UNIVERSITYOFREADINGANDRUTGERSUNIVERSITY › content › pdf › 10.1057 › palgrave.jibs.8490… · JOHNH.DUNNING inDunning,1998).Intheperiod1991-1996,64percentofglobalFDIinflows

Location and the Multinational

Enterprise: A Neglected Factor?

JohnH. Dunning*UNIVERSITYOF READING AND RUTGERS UNIVERSITY

This article first traces thechanging world economicscenario for international busi-ness over the past two decades,and then goes on to examine itsimplications for the location offoreign direct investment andmultinational enterprise activ-ity. It suggests that many of theexplanations of the 1970s andearly 1980s need to be modifiedas firm-specific assets havebecome mobile across naturalboundaries. A final section ofthe article examines the dy-namic interface between thevalue-added activities of multi-national enterprises in differentlocations.

INTRODUCTION

In1986, the economist WilfredJ.Ethier,in seeking to explain the existence of

multinational enterprises (MNEs), con-cluded that "internalization appears to beemerging as the Caesar of the OLI tri-

umvirate"(Ethier,1986, p. 803). I did notagree with this statement then; nor do Ido so now. The OLItriad of variables(ownership, location and internalization,discussed below) determining foreigndirect investment (FDI)and MNEactivitymay be likened to a three-legged stool;each leg is supportive of the other, andthe stool is only functional if the threelegs are evenly balanced. In so faras thethird leg completes this balancingit maybe regarded as the most important, butthere is no reasonto suppose one leg per-formsthis taskbetterthan another.

In the case of the eclectic paradigm,Iwould accept that the I component is thecritical leg, if, given the 0 advantagesoffirms and the L advantagesof countries,one is trying to explain why firms inter-nalize the cross-bordermarket for theseadvantages,ratherthan sell them or theirrightsto independent firms. But, I wouldaver it is no less correct to argue that,given its 0 specific advantages,the criti-cal choice of a multi-activity firm iswhether it should internalizeits interme-diate product markets within its homecountry or in a foreign country;and thatthe outcome of this choice is primarilydetermined by the costs and benefits of

* State of New Jersey Professor of International Business at Rutgers University, United States,and Emeritus Research Professor of International Business, University of Reading, UnitedKingdom.

JOURNAL OF INTERNATIONALBUSINESS STUDIES, 29, 1 (FIRST QUARTER 1998): 45-66. 45

Palgrave Macmillan Journalsis collaborating with JSTOR to digitize, preserve, and extend access to

Journal of International Business Studieswww.jstor.org

®

LOCATION AND THE MNE

adding value to these products in the twolocations. I say primarily because thegeography of international business activ-ity is not independent of its entry mode;nor, indeed, of the competitive advan-tages of the investing firms. This interde-pendence is particularly apparent whenone examines the dynamics of knowl-edge-intensive MNE activity.

In the 1960s, scholars, such as Ray-mond Vernon and his colleagues atHarvard (see especially Vernon, 1966,1974 and Wells, 1972) working on thedeterminants of FDI, gave pride of placeto locational variables, particularly thosedetermining the siting of U.S. marketseeking FDI by U.S. firms in advancedindustrial countries (see also the work ofsome European scholars, such as Banderaand White, 1968 and Scaperlanda andMauer, 1969). In the mid-1970s - apartfrom research on the internationalizationprocess of firms (see, for example,Johanson and Vahlne, 1977) - attentionswitched from the act of FDI per se to theinstitution making the investment. Herethe main focus of interest was why firmsshould choose to set up or acquire foreignvalue-adding activities, rather than exportthe intangible assets (or the right to usethese assets) underpinning such activi-ties, directly to foreign firms (see espe-cially the writings of Peter Buckley andMark Casson, J. C. McManus, Jean-

Frangois Hennart, Alan Rugman, andBirgitta Swedenborg, all of which arecited in Caves, 1982 and 1996).

While I would be the first to acknowl-edge the value of this approach inadvancing our understanding of MNE,qua MNEs, I believe that the contributionof the internalization school has donemore to explain the existence and growthof the multi-activity firm than that of theMNE per se. This is because, with rela-tively few exceptions,1 the transaction

and coordination costs identified witharm's-length intermediate product mar-kets have not, in general,been specific tocross-bordermarkets, or, indeed, to tra-versingspace.

The emphasis on the firm-specificdeterminants of international economicactivity,while still drivingmuch academ-ic research by scholars in businessschools, is now being complementedby arenewed interest in the spatial aspects ofFDI; and of how these affect both thecompetitiveadvantagesof firmsand theirmodes of entry into, and expansion in,foreignmarkets. Webelieve therearetwomain reasonsforthis. The firstis thatthechangingextent, characterand geographyof MNE activity over the past twodecades - itself a reflection of a series ofpath-breaking technological, economicand political events - is demanding anexplanation by international businessscholars. The second is that newresearch agendas, particularly those ofeconomic geographers, trade theoristsand international political economists,arenot only payingmore attentionto thespatial aspects of value-added activity,but are also seeking to incorporatetheseaspects into the mainstream thinkingabout the growth and competitiveness offirms,the relationshipbetween tradeandFDI, and the economic structure anddynamic comparative advantage ofregionsand countries.

This paper seeks to review some ofthese happenings, most of which comeinto prominence between the two edi-tions of the publicationof RichardCaves,Multinational Enterprise and EconomicAnalysis (1982 and 1996). To his credit,Richard Caves acknowledges many ofthese in his second (1996) edition. But,since much of his analysis relates to thework of scholars in the 1980s,2 his chap-ter on the internationalallocation of eco-

46 JOURNAL OF INTERNATIONALBUSINESS STUDIES

JOHN H. DUNNING

nomic activity (Chapter2) does not fullyembrace the events and academicresearch of the last decade or so. It isthese which will be the main concern ofthis contribution. The paperwill proceedin the following way. First it will brieflydescribe the changing global economicscenarioin which MNEactivityhas beenconducted since the mid-1970s, and alsothe variousstrandsof intellectualthoughtwhich have sought to explain this.Second, it will examine how the micro-locational determinants of internationalproduction have changed; and how thelocation portfolio of MNEs may itselfhelp promote their dynamic competitiveadvantages. Third, it will consider how,from a more macroeconomicstandpoint,the emergence of the MNE as a leadingvehicle of cross-bordertransactions hasaffected our thinking about the determi-nants of tradeand othernon-MNErelatedtransactions.

THE CHANGINGWORLD SCENARIOFOR INTERNATIONALBUSINESS

ACTIVITY

The last two decades have witnessed agradualmovementtowardsa world econ-omy characterizedby three features. Thefirst is the emergenceof intellectual capi-tal as the key wealth creating asset inmost industrialeconomies. In the l990s,the market value of industrial corpora-tions has been variously calculated (e.g.,by Blair, 1995, Handy, 1989 andEdvinsson, 1997) at between 21/2 and 5times the value of their tangible assets,compared with 11/2 times in 1982. Theannual capital expenditure on informa-tion technologyby U.S. corporationsnowexceeds that on production technology(Stewart,1997). The knowledge compo-nent of the output of manufacturinggoods is estimatedto have risen from 20percent in the 1950s to 70 percent in

1995 (Stewart, 1997); while those work-ers whose main task is to create newknowledge or disseminate information(viz. professional and technical workers,managers, sales and clerical workers -

the so called "white collar" workers)increased their share of the Americanlaborforce from42 percent in 1960 to 58percent in 1990, and this share is expect-ed to rise to more than 60 percent by2000.

A furtherindicatorof the rising signifi-cance of non-materialassets as creatorsorfacilitatorsof wealth is the growthof ser-vices, and particularly those which are,themselves, knowledge or informationintensive. In 1995, on average, servicesaccounted for 63 percent of the world'sgross national product (GNP),comparedwith 53 percent in 1980 and 45 percentin 1965 (World Bank, 1997). Insofar asknowledge intensive and knowledge sup-portingproductionhas its unique spatialneeds, and tends to requireresourcesandcapabilitieswhich MNEsare particularlywell suited to provide, it is not unreason-able to hypothesize that both these fea-tures will impinge on the geographicaldistributionof FDIandrelatedactivities.

Secondly, and even more transparent,is the increasingglobalizationof econom-ic activity, made possible, inter alia, byadvances in transport and communica-tions technologies and the reduction intrade and investment barriersthroughoutthe world (UNCTAD,various issues andWorld Bank, various issues). Over thelast two decades, the growth of worldtradehas consistently outstrippedthat ofworld output,while in the mid-1990s thesales of the foreign affiliates of MNEsexceeded the value of world tradeby 27percent (UNCTAD, 1997). Moreover,between one-third and one-half of tradein non-agriculturalproductsand betweenone-half and three-fifths of capital and

VOL. 29, No. 1, FIRST QUARTER,1998 47


knowledge flows are currently internal-ized within MNEs.3

At the same time, the ease at whichMNEs can transfer intangible assetsacross national boundaries is being con-strained by the fact that the location ofthe creation and use of these assets isbecoming increasingly influenced by thepresence of immobile clusters of comple-mentary value-added activities. This isparticularly the case where the transac-tion costs of traversingdistance are high,orwhere the transactionalbenefits of spa-tial proximity are significant.4 Thuswhile globalizationsuggeststhatthe loca-tion and ownership of production isbecominggeographicallymore dispersed,other economic forces are making for amore pronounced geographical concen-tration of such activity both within par-ticular regions and countries.5 In thewords of Ann Markusen (1996), theseevents arepresentingscholarsand policymakerswith a paradox of "sticky placeswithin slipperyspace."

The third feature of the contemporaryglobal economy is the emergenceof whatmay be called "alliance" capitalism(sometimes called relational, collective,stakeholderand collaborativecapitalism

see Dunning, 1995). While retainingmany of the characteristicsof hierarchicalcapitalism, the distinctive feature ofalliance capitalism is the growing extentto which, in orderto achieve theirrespec-tive objectives, the main stakeholdersinthe wealth-seekingprocess areneeding tocollaboratemoreactively and purposeful-ly with each other. Such collaborationincludes the conclusion of closer, contin-uing, and more clearly delineated intra-firmrelationships,e.g.,between function-al departmentsand between managementand labor;the growthof a varietyof inter-firm cooperative agreements, 6 e.g.,between suppliers and customers and

among competitors;and the recognitionby governments and firms alike of theneed to work as partnersif the economicgoals of society (forwhich the formerareultimately responsible) are to be bestachieved.

Once again, the growing propensity offirms to engage in cross-borderallianceshas implicationsnot just forthe modalityby which knowledgeand otherintangibleassets are transferred across nationalboundaries,but for the location of value-added activities - especially high valueasset-augmentingactivities.

Underpinning and reinforcingeach ofthe events just described are two otherfactorswhich also have had a profoundeffect on both the macroand micro geog-raphyof MNEs. The firstis the advent,inthe 1980s, of a new generationof techno-logical advances which, according toAlan Greenspan (in a speech given toNew Yorkbankersin April 1997)areonlynow, in the later 1990s, fully bearingfruit.7 The second factor is the renais-sance of the market economy, and theconsequential changes in the macroeco-nomic policies and macro-organizational(micro-management)strategies of manynational governments. This is mostvividly demonstratedby the happeningsin Chinaand Centraland EasternEurope,but, almost as far reaching, is the reap-praisal of the role of the State and mar-kets in economic developmentnow beingplayed out in India, and in severalAfrican and Latin American economies(World Bank, 1997). Both factors havehad a majorimpact on the economic andpolitical riskassessmentof FDIby MNEs.

THE CHANGINGGEOGRAPHYOFMNE ACTIVITY

The developments just describedhaveall impactedon the geographyof FDIandMNEactivity (as describedin more detail


JOHNH. DUNNING

in Dunning, 1998). In the period 1991-1996, 64 percent of global FDI inflowswere received by the developed coun-tries, 33 percentby developing countriesand 3 percent by Central and EasternEuropeancountries. The correspondingpercentages for the period 1975-1980were 77 percent,23 percentand less than0.1 percent (UNCTAD, 1997). No lessnoticeable have been the changes in thedistributionof inbound FDIwithin theseregions. While the shares of WesternEuropeand the United States,c.f. all FDIin developed countries, have remainedbroadlythe same,8 those within develop-ing countries have markedly changed.For example, between 1975-1980 and1991-1996, South, East and SoutheastAsia (including China and India)increased their share of inbound invest-ment to developing countries from 26percentto 62 percent,while that of LatinAmerica and the Caribbeanfell from 53percentto 34 percent.

It is perhaps worth observing thatalthoughthe shareof inbound FDIto thegross fixed capital formationof the coun-tries more than doubled between the sec-ond half of the 1970s and the firsthalf ofthe 1990s (UNCTAD,1988, 1996a), thechanging geographyof FDIparallels rea-sonably well that of all investment, inde-pendently of its ownership. Between1975 and 1980, and 1990 and 1995, forexample, the shareof world inbound FDIaccountedforby developed countriesfellfrom 78 percentto 70 percent,while thatof world gross fixed capital formation(including that part financed by foreignfirms)fell from 84 percent to 73 percent.The corresponding figures for all devel-oping economies were 21 percent and 30percent, and 15 percent and 26 percent;and for Asia 7 percent and 19 percent,and 7 percent and 19 percent. Althoughthere are differences in the geographyof

FDIwhich can be specifically attributedto the political or economic conditions inthe host country9 - and it is most cer-tainly the case that the geographyof out-wardFDIis quite stronglycountryspecif-ic10 - the data suggest that many of thefactorswhich explain the location of FDImay not be unique to its country of ori-gin. We shall not elaborateon this pointhere; but it is, perhaps, worthy of moreattention.

THE MICRO-ECONOMICSOF THELOCATIONOF MNE ACTIVITY

With the caveat of the last paragraph,we now consider how scholarly thinkingabout the location of MNEs has evolvedover the last two decades. Incidentally,we suspect that the fact that this subjecthas not been given much attention byinternational business scholars is partlybecause scholars have believed that theprinciples underlyingthe locational deci-sions of firmswithin nationalboundariescan be easily extended to explain theircross-border locational preferences;11and partly because economists wereeither generally satisfied with existingexplanations,or just not interestedin thesubject. Certainly,until the early 1990s,there was little in common between themethodologies of international tradeeconomists and locational economists,excepting the work of BertilOhlin (1933)and his successors. This was primarilybecause the formerwere concerned withcountry-specific general equilibriummodels or models under very restrictiveconditions, whereas the latterwere main-ly interested in firm or industry-specificpartial equilibrium models with fewerconstraints (Krugman, 1993).

Earlier in this paper, we identifiedthree major developments in the globaleconomy which have impinged uponboth the capabilities and strategies of

VOL. 29, No. 1, FIRSTQUARTER,1998 49


MNEs, or potential MNEs, and the loca-tional attractions offered by particularcountriesto mobile investors. In particu-lar, we emphasized first the growingsignificance of firm-specific knowledge-intensive assets in the wealth-creatingprocess, and the kind of customizedassets, e.g., skilled laborand public infra-structure, which needed to be jointlyused with these assets if they were to beeffectively harnessed and deployed;12secondly, the reduction of many naturaland artificial impediments to trade, butthe rise of other spatiallyrelatedtransac-tion costs; and thirdly, the growingneedand ease with which firms are able tocoordinate their cross-border activitiesand formallianceswith foreignfirms.

Some of these factorshave led firmstoown and concentrateparticulartypes ofvalue-added activities within a limitednumber of locations; others have ledthem to disperse such activities acrossseveral locations. Some have favored arealignment of MNE activity towardsadvanced developed economies; othershave favoreda location in emergingmar-ket economies. All are symptomaticof achanging internationaldivision of labor,which, because of their increasingrole inthe world economy, and their need tocapturethe economies of interdependentactivities,MNEshave helped to fashion.

The literatureon the locational prefer-ences of foreigndirect investors has longacknowledgedthat these will not dependon the types of activities in which theyare engaged, but on the motives for theinvestment and whether it is a new or asequentialone. Differentkinds of invest-ment incentives are needed to attractinbound MNE activity of a natural-resource-seeking,c.f. that of a market-orefficiency-seeking,kind. Export-orientedFDIis likely to be less influenced by thesize of local marketsthan is import-sub-

stituting FDI. Investment in R&D facili-ties requires a different kind of humanand physical infrastructure than invest-ment in assembling or marketing activi-ties, and so on.

But, perhaps, the most significantchange in the motives for FDI over thelast two decades has been the rapidgrowth of strategic asset-seeking FDI,which is geared less to exploiting anexisting 0 specific advantage of aninvesting firm, and more to protecting, oraugmenting, that advantage by the acqui-sition of new assets, or by a partneringarrangement with a foreign firm. In someways, such FDI is similar in intent to thatof a natural resource-seeking investmentin earlier times but, its locational needsare likely to be quite different. Partly thisis because it is frequently motivated bystrategic considerations (especially in oli-gopolistic industries),and partlybecausethe availability of the assets sought, viz.technical knowledge, learning experi-ences, management expertise and organi-zational competence, tend to be concen-trated in advanced industrial countries,or the largerdeveloping countries. Thegrowth of strategic asset-seeking FDI inrecent years is best demonstrated by theincreasing role of mergers and acquisi-tions as modalities of FDI. According toUNCTAD (1997), between 55 percent and60 percent of FDI flows over the period1985-1995 was accounted for by mergersand acquisitions. Most of these were con-centrated within North America, Europeand Japan, and in knowledge and infor-mation-intensive sectors.

The locational preferences of firmsmaking more traditional forms of FDIhave also changed - as, indeed, have theattitudes of recipient countries to theseinvestments. We might mention two ofthese. First, as foreign affiliates havebecome more embedded in host coun-


JOHN H. DUNNING

tries, this has led to a deepening of theirvalue chains, and a propensity for themto engage in higher-order(e.g., innovato-ry) activities. This fact has been docu-mented in numerous studies both on thegeographicaldistributionof R&Dand onthat of patents registered by MNEs (asrecent examples of these, see Dalton andSerapio (1995), Almeida (1996), Dunning(1996), Kuemmerle (1996), Shan andSong (1997), and various studies of JohnCantwell and colleagues, e.g., Cantwelland Harding(1997), and Bob PearceandMarinaPapapanastassiou,e.g., Papanas-tassiou and Pearce (1997), of theUniversity of Reading). Inter alia, theCantwelland Hardingstudy showed thatbetween 1991 and 1995, 11 percentof theU.S. registered patents of the world'slargestfirmswere attributableto researchlocations outside the home countryof theparent company. Only in the case ofJapanthere was not a rise in the propor-tion of patentsregisteredby foreignaffili-ates since the early1970s (fora moregen-eral discussion on asset-augmentingFDI,see an interestingdissertationby Wesson,1993). So far,however, this tendency ofengaging in higher-order activities hasbeen largelyconfined to developed coun-tries. In 1994, for example, some 91 per-cent of the foreign R&Dundertaken byU.S. MNEs was located in developedcountries, compared with 79 percent oftheir total foreign sales (Mataloni andFahim-Nadar,1996).

Second, the location-specific assetswhich MNEs perceive they need to addvalue to the competitive advantagestheyare exporting (via FDI) are changing astheir downstreamactivities arebecomingmore knowledge intensive. Various sur-veys have demonstratedthat, except forsome labor or resource investments indeveloping countries, MNEs are increas-ingly seeking locations which offer the

best economic and institutional facilitiesfor their core competencies to be effi-ciently utilized. For example, in a fieldstudy by Fabrice Hatem (1997), apartfrom market access and market growth,economic and institutionalfacilities werenot only valued much higher than tradi-tional criteriaof access to raw materials,cost of labor, and fear of protectionism,but in all cases they were also thoughttoincrease in significanceover the five yearperiod 1996-2001. There is a suggestion,too, that the presence of other foreigninvestors in a particular country isbecoming more significant, both as an"investment-stalk"or signaling effect tootherforeignfirms less familiarwith thatcountry(Srinivasanand Mody, 1997;Liu,1998), and as an agglomerativemagnetbywhich firms benefit from being part of ageographicalnetworkor cluster of relatedactivities and specialized support ser-vices. In a study of the location patternsof U.S. MNEs between 1982 and 1988,Wheeler and Mody (1992) identifiedthree agglomeration benefits, viz. infra-structurequality, degree of industrializa-tion, and existing level of FDI. Theyfound that these exhibited a high degreeof statistical significance and had largepositive impacts on investment (p. 66).In an analysis of Swedish outbound FDI,over the period 1975-1990, Braunerhjelmand Svensson (1995)confirmeda positiveand significant statistical relationshipbetween thatvariableand the presence ofpecuniary externalities associated withdemand and supply linkages, includingthe diffusion of knowledge, e.g., spillovereffects, resulting from a clustering ofrelatedfirms.

A more formal examination of thechangingnatureand significanceof exter-nal economies, and of how these arelead-ing to a more concentratedpatternof cer-tain kinds of FDI - particularly that of

VOL. 29, No. 1, FIRST QUARTER, 1998 51


strategic asset-seeking investment inknowledge-intensive-sectors - is set outin Krugman (1991). Indeed, it was hisstudy which helped spark off the fruitfuldialogue now taking place betweenindustrial geographers, economists andbusiness analysts. Though this dialogueis principally concerned with the role ofsub-national spatial units as repositoriesfor mobile investment, it is also offering anumber of valuable insights on thechanging role of transportation and com-munication costs as they affect the coor-dination and supply of end productsfrom existing agglomerations, and thedecentralization of intermediate produc-tion;13 and also on the changing compet-itive advantages of regions - particularlyas they impinge upon spatial transactioncosts and dynamic external economies,such as those to do with complex tech-nologies, uncertain or unpredictable mar-kets, inter-active learning, face-to-facediscussions and the exchange of uncodi-fiable knowledge (Florida, 1995; Storperand Scott, 1995).

Certainly the incentives offered byregional authorities within the EuropeanUnion (EU) and of states within theUnited States, have been shown to be adecisive factor in influencing the intra-regional location of inbound MNE activi-ty (for some interesting case studies, see,for example, Donahue, 1997 and Ohmae,1995). There is also a good deal of casualevidence to suggest that the promotionalcampaigns and incentives - in the formof the speedy processing of planningapplications, land grants, subsidizedrents, tax holidays and generous invest-ment allowances - offered by local orregional development agencies to attractFDI tend to resemble those of "locationtournaments"14 (Taylor, 1993, UNCTAD,1996b). Again, the experiences of theUnited States and the EU - or, indeed, of

some of the larger countries in the EU,e.g., the United Kingdom - are salutaryin this respect.

In Table 1, we attempt to summarizesome of the differences between the kindof variables posited to influence the loca-tional decisions of MNEs in the 1970s -

most of which are well documented inChapter 2 of Multinational Enterpriseand Economic Analysis - and thosewhich scholars are hypothesizing, andfield research is showing to influencethese same decisions of MNEs in the1990s. In doing so, we have separatelyclassified the four main kinds of FDIidentified earlier. However, we readilyaccept that other contextual variables,e.g., size of firm, degree of multinationali-ty, country or region of origin and desti-nation and industry, insofar as these havedifferent situational needs, may be noless significant.

The contents of the table are largelyself-explanatory, but we would highlightjust four main findings. The first is thechanging role of spatial transaction costs,which reflect both the liberalization ofcross-border markets and the changingcharacteristics of economic activity.While, in general, the reduction of thesecosts has led to more aggressive market-seeking FDI, and has promoted a welfare-enhancing international division of labor,it has also favored the spatial bunching offirms engaged in related activities, so thateach may benefit from the presence of theother, and of having access to localizedsupport facilities, shared service centers,distribution networks, customizeddemand patterns and specialized factorinputs (Maskell, 1996; Rees and McLean,1997).

The second finding is that the comple-mentary foreign assets and capabilitiessought by MNEs wishing to add value totheir core competitive advantages are


JOHN H. DUNNING

TABLE1Some Variables Influencing the Location of Value Added Activities

by MNEs in the 1970s and 1990s

Type of FDI In the 1970s In the 1990s

A. Resource 1. Availability,priceand qualityof nat- 1. As in the 1970s, but local opportunitiesforSeeking uralresources. upgradingqualityof resourcesand the process-

2. Infrastructureto enableresourcesto ing and transportationof theiroutputis a morebe exploited, and productsarising importantlocationalincentive.

fromthem to be exported. 2. Availabilityof local partnersto jointlypromote

3. Governmentrestrictionson FDI knowledgeand/orcapital-intensiveresource

and/oron capitaland dividend remis- exploitation.sions.

4. Investmentincentives, e.g.. tax holi-days.

B. Market Seeking 1. Mainlydomestic,and occasionally 1. Mostlylargeandgrowingdomesticmarkets,and(e.g., in Europe)adjacentregional adjacentregionalmarkets(e.g.,NAFTA,EUetc.).markets. 2. Availabilityand priceof skilled and professional

2. Realwage costs;materialcosts. labor.

3. Transportcosts; tariffandnon-tariff 3. Presenceand competitivenessof relatedfirms,tradebarriers. e.g., leadingindustrialsuppliers.

4. As A3 above,but also (whererele- 4. Qualityof nationaland local infrastructure,andvant)privilegedaccess to import institutionalcompetence.licenses. 5. Lessspatiallyrelatedmarketdistortions,but

increasedrole of agglomerativespatialeconomies and local servicesupportfacilities.

6. Macroeconomicand macro-organizationalpoli-cies as pursuedby host governments.

7. Increasedneed forpresenceclose to users inknowledge-intensivesectors.

8. Growingimportanceof promotionalactivitiesbyregionalor local developmentagencies.

C. Efficiency 1. Mainlyproductioncost related(e.g., 1. As in the 1970s,but moreemphasis placed onSeeking labor,materials,machinery,etc.). B2, 3, 4, 5 and 7 above,especially forknowl-

2. Freedomto engagein tradein inter- edge-intensiveand integratedMNEactivities,mediateand final products. e.g., R&Dand some office functions.

3. Presenceof agglomerativeeconomies, 2. Increasedrole of governmentsin removinge3g.,exportprocessingzones. obstaclesto restructuringeconomic activity,and

facilitatingthe upgradingof humanresourcesby4. Investmentincentives, e.g., tax breaks, appropriateeducationaland trainingprograms.

accelerateddepreciation,grants sub- 3. Availabilityof specialized spatialclusters,e.g.,sidized land. science and industrialparks,service support

systems etc.;and of specialized factorinputs.Opportunitiesfornew innitiativesby investingfirms;an entrepreneurialenvironment,and onewhich encouragescompetitivenessenhancingcooperationwithin and between firms.

D. Strategic Asset 1. Availabilityof knowledge-related 1. As in the 1970s,butgrowinggeographicaldisper-Seeking assets and marketsnecessaryto pro- sion of knowledge-basedassets,andneed of firms

tect or enhance0 specific advantages to harnesssuch assetsfromforeignlocations,of investing firms- and at the right makesthis a moreimportantmotiveforFDI.price. 2. The price and availabilityof "synergistic"assets

2. Institutionaland othervariablesinflu- to foreigninvestors.encing ease ordifficultyat which..such assets can be acquiredby foreign 3. Opportunitiesoffered(oftenby particularsub-ssrmsc nationalspatialunits) forexchangeof localizedfirms. tacitknowledge,ideas and interactivelearning.

4. Access to differentcultures,institutionsand sys-tems;and differentconsumerdemandsand pref-erences.

VOL. 29, NO. 1, FIRSTQUARTER,1998 53


increasingly of a knowledge-facilitatingkind, and that this is particularly the caseas their affiliates become more firmlyrooted in host economies (Grabher,1993).Examples include the deepening ofvalue-added activities by Japanese manu-facturing subsidiaries in Europe andNorth America. An exception to this find-ing is some low value-adding activities inthe least developed areas of the world.

The third finding is that as strategicasset-acquiring investment has becomemore important, the locational needs ofcorporations have shifted from those todo with access to markets, or to naturalresources, to those to do with access toknowledge-intensive assets and learningexperiences, which augment their exist-ing 0 specific advantages.

The fourth finding is that much of therecent FDI in developing countries isprompted either by traditional market-seeking motives (e.g., as in the case ofChina, Indonesia and India), or by thedesire to take advantage of lower (real)labor costs, and/or the availability andprice of natural resources. Yet, eventhere, where firms have a choice, thephysical and human infrastructure,together with the macroeconomic envi-ronment and institutional framework ofthe host country, tend to play a moredecisive role than they once did.

MACROECONOMICASPECTS OF THECHANGINGINTERNATIONAL

ALLOCATIONOF ECONOMICACTIVITY

In the previous section, we set outsome of the reasons for the changing loca-tional patterns of MNE activity over thepast two decades. We concluded thatdevelopments in the global economy overthese years had not only opened up orenlarged markets for products normallysupplied by MNEs, but, by affecting theproduction and transaction costs of FDI,

had markedly influenced its industrialstructure and geography. In general, the1990s are witnessing a closer integrationin the international value-added activitiesof MNEs. In the case of some kinds ofFDI, falling material, transportation andcommunication costs, and rising transac-tional benefits arising from the commongovernance of interdependent activitieshave made for a more concentrated pat-tern of FDI, both between and withinregions and/or countries. In other cases,however, the emergence of new - andoften important - markets, and the lower-ing of tariff and non-tariff barriers havemade for a more dispersed pattern of FDI.

We now turn to consider some macro-economic, or country specific, issues. Inparticular, we wish to address two ques-tions. First, to what extent is the chang-ing locational pattern of FDI affecting ourunderstanding about the determinants ofthe optimal international allocation ofeconomic activity; and second, how far,in light of the growing significance andintegration of MNE, does one need toreconsider the policy implications fornational and regional governments asthey seek to advance their particular eco-nomic and social objectives?

Until the 1950s, most explanations ofthe allocation of economic activity werebased on the distribution of naturalresources - especially labor, land andfinance capital. The principle of compar-ative advantage espoused that countriesshould specialize in the production ofthose products which required theresources and capabilities in which theywere relatively the best endowed; andtrade these for those which requiredresources and capabilities in which theywere relatively poorly endowed. Thiswas the basis for a general equilibriummodel of trade. Its restrictive assump-tions - viz. perfect competition, the


JOHN H. DUNNING

immobility of factors, homogeneity oftraded products, constant returns to scaleand zero transportation costs - as recent-ly reiterated by Krugman (1993), are wellknown. In that model, there was little orno room for innovatory activities, or forthe deployment of such created assets asintellectual capital, organizational exper-tise, entrepreneurship and interactivelearning, either by countries or firms; andeven less for the distinctive characteris-tics of MNEs.

Over the last four decades, theserestrictions have been gradually relaxedin three main ways. First, independentlyof the work of scholars on FDI and MNEactivity, there has been a growing appre-ciation by trade economists of the need toincorporate such variables as economiesof scale, fabricated assets, learning experi-ences and market structure into theirmodels, and to recognize that the role ofthese varies with types of economic activ-ity. It is, for example, now generallyaccepted that different parts of the valuechain may be distributed between coun-tries, or regions within countries, accord-ing to their knowledge, capital, naturalresource and labor content, and to thegeography of these inputs. Second, moreattention is now being paid to the extentto which the external economies whicharise from the clustering of related activi-ties, may lead to a concentration of eco-nomic activity in certain countries orregions. Third, more recognition hasbeen given to the differences in consumertastes between countries, while, verygradually, institutional factors, such asthose specific to the multi-activity ormulti-firm, and to the role of govern-ments have begun to be acknowledged.

In incorporating these changes intotheir thinking, the proponents of the posi-tive theory of trade are now able to offer amore realistic explanation of the interna-

tional allocation of economic activity;while, from a normative viewpoint,though dented, the principle of compara-tive advantagestill has much going for itas a guiding light as to how best to allo-cate scarce resources between countries(Wood, 1993). This is particularly thecase when it is widened to embracecreat-ed assets, including those which areinsti-tutional, policy and culturally related(Lipsey, 1997).

However,a second intellectual lacunaeremains, which makes it difficult to rec-oncile the approaches of location theo-rists and international trade economistsin explaining the internationalallocationof production. This is the presence- andthe increasing presence - of the MNE,whose centralfeatureis its common own-ership of cross bordervalue-addingactiv-ities. Herewe need to turn once againtothe work of the internalization scholars.For to explain how MNEs, qua MNEs,affect the international location of eco-nomic activity, we need to consider howthey differ from uni-national firms.Otherwise,one should be able to use thetenets of contemporarytradeand/or loca-tion theory to explain such activity. It ishere that researchby internationalbusi-ness scholarsis particularlyrelevant.

An earliersection of this papersuggest-ed that the changes in the geography ofFDI over the last two decades have beenbroadly in line with that of the capitalexpenditures of all firms. This couldmean that the ownership or multination-ality of firms was not a significant vari-able in explaining such changes;and thattrade in intermediate or final productsinternalized,and/orcontrolled,by MNEs,is no differently determined than tradebetween independent firms, i.e., arm's-lengthtrade.

However, as copious research shows(asreviewed, for example, in Caves,1996

VOL. 29, No. 1, FIRSTQUARTER, 1998 55


and Dunning, 1993), the main impact ofthe foreign-ness, or multi-nationality, offirms has not been on the level of eco-nomic activity and/or trade of the coun-tries in which they operate, but on thestructure of these variables. From thevery earliest of studies on FDI, scholarshave shown that the foreign affiliates ofMNEs tend to be concentratedin differ-ent industrial sectors than do their in-digenous counterparts. Since each sectoris likely to have its distinctive locationaland trading propensities, it follows thatFDIwill have a differentialimpacton thegeography of economic activity. Some-times, this impact will reflect the charac-teristics of the country of the investingfirms, e.g., JapaneseFDIin the Europeanauto and electronics industries in the1980s; sometimes a very unique competi-tive advantage,or set of advantages;andsometimes their pattern and degree ofmultinationality. For it is the geographi-cal diversityof asset-agmentingand asset-exploitingFDI,and the costs and benefitsassociated with their common gover-nance, which is one of the singular fea-tures of contemporary MNE activity -especially in developed countries.

Scholars, such as Bruce Kogut, recog-nized these specific attributes of MNEsmany years ago (see, for example, Kogut,1983 and 1985), but as the degree,scopeand intensity of the foreignoperationsoffirms has increased over the last decade(as demonstrated, for example, in theannual World Investment Reports ofUNCTAD),and as these are now used toharness new resources, capabilities andmarkets,as well as to exploit the existingOadvantagesof firms,so have these par-ticular qualities of multinationalitybecome moreprominent.

Though such qualities can be readilyembraced into location theory, they areless easily incorporatedinto generalequi-

librium trade models. Primarily,this isbecause, unlike industrial organizationtheory,tradetheoryhas not come to gripswith the multi-activity firm, or multi-plant production, or has embracedinno-vation into its thinking (one notableexception is that of Grossman andHelpman,1991). Recentpapersby JamesMarkusen (1995) and Markusen andVenables (1995) have made a braveattempt to integrate the OLIframeworkparadigmof internationalproductionandthe newer models of trade (i.e., thoseembracing firm-specific economies ofscale, product differentiationand imper-fect competition), but they tend to con-centrateon how the cross-borderspecial-ization of specific knowledge intensiveactivities may differfromthat predicatedby traditionaltrade theory. In a similarvein, research by Brainard (1993) andHorstmanand Markusen(1992) has con-cluded that MNE-relatedproductionwillbe in equilibrium when firm-level fixedcosts and spatial transaction costs arelargerelative to plant level economies.15None of these approaches,however, fullytakesaccountthe key propertiesof multi-nationality, as distinct from the foreignownership of firms. While embracingsome of the characteristicsof internalizedmarketsfor 0 specific assets, they ignoreothers- and especially those which else-where we have referredto as transaction-cost-minimizing0 advantages.16

Considering the normative implica-tions of the workof Markusenand others,and using the language of traditionaltradetheory,we might say that it will beto the benefit of countries if their firmsengagein outwardFDIin two very differ-ent situations. The first is where the uti-lization of their 0 specific advantages,the productionof which is relativelywellsuited to the resourcesand capabilitiesofthe home countries,is best undertakenin


JOHNH. DUNNING

a foreign country (or countries)17 andwithin the same firm (i.e., the benefits of"first best" internalized intermediateproduct markets exceed those of "firstbest"arm's-lengthtransactions).The sec-ond is when to protect or augment theirglobal competitive advantages, firmsengagein buying assets in a foreigncoun-try (or countries)more favorableto theircreation,but not to their deployment. Bycontrast, a country will benefit frominward direct investment when it has acomparativeadvantagein addingvalue tothe services of the imported createdassets - againwithin the investing entity- rather than producing these assetsitself, or where a foreignfirm chooses tobuy assets created in the country (at theright price) and utilize these assets in aforeigncountry(orcountries).

In most cases, given the presence ofMNEs, the recipe for an optimal alloca-tion of economic activity between coun-tries is quite similarto that in a world inwhich there is no FDI. But, the relativeroles of markets,hierarchiesand govern-ments in this recipe arelikely to be differ-ent. In conditions otherthan that of per-fect competition, hierarchies, or heterar-chies - in the guise of multi-activityand/or multi-national firms - may be amore efficient coordinator of resourcesand capabilities than arm's-length mar-kets (Caves, 1996). This is particularlylikely to be so in a dynamic knowledge-based economy in which some of theingredients of endemic market failure,and particularly those of uncertainty,irregularity, complexity, externalities,scale economies, vertical integration,andthe interdependenceof markets,are pre-sent, as it is these which tend to generatethe kind of value-added activities whichcan be coordinated more efficientlyundera single governance. In such cases,and providing that the final goods' mar-

kets servedby MNEsare contestable,andnational governments pursue positiveand non-distorting market facilitatingmacro-organizationalpolicies (Dunning,1997b), MNEs may act as surrogatesformarkets. By internalizing intermediateproduct markets, they may help protector enhance, ratherthan inhibit, the effi-ciency of final goods'markets.

While not wishing to undervalue therole of governments in curtailing theanti-competitive behavior of firms, andin pursuing market friendly macro-orga-nizational strategies,we believe that con-temporarychanges in the ways in whichresources and capabilities are managedare facilitating a more appropriate bal-ance between cross-border hierarchical(i.e., internalized) and external markettransactions. Perhaps, the one area forpotential concern is the widespreadgrowth of international mergers andacquisitions and strategic alliances(UNCTAD,1997). Insofar as these mayassist firms to be more innovatory,entre-preneurialand competitive in globalmar-kets, they areto the good;but, where theymake it easier for companies to engageinstructurallydistortingbusiness practices,they need to be carefullymonitored.

The unique impact of MNEs on theinternational allocation of productionrests on the extent to which the internal-ization of cross-borderintermediateprod-uct markets produces a different andmore efficient structure of economicactivity than would otherwise haveoccurred. Hereinlies an interestingpara-dox. On the one hand, the liberalizationof markets and the reduction of somekinds of spatial costs areeasingthe trans-border movement of goods, intangibleassets and services. On the other,techno-logical and organizationalchange, when-ever it enhances the interdependence ofvalue-added activity, is encouraging



internationalproduction to be undertak-en within plants and firms under thesame ownership, and for at least some ofthis productionto be spatiallyconcentrat-ed. Itwould seem thatas fastas structur-al and distance-relatedmarketfailuresareremoved, others, making for internalizedintermediate product markets anduntraded spatial interdependencies arebecomingmoreimportant.

Hints of this "new"internationaldivi-sion of labor are shown not only by thegrowing participationof MNEsin globalproduction - as described earlierin thispaper- but also by their increasingshareof world export markets, at least in themanufacturing sector (documented by,inter alia, Dunning, 1993, UNCTAD,1996a and Caves, 1996). Other data alsosuggest that the export propensity ofMNEs,or their affiliates,in the sectors inwhich they are most active exceeds thatof indigenous competitors. Except in thecase of a few countries,notablyJapan,thepayments for the services of knowledge-intensive assets received by U.S. MNEsfrom their foreignaffiliates, expressed asa proportionof their total exports,is con-siderablygreaterthan the equivalentpro-portionbetween U.S. and independentlyowned firms. For example, in 1996, roy-alties and fees received by U.S. firmsfromtheirforeignaffiliatesamountedto 6percentof their exportsto these affiliates.The corresponding proportion of non-affiliation royalties and fees received byall U.S. firmsas a proportionof totalU.S.exports was 3 percent (U.S. Departmentof Commerce,1997). Furthermore,of allroyalties and fees received by U.S. firmsfrom foreign-based firms in the years1993 to 1996, 79 percentwere internaltoU.S. MNEs.18

The extent to which MNEspromote,orgravitate to, spatial clusters within acountry or region is an under-researched

area. Clearly, some older clusters, e.g.,the Portuguese cork industry, the Swisswatch industry, the North Italy textileindustryand the Cityof Londonfinancialdistrict, developed without much MNEparticipation. But, many of the newlyestablished clusters, which are gearedmore to accessingthe externaleconomiesof knowledge creation, interactive learn-ing and the upgradingof the competitiveadvantage of the constituent firms, areinfluenced by a rather different set ofcosts and benefits;and a casual examina-tion of the membership of science andtechnology parks, export processingzones, researchand developmentconsor-tia and service support centers, wouldcertainly suggest that MNEs are activelyinvolved - often as flagship firms.Certainlyamong developed regions (e.g.,the EuropeanUnion) and countries (e.g.,United States), knowledge-intensive andexport-orientedactivities tend to be moregeographically concentrated than otherkinds of activities (see, forexample, illus-trations given in Porter, 1990, Dunning,1997b,Chap.3 andDunning,1997c).

Any modern theory of internationaleconomic activitymust then takeaccountof how the common ownership of cross-border production and transactionsmayresult in a differentvalue added and spa-tial configurationthan that which wouldarise if such functions were separatelyundertakenby uni-national firms. Interalia, the extra attributes comprise thespreadingof firm specific overheadsandrisks;the intra-firmsharingand transfer-ence of knowledge, experience and mar-kets; and the external economies arisingfrom jointly organized innovatory, pro-duction and marketing activities. Formany of these activities,thereis no exter-nal market;the output of one part of thefirm can only be sold as an input toanotherpartof the same firm. However,


JOHNH. DUNNING

these interdependent activities may notneed to be undertakenin the same regionor country. For other activities, internalmarketsmay offermorecoordinatingben-efits and/or less transactionalcosts thanarm's-lengthmarkets. Inboth cases, how-ever, it may be preferableto think of theMNE not as a second best substitute forthe market,but as a partnerwith the mar-ket to promote first-best allocative effi-ciency throughout and across valuechains.

The notion of efficiency-promotinginternalmarketsneeds to be more formal-ly built into both posit'iveand normativemacro models of internationaleconomicactivity. In addition to acknowledgingthe differentgeographicalneeds of asset-producing and asset-exploitingactivities,models of trade need to incorporatethebenefits of organizing the two sets ofactivity under common ownership vis-&-vis that of the external market. This, inprinciple, is not a difficult thing to do.Essentially, it comes down to an identifi-cation and evaluation of the country,activity and firm-specificvariableswhichdetermine whether the different transac-tional and coordinatingfunctions arebestorganizedwithin marketfriendly hierar-chies, or by the marketper se. We havealready argued that markets for createdassets, and for the goods and servicesarisingfromthem, are likely to be intrin-sically more imperfectthan those fornat-ural assets and forthe goods and servicesarising from them. In some instances,too, it may be efficiency enhancing forthese marketsto be internalized. We alsocontend, with Behrman and Grosse(1990) and Meyer (1998), thatmost cross-border markets are likely to be moreimperfect than their domestic equiva-lents, and that, because of this, MNEactivity may be more welfare enhancingthan multi-plantactivity within an econ-

omy. We say "most" cross-border mar-kets, because some domestic markets,particularly in emerging developingeconomies, are likely to be more imper-fect than those in developed countries.But issues such as foreign-exchangeuncertainty,institutionaland culturaldif-ferences, and the differentialrole of gov-ernments are obviously likely to play amore important role in, affecting theworkings of cross-borderthan domesticmarkets. And we say "may" be morewelfare enhancing, because much willdepend upon the conditionsunder whichforeigninvestmenttakesplace.

At the same time, the extent to whichcross-bordermarketsare internalized viaMNE activity will itself depend on thecharacteristicsof the tradingpartnersandthe countries involved, as.well as on thetypes of assets, goods and services beingexchanged. In their attempts to explainthe alternative forms of trans-bordertrade, and to advance both the positiveand normativetheories of trade, interna-tional economists need to delve deeperinto the structure of country-specificadvantages in organizing trade (particu-larly in knowledge-related products),through FDI and inter-firmalliances, ascomparedwith arm's-lengthmarkets.

CONCLUSIONSThe previous two sections of this paper

have examinedhow changesin the globaleconomy over the past two decades areaffecting scholarly thinking about boththe micro-economic geography of FDIand MNE activity, and the more macro-economic explanationsof the internation-al allocation of all value-added activity.In particular, we focussed on threepoints. The first is the growing impor-tance of intangible assets - and particu-larly intellectual capital - in the wealth-creatingprocess, and of the need of com-



panies to harness, as well as to exploit,these assets from a variety of locations.Secondly, we emphasized the changingrole of location-bound assets, whichmobile investorslook foras complementsto their own core competencies. In doingso, we again underscoredthe increasingsignificance of createdassets (and partic-ularly those which governments,in theirmacro-organizationalpolicies, can and doinfluence), and, also, the benefits whichspatial clusters offer whenever distance-related transactions and coordinationcosts arehigh.

Thirdly, we arguedthat, to adequatelyincorporatethe activities of MNEswithinexistingtrade-typetheoriesof the interna-tional allocation of economic activity,more attentionneeds to be given both tothe specific motives, determinants andconsequences of the common governanceof related cross-borderactivities, and tothe conditions in which internalizingintermediate product markets mightmake for a more efficient (in the sense ofthe "next best" realistic alternative,assuming that all cross-borderavoidablestructuralmarketimperfectionhave beenremoved) spatial configuration of eco-nomic activity in the contemporaryglob-al and innovatory economy. We havealso suggested that any paradigmof thegeographyof FDI,as contrastedto that ofthe investments of all firms, needs to beconstructedon similarlines.

Whatarethe implications of our analy-sis and findings for future internationalbusiness research? First, to returnto thestarting point of this paper, and in linewith the thinking of Michael Porter(1994, 1996), I believe more attentionneeds to be given to the importance oflocation per se as a variableaffectingtheglobalcompetitivenessof firms.Thatis tosay, the locational configurationof a fir-m's activities may itself be an 0-specific

advantage,as well as affect the modalityby which it augments, or exploits, itsexisting 0 advantages. With the gradualgeographicaldispersion of createdassets,and as firms become more multinationalby deepeningorwidening theircross-bor-der value chains, then, both from theviewpoint of harnessingnew competitiveadvantagesand more efficiently deploy-ing their home-basedassets, the structureand content of the location portfolio offirmsbecomes more criticalto their glob-al competitivepositions.

Second, in seeking to make an opti-mal use of the existing location-boundassets within this jurisdiction, and topromote the dynamic comparativeadvantage of their resource-capabilities,governments need to give more atten-tion to ensuring that their actions helpfashion, support and complement thoseof efficient hierarchies and markets.This involves a greater appreciationboth of the changing locational require-ments of mobile investments, and ofhow, in the case of those marketswhereendemic failure is the most widespread,governments may work in partnershipwith firms either to improve markets(i.e., by a "voice" strategy19), or toreplace these markets (by an "exit"strategy). With the growing importanceof knowledge-related infrastructure,andaccepting the idea of sub-national spa-tial units as nexus of untraded interde-pendencies (Storper, 1995),20 this pre-sents both new challenges and opportu-nities to both national and regional gov-ernments in their macro-organizationaland competitive enhancing policies.

NOTES

1. Such international-specific transac-tion costs have recently been explicitlyidentified by Klaus Meyer in a volume(Meyer, 1998) based upon his doctoral


JOHN H. DUNNING

dissertation at the London BusinessSchool.

2. For example, of the 1,150 or so pub-lications cited in his volume, only 13 per-cent are to monographs or articles pub-lished after 1990.

3. Author's estimate, based on data onthe royalties paid for managerial knowhow; and of the relationship between for-eign portfolio and foreign direct invest-ment.

4. There have been only a few attemptsto use transaction cost analysis to explainthe spatial distribution of economic activ-ity. One example is that of the industrialgeographers Michael Storper and AllenScott. See, for example, Storper (1995),Storper and Scott (1995) and Scott (1996).Yet, such analysis offers a powerful toolfor explaining why firms requiring idio-syncratic inputs, e.g., tacit knowledge ofvarious kinds, and/or those supplyingidiosyncratic and uncertain markets tendto value proximity with their suppliersand/or customers. Perhaps the best illus-tration of a spatial cluster, or agglomera-tion, of related activities to minimize dis-tance-related transaction costs, and toexploit the external economies associatedwith the close presence of related firms isthe Square Mile of the City of London.

5. Scott (1996) gives some examples,including the growing concentration andspecialization of both manufacturing andservice activities in large metropolitanareas within both developed and devel-oping countries. In an interesting recentpaper, Davis and Weinstein (1997) con-clude that intra-national concentration ofvalue-added activity is likely to obey thedictates of economic geography morethan that of the inter-national concentra-tion of such activity.

6. Estimates of such ventures varyenormously. A recent study by Booz,Allen and Hamilton (1997) has put the

number of cross-border alliances (includ-ing mergers and acquisitions) formed in1995 and 1996 to be as high as 15,000.Another assessment by Hagedoorn (1996)suggests that between 1980 and-1994, thenumber of newly established cross-bordertechnology-related inter-firm agreementsrose by over three times. Finally, thevalue of international mergers and acqui-sitions over the same period were esti-mated to have accounted for between 50percent and 60 percent of all new FDI(UNCTAD, 1997).

7. For a detailed exposition of thedevelopment of a new trajectory of tech-nological advances, see Lipsey (1997) andRuigrok and Van Tulder (1995).

8. Though there have been markedfluctuations in the shares within andbetween these periods, which reflect,inter alia, changes in exchange rates andthe positioning of countries in theircycles of economic development. Forexample, during 1975-1980, the UnitedStates attracted 32 percent of FDIreceived by developed countries; by1985-1990 that share had risen to 42 per-cent. However, it fell again to 18 percentin 1991 and 1992; but since then it hasrecovered, and in 1995-1996 it stood at 35percent.

9. Japan is a classic case in point. Inthe period 1990-1994 it accounted for 29percent of the world's gross fixed capitalformation, but only 0.8 percent ofinbound FDI flows.

10. To give just one example; in theperiod 1990-1994, 49 percent of U.S.direct investment flows were directed toWestern Europe, 10 percent to Asia and25 percent to Latin America. The corre-sponding percentages for Japanese directinvestment flows were 20 percent, 19percent and 10 percent (UNCTAD, 1997,Dunning, 1998).

11. Unlike the theory of the firm;

VOL. 29, No. 1, FIRSTQUARTER, 1998 61


although if there had been a well devel-oped theory of the multi-activity firmprior to the work of scholars such asBuckley, Casson and Hennart,one won-ders if this aspect of international busi-ness activity would have attracted somuch attention!

12. We use the word "customized"deliberately, following the contention ofPeck (1996) that host governments maysometimes need to individualize or cus-tomize the upgrading of their physicaland human infrastructureboth to meetthe specific needs of mobile investors,and promote the competitive dynamicadvantage of the location-boundresourceswithin theirjurisdiction.

13. I am indebted to the reviewer ofthis paperformakingthis point.

14. An expression first used in David(1984), and since taken up by Wheelerand Mody (1992) and Mytelka (1996).

15. In Markusen'swords "multination-al enterprises in this framework areexportersof the services of firm specificassets ... subsidiaries import these assets"(Markusen, 1995 p. 175).

16. Abbreviated, Ot transaction (orcoordinating) cost-minimizing advan-tages,c.f. Oa= asset-specificadvantages.

17. Which foreign country, or coun-tries, is decided by the normal locationalcriteria.

18. Other data on royalties and man-agementfees received by U.S. firms fromforeign firms are regularlypublished bythe United States Department ofCommerce in the Survey of CurrentBusiness, and in the Benchmark Surveysof U.S. Direct Investment Abroad. Seealso UNCTAD(1995, 1996aand 1997).

19. The concepts of "voice"and "exit"strategies as applied to MNE-relatedactivity are explained in Dunning(1997a).

20. The idea of a regionas a spatialunit

which internalizes distance-related trans-action costs which otherwise would fallupon its constituent firms is an interestingnotion worth pursuing by internationalbusiness scholars. For, like a firm, thestrategies pursued by a region to provide aset of unique, non-mobile and non-imitat-able locational advantages for its firmsmay well determine its own competitiveadvantages relative to those of otherregions. At the same time, regions, likefirms, may decline as well as prosper; butour knowledge about the focus leading tothe spatial dis-agglomeration of relatedactivities is woefully inadequate.

REFERENCES

Almeida, P. 1996. Knowledge sourcing byforeign multinationals: patent citationanalysis in the U.S. semi-conductorindustry. Strategic ManagementJournal, 17 (Winter): 155-65.

Bandera, V. N. &J. T. White. 1968. U.S.direct investments and domestic mar-kets in Europe. Economia Inter-nazionale, 21 (February):117-33.

Behrman, J. N. & Robert Grosse. 1990.International business and govern-ments. Columbia, South Carolina:University of South Carolina Press.

Blair, M. M. 1995. Ownership and con-trol: rethinking corporate governancefor the 21st century, Washington DC:The Brookings Institution.

Booz, Allen and Hamilton. 1997. Crossborder alliances in the age of collabo-ration. Los Angeles, CA: Booz Allenand Hamilton.

Brainard, S. L. 1993. A simple theory ofmultinational corporations and tradewith a trade-off between proximity andconcentration. Cambridge, MA: NBERWorking Paper No. 4269, February.

Braunerhjelm, P. & R. Svensson. 1995.Host country characteristics andagglomeration in foreign direct invest-


JOHN H. DUNNING

ment. Stockholm: Industrial Institutefor EC and Social Research (mimeo),October.

Cantwell, J. & R. Harding. 1997. The in-ternationalization of German compa-nies R&D. Discussion Paper in Inter-national Investment and ManagementNo. 233, University of Reading.

Caves, R. 1982 & 1996. Multinationalfirms and economic analysis. Cam-bridge: Cambridge University Press.First and second editions.

Dalton D. H. & M. G. Serapio. 1995.Globalizing industrial research anddevelopment. U.S. Department ofCommerce, Office of TechnologyPolicy, Washington, DC: U.S. Depart-ment of Commerce.

David, P. 1984. High technology centersand the economics of locational tour-naments. Stanford, CA: StanfordUniversity (mimeo).

Davidson, W. 1970. The location of for-eign direct investment activity: coun-try characteristics and experienceeffects. Journal of InternationalBusiness Studies, 11(2): 9-22.

Davis, D. R. &D. E. Weinstein. 1997. Eco-nomic geography and regional produc-tion structure: An empirical investiga-tion. Cambridge, MA: National Bureauof Economic Research, Working PaperSeries No. 6093 (July).

Donahue, J. D. 1996. Disunited States.New York: Basic Books.

Dunning,J. H. 1993. Multinationalenter-prises and the global economy. Wok-ingham, England and Reading, Mass.:Addison Wesley.

. 1995. What's wrong - andright - with trade theory. InternationalTrade Journal, 9(2): 153-202.

_. 1996. The geographicalsources of competitiveness of firms:some results of a new survey. Trans-national Corporations,5(3):1-30.

I._.. 1997a. Alliance capitalismand global business. London and NewYork: Routledge.

, editor. 1997b. Governments,globalization and international busi-ness. Oxford: Oxford University Press.

_. 1997c. The European inter-nal market program and inbound for-eign direct investment. Journal ofCommonMarketStudies, 35 (1 and 2):1-30 and 189-223.

. 1998. The changing geogra-phy of foreign direct investment. In N.Kumar, editor. Internationalization,foreign direct investment and technol-ogy transfer:Impact and prospects fordeveloping countries. London andNew York:Routledge.

Edvinson, L. 1997. Intellectual CapitalDevelopment. Stockholm: Skandia.

Ethier, W. J. 1986. The multinationalfirm. QuarterlyJournal of Economics,101:806-33.

Florida, R. 1995. Towards the learningregion. Futures,27(5): 527-36.

Fujita, M. & J. R. Thisse. 1996. Econ-omics of agglomeration. Kyoto: KyotoUniversity, Institute of EconomicResearch Discussion Paper No. 430,January.

Grabher, G., editor. 1993. The embeddedfirm. London and New York: Rout-ledge.

Grossman, G. M. & E. Helpman. 1991.Innovation and growth in the globaleconomy, Cambridge, MA: MIT Press.

Hagedoorn, J. 1996. Trends and patternsin strategic technology partneringsince the early seventies. Review ofIndustrialOrganization,11: 601-16.

Handy, C. 1989. The age of unreason.London: Hutchinson.

Hatem,F. 1997. Internationalinvestment:Towards the year 2001. Geneva:United Nations.

Helpman, E. &P. R. Krugman. 1985. Mar-

VOL. 29, No. 1, FIRST QUARTER, 1998 63


ket structure and foreign trade.Cambridge, MA, MIT Press.

Horstman, I. J. & J. R. Markusen. 1992.Endogenous market structures in inter-national trade. Journal of InternationalEconomics, 32: 109-29.

Johanson, J. & J. E. Vahlne. 1977. Theinternationalization process of the firm- a model of knowledge developmentand increasing market commitments.Journal of International BusinessStudies, 8: 23-32.

Kogut, B. 1983. Foreign direct investmentas a sequential process. In Kindle-berger, C. P. & D. Audretsch, editors.The multinational corporation in the1980s. Cambridge, Mass.: MIT Press.

_. 1985. Designing global strate-gies: corporate and competitive valueadded chains. Sloan ManagementReview, 25: 15-28.

Krugman, P. editor. 1986. Strategictrade policy and the new internation-al economics. Cambridge, MA: MITPress.

Krugman, P. R. 1991. Geography andtrade. Cambridge, MA: MIT Press.

. 1993. On the relationshipbetween trade theory and location the-ory. Review of International Eco-nomics, 1(2): 110-22.

Kuemmerle, W. 1996. The drivers of for-eign direct investment into researchand development: An empiricalinvestigation. Boston, Harvard Busi-ness School Working Paper No.96:062.

Lipsey, R. G. 1997. Globalization andnational government policies: Aneconomist's view. In John H. Dunning,editor. Governments, globalization andinternational business. Oxford: OxfordUniversity Press.

Liu, S. X. 1998. Foreign direct investmentand the multinational enterprise. A re-examination using signaling theory.

Westport, Conn.: Greenwood Publish-ing.

Loree, D. W. & S. E. Guisinger. 1995.Policy and nonpolicy determinants ofU.S. equity foreign direct investment.Journal of International BusinessStudies, 26(2): 281-300.

Malmberg, A., 0. Slovell & I. Zander.1996. Spatial clustering, local accu-mulation of knowledge and firm com-petitiveness;. Geografiska AnnalerSeries B, Human Geography, 78(2):85-97.

Markusen, A. 1996. Sticky places in slip-pery space: A Typology of industrialdistricts. Economic Geography, 72(3):293-313.

Markusen, J. R. 1995. The boundaries ofmultinational enterprises and the theo-ry of international trade. Journal ofEconomic Perspectives, 9(2): 169-89.

&A. Venables. 1995. Multina-tional firms and the New TradeTheory. Cambridge, Mass.: NBERWorking Paper No. 5036, February.

Maskell, P. 1996. Local embeddednessand patterns of international special-ization. Copenhagen, CopenhagenBusiness School (mimeo).

Mataloni, R. & M. Fahim-Nader. 1996.Operations of U.S. multinational com-panies: preliminary results from the1994 benchmark survey. Survey ofCurrentBusiness (December): 11-37.

Meyer, K. 1998. Direct investment ineconomies in transition. Cheltenham,UK, Lyme U.S.: Edward Elgar.

Mytelka, L. K. 1996. Locational tourna-ments, strategic partnerships and thestate. Ottawa: Carleton University(mimeo).

Ohlin, B. 1933. Inter-regional and inter-national trade. Cambridge, MA:Harvard University Press, revised edi-tion 1967.

Ohmae, K. 1995. The end of the nation


JOHN H. DUNNING

state: The rise of regional economies.London: Harper.

Papanastassiou, Marina & Robert Pearce.1997. Technology sourcing and thestrategic role of -manufacturing sub-sidiaries in the UK: local competenciesand global competitions. ManagementInternational Review, 37 (forthcom-ing).

Peck, F. W. 1996. Regional developmentand the production of space: the roleof infrastructure in the attraction ofnew inward investment. Environmentand Planning, 28: 327-39.

Porter, M. E. 1990. The competitiveadvantage of nations. New York: TheFree Press.

. 1994. The role of locationin competition. Journal of Economicsof Business, 1(1): 35-39.

.._._.....__ 1996. Competitive advan-

tage, agglomerative economies andregional policy. International Re-gional Science Review, 19(1 and 2):85-94.

Rees, D. & T. McLean. 1997. Trends inlocation choice. In A. Jolly, editor.European Business Handbook 1997.London: Kogan Page (for CBI).

Ruigrok, W. & R. Van Tulder. 1995. Thelogic of international restructuring.London and New York: Routledge.

Scaperlanda, A. & L. J. Mauer. 1969. Thedeterminants of U.S. direct investmentin the EEC. American EconomicReview, 59 (September): 558-68.

Scott, A. J. 1996. Regional motors of theglobal economy. Futures, 28(5): 391-411.

Shan, W. & J. Song. 1997. Foreign directinvestment and the sourcing of techno-logical advantage: evidence from thebiotechnology industry. Journal ofInternational Business Studies, 28(2):267-84.

Srinivasan, K. & Ashoka Mody. 1997.

Location determinants of foreign directinvestment: an empirical analysis ofU.S. and Japanese investment. Cana-dian Journal of Economics (forthcom-ing).

Stewart, T. A. 1997. Intellectual capital.London: Nicholas Bradley.

Storper, M. 1995. The resurgence ofregion economies: ten years later: theregion as a nexus of untraded interde-pendencies. European Urban andRegional Studies, 2(3): 191-221.

________* & A. J. Scott. 1995. Thewealth of regions. Futures, 27(5): 505-26.

Taylor, J. 1993. An analysis of thefactors determining the geographi-cal distribution of Japanese manufac-turing investment in the UK, 1984-91.Urban Studies, 30(7): 1209-24.

U.S. Department of Commerce. 1997. U.S.international sales and purchases ofprivate services. Survey of CurrentBusiness, October: 95-138.

UNCTC. 1988. Transnational corpora-tions and world development. NewYork: UN.

UNCTAD. 1995. World investment report1995: Transnational corporations andcompetitiveness. New York andGeneva: UN.

__,__._.i.1996a. Worldinvestmentre-

port 1996: Transnational corporations,investment, trade and internationalpolicy arrangements. New York andGeneva: UN.

. 1996b. Incentives and for-eign direct investment. Geneva andNew York: UN.

.,___ . 1997. Worldinvestmentre-port 1997: Transnationa] corporations,market structure and competition poli-cy. Geneva and New York: UN.

Vernon, Raymond. 1966. Internationalinvestment and international trade inthe product cycle. Quarterly Journal of



Economics,80: 190-207.. 1974. The location of eco-

nomic activity. In John H. Dunning,editor, Economic analysis and themultinational enterprise. London:Allen and Unwin.

Wells, L. T. editor. 1972. Theproductlifecycle and international trade.Cambridge, MA: Harvard UniversityPress.

Wesson, T. J. 1993. An alternative moti-vation for foreign direct investment.Ph.D. dissertation, Harvard University.

Wheeler, K. & Ashoka Mody. 1992.International investment and locationdecisions: the case of U.S. firms.Journal of International Economics,33:57-76.

Wood, A. 1993. Give Heckscher and Oh-lin a chance. Sussex: University ofSussex, Institute of DevelopmentStudies (mimeo).

World Bank. 1997. World developmentreport:The state in a changing world.Oxford and New York: OxfordUniversity Press.


Documents

UNIVERSITYOFREADINGANDRUTGERSUNIVERSITY › content › pdf › 10.1057 › palgrave.jibs.8490… · JOHNH.DUNNING inDunning,1998).Intheperiod1991-1996,64percentofglobalFDIinflows