The flexible developmental state: globalization, information technology and the "Celtic Tiger"

POLITICS & SOCIETYSEÁN Ó RIAIN The Flexible Developmental State: Globalization, Information Technology, and the “Celtic Tiger” SEÁN Ó RIAIN REMAKING THE DEVELOPMENTAL STATE For a number of years, commentators have looked to the “developmental states” of East Asia as a counterweight against the forces of economic globalization. These states have been seen as cause for hope that nations could promote economic development, even from a subordinate position in the international division of labor. Until the 1990s, this developmental state presided over one of the most spectacular economic success stories of this century, while other economies(e.g.,inLatinAmericaandtheEuropeanperiphery)languished,oftenaslittle more than export platforms.1 In the 1990s, however, the developmental state seemed to have had its day in thesun. The Japanese andKorean states appeared too inflexibletocope with rapidly changing informational industries and decentralized “post-Fordist” industrialstructures.KoreanandJapanesefirmsranintoincreasingdifficultiesthrough the 1990s in rapidly changing high-technology markets. The developmental statesalsoappearedtooweaktomanagetheincreasinglyinternationalizedeconomies over which they presided, as evidenced in the financial crisis of 1997 and 1998.2Theemergenceofflexibleregionaleconomiesanddispersedtransnational networksoftechnologyandfinanceappearedtohavesimultaneouslyundermined the developmental state from below and from above. Many thanks to Gerald Autler, Ted Egan, Michael Burawoy, Peter Evans, Séamus Grimes, Becky King, Seán Ó Foghlú, Denis O’Hearn, Balaji Parthasarathy, Bill Roche, Anno Saxenian, Andrew Schrank,KathleenSchwartzman,JohnSterne,MattZook,andthePolitics&Societyeditorialboard. POLITICS & SOCIETY, Vol. 28 No. 2, June 2000 157-193 © 2000 Sage Publications, Inc. 157 However, a different set of countries has emerged in the 1990s that appear likely candidates for mobility in the hierarchy of the international economy. The Republic of Ireland is among those economies that have begun to change their relation to the global economy based on growing strengths in information technology(IT)industries.EconomicgrowthratesinIreland3haveimproveddramatically since the economically and socially disastrous 1980s and have been among the most rapid in the world in the mid-1990s. There has been a significant improvement in export performance, and domestic demand has increased since 1994afterthemassunemploymentandemigrationofthe1980sandthecautionof the early 1990s. Consumer prices, despite some inflationary pressures in recent years, have risen slowly. All of the above indicators compare favorably with the performanceofotherleadingeconomiesintheOrganizationforEconomicCooperation and Development (OECD), including even comparatively “successful” economies such as the United States, the Netherlands, and even the Asian Tigers in the 1990s.4Furthermore, the “jobless growth” of earlier decades has been replaced by the highest rate of employment growth and the highest proportion of the population employed in the history of the state. The mass emigration of the 1980s has given way to a substantial net inflow of migrants, largely consisting of emigrants returning to Ireland but also including non-Irish-born immigrants. Unemploymenthas fallenmoreslowlybuthas recentlyreached itslowestlevelin seventeen years. Growth in Ireland is driven not just by foreign investment but by two relatively distinct modes of integration into the global economy—the partial local embedding of global corporate networks and the increasingly successful integration of local networks of indigenous firms into global business and technology networks. These two globalizations are further embedded within a set of national neocorporatist institutions that have managed the relation to the global economy of both the macro economy and of unionized workers. Since 1987, a series of national “social partnership” agreements have negotiated wage restraint, public spending limits, and some efforts to bridge social exclusion at the local level. The most effective developmental states of the 1990s have been transformed from the bureaucratic models that characterized the East Asian economies to the flexible versions we find in economies such as Ireland and Israel. The flexible developmental state (FDS) is defined by its ability to nurture post-Fordist networks of production and innovation, to attract international investment, and to link these local and global technology and business networks together in ways that promote development. This ability is sustained by the multiple embeddedness of the state in professional-led networks of innovation and in international capital, and by the state’s flexible organizational structure that enables the effectivemanagementofthismultiplicity.Thisuneasystructureofmultiplealliancesis threatened,however,bythedifficultiesofafragmentedstatestructureinhandling 158 POLITICS & SOCIETY the tensions arising from uneven internationalization of society and growing inequality. The Irish software industry is an informative case study through which to analyze the key characteristics of this emerging development strategy.5The industry has grown rapidly, is one of the most dynamic software industries outside the UnitedStates,andisdividedrelativelyequallyinemploymenttermsbetweenforeignandIrish-ownedfirms.ItoccupiesanincreasinglysignificantroleintheIrish economy. Software is a significant employer in its own right, behind only electronic equipment, food, and chemicals as industrial employers in 1997. It is also growing more quickly in terms of employment than any manufacturing sector.6 The industry accounted for 12 percent of Irish exports in 1998 and, perhaps more important, was rapidly becoming a model for industrial policy in other sectors of the economy. The article begins by critiquing accounts that overemphasize either local or globalprocessesthroughananalysisofthedynamicsofrecentindustrialandeconomic development in Ireland. I then develop the concept of the FDS in some detail as an alternative theoretical approach to understanding the central features of Irish economic development. The remainder of the article describes the structureofthesoftwareindustryinIrelandandappliesthethreedimensionsofthetheorytoexplainthedevelopmentanddilemmasoftheIrishsoftwareindustry—how the state intervened in the process of globalization itself, the social and institutional conditions of this state role, and the threat posed to development by social inequality.Finally,someissuesconcerning thesustainabilityandgeneralizability of the model are considered. NATIONAL DEVELOPMENT AND THE GLOBALIZATIONS OF IRELAND For some authors, the success of the Irish economy since the early 1990s is based on its mastery of localized flexible production and innovation. Charles SabelarguesthatIrishsuccessisbasedontheabilityofIrishfirmstolearnthelessons of decentralized post-Fordist production and create local alliances that support learning, efficiency, and innovation.7Local strength becomes the basis of global competitiveness. However, a variety of commentators from different endsofthepoliticalspectrumemphasizeIreland’slocationwithintheprocessof globalization as the critical feature of this turnaround. A persistence with export orientation,freetrade,andforeigninvestmentappearstobethekeytoIrishdevelopment in the 1990s, for better or worse. Denis O’Hearn explains the emergence of the Celtic Tiger in the following terms: Americancorporations mademajormovestoincrease theirpresencein the Europeanmarket; they tended to agglomerate their new projects to take advantage of the flexibility this allowed; and Ireland was fortunate enough to receive a major share of American-owned electronicsprojectsinEuropebecausetheyagglomeratedaroundothermajorfirmssuchas Intel.8 SEÁN Ó RIAIN 159 The similarities in the causal explanations provided by a more conservative analyst, Paul Krugman, are striking (although their evaluations of the process differ markedly): Thanks in part to luck, in part to policies . . ., Ireland got a head start over other European locations in attracting what became a surge of inward foreign direct investment; the early investments both generated a cascade through informational effects and, eventually, created external economies that further reinforced Ireland’s advantages.9 In fact, it will prove necessary to combine these “local” and “global” theories toprovidea full account of themultiplewaysin which localandglobal processes shape different modes of integration into the global economy and how they are sustained by particular institutions. Irish industry is characterized by two such broad modes of integration into the global economy, combining local and global networksindifferentways.Thefirstisbasedonattractingforeigninvestmentand attempting to embed it in the local economy, and the second and more recent modelisbased onthe growthofindigenousIrish-ownedfirmsthatcompeteinternationally and are increasingly closely integrated into international technology and business networks.10 Employment in foreign-owned industry grew 24.8 percent in manufacturing and 384.5 percent in financial and internally traded services (including software) between 1989 and 1998.11 Denis O’Hearn provides a cogent critique of reliance on such transnational corporation (TNC) investment as a model of economic development.12 He argues that Irish economic growth is largely an illusion as it is dominatedbyinvestmentbyTNCs,andeconomicstatisticsaregrosslyinflatedby the exports of a very narrow band of economic sectors and even key firms within those sectors—such as Intel in computers and Microsoft in software. Indeed, foreign profits dominate the Irish economy to such an extent that the gross domestic product (GDP) overstates the size of the Irish economy by some 13 percent, as it does not account for the repatriation of transnationals’ profits. This is further aggravatedbythepresenceofcertainsectors(includingsoftwareduplication)that are essentially “entrepôt” sectors, sectors that account for approximately 10 percent of manufacturing employment but around 30 percent of output.13 While foreign investment is certainly open to a wide range of criticisms, there are some indications that at least pockets of more sophisticated work have emerged among TNCs in Ireland. Intramural spending on research and development (R&D) among foreign-owned companies increased from £180m in 1993 to £256min1995andfrom2.1percentofsalesin1993to2.5percentin1995.14 TNC salesarecertainlyinflatedbytransferpricing15 andvariousotherformsofcreative accounting, but this, in turn, may obscure some of the resources they generate withintheeconomy.Inparticular,statisticsbasedonpercentagesofTNCsalesare likely to understate the sophistication of their operations, given the inflation of these sales figures. The Irish economy expenditures of TNCs declined as a percentage of (artificially inflated) sales between 1990 and 1996 (from 37.9 percent 160 POLITICS & SOCIETY to35.2percent),buttheabsolutespendingofTNCsintheIrisheconomyrose55.3 percent during that period.16 There is some evidence therefore of a local upgradingand embedding of some TNC operations, although this is limited, as O’Hearn rightly points out. Nonetheless, O’Hearn is too quick to reduce the dynamics of Irish developmenttoaTNC-led“entrepôt”model,dominatedbycompaniessuchasMicrosoft, Intel,and Dell. This cannot explain the existenceof somerelativelysophisticated TNC operations. Nor is the growth of indigenous industry simply derivative of TNC investment, as evidenced by the data on the relatively limited local linkages of TNCs. Critics of Ireland’s reliance on foreign investment cannot have it both ways. Given the legitimate aspects of the critique of the developmental impact of foreign investment, it then cannot be argued that Irish development has derived almostcompletelyfromkeyinvestmentssuchasIntelandMicrosoft.Evenwithin theTNCsector,thesefirmsremainrelativelyisolatedandhavelittletodowiththe more interesting TNCs such as Ericsson, Digital, and so on. Their disproportionateimpactonIrishexportandgrowthfiguresshouldnotblindustotheirrelativelackofcontributiontotheveryrealtransformationofproductionandinnovationcapabilities within parts oftheIrisheconomy.Finally, the neocorporatist “social partnership” arrangements in place since 1987 cannot be explained by O’Hearn’s model as they are only of tangential relevance to most of the TNCs. Most TNCs are nonunion employers for whom wage costs are a relatively low priority, and they have generally paid well above the terms of the national wage agreements. The spectacular growth in the output of TNCs can obscure the impressive performanceof Irish indigenous industryinwhichemploymentgrew10.4 percent in manufacturing and 197.7 percent in internationally traded services. This indigenous industry performance is very impressive in comparative terms, all the more sosinceitiscurrentlyonaveryrapidgrowthtrajectoryasmostoftheemployment growthhasoccurredsince1994.Furthermore,thisisnotgrowththatisdependent on links to TNCs in Ireland (which are generally weak)—exports of Irish-owned manufacturingfirmsgrew fasterthan both output and employmentbetween1990 and 1996.17 Industrial R&D in the whole economy increased to 1.02 percent of GDP from 0.48 percent in 1986 and 0.53 percent in 1990. This was achieved through a per annum real rate of growth of R&D spending of 15 percent among foreign firms and 16 percent among Irish-owned firms.18 The claims of authors such as Charles Sabel that the unexpected success of industry in Ireland is due to an ability to master flexible production contains an elementoftruth.Indigenousfirmstendtobesmallandmediumsizedwithatleast some evidence of networking among firms.19 There is at least some evidence of significant diffusion of teamwork at the level of the organization of production, although not sufficient to warrant the more enthusiastic claims made on Irish industry’s behalf.20 The picture that emerges of indigenous industry is of an overSEÁN Ó RIAIN 161 allupgradingintheorganizationofproductionandofmanagementandmarketing capabilities, resulting in growth spread across a range of sectors. Investment, although clearly not the driver of development, has picked up significantly, particularly in the most dynamic sectors.21 There are a number of sectors, in particular financial services and IT, that exhibitacombinationofbuoyantinvestment,highlevels ofR&D,significantlevels of “collaborative production,” and “clustering” or dense networking and that are the most dynamic in the Irish economy.22 The software industry is clearly located at this “post-Fordist leading edge” of the indigenous sector, within a context of overall upgrading of capabilities across a broad range of industries. Industrial growth has translated into a much improved employment performance. Between 1986 and 1995, full-time employment increased by 2.5 percent among men and 21.2 percent among women, while part-time employment increased by 24.4 percent among men and 36.6 percent among women. The Irish economy has indeed been creating part-time jobs rapidly, but it has also created significant numbers of full-time jobs, even more so in recent years. Furthermore, thejobscreatedhave notsimply been“bad”jobs.Infact,employmentgrowthhas been characterized by an overall upgrading of the occupational structure combined with a significant polarization of occupations and wages. The most rapidly expanding occupational categories between 1981 and 1995 include catering, personal services, sales workers, and security workers—which added 59,000 employees to their numbers during these years (an increase of 35.5 percent). However, by 1995, there were also 94,000 more managers, professionals, and associateprofessionals—anevenbiggerincreaseof44percent.Meanwhile,more classically “Fordist” occupations such as clerks, typists, operatives, transport workers, and skilled and craft workers added only 7,000 extra jobs, and farmers and laborers decreased in number by 68,000 or 27.3 percent.23 Inequality has been rising rapidly within the Irish economy, in terms both of the share of national income going to capital rather than labor24 and of wage inequality.Irelandhasseenwhatisprobablythemostrapidriseinwageinequality within the OECD, despite being second only to the United States in inequality in the 1980s.25 Private consumption’s share of GDP has decreased from a relatively stable figure of around 59 percent from 1985 to 1992 to 55.4 percent in 1996.26 Furthermore, the bottom half of the income distribution has seen its share of national household consumption reduced from 28.7 percent to 27.2 percent. Although GDP has grown rapidly during that period and absolute consumption has increased in real terms for all deciles of the income distribution, the increases have been meager indeed for the poorest 40 percent, a trend reflected in the continuing poverty (even by relatively restrictive definitions) of 15 percent to 16 percentoftheIrishpopulationbetween1987and1994.27 Nonetheless,whiletheIrish economy has undoubtedly created many “bad” jobs, it is also creating a significant number of “good” jobs. 162 POLITICS & SOCIETY The software industry has generated a disproportionate number of these “good” jobs and is one of the leading sectors within both foreign and indigenous industry. Before turning to a detailed investigation of the development of the industry, I will develop the central concept applied in that analysis—the flexible developmental state. BUREAUCRATIC AND FLEXIBLE DEVELOPMENTAL STATES “Local” and “global” perspectives on industrial development are onedimensional and insufficiently dynamic in their analyses. Combining these perspectives to investigate the multiple ways in which local and global processes shape one another over time allows us to identify spaces and opportunities for development strategies, even within an increasingly integrated global economy. Choices remain as to how best to integrate into global flows and networks. The state remains a central actor in the process of shaping these choices, building better or worse development paths out of these fragments of local and global networks. However, the ways in which the state mediates between local and global networks and the institutional foundations of this role in Ireland are different in a number of significant respects from the East Asian bureaucratic developmental states. It is necessary then to explain the emergence of each set of global connections that characterize the Irish economy and its software industry, and their surprising ability to coexist. Furthermore, we need to provide a comprehensive theoretical frameworkthatcanexplainboththepositiveexperiencesofeconomicgrowthand occupationalandindustrialupgrading andthe disturbingmassivesocialpolarization that accompanies them. Itistoaddressthesetheoreticalweaknessesandtoaccomplishtheseanalytical goals that I develop the concept of the flexible developmental state that is applied to the software industry case through the rest of this article. The most influential analyses of the role of the state in development have analyzed the success of the Asian Tiger economies such as Japan and Korea. Developmental states are the most likely candidates for boosting an economy beyond the narrow logic of privateinvestmentandtherebyimprovingitspositionin theinternational divisionof labor. They achieve their goals in the contemporary era not by taking on the tasks ofdevelopmentthemselvesbutbyshapingthecapabilitiesofsocietyandthemarkettodoso.Researchers havemovedfromanemphasisonanauthoritariandirective style of state intervention to an understanding of how the state pokes and prodsdomesticfirmstocompeteintheglobaleconomyandtoconstantlyupgrade their organizational and technical capabilities to that end. The state assists in the birth and growth of domestic, national firms through its role as “midwife” of new firmsandsectorsandbytendingtothe“husbandry”ofthesegrowingindustries.28 Through policies such as selective and strategic use of protectionism, the provision of industrial subsidies, and programs tied to performance, as well as the creSEÁN Ó RIAIN 163 ationofclosetiesbetweenfinancialcapital,industrialcapital,andthestate,economies such as Japan, Korea, and Taiwan were able to industrialize rapidly based onimprovedproductivityinmanufacturingand“industrializationbylearning.”29 However,the statecannotachievetheseendsinisolation;infact,it dependson its relation to society for its success. In particular, developmental states are characterized by what Peter Evans calls “embedded autonomy.” Such states are embedded in local capital through the close social ties between state bureaucrats anddomesticbusinessownersandmanagers.Whileaneducatedlaborforceorganized for learning is critical, labor is excluded from the key institutions of the bureaucratic developmental state.30 These states retain their autonomy due to the presence of a classic Weberian bureaucracy—based on meritocratic recruitment and promotion and norms of objective, procedural rationality.31 While embeddedness allows the state to gather information and mobilize resources, autonomy safeguarded by bureaucracy guarantees that national development goals remain central to state action.32 For this reason, I label the developmental states of East Asia bureaucratic developmental states (BDSs).33 But success can bring its own dilemmas. The developmental state promotes local firms and encourages them to compete globally. In becoming global firms, however, their alliances with the state are undermined as they become more and more closely aligned with the interests of their international partners. In the late 1990s, the Asian development project’s basis in the domestic alliance between state and capital was undermined as Asian firms internationalized, became more integrated into international financial markets, and ran into a massive debt crisis as Asian and U.S. capitalist institutions clashed, undermining the high debt financing of the Asiancorporations.Themuch-admiredcoherenceofthebureaucracy now became a liability as its responsiveness to changing local and global conditions proved to be poor. The theory of the developmental state has been a valuable one that has shed a great deal of light on the conditions of effective national development strategies. Nevertheless, the theory is relatively weak in explaining the role of the state in newly emerging economies such as Ireland and Israel, as well as in what Evans describesas“intermediatecases”suchasIndiaandBrazil.34 Therefore,itmustbe extendedalongitsthreeunderlying dimensions—state interventionin the globalization process, conditions of embedded autonomy, and threats to state interventionthatemergedynamically duringthedevelopmentprocess—todevelopaconception of a “remade” FDS. Table 1 outlines the differences between BDSs and FDSs along each of these three dimensions. First, BDS theorists analyze state intervention in the economy largely as a process of the creation of new domestic capabilities that can then be brought to bear in global markets. The global economy is theorized largely as a context in which states and national capitals can compete. However, under the globalizationproject,transnationalfirms,networksandflowsofmoney,informa164 POLITICS & SOCIETY tion, and resources have deeply penetrated the most successful localities and nations. Therefore, the role of the state is increasingly to mediate between the global and the local, to connect them, and to shape the nature of the relationship between them. The state may also be involved in creating the very actors that it hopeswillparticipateintheseglobal-localconnectionsandthedevelopmentproject more generally. While the BDS also promoted the creation of economic actors,itfocusedonhierarchicalbusinessgroupsas“nationalchampions”instead of transnationally integrated local networks.35 This is the first extension of BDS theory—the understanding of the role of the developmental state as mediating local and global connections. The FDS is defined precisely by its ability to create and animate post-Fordist networks of production and innovation and international networks of capital, and to link them together in ways that promote local and national development. The FDScanattempttodothisintwoprimaryways.Itcanconnecttoexistingflowsof capital by attracting foreign direct investment (FDI) and then building local networks of production (typically subsupply) and innovation (much more rarely) aroundthis “imported” industrialorganization.It can alsoattempttofoster indigenous networks of innovation and then encourage them to internationalize, but from a position of relative strength. In either case, the state development strategy is to connect the local to the global economy in such a way that local industrial transformation, accumulation, and development can take place. The FDS plays a key role in fostering “better” connections to the global. The Irish state, in fact, played a critical role in “scaling up” social networks within local technical communities into an innovative and growing industry. These efforts were organized largely through a variety of industrial development agencies that promoted an indigenous development alternative to reliance on forSEÁN Ó RIAIN 165 Table 1 Bureaucratic and Flexible Developmental States Bureaucratic Developmental State Flexible Developmental State Shaping globalization Managing dependency Mediating global connections – Strategic use of protectionism – Building local networks around global capital – Industrial subsidies – Taking local innovation networks global – Domestic banking system Sources of embedded Coherence of state bureaucracy Flexibility of state structure autonomy Embedded in Embedded in – Domestic capital – Professional labor – Foreign capital Threat to sustainability Internationalization of capital Internationalization of society Rigidity of state bureaucracy Fragmentation of state States and Globalizations Protectionist measures were eased after the 1950s for foreign investors in the Republic of Ireland and were removed entirely in 1964. The state became the key actor in attracting foreign investment, and attracting mobile investment became a dominantpolicygoalforthefollowingfortyyears.Amongthecriticalelementsof Ireland’s locational advantage were very generous tax incentives and grants, as wellasatransnational-friendlyenvironmentincludingnorestrictionsontherepatriationofprofits.Afterprovidingalow-orno-taxenvironmentsincethe1950s,a new policy in 1978 provided 10 percent corporation tax for all manufacturing from 1981 to 2000 inclusive. This was extended to firms in the internationally tradedservices sector (including softwareand data processing) in 1981andguaranteed until 2010 the same year. Recently a 12.5 percent corporate tax rate has been guaranteed until 2025. In later years, a world-class telecommunications system and, in particular, a young and cooperative labor force have arguably been as or more important. However, the state’s greatest efforts were dedicated to the mobilization and restructuring of local society through the creation and shaping of a “suitable” labor force. A state-led reorientation of the education system toward technical education was reinforced over time by the increasing influence of the new collegesfocused ontechnology,whosesuccessputpressureontheexistinguniversities to change their orientation to technical and scientific education and to fostering links with industry.51 Thestatewasplacedatcenterstageinindustrialpolicybyitseffortstocontinuallyupgrade these “factorsof production.” Ittookon the role of“midwife”toforeigninvestment.52 Thesepolicyandinstitutionalchangeshadthedesiredeffectof attracting extensive amounts of foreign investment, including such software firms as Microsoft, Lotus, Novell, and Corel. There has been some local embedding of these firms in a territorial complex around the Dublin region based on a subsupplier base in turnkey services and software manual printing. The state has played a central role in supporting the growth of these subsupplier industries, particularly by providing finance and advice to the entrepreneurs in these newly minted sectors and by coordinating relations among Irish-bornmanagersofTNCsandtheseemergingfirms.Printingfirms,forexample, received significant funding from the IDA for the expensive new Web print technologies required for software manual printing.53 Furthermore, the IDA began to target hardware companies in the early 1980s, encouraging them to expand their software operations. Some TNC subsidiaries have been able to develop such operations. This is particularly the case in more general IT and telecommunications companies such as Digital, Amdahl, IBM, Siemens Nixdorf, Phillips, Ericsson, or ATT/Lucent Technologies. As the manager of one of these operations said, “We have got more than the Microsoft and Lotus subsidiaries. The diversity of our parent company helps us a lot in that.” 172 POLITICS & SOCIETY Many policymakers did indeed hope that the TNCs attracted to Ireland would intimespinoffarangeofIrish-ownedfirms.Thishashappenedtoalimitedextent in the software industry, although it provides only one piece of the puzzle of the industry’sgrowth.One-thirdoftherespondentsinastudyofthirty-sixindigenous software entrepreneurs had worked in a TNC directly before starting their own company—either in TNCs in the IT sector or in the IT sections of TNCs in other sectors.Two-thirdsof indigenous entrepreneurshad workedfor a TNC in Ireland atsomestageoftheircareers,halfhadworkedabroadinsoftwareorarelatedsector,andhalfhadworkedin asector thatwasnowacustomeroftheirfirm.54 Direct spin-offs of software firms from software TNCs are relatively rare, however. While this international and customer experience might be expected given the dominanceofTNCsintheIrisheconomyandthemassemigrationofthe1980s,it has been a valuable source of experience for many technical entrepreneurs. Nevertheless, these figures alone cannot explain why these individuals did indeedultimatelyleavetheTNCsandstartoutonthepathofindigenousentrepreneurshiprather than taking the more common routesuntilthe early 1990s of emigration or long-term TNC employment. Return emigration did not become a significant factor until after the growth dynamic of Irish software was well under way, although it was clearly important in sustaining that growth. Nor can it explaintheinternalgrowthdynamicoftheindigenousindustry—themostsignificant part of the indigenous story. The same study showed that another third or so hadcomedirectlyfrom indigenous softwarefirms,while thefinalthird hadcome from indigenous firms in other sectors, most notably software distribution and computer hardware. Therefore, the dynamic of generating new companies is strongest in the indigenous sector and especially within the software industry itself. Sohowdidtheindigenous softwareindustry emergefromthevicious circle of TNC dependence and professional emigration? The state, through its heavy investment in education, had created a new class basis for an indigenous technologypromotionandbusinessexpansionagenda.Thedangerherewas,ofcourse,a massive brain drain that did indeed take place. However, of those who stayed, a number started their own companies and combined with a trickle of people from the foreign-owned sector and from user organizations to form a constituency that could support state agencies promoting indigenous development. What did the state agencies do to support the development of this new constituency? Thefirst element of the agencies’approach wasto definethe orientation of the Irish software industry. In its strategic review of 1992, the National Software Directorate (NSD) identified software products clearly as the major focus of the Irish industry, making official what had been the attitude of the state agencies for some time.55 Furthermore, the NSD identified the need of software product firms to export early in their development, due to the small size of the Irish domestic market, as a critically important issue for the Irish industry. The state agencies SEÁN Ó RIAIN 173 focusedthenonencouragingfirmswithsoftwareproductsforexport,beingmuch more receptive to such firms in their grant applications and designing some state supports(especiallyintheareaof marketing)sothattheywere orientedmainlyto the problems of product exporting. Industrial policy also began to shift through the 1980s toward a greater focus on indigenous industry and a greater selectivity in grant giving.56 However, the stateagenciesultimatelyfocusednotonlyon“pickingwinners,”whichtheydidto some extent, but also on a more generalized strategy of “making winners” out of the many firms they dealt with by upgrading the capabilities of the industry as a whole. Total grant payments by state agencies to indigenous software companies increased from IR£ 3 to 3½ million per annum between 1988 and 1990, up to 5.4 million in 1991, and 5.66 million in 1992.57 Forbairt, the primary agency dealing withindigenousfirmsatthetime,alsobecamemoredemandingoftheindigenous companies presenting proposals to them. As one Irish emigrant stockbroker in Silicon Valley put it to me: “I think Forbairt have been great. There’s a lot of money for companies, but you have to be good; you almost have to be selling in BritainorEuropebeforeyougetthegrants.Forbairtareverytough.”Thislevelof financing of the indigenous software industry was very significant in the context of a severe shortage of other capital sources for the industry. In fact, much of the early venture capital that did become available to the industry was induced into the industry by state actions. Forbairt, now part of Enterprise Ireland, ties the finance it provides to firms to various aspects of company development—in particular, R&D, marketing, and management development. A striking aspect of R&D funding is the critical importance of state and EU funding in the earlier period—accounting for almost 30 percent of funding directly and, since many of these funds were provided on a matching-funds basis, for around half of all R&D funding in the indigenous software companies in 1993. These governmental policies, therefore, were critical in stimulatingR&Dintheearlierstagesoftheindustry’sdevelopment.Anexecutive in Forbairt working closely with software companies told me that Forbairt attempts to provide for each stage of the development of a company—feasibility grantsasseedcapitalforstart-ups,employmentgrantsfortheearlygrowthphase, and then funding for R&D, training, and management development for the company development phase. Theimpactofthisapproachonfirmsseemsto havebeenpositive.Whilemany firmscomplainabouttheadministrativedemandsoffillingoutgrantapplications, analysis of grant-aided and non-grant-aided firms shows that through the early 1990s, at least grant-aided firms significantly outperformed non-grant-aided firms—both in the software industry and in the economy as a whole.58 Furthermore, many firms themselves say that they have found state support helpful, whichisimpressivegivenmanymanagers’skepticismofstateaction.59 Therefore, the agencies are clearly taking a self-consciously developmental approach to 174 POLITICS & SOCIETY companiesaswellastotheindustryasawhole,eventhoughthisapproachisoften couched in the more neoliberal language of enterprise.60 The Irish development agencies learned how to upgrade the capabilities of firmswithinadecentralizedindustry,askillquitedifferentfromthoseoftheBDS bureaucrats. However, they also went beyond this role and played a significant part in fostering and developing the social networks and associational life underpinning innovation in the industry. They have instituted “softer” supports such as the mentoring scheme, where small companies are put in touch with experienced industryfigureswhoforarelativelysmallsumbecomea“guide,philosopher,and friend” to the company, sometimes ultimately becoming company directors. The agenciesalsousedthegrant-aidprocessasanopportunitytoinformallystimulate connections among firms within the industry, drawing on their detailed knowledgeofotherfirmsaccumulatedthroughthegrant-aidprocess.ThesameForbairt executive mentioned above argued that the development agencies learned that the lone ranger approach doesn’t work. A company needs partners, the mentoring program helped with that. The agencies also played aninformalroleinintroducingpeople.That’ssomethingwewouldhavepushed,thepartnerships. The other thing the agencies pushed was the capital issue, Forbairt has been priming venture capital funds....Ourrole is not picking winners but “helping to make winners.”It’stoodifficulttopredictwinnerstopickthem,otherwisewewouldn’tneedto be working! A lot of our discussions with companies are not regarding the money but regardingbusinessissues....Thereisalotofcompanycapabilitydevelopmentwork;we werelessproactivethanthatin thedaysofCOPS[anindigenouscompanythatwentbust in 1990]. The agencies have gone from being funding agencies to doing company development. Furthermore, a network of industry and trade associations, universities, innovation and technology centers, and other forums and groups have been created that provide an associational infrastructure for information sharing, cooperation, and innovation in Irish software. While these bodies are outside the state or semiautonomous from it, in most cases they have been founded through state initiatives and underwritten by state guarantees and funding. Typically, they are located in universities and have representatives from the state, business, and education on their boards. Some brief examples should illustrate the key role of these institutions. The National Microelectronics Application Centre (MAC) was founded in the early 1980s and was charged with bringing electronics to Irish industry. Based in the UniversityofLimerick, it providestechnicaladvice and housesup-to-dateequipment for industry use and is reported by interviewees to have had close relations with many software firms. The director of MAC told me the following: Ourcoreskill isthedevelopmentcycle,rather thananyparticulartechnicalarea.Wespend a lot of time talking with entrepreneurs—90 percent may decide that their idea is not on, usually because the idea is out there already or they are too soon or too late vis-à-vis the SEÁN Ó RIAIN 175 technology curve. We get a lot of business from the remaining 10 percent. Typically, the people we are dealing with, it’s their first time through the cycle, they get upset at any shocks. Thecriticalpointhereisthatthereisaneducationalreturntotheindustryinterms ofthe90percentwhodonotpursueadoomedproject,eventhoughMACdoesnot receive anypayment from these companies. Indeed,theprimaryrole of MAC is a more specialized educational contribution than the development agencies can provide: We can lay out the potential of smart products for the entrepreneurs. We’ve helped in that regard,we have aneducation role,wecanguideentrepreneurs, wepush themonfirst-class global standards, let them know what is expected. In the late 1980s, the Programme in Advanced Technologies (PAT) was established, including software. The software PAT consisted of three centers initially. Perhaps the center with the most impact on the software industry itself has been the Centre for Software Engineering (CSE), located at Dublin City University. Although the initial PAT proposal called for a certain amount of research within the PATs, the CSE’s role has been mainly promotional and informational. In particular, it has promoted quality in software production and the ISO 9000 quality standard and has had a major impact on the software industry through its courses and various forms of information dissemination.61 In this it was no doubt helped byitsgoodrelationshipwithDublinCityUniversity,whereitislocatedandwhere ithadactuallybeenfoundedin1989beforebeingbroughtintothesoftwarePATin 1990.Furthermore,there arecloserelationswithindustryastheCSEboard spans academia, industry, and state agencies. The other two PAT centers—Multimedia Technologies Ireland and the National Institute for Management Technology— ran into more difficulties. Both of these centers were more focused on smaller constituencies, the small multimedia industry, and executives in business outside of high tech, respectively, and this made their relation to the software PAT somewhat problematic. Within these constituencies, however, they were reasonably wellevaluated.62 Ingeneral,infact,mostofthePATschemeshavereceivedfavorableevaluations,nomatterwhatsectortheyarein.Nonetheless,bothofthesecenters were closed or sold in a reorganization of the PAT scheme in 1997. Thestatehasalsobeeninstrumentalinestablishingindustryassociations. One such innovation center and industry association is the Localization Resources Centre(LRC),establishedatUniversityCollegeDublin(UCD)underthetechnologycentersscheme.The LRCpromotes qualityandresourcesharingin thelocalization sector of the software industry and has been instrumental in setting up the successfulindustryforum,theSoftwareLocalizationInterestGroup(SLIG).This is not the only industry association set up by a state agency—the Multimedia TechnologyInstitute (MTI) helped to start the Irish InteractiveMultimediaAssociation, which persists after the demise of MTI, while the NSD has a close rela176 POLITICS & SOCIETY tionshipwiththeIrishSoftwareAssociation(althoughthatbodyhas beeninexistence since the 1970s). Another agency funded by the EU but playing a somewhat different role is the NSD.TheNSDwasestablishedwithintheIDA,althoughitmovedtobecomepart ofForbairtwhentheagencieswere reorganizedin1994. Its role,whilesomewhat unclear at first, has evolved so that it has become a central focal point within the state agencies for dealings with software. Although the everyday work of grant assessment is carried out by the International Services Programme within Forbairt,the NSDisrepresented ongrant-givingcommittees.Italso fulfillsmany of the educational and networking roles discussed above in relation to the state agencies.However,theNSDisprimarilyastrategiclinkbetweentheindustryand the state development agencies. Furthermore, the NSD constantly monitors and reviews the key issues for the software industry and the operation of the institutionsandinfrastructuresupportingtheindustry.Ithasbeenaprimemoverinrefocusing the activities of the PATs and in the debates over a growing pressure on skills and the issue of access to capital. There is therefore a network of institutions providing technical and business information of various kinds and degrees of specialization to different sectors within the software industry. These institutions provide a way in which the state can provide an everyday impetus toward world-class technical, business, and qualitystandardswithout involvingitself directlyin regulatingthefirms—ameasurethat wouldbe deeply unpopular.The staffofthesecenters become the everydayteachersandadvisersoftheindustry.However,theyalsogatheragreatdealof up-to-the-minute information from the industry and can be a valuable source of information for the central state agencies when they need to make rapid shifts or adjustments in policy. They are largely staffed by people with backgrounds in industry themselves who sometimes maintain that involvement on a part-time basis.63 Various state agencies have played critical roles in developing the Irish softwareindustry.IDAIrelandhasbeenveryeffectiveinattractingforeigninvestment andsomewhatsuccessfulinpromotingtheupgradingofTNCactivitiesanddevelopmentof subsuppliers. Other state agencies have played key roles in developing a high value-added indigenous software sector. In particular, state actions have beenabletosteerIrishtechnicalprofessionalsawayfromthetemptationsofTNC employment or emigration and into the perils, and potentially lucrative rewards, of small-firm entrepreneurship. The state has also been able to shape the capacities and culture of the industry through its emphasis on promoting innovation, business development, and networking across the industry. The two modes of integration into the global economy of the Irish software industry are only tangentially linked to one another. However, the state has been criticaltotheemergenceofeachglobalizationofIrishsoftware.Howwasitpossible for the state to play this role for each of these globalizations, and how has it SEÁN Ó RIAIN 177 been possible for them to be pursued simultaneously without one overwhelming the other? Here we turn to our understanding of the multiple embedded autonomieswithintheIrishstateandtheflexibleorganizationalstructurethatreconciles them within the state itself. Embedded Autonomy through Flexibility The state agency most closely linked to TNCs is the IDA, restructured as IDA Ireland since 1994. The IDA was founded in 1949, took on a more central role in attractingforeigninvestmentintheearly1960s,andbythelate1960shadbecome the central agency in formulating and implementing industrial policy. IthasdevelopedclosetieswithmanyTNCs—abroadandinIreland.Theseties are first formed through the IDA’s international offices. The project officers in thesefourEuropean,sixU.S.,andfiveAsia-Pacificofficesestablishanearlyrelationshipwithmanycompaniesintheirareaandultimatelyhandthecompaniesoff to their counterparts in Ireland. The IDA also plays a critical role in providing for these firms’needs once they are established in Ireland, providing a “one-stop shop” of sorts for them within the Irish political economy. As one longtime TNC managersaid,“We’re a little cut offoutherebutwehavea good relationship with theIDAofficers.IcanalwaysringKieran[McGowan—thenCEOofIDAIreland] anyway.”TheseclosetiestolocalmanagersofTNCsareimportanttowhatefforts are made to pressure the TNC subsidiaries to continue to make efforts to upgrade their operations. Often these local and international ties are combined to good effect. For example, ties to a TNC head office might be useful in developing a closerrelationshipbetweenanIrishsupplierfirmandtheparentfirmintheUnited States—as one IDA officer told me: “Sometimes we have to use our contact with theU.S. to get around thelocalTNCmanagerwho’stryingtoget all the credit for himself and won’t tell head office about this supplier in Ireland.” The IDA is a quasi-independent state agency, although it receives its funding from, and must report to, the Department of Enterprise, Trade and Employment and the cabinet minister in that department. In 1969, the IDA became an agency outside of the civil service structure and it has become increasingly independent, untilrecent years, of the rest ofthestate economic development regime.Until the early1990s,mostoftheinitiativesinindustrialpolicyweredevelopedthroughthe IDA. Although its board membership changes, its executive leadership has been very stable and the IDA has developed a strong corporate identity of its own. Its placeoutside theconventionalcivilservicegives it aflexibilityininternal organization and a freedom of maneuvering that is unusual for state agencies. With strong internal promotion patterns and little mobility out of the IDA, it forms a verycohesiveunitwithin theIrishstate with astrongfocus onitsgiven objectives of attracting FDI. It has also been able to build up significant skills internally in terms of knowledge of international markets and the politics and economics of industrial location. The IDA is embedded in the TNCs and autonomous from the 178 POLITICS & SOCIETY rest of the state, indicating the difficulties of attempting to shape the actions of TNCs. The embedded autonomy of the state agencies promoting the indigenous industryis quitedifferent,andthese agencieshavealwayshad adifficultrelationshipwiththeIDAanditsdominantpositionwithinthestate.TheIDA’sautonomy within the Irish state institutionalized an industrial policy focus on attracting foreign investment. Such institutionalization appears to be a crucial element of the effectiveness in the medium to long term of any policy agenda, as it allows the accumulation of resources, skills, and political legitimacy and alliances. However, this also means that an existing agenda is likely to “crowd out” other agendas, unless they themselves can create “safe” institutional spaces where they can avoid an attack from the existing regime. Such spaces did not emerge for the promotion of indigenous industry until the 1980s. Nonetheless, there had always been elements within the Irish state that had concerned themselves with indigenous development. Within the IDA itself, programswerepursuedinthe1970srelatingtoindigenousindustry,althoughattracting FDI dominated the organizational objectives, culture, and skills of the IDA. There were, however, a series of other state bodies that carved out some space around the agenda of promoting science, technology, and innovation and that were largely oriented toward indigenous industry. These state institutions relied heavily on EU funds for their activities, and many of their programs were funded in large part by European Structural Funds. These funds were designated for Ireland as part of an effort to develop the peripheral regions of the EU in the face of the upcoming Single European Market in 1992. Many commentators have attributedIreland’sgrowthtothisinfluxofEUcapital.However,thesignificanceofthe Structural Funds was that they were the means by which a variety of new, sometimesexperimental,measurescouldbetakenwithouthavingtofighttherestofthe stateagenciesforfunding.Thenewdevelopmentregimecoulddevelopalongside theoldanddidnothavetochallengetheolddevelopmentmodeldirectlyforfunds and priority, except in rare cases. Such are the opportunities within a “networked polity”64 forlocal and nationalandtransnational actors for forgingnewcoalitions that bypass existing national power structures—even, as in cases such as this, where the ultimate goal is national economic transformation. Agenciessuch astheNational Board forScienceand Technology(NBST)and the NSD were embedded in the emerging technical professional class through personal and associational networks. Key individuals pursued careers that spanned private, public, and semipublic organizations, making the boundaries betweenthese differentspheres porous and fostering increasedinteraction across thoseorganizationalboundaries.This,inturn, improvedthe agencies’capacity to make and implement effective policy. The first director of the NSD,65 Barry Murphy, had been the managing director of Insight Software—an Irish company sold to Hoskyns from the United Kingdom in 1988. Director of the NSD until SEÁN Ó RIAIN 179 1996,he moved on to work with theCullinaneGroupIreland(CGI), an organization founded by U.S. software industry veteran John Cullinane to invest in Irish softwarecompaniesbutthatalsocarriesoutresearchonthesoftwareindustryand boosts Irish high-tech. Murphy was also involved in writing the strategy document published by the Irish Software Association in March 1998. The second directorofNSD,JenniferCondon,camefromtheindustrytoworkonthemarketing side of the National Software Centre (a state-sponsored center supporting industry development and carrying out research) in the mid-1980s. She was then managing director of ICL’s IT Centre in Dublin from 1988 until 1996, when she joined the NSD. The post of director of the NSD is itself a temporary one that ensures that the directorultimatelyreturnstotheindustryinsomeguise.ThestaffoftheNSDconsist of the director and three other staff members—two of whom are from within thestatedevelopmentagenciesandathirdwithabackgroundintheindustry.This also ensures a combination of embeddedness in the industry and a certain autonomy of focus and objectives. There also is a group of prominent software managerswhohavebeeninvolvedinpolicyinitiativeswithintheindustry,eitherthrough theindustrytradeassociation,theIrishSoftwareAssociation,orasparticipantsin policy consultation groups such as the SoftwareIndustryWorkingGroupin1989 ortheSoftwareConsultativeCommitteeatthepresenttime.Furthermore,thereis a significant overlap with the educational sector, as prominent computer science and engineering academics have liaised closely with industry and the state in a variety of forums. Autonomy from their business constituency is maintained by the performance requirements placed on the various agencies, centers, and programs. These requirementsaregivenacuttingedgebytheregularevaluationofprograms—and publication of those evaluations.66 Furthermore, poor evaluations can result in closure,sale,orreorganizationofsuchprograms.Forexample,aspartofthisprocess of constant shifting and experimentation with state and semistate agencies, the National Institute of Management Technology was sold, MTI restructured, and the Centre for Software Engineering given new responsibilities. This can occuratthelevelofreorganizationoftheagenciesthemselves.Theseincludedthe NationalScienceCouncil,foundedinthe late 1960s, and its successor, the NBST founded in 1977. In the mid-1980s, science and technology policies were reorganized significantly with a new Office of Science and Technology being formed under a minister of state (a junior cabinet minister) in what was then the Department of Industry and Commerce. The NBST was merged with the Institute for IndustrialResearchand Standards in 1987 to formEolas,a new science and technology agency. Around the same time, various agencies were consolidated into one agency to deal with active labor market policy (FÁS) and into another to deal with export marketing (An Bord Tráchtála). In 1994, Forbairt, an agency focused solely on indigenous industry, was created incorporating most of the staff and 180 POLITICS & SOCIETY functions of Eolas and the Irish Industry section of the IDA. In 1998, An Bord Tráchtála was merged into Forbairt, although under some protest from local firms,whichfearedalossofmarketingassistance.Finally,thesevariousagencies werefolded intoEnterpriseIreland,which consolidated theseagenciesunder one institutional roof. This experimentation can also operate within an existing agency that changes its focus over time. Due to the requirements of the EU funders, centers such as LRC, CSE, and NSD are subject to significant and detailed external evaluation. They are also ultimately subject to the control of the state and will therefore also reflect broad policy concerns. For example, the CSE has been setting targets for theuptakeofquality procedureswithin theindustry asa whole.In 1997,the CSE, at the request of the NSD, instigated programs aimed at diffusing quality proceduresintosmallsoftwarefirms,basedonitspriorsuccessatdiffusingsuchprocedures among the larger firms. This loosely coupled and flexible organizational structure has not emerged without ongoing tensions. In particular, the IDA was peculiarly powerful within theIrishstate,andthishelpstoexplaintheweaknessoftheinstitutionssupporting indigenous industry until the late 1980s. It took the massive social and economic crisis of the 1980s to delegitimate the IDA’s role as the sole bearer of the task of Irish industrial transformation. It was into this restricted institutional space that the alliance of Irish technical professionals and the previously marginalized “scienceandtechnology”stateagenciessteppedtosupportindigenousindustry,inthe process creating the delicate compromise at the heart of the FDS. Inequality, Fragmentation, and Sustainability Despite the relatively successful adaptation of state structures to the multiple globalizationsof Irish industry,andsoftwarein particular, thisisnotastorywithout its own tensions and darker sides. These globalizations have occurred within the context of a macroeconomic stabilization secured since 1987 by a national neocorporatist “social partnership” compromise. Rising inequality has created significant tensions between the institutionalized globalizations of software (and similar industries) and the institutions of national neocorporatism. Wage inequality has increased drastically in Ireland since the late 1980s, at whichpointitwasalreadyoneofthemostunequalOECDnationsinanycase.The major component of that increase in inequality has been a growing gap between themiddleandthetopoftheincomedistribution.Whiletheaverageincomeofthe top 10 percent was 195 percent of the median income in 1987, by 1994, that percentage had grown to 224 percent. Although the supply of skilled labor has increased rapidly, it has still been outstripped by demand, and increasing returns toeducationaccountforasizeableproportionofthisgrowthinwageinequality.67 The Irish case is clearly an exception to the tendency internationally for neocorporatism to lessen wage inequality. This is explained, as far as can be told SEÁN Ó RIAIN 181 tion of national economies, limiting the destructive tendencies toward regulatory competition. Any such political project that does emerge is increasingly likely to wage its battles on the terrain currently staked out by the FDS. NOTES 1. See Chalmers Johnson, MITI and the Japanese Miracle (Stanford, CA: Stanford UniversityPress,1982);AliceH.Amsden,Asia’sNextGiant:SouthKoreaandLateIndustrialization (Oxford, UK: Oxford University Press, 1989); Robert Wade, Governing the Market (Princeton, NJ: Princeton University Press, 1990); Stephen Haggard, Pathways from the Periphery: The Politics of Growth in the Newly Industrializing Countries (Ithaca, NY: Cornell University Press, 1990); Richard Applebaum and Jeffrey Henderson, eds., States and Development in the Asian Pacific Rim (London: Sage, 1992); Peter Evans, Embedded Autonomy (Princeton, NJ: Princeton University Press, 1995); Meredith Woo-Cumings,ed.,TheDevelopmentalState(Ithaca,NY:CornellUniversityPress,1999). 2. See Jason Dedrick and Ken Kraemer, Asia’s Computer Challenge: Threat or Opportunity for the United States and the World? (Oxford, UK: Oxford University Press, 1998) regarding competitive difficulties. For more on the debt crisis, see Nicole Biggart, “DeepFinance—TheOrganizationalBases ofSouth Korea’sFinancialCollapse,”Journal ofManagementInquiry7(1998):311-320;andaseriesofarticlesbyRobertWade,including“TheComingFightoverCapitalFlows,”ForeignPolicy113(1998):41-54;“TheAsian Debt-and-DevelopmentCrisisof1997-?:CausesandConsequences,”WorldDevelopment 26 (1998): 1535-53; “From ‘Miracle’to ‘Cronyism’: Explaining the Great Asian Slump,” Cambridge Journal of Economics 22 (1998); “The Asian Crisis and the Global Economy: Causes, Consequences, and Cure,” Current History 97 (1998): 361-73; Robert Wade and FrankVeneroso,“TheAsianCrisis:TheHighDebtModelversustheWallStreet-Treasury-IMF Complex,” New Left Review 228 (1998): 3-24. 3. Ireland refers here to the twenty-six counties of the Republic of Ireland. 4. Paul Sweeney, The Celtic Tiger (Dublin: Oak Tree Press, 1998). 5.ThisarticledrawsheavilyontheanalysisinSeánÓRiain,“AnOffshoreSiliconValley?”CompetitionandChange2(1997):175-212;SeánÓRiain,“RemakingtheDevelopmental State: The Irish Software Industry in the Global Economy” (Ph.D. diss., Department of Sociology, University of California, Berkeley, 1999). 6. McIver Consulting, Manpower, Education and Training Study of the Software Sector (Dublin: Forás Áiseanna Saothair, 1998). 7.CharlesSabel,Ireland:LocalPartnershipsandSocialInnovation(Paris:Organization for Economic Cooperation and Development, 1996). 8. Denis O’Hearn, Inside the Celtic Tiger: The Irish Economy and the Asian Model (London: Pluto, 1998), 153. 9. Paul Krugman, “Good News from Ireland: A Geographical Perspective,” in Alan Gray, ed., International Perspectives on the Irish Economy (Dublin: Indecon, 1998), 51. See also chapters by Jeffrey Sachs and Kenneth Arrow in the same volume. 10. See Ó Riain, “An Offshore Silicon Valley?” for an account of how local and global processesshapedeachotherovertimetocreateeachofthesetwoglobalizationsofthesoftware industry. 11. Forfás, Survey of Employment (Dublin: Forfás, 1998). 12.O’Hearn,InsidetheCelticTiger;DenisO’Hearn,“Globalization,‘NewTigers’and the End of the Developmental State? The Case of the Celtic Tiger,” Politics & Society 28 (2000). 188 POLITICS & SOCIETY 13. Patrick Honohan, Bertrand Maitre, and Charles Conroy, “Invisible Entrepôt Activity in Irish Manufacturing,” Irish Banking Review, summer 1998. 14. Forfás, Research and Development in the Business Sector: Findings from the 1993 Census of R&D Performing Enterprises in Ireland (Dublin: Forfás, 1995); Forfás, Survey of Product and Process Innovation in Irish Industry 1993-1995 (Dublin: Forfás, 1997). 15.JimStewart,“TransferPricing:SomeEmpiricalEvidencefromIreland,”Journalof Economic Studies 16 (1989): 40-56. 16. Central Statistics Office, Census of Industrial Production 1990 (Dublin: Central Statistics Office, 1991); Central Statistics Office, Census of Industrial Production 1996 (Cork, Ireland: Central Statistics Office, 1997). 17. Ibid. 18. Forfás, Survey of Product and Process Innovation, 30-32. 19. National Economic and Social Council, Sustaining Competitive Advantage: Proceedings of NESC Seminar, Research Series no. 4 (Dublin: National Economic and Social Council, March 1998). 20.JohnGeary, “TheNewWorkplace: Changeat WorkinIreland,”InternationalJournal of Human Resource Management 10 (1999): 879; William K. Roche and John Geary, “ ‘Collaborative Production’and the Irish Boom: Work Organization, Partnership and Direct Involvement in Irish Workplaces,” Economic and Social Review (2000), forthcoming. 21. National Economic and Social Council, Private Sector Investment in Ireland, Reportno.103(Dublin:NationalEconomicandSocialCouncil,February1998),chap.2. 22. On investment, see National Economic and Social Council, Private Sector Investment; on research and development (R&D) and indigenous industry more generally, see Mary O’Sullivan, “The Sustainability of Industrial Development in Ireland” (paper presented to Dublin Economics Workshop, Dublin, April 1999); on sectoral and other factors affecting collaborative production, see Geary, “The New Workplace,” 879-86; on clustering, see National Economic and Social Council, Sustaining Competitive Advantage. 23. Figures on occupational change from Paul Tansey, Ireland at Work: Economic Growth and the Labour Market 1987-1997 (Dublin: Oak Tree Press, 1998), 41. 24. O’Hearn, Inside the Celtic Tiger, 125. 25. Alan Barrett, Tim Callan, and Brian Nolan, “Rising Wage Inequality, Returns to Education and Labour Market institutions: Evidence from Ireland,” British Journal of Industrial Relations 37 (1999): 77-100. 26. O’Hearn, Inside the Celtic Tiger, 126. 27. Tim Callan et al., Poverty in the 1990s: Evidence from the Living in Ireland Survey (Dublin: Oak Tree Press, 1996). 28. Evans, Embedded Autonomy. 29. Amsden, Asia’s Next Giant. 30.Amsden,Asia’sNextGiant; FredDeyo,BeneaththeMiracle:Labor Subordination in the New Asian Industrialism (Berkeley: University of California Press, 1989). 31. Evans, Embedded Autonomy; see Peter Evans and James Rauch, “Bureaucracy and Growth: A Cross-National Analysis of the Effects of ‘Weberian’States Structures on Economic Growth,” American Sociological Review 64 (1999): 748-765 for cross-national quantitative evidence regarding the beneficial growth effects of bureaucracy. 32. Evans, Embedded Autonomy; Sylvia Maxfield and Ben Schneider, eds., Business and the State in Developing Countries (Ithaca, NY: Cornell University Press, 1997). 33.FordifferencesamongtheAsianTigereconomies,seeMarcoOrrù,NicoleBiggart, and Gary Hamilton, The Economic Organization of East Asian Capitalism (Thousand Oaks, CA: Sage, 1997). The Korean model has the most authoritarian state and corporate SEÁN Ó RIAIN 189 structures,Japancombines“corporatismwithoutlabor”with“enterpriseunionism,”while Taiwan has a much less hierarchical political and corporate structure. 34. Erik Olin Wright, “Review of P. Evans ‘Embedded Autonomy,’” Contemporary Sociology 25 (1996): 176-77. 35.ForcritiquesofPeterEvans’sinattentiontothestate’sroleinconstitutingeconomic actors as well as guiding their actions, see David Stark and Laszlo Bruszt, Postsocialist Pathways: Transforming Politics and Property in East Central Europe (Cambridge, UK: Cambridge University Press, 1998) and Nicole Biggart and Mauro Guillen, “Developing Difference: Social Organization and the Rise of the Auto Industries of South Korea, Taiwan, Spain, and Argentina,” American Sociological Review 64 (1999): 722-47. 36. The position of prime minister. 37. Chris Ansell, “The Networked Polity: Regional Development in Western Europe” (DepartmentofPoliticalScience,UniversityofCalifornia,Berkeley,1999,mimeographed). 38.Neil Brenner, “GlobalCities,Glocal States: GlobalCityFormationand StateTerritorialRestructuringinContemporaryEurope,”Reviewof International Political Economy 5 (1998): 1-37; “Beyond State-Centrism? Space, Territoriality and Geographical Scale in Globalization Studies,” Theory and Society 28 (1999). 39.Onloosecoupling,seeJ.DouglasOrtonandKarlE.Weick,“LooselyCoupledSystems: A Reconceptualization,” Academy of Management Review 15 (1990): 203-23; Charles Perrow, Complex Organizations (New York: Random House, 1986), chap. 4. 40. On modes of control in decentralized organizational forms, see Wolf V. Heydebrand, “New Organizational Forms,” Work and Occupations 16 (1989): 323-57. 41. On autocentric development, see Samir Amin, Unequal Development (New York: Monthly Review Press, 1976); Dieter Senghaas, The European Experience: A Historical Critique of Development Theory (Dover, NH: Berg, 1985). 42.SeánÓRiain,“AnOffshoreSiliconValley?”;NeilCoe,“USTransnationalsandthe Irish Software Industry,” European Urban and Regional Studies 4 (1997): 211-30. 43. Localization of software refers to the process of customizing existing software packagesforspecificnationalandlinguisticmarkets.Themainactivityisthetranslationof the text, but it may also involve changing date formats, letter formats, and other culture-specific aspects of the software. In U.S. software companies, the work of designing thesoftwareprogramso thatitcan be customizedin this wayis calledinternationalization andisgenerallycarriedoutin theUnited States.The workofactuallycustomizingtheprogramforspecificmarketsiscalledlocalization.Thisworkisrelativelyuncomplicated.See Honohan, Maitre, and Conroy, “Invisible Entrepôt Activity,” for an analysis of industrial statistics relating to these activities. 44. Ó Riain, Remaking the Developmental State, 81. 45. The Israeli software industry was estimated to earn approximately $700 million in exports in 1997 (including indigenous and transnational corporation [TNC] firms), while the value of Indian software industry exports rose to $1,750 in 1997-98. Exports of Irish indigenous firms came to $713 million and those of TNCs to (an inflated figure of) $5,615 million.Careshouldbetakeninusingthesefigurestocomparethethreeindustriesasthere area number ofdifferencesin howthefigureswere calculated. However,it is clearthatthe Irish industry is of a similar order of magnitude as the industries in India and Israel. The Irish industry’s export orientation is similar to that of the Indian industry, although the Indian industry’s exports consist largely of offshore development and international labor contracting. See Balaji Parthasarathy, “Industrializing Countries and the Changing International Division of Labor in the Computer Software Industry: Lessons from the Indian Case 1977-1997” (paper presented to the Sloan Foundation Globalization Workshop, DukeUniversity,Durham,NC,26-28April1998).TheIrishindustryisclosertotheIsraeli 190 POLITICS & SOCIETY and indeed the U.S. industry in its greater focus on software products and avoidance of large-scale labor contracting. Furthermore, these are the leading international software industries outside the G-7 (the leading capitalist economies, which meet as a group on a regular basis). The other leading contenders such as Taiwan, Singapore, and Australia are wellbehindthe“threeI’s”ofglobalsoftware.SeeDedrickandKraemer, Asia’sComputer Challenge. Other emerging industries such as Hungary and China that are likely to prove significant in the long term are still relatively underdeveloped. 46.Thedatainthisarticlecomefrom120interviewscarriedoutinSiliconValleyandin IrelandbetweenOctober1995andSeptember1997,industryjournalsandotherdocumentary sources, and a survey of software firms in Ireland carried out in June and July 1997. The response rate to the survey was an above-average 54.5 percent, the sample was reweighted to reflect the size and ownership of the survey population, and six companies were removed from the sample because they received more than 60 percent of their revenues from hardware. 47. National Software Directorate, 1995 Software Industry Survey Results (Dublin: Forbairt, 1995). 48. Gary Gereffi, “The International Economy,” in Neil Smelser and Richard Swedberg, eds., The Handbook of Economic Sociology (Princeton, NJ: Princeton University Press/Russell Sage Foundation, 1994); Bennett Harrison, Lean and Mean (New York: Basic Books, 1994). 49.DavidJacobsonandDavidO’Sullivan,“AnalysinganIndustryinChange:TheIrish Software Manual Printing Industry,” New Technology, Work and Employment 9 (1994): 103-14; “The Irish Software Manual Industry: Globalization through Local Supply?” (paper presented at “Ireland, Europe & the Global Information Society: A Conference for Social Scientists,” Dublin, 24-25 April 1997); Industrial Development Authority, “The Irish Software Industry,” information pamphlet, n.d. 50. Union density fell from 48.3 percent in 1992 to 43.5 percent in 1997. The only sector that increased union density is Public Administration and Defense, with a union density of 74.1 percent in 1997. By contrast, Commerce, Insurance, Finance and Business Services (including software) fell from 31.6 percent to 29.0 percent in 1997. Data generously provided by William K. Roche, Graduate School of Business, University College, Dublin. 51.PatrickClancy,“TheEvolutionofPolicyinThird-LevelEducation,”inD.Mulcahy and D. O’Sullivan, eds., Irish Educational Policy: Process and Substance (Dublin: InstituteofPublicAdministration,1989);RobertOsborne,HigherEducationinIreland:North and South (London: Jessica Kingsley, 1996). 52.Evans,EmbeddedAutonomy;seeO’Hearn,“Globalization,NewTigersandtheEnd of the Developmental State” for application of the concept of the state as “midwife” to attraction of foreign investment. 53. Jacobson and O’Sullivan, “Analysing an Industry in Change.” 54. Colm O’Gorman, Eoin O’Malley, and John Mooney, Clusters in Ireland: The Irish Indigenous Software Industry: An Application of Porter’s Cluster Analysis (Dublin: National Economic and Social Council, 1997), 35. 55. National Software Directorate, The Software Industry in Ireland: A Strategic Review (Dublin: IDA Ireland, 1992). 56. Eoin O’Malley, “Developments in Irish Industrial Policy since the Mid-1980s” (paper presented at the special conference of the Political Studies Association of Ireland “TheStateof the Irish Political System,”Cork, Ireland, 28-30 May 1992); Eoin O’Malley, KieranA.Kennedy,andRoryO’Donnell,ReporttotheIndustrialPolicyReviewGroupon the Impact of the Industrial Development Agencies (Dublin: Stationery Office, 1992). SEÁN Ó RIAIN 191 57. Ann Clarke, Software Support Programme: Final Report (Dublin: EU Structural Funds Operational Programme for Industrial Development Evaluation Unit, 1995). These arethelatestyearsforwhichfiguresareavailable,although totalstateexpenditureonbusiness development has continued to increase. 58.Onsoftware,seeClarke,SoftwareSupport Programme. Onallfirms,seeO’Malley, “DevelopmentsinIrishIndustrialPolicy,”andO’Malley,Kennedy,andO’Donnell,Report to the Industrial Policy Review Group. 59. Clarke, Software Support Programme; O’Gorman, O’Malley, and Mooney, Clusters in Ireland. 60. Patricia Carr, “‘Riding the Juggernaut’: Selectivity and Entrepreneurship in Ireland,” Irish Journal of Sociology 5 (1995): 67-88; “The Cultural Production of Enterprise: Understanding Selectivity as Cultural Policy,” Economic and Social Review 29 (1998): 133-55. 61. Clarke, Software Support Programme, chap. 7. 62. Ibid., chaps. 8-9. 63.Foramorecompleteanalysisoftheseandotherassociations,agencies,andforums, see Ó Riain, Remaking the Developmental State, chap. 4. 64. Ansell, “The Networked Polity.” 65.Thehead of the National Software Directorate, chargedwith developingpolicyfor, and promoting, the software industry in Ireland, particularly indigenous firms. 66. See, for example, Clarke, Software Support Programme; give Forfás references. 67.AnthonyAtkinson, Lee Rainwater,andTimothySmeeding, Income Distributionin OECD Countries: Evidence from the Luxembourg Income Study (Paris: Organization for Economic Cooperation and Development, 1995); Barrett, Callan, and Nolan, “Rising Wage Inequality.” 68. John Geary, “Initial Results from the National Survey of Employee Relations and HumanResourcePracticeinIreland,”workingpaper,SmurfitBusinessSchool,University College Dublin, 1998. 69. Summary quote from Barrett, Callan, and Nolan, “Rising Wage Inequality,” 95. Neither have recent changes in tax and social welfare done anything to mitigate these inequalities.Although redistributionwassignificant between 1987 and 1994, tax and welfarechangesfrom1995to1998resultedinthepoorest30percentofthepopulationmaking 2 percent less than if their income had increased at the same rate as the average increase in incomefor the population. During the same period,the top 10 percent of earners benefited fromtaxchangestotheextentthattheymade4percentmorethanwouldhavebeenthecase if they had simply followed the average income increase. For the 1987-94 period, see Tim Callan and Brian Nolan, “Income Inequality in Ireland in the 1980s and 1990s,” in Frank Barry, ed., Understanding Ireland’s Economic Growth (London: Macmillan, 1999), 167-92. For the 1995-98 period, see Tim Callan, Brian Nolan, and John Walsh, “Income TaxandSocialWelfarePolicy,”inT.Callan, D.Duffy,T.Fahey,B.Feeney,B.Nolan, P. O’Connell, S. Scott, and J. Walsh, eds., Budget Perspectives (Dublin: Economic and Social Research Institute/Oak Tree Press, 1998). 70.SeánÓRiain,“Time-SpaceIntensification:IrishSoftwareDevelopersintheGlobal Workplace,”inM.Burawoy,J.Blum, S.George,Z.Gille, T. Gowan,L.Haney, M.Klawiter, S. Lopez, S. Ó Riain, and M. Thayer, Global Ethnography (Berkeley: University of California Press, forthcoming). 71. Pat O’Connor, Emerging Voices: Women in Contemporary Irish Society (Dublin: Institute of Public Administration, 1998). 72.ProportionscalculatedonbasisofdatainCentralStatisticsOffice,HouseholdBudget Survey 1987 (Dublin: Central Statistics Office, 1988), 44-45, and Central Statistics 192 POLITICS & SOCIETY Office, Household Budget Survey 1994 (Cork, Ireland: Central Statistics Office, 1997), 46-47. 73.Thiscanonlybeaggravatedbythegrowinginequalitybetweentheshareofnational income going to capital and to labor and the ability of owners of capital to shelter that income from tax through international financial instruments, as partially revealed in a series of recent financial and tax scandals. 74.Fora detailed discussionof these issues,seeÓ Riain,Remakingthe Developmental State, chap. 8. On trends toward fragmentation in industrial relations, see William K. Roche, “Between Regime Fragmentation and Realignment: Irish Industrial Relations in the 1990s,” Industrial Relations Journal 29 (1998): 112-25. 75. World Bank, The East Asian Miracle (Washington, DC: World Bank, 1993); O’Hearn, “Globalization, New Tigers and the End of the Developmental State?” 76. O’Hearn, “Globalization, New Tigers and the End of the Developmental State?” 77. Ansell, “The Networked Polity.” 78. On Dutch regional policy, see Brenner, “Global Cities, Glocal States.” On neocorporatism in the Netherlands, see Jelle Visser and Anton Hemerijck, A Dutch Miracle? (Amsterdam: Amsterdam University Press, 1997). 79. Giovanni Arrighi, The Long Twentieth Century (London: Verso, 1994). 80. On local state corporatism, see Jean Oi, Rural China Takes Off (Berkeley: University of California Press, 1999). On the central and local states in China, see Michael Burawoy, “The State and Economic Involution: Russia through a China Lens,” in Peter Evans, ed., State-Society Synergy: Government and Social Capital in Development (Berkeley: International and Area Studies, University of California, Berkeley, 1997), 150-77. 81. On interfirm networks, see Dedrick and Kraemer, Asia’s Computer Challenge.On transnationalcommunities,seeAnnaLeeSaxenian,SiliconValley’sNewImmigrantEntrepreneurs (San Francisco: Public Policy Institute of California, 1999). On corporate finance, see Biggart, “Deep Finance.” 82. Philip McMichael, Development and Social Change: A Global Perspective (Thousand Oaks, CA: Pine Forge Press, 1996). 83. Philip Cerny, “Globalization and the Changing Logic of Collective Action,” International Organization 49 (1995): 595-625. 84.FritzScharpf,GoverninginEurope(Oxford,UK:OxfordUniversityPress,1999). SEÁN Ó RIAIN 193