science and Technology Development(2006–2020).

50 See China Daily, China to adjust export tax rebates,6 June 2006 (http://en.ce.cn/Business/Macro-economic/200606/15/t20060615_7365602.shtml).

51 See Shanghai Securities Daily, Merger of the dualtax system has reached the Ministry of Finance, 6June 2006 (www.china.org.cn/chinese/zhuanti/2006ssgclt/1231761.htm).

52 Loriaux (2003: 108–109) argues that the move fromfixed to flexible exchange rates in 1973 rendered apolicy of State-controlled bank credit very costlybecause it led to an “overdraft economy” in whichinterest rate hikes had little or no impact on the de-mand for credit by businesses.

53 Jacquet and Darmon (2005: 86) point out that thecreation of clusters of competitiveness is the “of-fensive” part of France’s industrial policy, whichhas been complemented by a “defensive” part con-sisting of tax credits and other fiscal benefits forindustrial sectors and geographical regions facingeconomic difficulties. For detailed informationon clusters of competitiveness, see www.competi-tivite.gouv.fr/.

54 Some of the 67 initially chosen projects weremerged, while the projects for an additional nineclusters were approved in December 2005 (Ministèrede l’Economie, des Finances et de l’Industrie, 2006).

55 The European Council meeting in Lisbon in March2000 set the objective of making Europe the mostcompetitive and dynamic knowledge-based economyin the world by the year 2010 by, inter alia, creatinga knowledge-based economy and enhancing com-petitiveness and innovation.

56 Apart from promoting the use of public procure-ment to stimulate research and innovation, the newinitiatives in the action plan include: a revised Stateaid policy, which aims to reduce State aid graduallywhile refocusing it on activities that are likely tohave the most sustainable impact on competitive-ness, jobs and growth, and that promote cross-


border cooperation in research; tax incentives forfirms to invest more in innovative areas; efforts to-wards the creation of an attractive single marketwithin the EU for researchers, structural cohesionand regional funding focused more on research andinnovation; and financial instruments to support re-search within SMEs (Commission of the EuropeanCommunities, 2005).

57 Contrary to the broad-based and virtually uncondi-tional protectionist measures that often accompa-nied inward-looking, import-substituting industri-alization strategies, export targets were the main per-formance standard imposed by East Asian govern-ments on business as a reciprocal control mecha-nism for public policy support. They were designedto help the supported production activities achieveinternational competitiveness and to minimize therisk of rent-seeking and other abuse of public policysupport.

58 However, in some instances, the mere threat of sanc-tions from other countries may have an impact on acountry’s policy-making.

59 The generally long time lapse between the adop-tion of a certain policy measure that potentially in-fringes rules and the ruling of a dispute settlementpanel may allow countries with strong administra-tive capabilities to achieve the intended goal anddiscontinue the policy measure before such a rulingand the potentially associated sanctions are adopted.

60 Another area is the negotiations on a multilateralinvestment agreement, now dormant, that aimed atremoving virtually all restrictions on FDI.

61 It may also induce an even greater use of anti-dump-ing measures and countervailing duties, which areinherently discriminatory and costly to implement.

62 According to Finger and Nogues (2002: 334), in-fluential developed countries had announced thatthey would withdraw from the GATT as soon as theWTO came into existence. This implied that a coun-try that did not accept the “grand bargain” of theURAs would not have enjoyed protection from dis-criminatory treatment, either from the new WTO orthe old GATT rules and regulations.

Notes for figure 5.2:The product categories are based on the InternationalStandard Industrial Classification (ISIC) Rev. 3 forMexico and the Republic of Korea; and on ISIC Rev. 2for Brazil. Resource-intensive manufactures include: 15,16, 20, 21, 23, 26, 27 and 28 in Rev. 3, and 311, 313, 314,331, 341, 353, 354, 362, 369, 371, 372 and 381 in Rev. 2.Labour-intensive manufactures include: 17, 18, 19, 22,25, 36 and 37 in Rev. 3, and 321, 322, 323, 324, 332, 342,355, 356, 361 and 390 in Rev. 2. Technology-intensivemanufactures include: 24, 29, 30, 31, 32, 33, 34 and 35 inRev. 3, and 351, 352, 382, 383, 384 and 385 in Rev. 2.This classification is based on the categories used in TDR2002, chap. III.


Aguayo Ayala F and Gallagher KP (2005). PreservingPolicy Space for Sustainable Development: TheSubsidies Agreement at the WTO. Winnipeg, Inter-national Institute for Sustainable Development.

Akyüz Y (2005). The WTO negotiations on industrial tar-iffs: What is at stake for developing countries? Ge-neva, Third World Network.

Amsden AH (1989). Asia’s Next Giant: South Korea andLate Industrialization. Oxford, Oxford UniversityPress.

Amsden AH (1999). Industrialization under new WTOlaw. Paper prepared for the UNCTAD High-levelRound Table on Trade and Development: Directionsfor the Twenty-first Century. TD(X)/RT.1/7. Geneva.1 December 1999.

Amsden AH (2001). The Rise of “the Rest”: Challengesto the West from Late-Industrializing Economies.New York, Oxford University Press.

Anderson K (2002). Economywide dimensions of tradepolicy and reform. In: Hoekman B, Mattoo A andEnglish P, eds., Development, Trade and the WTO.A Handbook. Washington, DC, World Bank.

Bairoch P (1993). Economics and World History. Myths andParadoxes. Chicago, University of Chicago Press.

Baldwin R (1969). The case against infant industry pro-tection. Journal of Political Economy, 77: 295–305.

Baunsgaard T and Keen M (2005). Tax revenue and (or?)trade liberalization. Working Paper no. WP/05/12,International Monetary Fund, Washington, DC.

Beffa JL (2005). Pour une nouvelle politique industrielle.Paris, La documentation française.

Bora B, Lloyd PJ and Pangestu M (2000). Industrial policyand the WTO. World Economy, 23 (4): 543–559.

Brander J and Spencer B (1985). Export subsidies andinternational market share rivalry. Journal of Inter-national Economics, 18: 83–100.

Bruton H (1998). A reconsideration of import substitu-tion. Journal of Economic Literature, 26: 903–936.

Burgess R and Venables AJ (2004). Toward a micro-economics of growth. In: Bourguignon F and

Pleskovic B, eds., Accelerating Development. Pro-ceedings of the Annual World Bank Conference onDevelopment Economics. New York, World Bankand Oxford University Press.

Cao C (2004). Challenges for technological developmentin china’s industry: Foreign investors are the mainproviders of technology. China Perspectives, no. 54July/August. Paris, French Centre for Research onContemporary China. Available at: www.cefc.com.hk/uk/pc/articles/art_ligne.php?num_art_ligne=5401.

Chang HJ (1996). The Political Economy of IndustrialPolicy. Houndsmill and London, Macmillan.

Chang HJ (2002). Kicking Away the Ladder: Develop-ment Strategy in Historical Perspective. London,Anthem.

Chang HJ (2005). Kicking away the ladder: “good poli-cies” and “good institutions” in historical perspec-tive. In: Gallagher KP, ed., Putting DevelopmentFirst. The Importance of Policy Space in the WTOand International Financial Institutions. Londonand New York, Zed Books.

Chang HJ (2006). Policy space in historical perspectivewith special reference to trade and industrial poli-cies. Economic and Political Weekly, 41: 627–633,18 February.

Cho AH and Dubash NK (2005). Will investment rulesshrink policy space for sustainable development?Evidence from the electricity sector. In: GallagherKP, ed., Putting Development First. The Importanceof Policy Space in the WTO and International Finan-cial Institutions. London and New York, Zed Books.

Christiansen H, Oman C and Charlton A (2003). Incen-tives-based competition for foreign direct invest-ment: the case of Brazil. Working Paper on Interna-tional Investment no. 2003/1. Paris, OECD Direc-torate for Financial, Fiscal and Enterprise Affairs.

CIADT (Comité Interministériel d’Aménagement et deDéveloppement du Territoire) (2004). Appel àprojet: pôles de compétitivité. Paris. Available at:

References


www.competitivite.gouv.fr/IMG/pdf/Appel_a_projets_poles_de_competitivite_VF.pdf.

Cimoli M et al. (2006). Growth, structural change andtechnological capabilities: Latin America in a com-parative perspective. Working Paper no. 2006/11.Pisa. Laboratory of Economics and Management(LEM), University of Pisa.

Commission of the European Communities (2005). Moreresearch and innovation – investing for growth andemployment: a common approach. Document COM(2005) 488 final. Brussels.

Correa CM (2000). International Property Rights, theWTO and Developing Countries. The TRIPS Agree-ment and Policy Options. London and KualaLumpur, Zed Books and Third World Network.

Correa CM (2005). An agenda for patent reform and har-monization for developing countries. Paper preparedfor the Bellagio Dialogue on Intellectual Propertyand Sustainable Development: Revising the Agendain a New Context, 24–28 September. InternationalCentre for Trade and Sustainable Development(ICTSD).

Correa CM (2006). Implications of bilateral free tradeagreements on access to medicines. Bulletin of theWorld Health Organization, 84 (5): 399–404.

Das BL (2003). The WTO and the Multilateral TradingSystem: Past, Present and Future. London, New Yorkand Penang, Zed Books and Third World Network.

De Negri J (2006). Innovation, technology and perform-ance of Brazilian industrial companies: major re-search findings. UNCTAD seminar presentation.Mimeo. Brasilia, Institute for Applied EconomicResearch (IPEA).

Di Caprio A and Amsden AH (2004). Does the new interna-tional trade regime leave room for industrializationpolicies in the middle-income countries? WorkingPaper no. 22 for the World Commission on the SocialDimension of Globalization. Geneva, Policy Inte-gration Department, International Labour Office.

Doh JP (2003). The Bombardier-Embraer dispute and itsimplications for Western Hemisphere integration.Policy Papers on the Americas, XIV (12). Wash-ington, DC, Center for Strategic and InternationalStudies.

Ernst D (2004). Internationalisation of innovation: Whyis chip design moving to Asia? Working Paper 64,East-West Center, Honolulu, HI, March.

Evans P (1995). Embedded Autonomy: States and Indus-trial Transformation. Princeton, NJ, Princeton Uni-versity Press.

Finger JM and Nogues JJ (2002). The unbalanced Uru-guay Round outcome: the new areas in future WTOnegotiations. World Economy, 25 (3): 321–340.

Francois JF and Martin W (2002). Binding tariffs: whydo it? In: Hoekman B, Mattoo A and English P, eds.,Development, Trade and the WTO: A Handbook.Washington, DC, the World Bank.

Gallagher KP (2005). Globalization and the nation-state:reasserting policy autonomy for development. In:Gallagher KP, ed., Putting Development First: TheImportance of Policy Space in the WTO and Inter-national Financial Institutions. London and NewYork, Zed Books.

Goldstein A (2002). EMBRAER: from national championto global player. CEPAL Review, 77: 97–115.

Gomory RE and Baumol WJ (2000). Global Trade andConflicting National Interests. Cambridge, MA, andLondon, MIT Press.

Hausmann R and Rodrik D (2003). Economic develop-ment as self-discovery. Journal of DevelopmentEconomics, 72 (2): 603–633.

Imbs J and Wacziarg R (2003). Stages of diversification.American Economic Review, 93 (1): 63–86.

Jacquet N and Darmon D (2005). Les Pôles de Compétitivité.Paris, La documentation française.

Kaldor N (1966). Causes of the Slow Rate of EconomicGrowth of the United Kingdom. Cambridge, Cam-bridge University Press.

Kowalski P (2005). Impact of changes in tariffs on devel-oping countries’ government revenue. Trade PolicyWorking Paper no. 18. Paris, Organisation for Eco-nomic Co-operation and Development (OECD).

Krueger AO (1990). Government failures in development.Journal of Economic Perspectives, 4: 9–23.

Kumar N (2003). Intellectual property rights, technologyand economic development: experiences of Asiancountries. Economic and Political Weekly, 23: 209–225, 18 January.

Kumar N (2005). Performance requirements as tools ofdevelopment policy: lessons from developed anddeveloping countries. In: Gallagher KP, ed., PuttingDevelopment First: The Importance of Policy Spacein the WTO and International Financial Institutions.London and New York, Zed Books.

Laird S, Vanzetti D and Fernandez de Cordoba S (2006).Smoke and mirrors: making sense of the WTO in-dustrial tariff negotiations. UNCTAD Policy Issuesin International Trade and Commodities Study Se-ries no. 30. United Nations publication, sales no.E.05.II.D.16, New York and Geneva.

Lall S (2004). Reinventing industrial strategy. The roleof government policy in building industrial com-petitiveness. G-24 Discussion Paper no. 28. UnitedNations document no. UNCTAD/GDS/MDPB/G24/2004/4, New York and Geneva.

Landes DS (1998). The Wealth and Poverty of Nations.Why Some are so Rich and Some are so Poor. NewYork and London, WW Norton.

Levine R and Renelt D (1992). A sensitivity analysis ofcross-country growth regressions. American Eco-nomic Review, 82: 942–963.

Linden G (2004). China standard time: a study in strate-gic industrial policy. Business and Politics, 6 (3),article 4. www.bepress.com/bap/vol6/iss3/art4.


Loriaux M (2003). France: a new ‘capitalism of voice’?In: Weiss L, ed., States in the Global Economy:Bringing Domestic Institutions Back In. Cambridgeand New York, Cambridge University Press.

Lu D (2001). Industrial policy and resource allocation:Implications on China’s participation in globaliza-tion. China Economic Review, 11 (4): 342–360.

Lugones G (2006). Recent policy experiences of struc-tural change in Latin America: selected case studies.Mimeo. Geneva, UNCTAD.

Maddison A (2001). The World Economy: A MillennialPerspective. Paris, Organisation for Economic Co-operation and Development (OECD).

Maskus KE (1997). Implications of regional and multi-lateral agreements for intellectual property rights.World Economy, 20: 681–694.

Mayer J (1996). Learning sequences and structural di-versification in developing countries. Journal ofDevelopment Studies, 33 (2): 210–229.

Mayer J (2004). Industrialization in developing countries:some evidence from a new economic geographyperspective. UNCTAD Discussion Paper no. 174.Geneva.

Meade JE (1955). Trade and Welfare: The Theory of In-ternational Economic Policy, Vol. 2. London, NewYork and Toronto, Oxford University Press.

Melo A (2001). Industrial policy in Latin America andthe Caribbean at the turn of the century. WorkingPaper no. 459, Inter-American Development Bank,Washington, DC.

Ministère de l’Economie, des Finances et de l’Industrie(2006). Tableau de Bord de l’Innovation. 14e édition.Paris. Available at: www.industrie.gouv.fr/observat/innov/pdf/tbi14.pdf.

MOFCOM (2006a). FDI Statistics by Ministry of Com-merce (China). Available at: http://kjs.mofcom.gov.cn/aarticle/bn/bs/200601/20060101437606.html.

MOFCOM (2006b ). FDI Statistics and Analysis of High-tech Imports and Exports of 2005. Department ofScience and Technology of the Ministry of Com-merce (China), January. Available at: www.mofcom.gov.cn, accessed 29 Jan. 2006.

Moreno-Brid JC, Rivas Valdivia JC and Santamaria J(2005). Mexico: economic growth, exports and in-dustrial performance after NAFTA. Serie Estudiosy Perspectivas, ECLAC, Santiago.

Morin JF (2003). Le droit international des brevets : en-tre le multilatéralisme et le bilatéralisme américain.Etudes Internacionales, 34 (4): 537–562.

Mortimore M and Barron F (2005). Informe sobre la in-dustria automotriz mexicana. Serie DesarrolloProductivo no. 162, ECLAC, Santiago.

Noland M and Pack H (2005). The East Asian industrialpolicy experience: implications for the Middle East.Working Paper 05–14, Institute for InternationalEconomics, Washington, DC.

OECD (2005). Science, Technology and Industry Score-board 2005. Paris. Organisation for Economic Co-operation and Development.

OECD (2006). Examens Territoriaux de l’OCDE : France.Paris, Organisation for Economic Co-operation andDevelopment.

Pack, H (2000). Industrial policy: growth elixir or poi-son? World Bank Research Observer, 15: 47–67.

Pack H and Saggi K (2006). The case for industrial policy:a critical survey. Working Paper 3839, the WorldBank, Washington, DC.

Peres W (2006). El (lento) retorno de las políticas indus-triales en América Latina y el Caribe. Serie DesarrolloProductivo no. 166, ECLAC, Santiago.

Puga D and Venables AJ (1996). The spread of industry:spatial agglomeration in economic development.Discussion Paper 1354, Centre for Economic PolicyResearch, London.

Puga, D and Venables AJ (1999). Agglomeration and eco-nomic development: import substitution versus tradeliberalization. Economic Journal, 109: 292–311.

Reichman JH and Hasenzahl C (2003). Non-voluntarylicensing of patented inventions: Historical perspec-tive, legal framework under TRIPS, and an over-view of the practice in Canada and the USA. IssuePaper no. 5, UNCTAD-ICTSD Project on IPRs andSustainable Development. Geneva, UNCTAD andITCSD.

Rodriguez F and Rodrik D (2000). Trade policy and eco-nomic growth: a skeptic’s guide to the cross-nationalevidence. In: Bernanke B and Rogoff KS, eds.,Macroeconomics Annual 2000. Cambridge, MITPress.

Rodrik D (2004). Industrial policy for the twenty-firstcentury. Discussion Paper no. 4767, Centre for Eco-nomic Policy Research, London. November.

Rodrik D (2006). Goodbye Washington Consensus, HelloWashington Confusion? Mimeo. January. Availableat: http://ksghome.harvard.edu/~drodrik/Lessons%20of%20the%201990s%20review%20_JEL_.pdf.

Ros J (2000). Development Theory and the Economics ofGrowth. Ann Arbor, MI, University of MichiganPress.

Safarian AE (2003). The use and impact of performancerequirements in developed countries. In: UNCTAD,The Development Dimension of FDI: Policy andRule-Making Perspectives. Proceedings of the Ex-pert Meeting held in Geneva from 6–8 November2002. Document no. UNCATD/ITE/IIA/2003/4,United Nations publication, sales no. E.03.II.D.22,United Nations, New York and Geneva.

Sala-i-Martin X, Doppelhofer G and Miller RI (2004).Determinants of long-term growth: a Bayesian av-eraging of classical estimates (BACE) approach.American Economic Review, 94: 813–835.

Scitovsky T (1954). Two concepts of external economies.Journal of Political Economy, 62 (2): 143–151.


Shadlen K (2005a). Exchanging development for marketaccess? Deep integration and industrial policy un-der multilateral and regional-bilateral trade agree-ments. Review of International Political Economy,12: 750–775.

Shadlen K (2005b). Policy space for development in theWTO and beyond: the case of intellectual propertyrights. Working Paper no. 05–06, Global Develop-ment and Environment Institute, Tufts University,Medford.

Singh A (2005). Special and differential treatment: themultilateral trading system and economic develop-ment in the twenty-first century. In: Gallagher KP,ed., Putting Development First: The Importance ofPolicy Space in the WTO and International Finan-cial Institutions. London and New York, Zed Books.

Stiglitz JE (2005). Development policies in a world ofglobalization. In: Gallagher KP, ed., Putting Devel-opment First: The Importance of Policy Space inthe WTO and International Financial Institutions.London and New York, Zed Books.

Toner P (1999). Main Currents in Cumulative Causation:The Dynamics of Growth and Development.Houndsmill and London, Macmillan.

Tsangarides CG (2005). Growth empirics under modeluncertainty: Is Africa different? Working Paper no.05/18, International Monetary Fund. Washington,DC. January.

UNCTAD (1994). Directory of Import Regimes. Part I:Monitoring Import Regimes. UNCTAD/DMS/2/Rev.1 (PART I). United Nations publication, salesno. E.94.II.D.6, New York.

UNCTAD (2000). World Investment Report 2000: Cross-border Mergers and Acquisitions and Development.United Nations publication, sales no. E.00.II.D.20,New York and Geneva.

UNCTAD (2005). Globalization of R&D and developingcountries. UNCTAD/ITE/IIA/2005/6, Geneva.

UNCTAD and ICTSD (2005). Resource Handbook onTRIPS and Development. Cambridge, CambridgeUniversity Press.

UNCTAD (various issues). Trade and Development Re-port, United Nations publication, New York andGeneva.

UNESCO (2005). Science Report 2005. Paris, UNESCO.UNIDO (2005). Industrial Development Report 2005.

Capability Building for Catching-up. Historical,

empirical and policy dimensions. United Nationspublication, sales no. E.05.II.B.25, Vienna. UnitedNations Industrial Development Organization.

United States Trade Representative (USTR) (2004). TheWork of USTR – Intellectual Property. Washing-ton, DC. Available at: www.ustr.gov.

Wade RH (1990). Governing the Market. Princeton,Princeton University Press.

Wade RH (2003a). Creating capitalisms. Introduction tothe 2003 edition of Governing the Market.Princeton, Princeton University Press.

Wade RH (2003b). What strategies are viable for devel-oping countries today? The World Trade Organiza-tion and the shrinking of ‘development space’. Re-view of International Political Economy, 10 (4):621–644.

Wade RH (2006). Bringing industrial policies back in:Breaking the policy monopoly. Mimeo. Backgroundpaper prepared for UNCTAD’s Trade and Devel-opment Report 2006.

Weiss L (2005). Global governance, national strategies:how industrialized states make room to move underthe WTO. Review of International PoliticalEconomy, 12 (5): 723–749.

Wood A and Mayer J (1998). Africa’s export structure ina comparative perspective. African Development ina Comparative Perspective, Study no. 4, UNCTAD,Geneva.

World Bank (1993). The East Asian Miracle: EconomicGrowth and Public Policy. New York and Oxford,Oxford University Press for the World Bank.

World Bank (2005). Economic Growth in the 1990s.Learning from a Decade of Reform. Washington,DC, World Bank.

WTO (2001). Implementation-Related Issues and Con-cerns. Decision of 14 November 2001. DocumentWT/MIN(01)/17, World Trade Organization, Ge-neva.

WTO (2004). The July 2004 Package. Available at: www.wto.org/english/tratop_e/dda_e/dda_package_july04_e.htm

WTO and UNCTAD (2002). Joint Study on Trade-Re-lated Investment Measures and Other PerformanceRequirements. WTO document G/C/W/307/Add.1,Geneva.

Zeira J (1997). Investment as a process of search. Jour-nal of Political Economy, 95: 204–210.

trade and development report, 2006

Documents