Global and Competitive Strategy

Regional Studies. Vol. 38.9, pp. 1085-1100, December 2004 | ^ Carfax Publishing* ^'^ milW Taylor 6. Francis Croup

Globalization and Competitive Strategy inEurope's Vulnerable Regions: Firm, Industry and

Country Effects in Labour-intensive Industries

GORDON L. CLARK*, THEO PALASKASf, PAUL TRACEYt andMARIA TSAMPRAf

* School of Geography & the Environment and Said Business School, University of Oxford, Mansfield Road,Oxford OXl 3TB, UK. Email: [email protected]

"fDepartment of Economics & Regional Development, Panteion University of Athens, 136 Syngrou Avenue, GR-i7671Athens, Greece. Emails: [email protected], [email protected]

^ fudge Institute of Management, University of Cambridge, Trumpington Street, Cambridge CB2 iAG, UK.Email: [email protected]

(Received November 2003: in revised form May 2004)

CLARK G. L., PALASKAS T , TRACEY P. and TSAMPRA M . (2004) Globalization and competitive strategy in Europe's

vulnerable regions: firm, industry and country effects in labour-intensive industries, Regional Studies 38, tO77—1092. Europeanintegration bas prompted great interest in tbe adjustment capacities of smail and medium enterprises in regions cbaracterizedby bigh levels of unemployment and lower-tban-average incomes. At issue, in tbese circumstances, is tbe extent to wbicbeconomic competitiveness can be enbanced by tecbnological cbange and tbe resources of European and national governments.Relying upon tbe results of a detailed survey of small- and medium-sized firm competitive strategies in selected vulnerableregions across Europe, tbe paper focuses upon firm adjustment strategies in four labour-intensive industries vulnerable tointernational competition. It draws togetber tbe results of tbese surveys providing econometric and statistical analyses tbatdemonstrate tbe commonalities and differences apparent in small- and medium-sized firms' responses to cbanging marketcompetition. Significant insigbts were gleaned from tbe pooling of tbese data. It is sbown tbat tbere are statistically significantfirm, country and industry effects in competitive strategies. Unlike many otber related studies, tbe derived results are quiteconsistent across Europe. Furtbermore, statistically significant relationsbips are establisbed between cbanges in market sales, localemployment and tbe adoption of process-specific technologies. Tbe findings provide robust conclusions about tbe significanceof country and industry determinants of European small and medium enterprises' competitive strategies in relation to tbeexpanding European and global economy. In sum, tbe paper raises doubts about tbe value of regional development strategiestbat rely exclusively upon clusters and geographical embeddedness in tbe face of globalization.

Competitive strategy Firm Industry Country effects

CLARK G. L., PALASKAS T , TRACEY P. et TSAMPRA M . (2004) La mondialisation et la strategie competitive dans les regions

defavorisees d'Europe: les retombees sur l'entreprise, l'industrie et le pays des industries a forte main-d'oeuvre. Regional Studies38, 1077—1092. L'integration europeenne a suscite un interet important a la capacite d'ajustement des petites et moyennesentreprises dans les regions qui se caracterisent par des niveaux de chomage eleves et par des niveaux de revenu qui sontinferieurs a la moyenne. Dans de teUes circonstances, la question qui se pose est la suivante. Jusqu'a quel point peut-on ameliorerla competitivite economique au moyen de revolution tecbnologique et a I'aide des ressources dont disposent l'administrationeuropeenne et nationale? Puisant dans les resultats provenant d'une enquete detaiUee de la strategie competitive des PME situeesdans certaines regions defavorisees a travers l'Europe, l'article porte sur les strategies d'ajustement dans quatre industries a fortemain-d'oeuvre qui sont exposees a la competition international. A partir des resultats de ces enquetes, on fournit des analyseseconometriques et statistiques, qui demontrent les similarites et les differences qui proviennent des reponses des petites etmoyennes entreprises a revolution du marcbe. La mise en commun de ces donnees a facilite l'obtention d'importants aper^us.On montre que les strategies competitives ont des retombees statistiquement importantes sur l'entreprise, le pays et l'industrie.A la difference d'autres etudes liees, les resultats se sont averes relativement bomogenes a travers l'Europe. D'autres rapportsstatistiquement plus importants ont ete etablis entre la variation des ventes, I'emploi local et l'adoption des tecbnologiesspecifiques aux procedes. Ces resultats fournissent de solides conclusions quant a l'importance des determinants nationaux etindustriels des strategies competitives des petites et moyennes entreprises europeennes par rapport au developpement deseconomies europeenne et mondiale. En somme, on doute de la valeur des politiques d'amenagement du territoire qui dependentexclusivement des regroupements et de l'ancrage geograpbique face a la mondialisation.

Strategie competitive Entreprise Industrie Effets nationaux

0034-3404 print/1360-0591 online/04/091085-16 ©2004 Regional Studies Association DOI: 10.1080/0034340042000292656

http://www.regional-studies-assoc.ac.uk

1086 Gordon L. Clark et al.

CLARK G. L., PALASKAS T , TRACEY P. und TSAMPRA M . (2004) Globalisierung und Wettbewerbsstrategie in wirtschaftlichwehrlosen Regionen Europas: Firmen-, Industrie- und Landereffekte in arbeitsintensiven Industrien, Regional Studies 38, 1077—1092. Die europaische Integration hat grosses Interesse an den Anpassungsfahigkeiten Ideiner und niittelgroBer Unternehmen(small and medium enterprises - SMSs) in Regionen geweckt, die durch hohe Erwerbslosigkeit und unterdurchsclinittlichnieddge Einkonimen gekennzeichnet sind. Unter diesen Umstanden erhebt sich die Frage, in welchem AusmaB diewirtschaftliche Konkurrenzfahigkeit durch technologischen Wandel und die Mittel der europaischen und einzelner Landerregie-rungen aufgebessert werden kann. Unter VerlaB auf die Ergebnisse einer ins Einzelne gehenden Untersuchung von Wettbew-erbsstrategien kleiner und mittelgrosser Unternehmen in ausgewahlten, wirtschaftlich schwachen Regionen in ganz Europakonzentrieren die Autoren sich auf Anpassungsstrategien von Firmen in vier arbeitsintensiven Industrien, die als wehrlosgegenliber der intemationalen Konkurrenz angesehen werden. Die Autoren betrachten die Ergebnisse dieser Untersuchungenim Zusanimenhang, wobei okonomische und statistische Analysen Gemeinsanikeiten und Unterschiede aufzeigen, die inAntworten kleiner und mittlerer Firmen auf die sich wandelnde Marktkonkurrenz erscheinen. Die Kombination dieser Datengewahrt wesentliche Einsichten. Es wird aufgezeigt, daB es statistisch signifikante Firmen-, Lander- und Industrieeffekte inWettbewerbsstrategien gibt. lm Unterschied zu vielen anderen, verwandten Untersuchungen stinimten die gewonnenenErgebnisse in ganz Europa uberein. Dartiberhinaus wurden statistisch signifikante Beziehungen zwischen Veranderungen beiMarktsabsatz, Erwerbstatigkeit am Orte und der Einfuhrung verfahrensspezifischer Techniken festgestellt. Diese Befundeverniitteln uberzeugende Schliisse auf die Bedeutung der Lander— und Industriedeterminanten der Wettbewerbsstrategien dereuropaischen kleinen und mittelgrossen Unternehmen im Verhaltnis zu der sich ausweitenden europaischen und globalenWirtschaft. Zusammenfassend werden Zweifel am Werte der regionalen Entwicklungsstrategien laut, die sich angesichts derGlobalisierung ausschlieBlich auf Cluster und geographisches Eingebettetsein verlassen.

Wettbewerbsstrategie Firmen Industrie- und Landereffekte

CLARK G. L., PALASKAS T , TRACEY P. y TSAMPRA M . (2004) Globalizacion y estrategia competitiva en las regiones europeasmas vulnerables: efectos empresariales, industrials y del pais en la industrias de niano de obra intensiva. Regional Studies 38,1077-1092. La integracion europea ha despertado un gran interes en las capacidades de ajuste de las pequeiias y medianasempresas en regiones caracterizadas por altos niveles de desenipleo e ingresos que estan por debajo de la media europea. Entales circunstancias, una de las cuestiones es el grado hasta el cual se puede mejorar la conipetitividad economica por medio delcambio tecnologico y de los recursos de los gobiernos tanto europeos conio nacionales. Contando con los resultados obtenidosde una encuesta detallada de las estrategias competitivas de las PIMEs en una serie de regiones vulnerables seleccionadas a travesde Europa, nos centranios en las estragias de ajuste de las empresas en tres tipos de industrias de mano de obra intensivavulnerables a la competicion intemacional. Reunimos los resultados de estas escuestas y ofrecemos analisis econometricos yestadisticos que demuestran las similitudes y diferencias que son aparentes en las respuestas de las pequenas y medianas empresascon respecto a la cambiante competencia de mercado. Un niimero de percepciones nuevas se ha derivado de estos datos.Mostramos que existen efectos empresariales, industriales y del pais significativos en las estrategias competitivas. A diferencia demuchos otros estudios similares, los resultados obtenidos eran bastante consistentes a traves de Europa. Ademas, se establecieronotras relaciones estadisticas significativas entre los cambios en las ventas al mercado, el empleo local y la adopcion de tecnologiasespecificas de proceso. Estos resultados ofrecen conclusiones robustas en lo que respecta a la significancia de los determinantesindustriales y del pais de las estrategias competitivas de las PIMEs europeas en relacion a la economia global y europea enexpansion. En resumen, planteamos una serie de dudas acerca del valor de las estrategias de desarroUo regional que dependenexclusivaniente de clusters y de una incrustacion geografica ante la globahzacion.

Estrategia competitiva Empressa Efectos industriales y del pais

JEL classifications: LIO, O33, R i l

INTRODUCTION

As the process of market integration reaches into themost remote corners of the European Union (EU) andbeyond, it is increasingly apparent that long-term struc-tural imbalances may be exacerbated by EU expansionand global competition ( H U D S O N , 2002). Not with-standing case studies of the successful 'exemplaryregions' of the Third Italy and Baden-Wurttenburg,Germany, many of Europe's non-metropolitan regionsare dominated by labour-intensive industries and highlevels of unemployment beset by competition fromwithin and without an enlarged Europe. In manycases, these firms are small family or community-based organizations with low levels of capitalization

and employment. Global competition has placed apremium on the adjustment potential of small- andmedium-sized enterprises (SMEs) in these 'vulnerableregions', suggesting a rather bleak map of future Euro-pean regional growth. Even so, those firms and theirregions that can enhance labour productivity throughthe adoption of technology, innovation and productdevelopment may be the winners in global markets,just as those ftrms not so adept may simply contributefurther to existing poor rates of local employmentcreation and out-migration ( M A R T I N , 2001).

The present paper looks closely at the adjustmentcapacity of these types of firms concentrating on SMEsin four labour-intensive industries located in vulnerable

Globalization and Competitive Strategy in Europe's Vulnerable Regions 1087

regions of Europe. Using EU Objectives 1 and 2' statusdesignations, a set of similarly placed regions was identi-fied for five countries including Greece, Italy, Ireland,Spain and the UK (GUERRIERI and IAMMARINO,

2002). As for the industries, these included apparel(18), leather and footwear products (19), electrical andelectronic assembly (30—32), and automotive compo-nents (34). Based on the results of a large multi-countrystudy of the competitive strategies of SMEs in suchsettings, a series of important policy-related questionswas evaluated. For instance, holding the type of regionconstant, can consistent patterns of strategic adjustmentbe identified across countries and within or betweenindustries? Most importantly, the data set allows usto test whether the adoption of labour productivityenhancing technology by SMEs facing heightenedmarket competition has changed the demand for labourin the home region of the firm, and whether govern-ment policies have affected the adjustment potentialsof such firms.^

At the outset, it should be recognized that this paperis less an exercise in the construction of theory and morea systematic analysis of a unique data set for policy-related purposes. Based upon a common questionnaireand survey methodology, a data set on SME competitivestrategy has been created covering a range of topicsincluding market conditions, employment, technologyand investment. The data set is quite extensive, beingdesigned to aid comparative statistical analysis. Theconstruction of the data set is discussed below and inAppendix 1. At this point, it should be recognized thatrobust relationships between changing market condi-tions, the adoption of technology and employment werederived from the present analysis. Furthermore, andunlike many related studies, consistent results wereestablished distinguishing between industry and countryeffects in driving firm competitive strategy across coun-tries within industries (cf HAWAWINI et al., 2000).Even so, it was found that there remain significant firm-specific factors that drive labour productivity.''

The paper is organized as follows. First, it makescommonplace but necessary observations about thenature of EU regional economic structure. Here, con-trasts are drawn between European and North Amer-ican patterns and expectations. This is followed withthe brief specification of a neo-Keynesian 'model' ofthe firm and its relationship to global commodity andlocal labour markets. The data set and empirical strategyare then described in detail before subsequent sectionsdiscuss the results and the implications of the econo-metric tests.

SMALL- AND MEDIUM-SIZEDENTERPRISES AND ECONOMIC

PERFORMANCE

The paper begins by assuming that SME performanceis very sensitive to market revenue and hence their

market share. This is an immediate implication ofrecent theoretical work by GREENWALD and ST IG-

LITZ (1995) and matches the present authors' intuitiveunderstanding of the predicament faced by many ofEurope's smallest firms. It could be also suggested thatSMEs are more or less 'competitive' if they expand ormaintain their sales in existing markets. Of course,there are more extensive definitions of competitivenessthat would take into account firm's strategic goals inadding new markets and/or expanding the range ofproducts produced so that their share of market demandmay increase over time. By beginning with the firm,its market revenue and market share, it was also assumedthat competitiveness is a characteristic of firms and theresources - including geography and history — at theirdisposal, and is the result of strategic choices made bytheir owners and managers.'' It is not the wish of theauthors to exaggerate or idealize the scope of firm-based strategic decision-making. If markets are highlycompetitive, firms' strategic options may be verylimited; strategic choice and decision-making in thiscontext may be simply an issue of internal flexibilityand adjustment capacity in accordance with marketsignals (CLARK, 1994).

By focusing upon the firm and its strategic choices,the authors do not mean to ignore the context inwhich such issues are considered and resolved (CLARK

and TRACEY, 2004). In point of fact, as many otherswould argue, the time and place of strategic decision-making may have significant implications for thoseoptions considered as relevant, those ignored and thoseultimately taken. But as is known, whatever the localfactors involved in decision-making, market scope andprospects have broadened enormously over the past25 years working up the spatial scale from local toregional and national, and now to Europe and theworld beyond. Indeed, just as 'local' firms have hadopportunities to expand into markets that have takenthem away from their local communities, so too haveother firms located in faraway jurisdictions come to thecommodity and consumer markets of the developedand developing worlds. Markets are increasingly opento rival producers whatever their location of production(CLARK et al, 2000).

There remain, however, considerable tensionsbetween wbere firms produce and the ultimate destina-tion of their products. If at some point in time in thepast, the geographical scale of production matched thegeographical scale of final markets, it could be arguedthat there was a symbiotic geographical relationshipbetween the organization of the production processand the configuration of fmal consumer markets. Theliterature on regional economic growth assumes thatthis was once the case and goes on to show thatthe increasing spatial disjunction between the site ofproduction and the geographical scope of marketshas increased the premium placed upon the strategiccapacity of those firms that have a distinctive and


committed place of origin. While much ofthe literaturefocuses upon the mobility of capital in relation to theconfiguration of final demand, in Europe at least, oneshould be less sanguine about the prospects of firmsrelocating in relation to the imperatives of marketcompetition given that in many industries SMEs arethe dominant unit of production. In many cases, factormobility is a less important option compared withdesigning and implementing changes to the organiza-tion of production and the technology of productionitself ( R I C C I , 1999),

It must be recognized, however, that conventionalmodels of regional economic growth begin with ratherdifferent assumptions. Being dominated by Anglo-American theoretical presumptions and a distinctiveheritage of empirical research, much of the literatureassumes high levels of factor mobility and ultimatelyspatial and economic convergence in measures ofemployment and welfare (BARRO and S A L A - I -

M A R T I N , 1995), By implication, assumed are nationalmarkets and relatively low transaction costs both withrespect to the distance to market and to the flow ofcommodities between firms within related chains oftransactions that produce the products brought to mar-ket. Obviously, the exemplar case is the US economy.Just as obviously, those that advocate a single integratedEuropean market have in mind an institutional config-uration that would at least mimic the US case particu-larly in terms of enhancing the efficient allocation ofcapital if not labour between the regions of Europe(KRUEGER, 2000), If European firms are currently'embedded' in their localities, in the end Europeanintegration may transform geographical constraintsinto extensive geographical and industrial opportunities(cf RoTEMBERG and SALONER, 2000, who use non-negligible transport costs to help sustain, in theory, adifferentiated economic landscape).

These theoretical and empirical expectations havebeen challenged in recent years by the new economicgeography allied with KRUGMAN (1991), but sharedwith many economic geographers whatever their dis-ciplinary heritage. If increasing returns to scale asopposed to constant returns to scale are introduced,then it is possible that individual firms may wish toconcentrate their productive capacity while developinglocal networks of intensive linkages that in effect sharebetw^een firms the benefits of increasing returns toscale. If it is assumed that 'learning by doing' character-izes many firms' experience in exploiting their inher-ited configuration of production, there may beconsiderable benefits in sticking with past investmentswhile adopting new forms of technology that reinforcetheir knowledge base. Finally, if knowledge spills overbetween firms within an industry-region by virtue ofthe movement of labour between related firms, a firm'slabour productivity may develop in accordance with itsco-location with other firms. This is a theoreticalrationale for linking the competitiveness of firms with

the attributes of industry—regions (COOKE andM O R G A N , 1997),'

At the limit, firms' competitiveness may be thoughtdependent upon their location in region- and industry-specific regimes of accumulation. And it is possiblethat firms' competitive strategies are both enabled andhmited by their region-industry setting. However, itshould be observed that there are at least two counter-vailing forces that may undercut long-term region—industry-specific regimes of accumulation. As marketsfor products have extended beyond communities,regions and even nations, the attributes of productsmust be seen as increasingly exogenous to region-industry regimes of accumulation. Commodity marketsmay effectively discount the value of region—industriesby homogenizing products, and perhaps even the meansby which they are produced, A region—industry com-plex may control the defmition of product quahty as itmay control the definition of product characteristics.In that case, it may dominate an industry. However,as the geographical scope of commodity marketsbroadens, there are increasingly more rival region-industry complexes that challenge this kind of hege-mony. In any event, it is apparent that, within anindustry, product characteristics and the means andmethods of producing those products diffuse betweenjurisdictions over the long term (BREZIS et al., 1993),

With respect to firms' competitiveness, one shouldexpect to find a variety of interaction effects includinginteraction between firms and regions, firms and indus-tries, and even firms and country settings. These effectsare the focus ofthe present country-specific case studies(CLARK et al., 2002), One should also expect to finddistinctive scale effects: as the geographical scope ofmarkets for labour-intensive industries has shifted fromregion to nation to Europe and the world, the industryaffiliation ofa firm as opposed to its country and regionof origin could now be the dominant and driving forcebehind strategic choices with respect to competitiveadvantage. This is the topic of this paper,

MODEL OF THE FIRM(THEORETICAL EXPECTATIONS)

To sustain the present empirical analysis, one shouldestabhsh an appropriate theoretical reference point.One way to do so would be to invoke a theory of thefirm and economy based on real prices and perfectcompetition. As indicated above, however, there aremany reasons to suppose that European productmarkets, capital markets and labour markets are subjectto significant 'imperfections', A more appropriate theo-retical reference point with which to inform empiricalanalysis of current circumstances would be the literaturegrouped under the heading of the 'new Keynesianeconomics' (MANKIW and R O M E R , 1991), Therein,due recognition is made of the structural imperfectionsof markets and institutions that sustain price rigidity.


the immobility of capital and labour, and the incre-mental costs of adjustment. Whether 'perfect markets'are the most appropriate reference point for long-termwelfare analysis in the European context is, of course,the subject of considerable intellectual and pohticaldebate.

Since much of the paper is devoted to the results ofempirical analysis, it does not intend to derive a formalmodel of the firm based on first principles. However,it is the authors' wish to record their initial expectationsabout the most important variables, the expected rela-tionships between those variables, and whether theyshould expect positive or negative signs on the esti-mated parameters. Also informing the authors' expecta-tions is the work of GREENWALD and STIGLITZ

(1995) and HART (1982), as well as the more recentwork of DAVIS and WEINSTEIN (1999) on the dis-puted significance of economic geography for Europeand Organization for Economic Co-operation andDevelopment countries. While there are some signifi-cant differences between these writers, especially asregards the status and importance of market efficiency,it is shown empirically that there remain importantcountry and industry effects that dominate the adjust-ment capacity of European SMEs,' ' The paper will firstconsider the competitive domain of its firms.

The unit of analysis is SMEs, While the definitionof SMEs is a topic of considerable academic research,it is apparent from the present data that many arevery small (with less than ten employees). Given theirsignificance as employers for many European countries(especially Greece, Italy and Spain), the competitivestrategies of these types of firms have enormousimphcations for the long-term growth potential ofEurope as a whole. In the present study, of course, thesefirms operate in labour-intensive industries subject toEuropean and global competition. As such, these firmsare assumed to adjust to market signals such as theindustry price of a commodity and the volume orquantity demanded of such a product demanded byconsumers (whether other businesses or final con-sumers). For these firms, market prices and the quanti-ties demanded may be quite volatile and subject toconsiderable uncertainty from year to year. Entry intodomestic and European industry markets by 'external'producers (notably firom South Asia and the Far East)is assumed to be an ever-present threat and reality. Inthis respect, while market competition may not be'perfect', it is amongst firms and potential rivals ofwhich none can effectively 'manage' prices andquantities.

On the input side ofthe equation, the SMEs in thisstudy are assumed to face short-term fixed prices forlabour and most inputs to the production process. Ifmarket demand rises over time with either constant orincreasing real prices, then it could be argued thatthe inherited configuration of production is probablyirrelevant to firms' competitive strategies. The issue

here is the extent to which the principals of firms wishto take advantage of market demand by increasingthe volume of output. But this depends upon whatassumption is made about the objective function ofsuch firms. Do they maximize sales, profits or someother related measured variable that represents theirutility? Or do they seek to protect the flow of incomeover time and the cash-out value of the firm (wealth)in conditions of market uncertainty? To discriminatebetween these rival behavioural propositions wouldrequire fine-tuned experimental psychology, which isnot the topic of the present study. At this juncture, itis sufficient to note that it is assumed that firms seekto maximize sales to achieve their other objectives,namely improved operational efficiency and profit(BAUMOL, 1959), By sales maximization, the presentpaper does not mean the number of physical units sold,but rather the total revenue achieved by the firmthrough transactions with customers. As a managementobjective, sales are more observable and, on one level,easier to manipulate than profits (AMIHUD andK A M I N , 1979; VICKERS, 1985), As Baumol notes, theresult is that sales and market share become crucialreference points for firms' strategic decision-making.This is consistent with the position, outlined above,that the performance of SMEs is closely related totheir market revenue.

Whereas the 'new economic geography' invokesscale economies as an incentive rewarding the growthofa firm and related networks of local firms, there arereasonable doubts about the existence of increasingreturns to scale amongst SMEs, FoUovving B L A N -CHARD and KiYOTAKi (1987), a more plausibleassumption would be to assume constant elasticity ofsubstitution where output is a linear function of inputsof labour and capital (the embedded capital stock andtechnology of production). By this logic, increasingdemand is met in the first instance by increasing outputthrough a fixed combination of labour and capital.Similarly, any decrease in demand (dechning sales) ismatched by a proportionate decrease in the use of thegiven inputs to production. In many circumstances, theembedded capital stock (and related technology ofproduction) is rudimentary at best. Declining demandis met, in these circumstances, by decreased hours oflabour and decreasing use of capital. Herein, it isassumed that quantity adjustments (rather than priceadjustments) dominate SMEs' responses to changingdemand.

The present paper is also concerned with the issueof technological adjustment — going beyond immediatecapacity adjustment to changing market demand, tolonger-term responses that may include changing theinherited configuration of production. The foUovvingassumptions are made. First, the adoption of newtechnology is a function of investment, which is itselfa function of the capacity of firms to finance that

1090 Gordon L. Glark et al.

investment. Given that SMEs are almost always under-capitalized, subsidized sources of finance may bethought an important enabling condition for invest-ment. Second, it is also assumed that the choice oftechnology is determined by both the expected demandfor products differentiated by quality and price, and therule that firms maximize total revenues and minimizecosts. In the end, the choice of technology is onedeterminant of labour productivity (there being other,perhaps more important, determinants of labourproductivity including the implementation and organ-ization of that technology in real production processes;GERTLER, 2001). Third, it is assumed that the adoptionof technology has significant implications for the com-position and volume of labour used in the productionprocess. Again, it is assumed that quantity adjustmentdoniinates price adjustment in this case (the embodiedskills of labour).

As the nature of the embedded capital and technol-ogy of production changes, one might expect to seefirms consider the advantages and disadvantages ofincreasing output in relation to expected demand. Ineffect, the issue here is one of firm size and commit-ment to growth. These are not issues of great signifi-cance for this paper. But one would suggest that growthin firm size is not necessarily desired by the owners ofSMEs if they are owner-managers and entrepreneurs.In this respect, there may be a significant differencebetween the value of cash flow (and all that entails forbuilding a firm) and profit (and all that entails forpersonal and family wealth as reflected in propertyand consumption). In this respect, the adoption oftechnology may be as much a defensive competitivestrategy designed to protect market share and retainexpected profits as it is an expansionary competitivestrategy designed to capture market share and therebyenable growth in employment and income. In otherwords, all things being equal, the adoption of technol-ogy could be associated with increasing or decreasingsales.

DATA AND EMPIRICALMETHODOLOGY

As noted above, this paper relies upon a unique dataset developed in collaboration with partners across theEU. The data set was produced through the design andexecution of a survey of SMEs in four labour-intensiveindustries located in vulnerable regions of Europe.Using common questions, a well-defined surveymethodology and an agreed coding system, the dataset allows for the pooling of data by question acrossindustries and regions. The results of econometric andstatistical tests reported below rely upon this pooleddata set recognizing that country-specific case studiestake advantage of the insights generated by the surveynot easily captured in a strictly quantitative approachto the topic. Here, the paper focuses upon quantitative

insights and the arguments generated by the pooleddata set are augmented, in conclusion, with relevantinsights from the case studies.

Given the significance of understanding competitivestrategy across the EU in labour-intensive industriesand vulnerable regions, the paper sets out to collectdata that was robust enough for comparative econo-metric and statistical analysis. Here, the authors' wishis to summarize the steps taken in this process so as toindicate both the strengths and weaknesses of thederived data (see also Appendix 1).

The first step involved defining 'vulnerable' regionsbeginning with three criteria: Objective 1 status, a non-metropolitan location (being outside of large centresof urban economic development), and with labour-intensive firms and industries. In the Italian case, theauthors' coUeagues developed a statistical method forthe classification of regions using a wide range ofindicators of economic well-being including incomeand employment opportunities (GUERRIERI andIAMMARINO, 2002). At this point in the procedure, itwas ensured that the identification of regions containedenough commonality across countries to make formeaningful comparison and pooling of data. For theprecise definition of the regions by country, see thepresent authors' report on this phase of the project(CLARK et al., 2002).

The second step involved establishing trends ofemployment, unemployment and incomes by industryand region, paying particular attention to the numbersand sizes of firms in the identified four labour-intensiveindustries. Those industries, as noted above, includedfootwear and leather products, apparel, electrical andelectronic equipment, and automobile componentparts. An assessment was made at this point about thecompetitive place of these industries in Europe and inthe global economy. These industries are significantemployers across Europe, and are especially sensitive tomarket competition accompanying global economicintegration.'

The third step was to design an appropriate question-naire, linking the firm and its ownership structure toa set of crucial issues having to do with competitivestrategy, market conditions, employment, investmentand technological change, and government policy. Theinitial questionnaire was designed to be a scaled attitu-dinal survey instrument built upon overlapping ques-tions allowing for cross-tabular representation of theresults. It was based, in part, upon a similar type ofsurvey of Asian SMEs reported by CLARK and K I M

(1995). Having developed the questionnaire, it wasthen tested in a number of countries and the relevantregions. It was found to be too long, too complicatedand too demanding in terms of the distinctions ofjudgement asked of the respondents. The questionnairewas then redesigned so as to identify actions taken asopposed to attitudes while deleting extraneous ques-tions (see Appendix 1).**

Globalization and Gompetitive Strategy in Europe's Vulnerable Regions 1091

The fourth step was to implement the questionnairein each country, region and industry over 6 months(during the first half of 2001). Note that respondentswere asked to consider their competitive strategies,market conditions and responses over the past 3—5years. In many cases, the surveys were executed throughon-site interviews with the most relevant firmrepresentative (in many cases being the principal andmanaging director of the firm). In almost all cases, atleast 20 completed responses were produced for eachregion—industry, being in combination a data set of over331 completed and consistent questionnaire responsescovering more than 25 specific questions regardingEuropean SME industry competitive strategy in theglobal economy'

The fifth step was to analyse the coded responses.The data set was compiled, cleaned of missing data andambiguities in response, and then set in a databaseconsistent with facilitating comparative econometricand statistical analysis. In the main, it was found thatthere were surprising commonalities amongst SMEs ineach industry across regions and countries of the EU.At the same time, note that exceptions had to be madeto the general rules driving the design and collectionof the data. For example, in the UK, the regionreference had to be shifted from an Objective 1 toan Objective 2 status region because in the former(Liverpool and surrounding areas), there were simplyinsufficient firms of an appropriate size in the relevantindustries to survey. Similarly, in Ireland, for differentreasons in one particular industry there were alsoinsufficient firms to survey. While in Spain, adjustmenthad to be made between the industries in terms of theproportional share of the industry interviews.

Having developed the database, having ensured itsintegrity, and having assessed its compatibility withinand between regions, within and between industries,and across Europe, the next step was to undertakean analysis of the relationships between competitivestrategy, employment, investment and technologicalchange. These tests are reported below.

ECONOMETRIC AND STATISTICALRESULTS

The results of the various econometric and statisticalanalyses performed using the entire database of 331completed survey questionnaires are reported. Theresults are dependent, of course, upon the integrity ofthe survey instrument and the collection of the data ineacb country and the related regions. Where possible,it was ensured that tbe data-collection process usedcommon protocols and procedures (as described inthe Appendix). Given the sparseness of data in someindividual country/industry sets (e.g. leather productsin Ireland and Italy), care must be taken in assessingthe significance of the results.

Analysis of the whole model

To begin, the complete model linking firm decision-making with variables such as changes in sales andemployment, as well as the relationships between sales,investment and technological cbange is analysed. Theserelationships are clearly bighly aggregated, beingbrought togetber across tbe total data set. Even so,knowledge of tbe robustness of tbese expected relation-ships is important in coming to grips witb tbe signifi-cance of the subsequent observations. At tbis point, ofmost concern were tbe statistical significance, parameterestimates and overall goodness of fit of tbe expectedrelationships. To summarize, tbe model estimated wasas follows:

= a,, -ha,2lNV,;--I-a

; = flj, + ai.'SALj, + a

;.-f dnXEUM^;,-h «

(1)

M2, (2)

TECHj, = fl3, + ^32INV . -1- 33 CACF , + u^, (3)

(4)

(5)

where SAL is sales and SAL is identified by exogenouselements, INV is investment and INV is identified byexogenous elements, XLRM is exports to local/regional markets, XEUM is exports to EU markets,EUPNT is EU programmes for new tecbnologies,EUPRT is EU programmes for retraining, LOQP is alack of qualified personnel, TECH is adopted technol-ogies and TECH is identified by exogenous elements,CACF is cost/access to finance (factors inbibiting tbeadoption of tecbnologies), EMPL is employment andEMPL is identified by exogenous elements, LBSK istbe need for adequately skilled employees and LBSK isidentified by exogenous elements, ATPT is the adoptedproduction process tecbnologies, and where / is theinterviewed firm in sector 5.'"

In most cross-section econometric models, the mostcommon estimation approacb bas been tbe spe-cification of single-equation models to identify andestimate tbe impact of predetermined variables onendogenous elements. However, tbe economic rela-tionships of the present model (involving SAL, INV,TECH, EMPL and LBSKL) are interrelated becausetbey are needed for determining tbe value of tbeendogenous variables included in the model. Tberefore,tbe present model is a system of simultaneous equationssince tbe adoption of tbe single-equation method ofestimation would lead to biased estimators. This arisesbecause tbe reduced form of tbe present specifiedmodel sbows explicitly tbat tbe endogenous variablesSAL, INV, TECH, EMPL and LBSK are jointlydependent upon tbe predetermined variables XLRM,


Table 1. Summary of results for the estimated relationships regarding firm-specific adjustment (sales, investment, technologicalchange, employment and labour force composition) using the total database (331 complete sets of observations)

Dependentvariable

SAL

INV

TECH

EMPL

LBSK

Constant2.338 (12.372)

2.052 (7.179)

2.635 (12.542)

2.326 (9.368)

2.595 (12.153)

INV0.148 (2.590)

SAL0.155 (2.748)

INV0.146 (2.414)

SAL0.157 (2.860)

EMPL0.175 (2.903)

Explanatory variables

•XLRM0.122 (2.104)EUPNT

0.135 (2.274)CACF

-0.138 (2.215)LBSK

0.203 (3.702)ATPR

0.393 (2.579)

XEUM0.137 (2.354)

EUPRT0.110 (1.862)

LOQP-0.185 (3.279)

F>uu«:c

5.633

7.847

4.794

9.761

8.953

Statistical tests

Degrees offreedom

292

291

250

318

318

0.234

0.312

0.192

0.240

0.294

Notes: (-statistics are in parentheses below the estimated values of the parameters.For an explanation of variables, see the text.

XEUM, EUPNT, EUPRT, LOQP, CACF and ATPR,and the disturbance terms MH, M21, M31, w , and M5] ofthe system. The model was estimated using the two-stage least-squares method (perhaps the best-knownsingle-equation method that can be used; cf. H A R T -

MANN and M A STEN, 2000). The estimates are, ingeneral, given as non-linear functions of the reducedform estimates and inherit all their asymptotic proper-ties. The results are shown in Table 1.

With respect to these results, four observations canbe made. First, the signs of the estimated parameterslinking, for example, sales, employment and labourcomposition were as expected (in a neo-Keynesiansense). This is an important finding since it helps oneappreciate the significance of commonalities amongstthe data and the logic by which the analysis wasapproached. Second, it is clear that both local andEuropean markets are important driving forces behindfirms' changing competitiveness: the European market'effect' is more important than the local tnarket 'effect'.Market integration and industry competition are asignificant issue behind European SME competitivestrategy. Third, with respect to the adoption of processtechnologies designed to enhance SME competi-tiveness. Table 1 demonstrates the expected relationshipbetween technology and investment (positive) and amost important relationship between technology adop-tion and the cost of finance (negative), as well as the(positive) relationship between investment and EUprogrammes designed to enhance access to new tech-nologies and labour skills. Finally, as expected, there isa strong (positive) relationship between technologicalchange and the skill composition of employed labour.While the coefficients of determination are modest,these tests of significance are robust.

The authors were, of course, concerned to test for'problems' or significant discontinuities in the data,particularly given the apparent differences between thecountry studies. Two different tests of the coherence

of the database were conducted. In the first instance,tests were undertaken on the stability of the estimatedparameters from the econometric model. For methodo-logical completeness, the geometric properties, i.e. themean and variance of the variables used in the econo-metric estimations and analysis, are presented in Table 2.The variance shows the distribution of the responsesaround the mean and indicates whether the responsesare normally distributed. The mean is calculated fortwo types of variables. The ftrst type (Table 2a) includesvariables derived from responses on the scale 1—5,where 1 = decreased greatly and 5 = increased greatly.The second type are dichotomous variables (Table 2b)with l=Yes and 0 = No. Note that variable TECHdoes not appear among the explanatory variables of theestimated equations (4) and (5) of Table 2; it was

Table 2. Summary of the means and variances of the variablesused in the econometric model (as estimated using the entire

RA STE I countries and industries database)

Variable Mean Variance

2a. Variables where the responses were on the scale 1—5(where 1 = decreased greatly, 2 = decreased, 3 = remained the same,4 = increased and 5 = increased greatly)SAL 3.11 1.69INV 2.67 3.32EMPL 3.38 1.38TECH 2.76 3.04LBSK 3.31 1.64

2b. Variables where the responses were 1 = yes and 0 = noXLRM 0.59 0.24XLEU 0.63 0.23EUPNT 0.17 0.14EUPRT 0.13 0.12LOQP 0.32 0.22CACF 0.46 0.25ATPR 0.32 0.22

Note: For an explanation of variables, see the text.

• Globalization and Competitive Strategy in Europe's Vulnerable Regions 1093

omitted because of collinearity with SAL in equation(4) and with ATPR in equation (5).

It was also sought to test whether the exclusionof a country would significantly change the derivedeconometric and results of the analysis of variance(ANOVA). Using Spain as a test case, the analyseswere run with and without this country's survey data.While there were minor differences in the estimatedparameter values, there were no instances of changedstatistical significance, nor were there any instances ofchanged signs on the estimated parameters. Furthermore, with respect to the ANOVA analyses reportedbelow, the exclusion of the Spanish data would nothave altered the statistical results significantly. As aresult, the authors are confident that the statisticalresults are robust.

Analysis of variance (ANOVA)

The aim ofthe ANOVA was to examine whether ornot the adjustment capacity of the examined firms —specifically investment in technology and its conse-quences for competitiveness and employment - differedstatistically significantly across the industries and coun-tries in question. The s variance was considered to bethe industry/sector, and f \vas considered to be theinterviewed firm located in region r of country h. Itwas assumed that r was identical with h (although thisis not strictly true for Greece)." The s variance isdescribed by WA for wearing apparel (18), LP forleather products and footwear (19), EL for electronics(30—32) and TE for transport equipment manufacturing(34). The h variance is described by GR for Greece,EN for England, IT for Italy, SP for Spain and IR forIreland.

More specifically given the research objectives, thestatistical analysis examines whether and to what extent:(1) the behaviour of the / firms across the h = GR,EN, IT, SP and IR countries differed statistically {HQhypothesis: do not differ), independently of the 5 = WA,LP, EL and TE industries. The s is the basis here forthe classification of the/into /; subsamples; and (2) thebehaviour of the/firms across the 5 = WA, LP, EL andTE industries differed statistically (Ho hypothesis: theydo not differ), independently ofthe r—1, 2, 3, . . .vulnerable regions, and the h = GR, EN, IT, SP and IRcountries. The h is the basis here for the classification ofthe/ in to s subsamples. ANOVA is equivalent to theestimation of the F* distribution with p^ = m — 1 andV2 = N— m degrees of freedom given by:

- 1 )

(6)

where «, is the size of the jth sample, N = L"L i«, is thesize of the 'pooled' (enlarged) sample and m is thenumber of samples.

The Ho hypothesis holds at a 95% level of significanceas long as the estimated F* statistic is smaller than thetheoretical specific level of significance given by thedegrees of freedom. If the variables s and h that are thebasis for the classification of the f's into h and ssubsamples, respectively, are important causes of vari-ation in / ' s behaviour, performance, etc., then thedifferences among the means of the subsamples will belarge: this will be shown by a large dispersion of themeans of the subsamples Yj-^ around a common meanX that is by a large variance ofthe distribution. On thecontrary, if/? or s is not an important source of variationofthe s and h subsamples, respectively, then the differ-ences among the means ofthe subsamples will be small.

The ANOVA results are suntmarized in Table 3.The F-statistic and the level of significance foraccepting the null hypothesis (Ho) are presented withrespect to (i) the mean of the 17 variables acrossfour sectors within the same country do not differstatistically (Table 3, a); and (ii) the mean of thevariables across five countries within the same sectordo not differ statistically (Table 3, b). More detailedstatistical results are available from the authors.'^

First, the variation in source of employees (variable1) of the firms is examined for the industries andcountries in question. Any identified differences wouldsuggest diversity in the role of firms within the local/regional labour markets. Hypothesis (ii) is examinedacross the five countries in question, and the test resultssupport the proposition that firms in the sample sectorsacross the different countries use similar sources ofemployees, with the exception of leather and footwearproducts. When hypothesis (i) is examined, the resultsindicate that the sources of employees across industriesdo not differ for any ofthe countries.

The competitive strategy of firms was examinedbased on change in sales (variable 13) and response tochange in sales (variable 3). In other words, whetherincreases or decreases in sales over the last 3 yearsare associated with the firm's response by changingproduction capacity, expanding into new markets,developing products, outsourcing, forming partnershipsand acquisitions. Hypothesis (i) is tested first, and theresults strongly support the proposition that the changein sales of firms in England, Italy and Ireland over thelast 3 years was similar across industries, but differentin Greece and Spain. The most important responseof firms to the change of sales across the examinedindustries was similar in Greece, Italy and England, butdiffered in Spain and Ireland. In other words, thecompetitive response of firms to changing sales is, ingeneral terms, both country and industry determined.

Tests for hypothesis (ii) suggest that firms in three ofthe four industries had significantly different salespatterns across the examined countries. In other words.


Table 3 . F-statistic results of analysis of variance

3a. F-statistic results testing HQ: (i) Mean of thevariables across four sectors within the same country do

not differ statistically

3b. F-statistic results testing H,,; (ii) Mean ofthe variables across five countries within the

same sector do not differ statistically

Question/variable

1. Source of employees2. Sales3. Response to change in sales4. Investment strategies5. Investment sources6. Effect on labour7. Desired results of investment8. Technologies adopted9. Sources of technological

knowledge10. Ways upgrading skills11. Development programs12. Factors inhibiting technology

adoption13. A Sales14. A Demand for skilled labour15. A Total employment16. A Per unit labour costs17. A Total labour costs

Greece

1.8870.5721.4524.803**1.1010.2481.3162.959**

0.4480.5300.935

2.2072.644*7.809**5.740**0.9380.326

England

0.2435.750**0.6680.4722.0040.4781.8742.080

8.835**3.460**8.002**

1.8301.7662.849**2.608*2.2853.285**

Italy

0.8821.2751.4040.4761.2640.7651.9772.791*

0.7100.5670.254

0.4840.1832.2290.5150.6911.145

Spain

0.8231.0612.758*3.287**3.462**5.369**0.4202.839**

0.2490.8481.594

0.6402.728*6.720**4.650**2.548*3.769*

Ireland

0.6311.6342.515*0.9573.741*0.3121.0190.531

—0.5420.240

0.7460.6240.4341.6851.9161.356

Wearingapparel

0.5925.941**2.850**9.375**2.672**2.430*2.568**3.359**

2.570**1.9731.466

0.82313.735**25.627**

7.713**9.198**0.476

Leatherand

footwear

2.453*1.7971.0220.4184.330**1.1531.2264.567**

0.9860.3980.065

0.8571.2071.2081.9880.5951.481

Electricaland

electronicassembly

0.8810.8810.7651.3802.432*0.2802.873**1.373

2.2661.2861.511

0.1728.909**

28.908**3.208**0.6181.891

Automotivecomponents

0.1070.1071.5552.0852.341*3.053**1.3301.503

0.6912.507*2.967**

1.13111.286**9.472**1.7382.279*2.929**

Notes: A = Change in.Levels of significance: **99%, *95%; otherwise not significantly different from zero (level of significance).

the economic performance of firms appears stronglydependent upon the country from which they operate.The exception is the leather products industry, whichappears to be independent from the business environ-ment of each country in terms of sales. With respectto the most important response of firms to changes insales, in three of four industries, similarity is sustainedacross countries. However, in the wearing apparelindustry, the response of firms to changing sales differsamong countries; in other words, the country contextappears to be more influential for firms in this industrythan for those in the other industries considered.

Next, the analysis focuses on the investment strategies(variable 4) ofthe firms, and more specifically on theirmost important investments either in new plants andequipment or in innovation and product development.The investment priorities of the firms are examinedalong with variable 5, their investment sources (internal,provided by banks, government or EU subsidy) andvariable 6, their effects on labour (either displacementor change in demand for skilled/unskilled and part/full-time labour). Moreover, the desired results of invest-ment (variable 7) are taken into account (increasedproductivity, increased market share, increased profit orincreased technological sophistication).

The results for hypothesis (i) show that in three offive countries, the most important investment strategiesadopted by the firms do not differ significantly acrossthe examined industries. However, in Greece and

Spain, investment decisions differ among the industries.In three of five countries, investment sources appearsimilar across industries. However, in Spain and Ireland,sources of investment seem to differ among industries.In four countries, the investment impact on the variableeffect on labour has a similar pattern across industries.The exception is Spain, where investment implicationson labour differ across industries. The most desiredinvestment results do not differ across industries, in allcountries.

Tests for hypothesis (ii) suggest that investment strat-egies are similar in three of four industries acrosscountries, the exception being wearing apparel, whereinvestment choices depend upon the business context;in other words, the place that regional agglomerationshold in the international network of this industrydetermines their investment strategies. The hypothesisthat investment sources are similar in the examinedindustries across countries is very weak. Investmentimplications on variable 6, effect on labour, are similarin the leather product and electronic assembly industriesacross countries, but in the automotive componentsand wearing apparel industries there are differencesbetween the various contexts. Finally, the results of themost desired investment outcome are mixed: similarityis found for the leather products and transport equip-ment manufacturing industries across countries, whiledifferentiation is depicted for the wearing apparel andelectronics industries among countries.


Having analysed the competitive and investmentstrategies of firms, the paper attempts to associate theresults with the technologies adopted (variable 8) inspecific firms. In other words, the analysis examinesthe most important technologies adopted by the firmsin the last 3 years (in product development, inventorycontrol, marketing and communication), and the mostimportant sources of technological knowledge (variable9) (internal personnel, suppliers and customers, institu-tional organizations). The test of hypothesis (i) suggestssimilarity across industries in the most important tech-nologies adopted by firms in Ireland and England, butdifferences in Greece, Italy and Spain. The sources oftechnological knowledge are similar in three countriesacross industries (Greece, Italy, Spain), but in Englanddifferentiation is depicted among industries (no dataare available for Ireland for this variable).

Tests for hypothesis (ii) suggest that in wearingapparel and leather products, the investment in technol-ogy differs across countries. In other words, it dependsupon the business context. In the electronics andautomotive components industries, investment in tech-nology is less context determined and rather industryfocused. With the exception of the wearing apparelindustry, the sources of technological knowledge aresimilar in all industries, independently of the countrycontext.

Next, the paper examines the impact of strategiesadopted by firms (response to change in sales, invest-ment, technology adoption) on labour, and more spe-cifically upon: change in demand for skilled labour(variable 14), change in total employment (variable 15),change in per unit labour costs (variable 16) and changein total labour costs (variable 17). Provided that firmsapply similar technology strategies (independently ofindustry and country), the effects of these strategies onlabour are interesting. Starting with hypothesis (i), theresults suggest that in three of five countries — Greece,England and Spain — the need for adequately skilledemployees over the past 3 years differs significantlyacross industries. The pattern of total employmentdiffers across industries in Greece, Spain and England.The labour cost per unit over the last 3 years evolvesalong a similar pattern across industries in four coun-tries, the exception being Spain. Finally, total labourcosts over the past 3 years differ at the 99% level ofsignificance across industries for England and at the95% level of significance for Spain, but are similarGreece, Italy and Ireland.

Tests for hypothesis (ii) indicate that the impact onthe demand for skilled labour is significantly differentacross countries in the examined industries — with theexception of leather products. The pattern of changein total employment is similar across countries in leatherproducts and automotive components, but differs forwearing apparel and electronic assembly. With regardto change in per unit labour cost, in wearing appareland automotive components the pattern differs among

countries, but is similar for leather products and elec-tronic assembly. Change in total labour cost is similaracross different countries and contexts in all industriesapart from automotive components.

The fmal issue concerns policy implications. Herethe paper examined variable ways of upgrading skills(in-firm training or institutionally provided), variabledevelopment programmes (loan and credit facilities, taxcredits and labour-cost subsidies, or R&D and traininggrants) and variable factors inhibiting technologyadoption (uncertain benefits, cost/access to finance,lack of information, size of market, lack of skilledlabour, the desire for a manageable firm size). Theresults for hypothesis (i) suggest that, in general, theimpact of the policy and institutional context on SMEsin the examined countries is similar across industries,with the exception of England, where the ways thatfirms upgrade skills and use development programmesdiffer among industries.

The results for hypothesis (ii) indicate that policymeasures and institutions affect firms similarly across allcountries in the examined industries. The exception isthe automotive components industry, where the mostimportant development programmes used by firms inthe past 3 years are different.

Summary of the ANOVA results

In an attempt to systemize the results of the statisticalanalysis of the research data, according to the researchhypotheses, the following are concluded:

• Firm economic performance is country defined.• Investment strategies are defmed by the role/place of

country-based networks in a world of integratednetworks of production.

• Technology adoption and the sources of techno-logical knowledge are country defined; but in lesstraditional firms, they are industry led.

• Change in a firm's total labour costs is industrydetermined.

• Total employment of the firm is industry determined.• Demand for skilled labour is both country defined

and industry led.• Investment impact on firms' workforces is both

country and industry led.• Policy and institutional impacts are country defined

and industry determined.• Competitive strategy is neither country defined nor

industry determined; firm-specific factors dominate.• Labour cost per unit change is neither country

defined nor industry determined; firm-specificfactors dominate.

CONCLUSIONS

For many years, the economic map of Europe has beencharacterized by regions of long-term unemployment


and concentrated centres of economic growth andemployment. A vast array of regional developmentpolicies has been used to ameliorate persistent highlevels of unemployment, from state aid to a systematicconcern with the competitiveness of firms and regions.As European and global integration gathers pace, theeconomic performance of 'vulnerable' regions andindustries in the EU has become more important thanbefore. Many communities and the whole EU have alarge stake in the adjustment potential of firms to theswirling winds of market competition. Furthermore, inthose non-metropolitan regions dominated by labour-intensive industries, the competitiveness of local SMEsin the global economy has enormous implications forthe long-term growth trajectories of those regions.

Competitiveness is commonly thought synonymouswith clusters of innovation. Whether associated withthe work of PORTER (1990) or the more subtleapproaches championed by FELDMAN (2000) andSTORPER (2000), amongst others, competitiveness isbelieved to be a process of enhanced labour produc-tivity through endogenous technological change organ-ized via local networks of information and exchange.While sympathetic to that approach, the present paperhas also focused upon the adjustment strategies ofSMEs to changes in market demand. The goal was toestablish the principal causal factors behind firm-basedcompetitive strategy while making an explicit connec-tion between the local employment consequences oftechnology-related adjustment strategies. This approachis consistent with a long tradition in economics and ingeography, linking the prospects of regions to theadjustment potential of local firms in the face of marketcompetition regionally, nationally and internationally.

By convention, studies of market-related adjustmentpotential distinguish between the size of the firm, thenature of firm ownership, regional identity, industryidentity and even country identity. For instance, it isoften contended that a small firm within a distinctiveindustry and regional milieu has a quite different adjust-ment capacity to similarly sized firms in different indus-try and regional settings. Similarly, larger firms arepresumed to have a rather different adjustment potentialthan smaller firms even within the same industry andregional setting. By this logic, 'identity' may be boththe objective of analysis and in various ways a meansof explaining observed variation in adjustment poten-tial. The present study has attempted to hold constantthe size of the firm, the nature of the region andindustry affiliation. It has been found, however, thatthere are significant variations in firms' responses toglobal competition that can be attributed to the countryand the industry of origin.

Remarkably, the paper derived robust econometricrelationships ofthe expected sign between employmentand output, between output and investment, betweeninvestment and technological change, and betweentechnological change and employment composition.

These econometric results across the entire data set areencouraging given the fact that much of the relatedliterature cannot establish such systematic effects(HAWAWINI et ah, 2000). The results are not surprisingif one takes as a reference point economic theory andcertain conventions regarding the nature and structureof firm decision-making in the context of global marketcompetition. But the results are surprising at anotherlevel: given the logistics involved in sustaining a com-prehensive and coherent fieldwork-based study acrossfive countries, the strength ofthe derived relationshipsindicates just how important market competition hasbecome for labour-intensive industries in the contextof European and global economic integration.

Having established a robust econometric framework,the paper looked more closely at the significance orotherwise of industry and country effects across a rangeof issues including competitive strategy and employ-ment composition. Here, a mix of effects and implica-tions were found. For instance, industry effects seem todrive changes in firms' labour costs, whereas investmentstrategies and the adoption of production technologyappear driven by the country setting of the firm. Atthe same time, there are a number of issues that appeardriven by a mixture of country and industry location.For example, the impact of investment on a firm'scomposition of labour is driven by both factors just asthe demand for skilled labour is driven by industry andcountry effects. These are important results, bearingupon the efficacy of country- and industry-specificoccupational and training systems. Even so, whateverthe significance of these macro-factors, there remainvital firm and region-specific factors that appear todrive competitive strategy and labour productivity (asmight be expected; witness the reports of country-based colleagues given in CLARK et ai, 2002, and thearguments of GERTLER, 2001, STORPER and SALAIS,

1997, and others).

Based upon a common questionnaire and surveymethodology, well-defined questions focused on theactions taken over the past 3—5 years were used to deter-mine what drives SME decision-making. The surveyinstrument has been described above and in Appendix 1.What is notable about the survey instrument is that'vulnerable' regions from very different nationalcircumstances were included — from the Celtic Tiger(Ireland) to Greece. A most important finding from theeconometric results was that government policy canmake a positive difference to SME market adjustmentpotential. The derived relationships were empiricallyrobust and indicate that, as might be theoreticallyexpect, SMEs rarely have the internal resources neces-sary to go it alone in European and global markets. Thisdoes not mean, of course, that government policy is asufficient ingredient in determining 'successful' adjust-ment potential. But it does indicate that governmentpolicy has played an active and positive role in enhancingthat capacity (cf. HARTMANN and MASTEN, 2000).

Globalization and Gompetitive Strategy in Europe's Vulnerable Regions 1097

The results of the present study have significantimplications for conceptualizing SME performance byregion and industry across Europe, as well as implica-tions for the role and status of government policydesigned to enhance European and global competi-tiveness. Just as importantly, the results have implica-tions for the long-term development of vulnerableregions in terms of the demand for labour (quality andquantity). There was a significant relationship betweenthe adoption of new technology and changes in theskills mix of employed labour within local SMEs. Thatrelationship may vary by industry, and there is nodoubt that the adoption of ne'w technology may havedistinctive 'local' effects not captured by the economet-ric analysis (as suggested by GERTLER, 2001). Never-theless, it is apparent that new production technologiesincrease the complement of skills required and decreasethe per unit produced volume of labour. In this respect,as small firms in labour-intensive industries rise to thechallenge posed by European and global competition,their role in regions characterized by high rates ofunemployment and low relative income will furtherdecline.

There is one further important implication: global-ization and market competition create winners andlosers. If the present econometric results hold for otherlabour-intensive industries across Europe, there may bemore commonalities amongst 'losers' than might havebeen imagined based on domestic EU politics. Aswhole industries respond to global competition, thecountry—regional concentration ofthe negative effectsof SME competitiveness may reinforce a political-cum-regional sense that integration itself is the driving forcebehind changing lifetime employment and incomeopportunities. The authors do not mean to suggest thatEuropean and global integration can in any sense be'stopped' or held in abeyance for a generation. In fact,as other Central and Eastern European countries jointhe EU and as their labour-intensive industries becomeexposed to the forces of market competition, it mightbe imagined that this type of study will be moreimportant than before.

Acknowledgements — The paper was made possiblethrough a European Union FP5 research grant (RASTEIHPSE-1999-00035) and the continuing interest of theauthors' European Union project managers Marshall Hsiaand Nikolaos Kastrinos. The authors recognize the commit-ment of their colleagues from Ireland/France (Mary O'SuUi-van and Manuela Giangrande), Italy (Paolo Guerrieri andSimona Iammarino), Spain (Pere Escorsa) and the UK (HelenLawton Smith) to the project including their survey work andcomments on a previous draft (at the Barcelona coordinationmeeting). Their work and that of their research teams madethe paper possible. In addition, the authors acknowledge theresearch assistance on the development and analysis of thedata set, including that of Jane Battersby and D. Boutzona.Finally, they thank the referees for comments and assistance.

APPENDIX 1: SURVEY INSTRUMENT

The structure ofthe questionnaire included the follow-ing dimensions specific to the project. In the firstinstance, the size ofthe firm (SME), its location (byvulnerable region) and its industry (labour intensive)were aspects of the survey held constant by virtue ofthe authors' prior agreement about the significance ofthose issues. The questionnaire began having identifiedthe firm, its location and industry and had eight separatesections as summarized below:

1. Background: the questionnaire began with theresponsibility of the respondent, as well as the sizeof the firm as indicated by sales and the numberof employees. The opportunity was taken to askquestions concerning the gender make up of thefirm's employees, the split between full- and part-time employment, and the geographical origin ofemployees.

2. Ownership: the questionnaire sought informationabout the formal ownership structure ofthe SME,recognizing that these types of organizations vary inform and in practice between jurisdictions and inrelation to related firms. In theory and in practice,the authors began w ith the assumption that thenature of ownership and management matters agreat deal in investment decision-making.

3. Supply and distribution links: it is widely recognizedthat SMEs are more often than not componentparts of extended supply and distribution networks.The questionnaire sought to elicit informationabout the geographical and functional origin anddestination of inputs to production and outputs ofproduction.

4. Competitive strategies: recognizing the rapidlychanging competitive circumstances of Europeanindustries, their regions and firms, questions wereasked so as to understand better the pattern of totalsales over the past 3 years as well as the nature ofSME response to changes in total sales.

5. Investment strategies: information about investmentstrategies w as needed on the nature of investment,the sources of funds used to finance investment, theeffect of investment on employment and the desiredresults of investment for the firm.

6. Technology adoption: having established the com-petitive circumstances of the firm, as well as theinvestment strategies of SMEs, the next sectionlinked explicitly to the adoption of technology(broadly defmed). Much is made of technology inthe economic development and regional develop-ment literature, while at the same time, there is atendency to exaggerate the nature and sophisticationof technology for SMEs. The questionnaire madeallowance for the simplest types of technology,including communication technologies.

7. Labour and technological change: the effect oftechnological change on employment was explicitly


tackled. The goal was to elicit information on thedemand for labour in terms of the quantity oflabour, skill and productivity.

8. Policy and pohcy institutions: the connection wasmade between decisions taken and identified inprevious sections and the relevance and usefulnessof existing regional, national and EU policyinstruments.

Reproduced below is the final version of a questionasked about the origin of firm employees. As initiallyframed, respondents ranked options using a 5 (mostimportant) to 1 (least important) and 9 (irrelevant)code. The revised question asked respondents aboutthe likely actions they may have taken, while neverthe-less asking for a judgement about the most significantsource of firm's employees:

8. Which of the following sources of employeesdoes the firm use?

(a) family memhers(h) local community(c) people from outside the region(d) joint venture partner(e) parent firm

9. Ofthe optionsmost important?

Yes No

selected (YES) for question 8, which is the

NOTES

1. Objective 1 regions are EU regions with a GrossDomestic Product per head less than 75% of the Com-munity average. More than two-thirds of the moniesallocated by the EU to regional development is investedin Objective 1 regions. Objective 2 regions are EUregions with a Gross Domestic Product per head closeto the Community average and which are experiencing'structural difficulties', often the decline of industrialactivity, leading to higher than average levels of un-employment (for a more detailed description ofthe EUregional policy context, see http://www.europa.eu.int/comm/regional_policy).

2. In sum, about 25 sets of questions and answers perrespondent, about 20 respondents per industry and coun-try. After ensuring the integrity of the derived data, thepooled database is comprised of 331 completed files withthe balance of observations drawn from Greece, Irelandand the UK. The completed data files from Italy andSpain are significantly smaller in number (see below).

3. HAWAWINI et al. (2000) observed that many studiesof firms' profitability find that firm-specific effects faroutweigh industry effects. They looked closely at a largeUS data set covering about 1000 firms from three-digitSIC industries over nearly 20 years. By identifying keyfirms within these industry groups, they distinguishedthe effects of firm and industry effects with and withoutthese firms. On balance, having excluded these types offirms from their sample, they identified a significantindustry effect. The present paper does not make anysuch adjustments but finds, nevertheless, significant

country and industry effects behind firm-competitivestrategies.

4. Competitiveness is a controversial issue in the socialsciences. In part, this is because the very idea of competi-tiveness contradicts the tenets of neoclassical economics:'in a world inhabited by "representative firms" operatingunder perfect information and with no scale effects . . .the term competitiveness is meaningless' (REINART,1995, pp. 26-27). WhUe KRUGMAN (1993) believesthat the term should be expunged from social science.PORTER (1990) claims that competitiveness can bethought as a proxy for labour productivity. In this regard,competitiveness is considered here to be about thecapacity of firms, industries, regions and countries (beingsites of production) to make profits in markets subjectto international competition. Competitiveness is animportant characteristic of places and industries in theglobal economy, a conceptual tool whereby one mayunderstand the generation of uneven patterns of wealthand living standards across regions and nations.

5. There is an enormous literature devoted to region-specific case studies and the lessons to be drawn therein(STORPER and SALAIS, 1997). It is also an epistemo-

logical issue, being the basis of arguments in favourof capitalist diversity (CROUCH and STREECK, 1997:

C R O U C H et al, 2001).

6. In their Introduction, M A N K I W and R O M E R (1991)

emphasize research that helps explain how the existenceof 'small' imperfections may accumulate into systemicbarriers of price flexibility. The European context doesnot need such subtle theories and inferences: see the com-pelling arguments made by BLANCHARD and SUMMERS

(1987) and KRUEGER (2000) regarding the institutionallogic of European labour and product markets.

7. There are, of course, considerable practical issues in-volved in using common industry classifications acrossthe EU. Most obviously, there may be significant differ-ences within industries between countries and regionsdue to differences in the precise composition of each'industry'. Clearly, leather and footwear products varybetween Italy and Spain, for example. Even so, for allthe apparent significance of such composition effects,their practical importance must be assessed against thederived results and the compromises one must make inany comparative study (on the inevitable qualificationsthat attach to industry studies, see HAWAWINI et al.2000).

8. The design of the questionnaire was based on thefollowing four principles of comparative studies andrelated fieldwork: simplicity, allowing for variationsbetween jurisdictions and industries in terms of theassumed knowledge of the respondents; standardization,ensuring that the questionnaire was relevant to the firms,industries and regions of Europe while recognizing thatits implementation in different settings is likely to bringforth complications that arise from different customsand traditions; coherence, beginning with the openingquestions through to the final questions and addingimportant pieces of information in a sequence of ques-tions that in combination contribute to the realisationof the project; and parsimony, being concerned to keepto a minimum the number of questions asked, not onlyto ensure that the questionnaires could be implemented


according to the available time for each respondent, butalso according to the logic ofthe' project.

9. Note that focus is on those firms that survived over thepast 5 years. Missing, of course, are those firms that'died' before implementation ofthe present study.

10. The variables SAL, EMPL, TECH and LBSK werescaled using a common structure of 1-5, where 1 isdecreased greatly and 5 is increased gready. The othervariables were dichotomous variables set as 0 = NO and1 =YES (for more details, see the Appendix).

11. The present paper uses country rather than regionalaffiliation as a means of distinguishing the relative effectsof industry and geography on firm competitive strategy.

12.

It is possible, of course, that the 'country effect' is aproxy for the region effect or even a firm-specific effect.Given the aggregation ofthe data and the nature oftheeconometric procedures, it is not possible to make amore sensitive analysis of these issues. Those are left forthe case studies reported by CLARK et al. (2002).The present paper, however, excludes from its analysisthe results for firm ownership and the nature and extentof supply and distribution networks. Neither is direcdyrelevant to the task, although they may contain certainuseful insights. Discussion of these and other relatedissues are left for another time (CLARK et al. 2004).

REFERENCES

AMiHtJD Y. and KAMIN J. (1979) Revenue and profit maximization: differences in behavior by the type of control and bymarket power. Southern Economic Journal 45, 838-846.

B A R R O R . and S A L A - I - M A R T I N X. (1995) Economic Growth. McGraw-HiU, New York.BAUMOL W. J. (1959) Business Behavior, Value and Growth. Macmillan, New York.BLANCHARD O.J. and SUMMERS L. H . (1987) Hysteresis in unemployment, European Economic Review il, 288-295.BLANCHARD O.J. and KlYOTAKl N. (1987) Monopolistic competition and the effects of aggregate demand, American Economic

Review 77, 647-666.BREZIS E . S., K R U G M A N P. and T S I D D O N D. (1993) Leap-frogging in international competition: a theory of cycles in

national technological leadership, American Economic Review 83, 1211-1219.CLARK G. L. (1994) Strategy and structure: the domain of competition and the characteristics of sunk costs. Environment and

Planning A, 26, 9-32.CLARK G. L. and KIM W. B . (Eds) (1995) Asian NlEs in the Global Ecooomy. Johns Hopkins University Press, Baltimore.CLARK G. L. and TRACEY P. (2004) Global Competitiveness and Innovation. Palgrave Macmillan, Basingstoke.CLARK G. L., FELDMAN M . and G E R T L E R M . S. (2000) Economic geography: transition and growth, in CLARK G. L.,

FELDMAN M . and G E R T L E R M . S. (Eds) The Oxford Handbook of Economic Geography, pp. 3—17. Oxford University Press,Oxford.

CLARK G. L., PALASKAS T , TRACEY P. and TSAMPRA M . (2004) Market revenue and the scope and scale of SME networks

in Europe's vulnerable regions. Environment and Planning A 36, 1305-1326.CLARK G. L., TRACEY P. and LAWTON SMITH H . (2002) RASTEL Report to the European Union. School of Geography and

the Environment, University of Oxford, Oxford.C O O K E P. and M O R G A N K. (1997) The Associational Economy. Oxford University Press, Oxford.C R O U C H C., LE GALES P., TRIGILIA C . and VOELZKOW H . (Eds) (2001) Local Production Systems: Rise or Demise? Oxford

University Press, Oxford.C R O U C H C . and STREECK W. (Eds) (1997) Political Economy of Modern Capitalism: Mapping Convergence and Diversity. Sage,

London.DAVIS D. R . and W E I N S T E I N D. E . (1999) Economic geography and regional production structure: an empirical investigation,

European Economic Review 43, 379—407.

FELDMAN M . (2000) Location and innovation: the new economic geography of innovation, spiU-overs, and agglomeration, inCLARK G. L., FELDMAN M . and G E R T L E R M . S. (Eds) Tlie Oxford Handbook of Economic Geography, pp. 373-394. OxfordUniversity Press, Oxford.

G E R T L E R M . S. (2001) Best practice? Geography, learning and the institutional limits to strong convergence, JoMma/ of EconomicGeography 1, 5-26.

GREENWALD B . and STIGLITZ J. (1995) Labour-market adjustment and the persistence of unemployment, American EconomicReview 85, 219-225.

GuERRiERi P. and IAMMARINO S. (2002) Vulnerability and regions in the European Union: an exercise on tfie ItalianMezzogiorno. Working Paper 02-17. School of Geography and the Environment, University of Oxford, Oxford.

H A R T O . (1982) A model of imperfect competition with Keynesian features. Quarterly Journal of Economics 97, 109-138.H A R T M A N N G. B. and MASTEN J. (2000) Profiles of state technology transfer and its impact on small manufacturers, JoMma/

of Technology Transfer 25, 83-88.

HAWAWINI G., SUBRAMANIAN V and V E R D I N P. (2000) Is profitability driven by industry- or firm-specific factors? A new

look at the evidence. Working Paper 2000/80/FIN. INSEAD, Fontainbleau.H U D S O N R . (2002) Changing industrial production systems and regional development in the new Europe, Transactions, Institute

of British Geographers, n.s., 27, 262—281.

K R U E G E R A. (2000) From Bismarck to main street: the march to European Union and the labour compact. Working Paper7456. National Bureau of Economic Research, Cambridge, MA.

KRUGMAN P (1991) Geography and Trade. MIT Press, Cambridge, MA.K R U G M A N P. (1993) What do undergraduates need to know about trade?, American Economic Review 83, 23-26.

1100 Gordon L. Glark et al.

M A N K I W N . G. and R O M E R D. (Eds) (1991) New Keynesian Economics. MIT Press, Cambridge, MA.M A R T I N R . (2001) EMU versus the regions? Regional convergence and divergence in Euroland, JowrMa/ of Economic Geography

1, 51-80.P O R T E R M . (1990) The Competitive Advantages of Nations. Macmillan, London.R E I N A R T E . S. (1995) Competitiveness and its predecessors — a 500-year cross national perspective. Structural Change and

Economic Dynamics 6, 23-42.R I C C I L. A. (1999) Economic geography and comparative advantage: agglomeration versus specialization, European Economic

Review 43, 357-377.ROTEMBERG J. J. and SALONER G. (2000) Competition and human capital accumulation: a theory of interregional

specialization and trade. Regional Science and Urban Economics 30, 373-404.S T O R P E R M . (2000) Globalization, localization, and trade, in CLARK G. L., FELDMAN M . and G E R T L E R M . S. (Eds) The

Oxford Handbook of Economic Geography, pp. 146—167. Oxford University Press, Oxford.S T O R P E R M . and SALAIS R . (1997) Worlds of Production. Harvard University Press, Cambridge, MA.ViCKERS J. (1985) Deregulation and the theory ofthe firm. Economic Journal 95 (Suppl.), 138-147.

Global and Competitive Strategy

Documents

Transcript of Global and Competitive Strategy