The Impact of ISO 9000 Diffusion on Trade and FDI: A New Institutional Analysis

7/29/2019 The Impact of ISO 9000 Diffusion on Trade and FDI: A New Institutional Analysis

1/22

The Impact of ISO 9000 Diffusion on Trade and FDI: A New Institutional AnalysisAuthor(s): Joseph A. Clougherty and Micha GrajekReviewed work(s):Source: Journal of International Business Studies, Vol. 39, No. 4, Institutions andInternational Business (Jun., 2008), pp. 613-633Published by: Palgrave Macmillan Journals

Stable URL:http://www.jstor.org/stable/25483290 .

Accessed: 20/02/2013 07:03

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of

content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms

of scholarship. For more information about JSTOR, please contact [email protected].

.

Palgrave Macmillan Journals is collaborating with JSTOR to digitize, preserve and extend access toJournal of

International Business Studies.

http://www.jstor.org

This content downloaded on Wed, 20 Feb 2013 07:03:50 AMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/action/showPublisher?publisherCode=palhttp://www.jstor.org/stable/25483290?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/25483290?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=pal


2/22

Journalf nternationalusinesstudies2008) 9,613-633 ^L? 2008 Academy f Internationalusiness All rights eserved 047-2506 $30.00_"T^www.jibs.net

The impact of ISO 9000 diffusion on tradeand FDI: A new institutional analysis

Joseph A Clougherty1'2 andMichaJGrajek37Wissenschaftszentrum Berlin (WZB), Germany;2CEPR, London, UK; 3ESMT European School of

Management and Technology, Berlin, GermanyCorrespondence:jA Clougherty, Wissenschaftszentrum Berlin(WZB), Reichpietschufer 0, Berlin 10785,Germany.Tel: +49 30 25491 427;Fax: +49 30 25491 444;E-mail: [email protected]

Received: 31 January 2006Revised: 14November 2006Accepted: 31August 2007Online publication date: 14 February 2008

AbstractThe effectsof ISO 9000 diffusionon trade and foreigndirect investment FDI)have gone understudied. We employ panel data reported by OECD nationsover the 1995-2002 period to estimate the impact of ISO adoptions oncountry-pair economic relations.We find ISO diffusion to have no effect indeveloped nations, but to positively pull FDI (i.e., enhancing inwardFDI) andpositively push trade (i.e., enhancing exports) indeveloping nations.journal of InternationalBusiness Studies (2008) 39, 613-633.doi: 10.1057/palgrave.jibs.8400368Keywords: theory of FDI and the MNE; trade flows; transaction cost analysis;institutional environment

INTRODUCTIONThe Geneva-based International Organisation for Standardisationwas constituted in 1946. However, unlike many of the otherpost-World War II international institutions (e.g., the UnitedNations, the IMF, the World Bank, the International HealthOrganisation, or the International Civil Aviation Organisation)this organisation has experienced itsmost striking success over thelast two decades. The ISO 9000 certification system for firm qualitywas introduced in 1987, and by the end of 2002 some 560,000certificates had been awarded to sites in 159 nations (Corbett,2003). In short, ISO 9000 (sometimes referred to simply as ISOhere) certification for firm quality has been far and away the mostsuccessful and widespread set of standards implanted by theInternational Organisation for Standardisation (Casper & Hancke,1999; Mendel, 2002).The remarkable spread of ISO has spawned a recent literature thatattempts to define what drives the global diffusion of certificationto ISO 9000 (e.g., Corbett, 2003; Guler, Guillen, & Macpherson,2002; Mendel, 2002) and to ISO 14000 (a sister standard forenvironmental processes: e.g., Christmann & Taylor, 2001; Corbett& Kirsch, 2001; Delmas, 2000, 2002; Prakash & Potoski, 2006,2007). With regard to ISO 9000 standards (the focus of this study),several diffusion drivers have been identified: the role of government directives (particularly the EU); government agencies mandating contractor adoption; multinational enterprises requiringsupplier adoption; and extensive adoption by trading partner andtrading competitor nations. Yet the macro-level economic impactof ISO 9000 certification has been an under-investigated topic.

http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp


3/22

^_Impact of ISO 9000 on trade and FDI Joseph Clougherty and MichatGrajek_614

We aim to look not at the macro-level drivers ofISO 9000 adoption and diffusion, but instead at theeconomic impact of ISO 9000 adoption on countrypair economic relationships. More specifically, weare interested in how the adoption of ISO 9000standards influences the two primary means (tradeand foreign irect investment FDI)) bywhich firmsconduct cross-border economic activity. ISO 9000standards - via their quality-signalling, commonlanguage, and conflict-settling properties - potentially alter the costs involved with trade and FDIactivity. In short, ISO 9000 may reduce the transaction costs and information asymmetries involvedwith arm's length interactions between businessesfrom different nations (thereby enhancing trade)and with engaging in international greenfields,joint ventures QVs) and acquisitions (therebyenhancing FDI).This paper is organised as follows in order tosupport the analysis. We begin by setting theresearch question in the international business (IB)literature, giving credit to new institutional economics (NIE) insights as conceptually fundamental, andsupporting the ISO 9000 system as representative ofan international institution. Next, we consider theproperties of ISO 9000, and generate formal propositions with respect to how diffusion may directlyimpact on country-pair trade flows and FDI stocks.Further, we analyse the potential for the ISO effect todepend on country-pair type: that is,whether theproposed effects hold systematically. We thendescribe the panel data, discuss estimation issues,and empirically test the impact of ISO 9000 adoption on country-pair economic relationships. Finally,we conclude and discuss.A NEW INSTITUTIONAL ANALYSIS OF ISO9000INAN IBCONTEXTIB researchers are inherently interested in thefactors that affect cross-border economic relationships.Dunning's (1981, 1988) OLI paradigm helpsboth contextualise and conceptualise the importance of institutional factors: in particular, thelocation-specific advantages of host-country environments represent the key influence as to whereforeign production takes place, as they createclimates that facilitate an MNE's exploiting ownership-specific and internalisation advantages. Bothnon-institutional factors (e.g., market size, marketgrowth, labour costs, production costs, transportation costs, natural resources, and distance) andinstitutional factors (e.g., governance structures,trade barriers, FDI-specific policies, macroeconomic

policies, political stability, and tax regulations)have been identified in the literature as prominentlocation advantages (Li & Resnick, 2003; Sethi,Guisinger, Phelan, & Berg, 2003). The range ofpotential institutional factors is, accordingly, extensive: from the well-studied importance of politicalconstraints on executive discretion that secureproperty rights (e.g., Henisz, 2002; Henisz & Zelner,2001) to exploratory notions on how antitrustpolicy might impact on FDI (Brewer, 1993; Li &Resnick, 2003).A striking omission in the above list of locationadvantages is the status of managerial practicein the host nation environment. Clearly, foreigninvestors will consider not only factors such as thenatural resources, human resources and politicalcontext in a country, but also the general excellence ofmanagerial practice in that same environment. While the empirical literature on FDI drivershas neglected the importance of managerial practice and standards, these factors havebeen explored conceptually. For instance, Dunning(2001) notes that commercial infrastructure andbusiness culture enhance location attractiveness.Consider also the IB scholarship that stresses theexistence of FDI motivated by MNEs looking togain access to resources and capabilities: hencesome studies (e.g., Anand, Capron, & Mitchell,2005; Anand & Kogut, 1997) examine the geographical pull of host country skills. While these skillsare typically considered to be embodied inworkercapabilities, Anand and Kogut (1997) observe that anation's capabilities also reside in firms andindustrial networks, and that IB researchers oftenignore the important role played by resourcesembedded in firms. Consider also the call by Ricart,Enright, Ghemawat, Hart, and Khanna (2004) forthe JIBS community to embrace the importance ofspecialised intermediaries: institutions that helpconsummate transactions by reducing the information asymmetries and transaction costs involvedwith identifying the relevant resources embeddedin firms.Our main interest resides in how the diffusion of

ISO 9000 practices - the internationally recognisedstandard for quality management - impacts on theeconomic relationships between countries. Byempirically considering the impact of ISO 9000diffusion on both FDI and trade in internationalcountry-pair environments, we hope to add tothe extensive IB literature on location advantages.Furthermore - and akin to the recent IB literaturethat identifies firm-based sources of capabilities - we

Journal f International Business Studies



4/22

_Impact of ISO 9000 on trade and FDI Joseph Cloughertyand Michat Grajek_T* 615

think it is important to consider firm properties andthe role of institutional intermediaries. In our case,the overall properties of firms in a particularnational environment and the intermediary roleplayed by the ISO 9000 system might impact onthe macro-level phenomenon of cross-countrytrade and FDI. In short, we argue that the diffusionof internationally recognised managerial practicesin a country - in addition to natural, human andpolitical resources - represents an important factorthat drives FDI and trade decisions.

Before delving further into our analysis of ISO9000's potential impact, it behoves us to groundour analysis institutionally via two additionalmeans. First, we must lay out how institutions areconsidered by NIE scholars, as that perspective(focusing on how alternative forms of governanceeconomise on transaction costs) represents our coretheoretical backdrop. Second, we must establishthat the ISO 9000 system is an institution - at leastan informal-decentralised international institution- in order to engage in such an analysis.New Institutional EconomicsThe work of Robert Coase and Oliver Williamson isfundamental to the NIE perspective. Coase (1937)identified transaction costs as explaining why someexchanges were done within firms and not bymarkets. Williamson (1975, 1985) expanded onCoase's work by identifying bounded rationalityand individual opportunism as heightening transaction costs. Institutions play a critical role inNIEscholarship, as they reduce uncertainty by providing dependable and efficient frameworks foreconomic exchange (North, 1990). The "new institutionalism" in organisational theory and sociology has one particular concern with the NIEapproach that is germane to our study: institutionsand organisations do not necessarily lead to beneficial outcomes (DiMaggio & Powell, 1991; Scott,1995). Instead, new institutionalism holds thatactors often construct institutions that never manifest the desired outcomes. For instance, DiMaggioand Powell (1983: 147) state that "forms of organization change occur as the result of processes thatmake organizations more similar without necessarilymaking them more efficient".In line with the NIE approach and new institutionalism critique, we base our theoretical justifications on the external - not the internal - benefits ofISO 9000 certification. Hudson and Jones (2003:1004) observe that ISO 9000 may generate benefits"both because quality ismore effectively signaled

and also because there may be an actual increase inquality [and/or] better cost control". We neglectthe internal benefits, in part because manyscholars (e.g., Juran, 1999; Simmons & White, 1999;Walgenbach, 1998) have found ISO 9000 certification not to impact on actual quality or efficiency:Delmas (2002) notes that firms seek ISO 9000irrespective of whether they expect actual qualityimprovements. Terziovski, Power, and Sohal (2003)begin tomake sense of the mixed internal performance record for ISO 9000 by observing that firmsat early stages of the quality process may findcertification to spawn internal benefits, but companies with good quality systems may find ISO togenerate internal losses owing to the added costs,delays, and burdensome documents necessary forcertification. Nevertheless, any internal benefits ofISO 9000 would seemingly be complementary tothe external benefits. Furthermore, in our empiricalset-up we cannot distinguish between internal andexternal benefits: hence the internal benefits mayfactor in our results. Yet neglecting the potentialinternal benefits of ISO adoption is consistent withour NIE approach (focusing on the external problems of transaction costs and information asymmetries), consistent with new institutionalism'scritique (efficiency is often not realised), andconsistent with the literature on ISO's internalbenefits (amixed performance record).In sum, we engage in an NIE analysis of the ISO9000 system as the NIE approach iswell suited tomaking sense of institutional remedies for transaction cost and information asymmetry problems - theheart of our interest in ISO's potential beneficialimpact on FDI and trade.We move forward then withthe idea that institutions (and the ISO 9000 system inparticular) emerge or take specific forms to solveproblems and facilitate commerce. Yet we remainopen to new institutionalism's charge that new formsmay not generate the intended consequences.ISO 9000 as an International Institution

Douglas North (1991: 97) defines institutions ashumanly devised constraints that structure economic interaction and consist of both informalconstraints (e.g., sanctions, taboos, customs, traditions, and codes of conduct) and formal rules tocreate order and reduce uncertainty in exchange.Hence recognising the importance of both formaland informal institutions represents an importantNIE breakthrough (Henisz & Williamson, 1999), asprior scholarship generally neglected informalinstitutions by taking a formal-legalism approach




5/22

^2__Impact f ISO 9000 on trade and FDI Joseph Cloughertyand Michai Grajek_616

to institutional analysis (Yarbrough & Yarbrough,1990). Accordingly, despite the voluntary anddecentralised nature of the ISO 9000 system, it isvery much an institution. King, Lenox, and Terlaak(2005) concur when they surmise that the ISO

management standards represent examples ofdecentralised institutions.Understanding the ISO 9000 system as an international institution can be enhanced by consideringthe discourse within international political econo

my (IPE). The domain of IPE focuses on the social,political, and economic arrangements that impacton the global systems of production and exchange:that is, which institutional conditions promoteinternational cooperation (Strange, 1988). For IPEscholars, international regimes (i.e., internationalinstitutions) represent multilateral agreements thatpromote cooperative behaviour, regulate relations,and stabilise the international order in particularissue areas. DiMaggio and Powell (1991) - insurveying new institutionalism's impact on IPE -note that some international institutions are formalorganisations, while others are complex sets of rules,standards and agencies. DiMaggio and Powell furthernote that regimes are institutions in that they buildupon, homogenise, and reproduce standard expectations. Accordingly, IPE scholars could consider theInternational Organisation for Standardisation andits ISO 9000 standard as an international regime.

Furthermore, ISO's decentralised nature yieldssome estimation benefits for empirical testing.Consider the dichotomous nature of most institutional conditions: that is, you are either in or out.Rose (2004) cites this dynamic as crucial to hisfinding theWTO to have no significant impact ontrade flows. One of the problems with a dichotomous measure for institutional membership isthat it cannot capture significant variation ininstitutional commitments: for example, nonmembers to an institution attempting to joinprobably exhibit different behaviour from nonmembers with no ambition of joining. A decentralised institution that depends on the choices ofmyriad local firms allows us to elicit a finermeasureof effective national commitments to institutionalprinciples. In short, the greater the diffusion of ISO9000 practices in a country, the greater thecommitment to those managerial principles in thatnational environment.

COUNTRY-PAIR TRADE FLOWSThe stated objective behind the formation of theInternational Organisation for Standardisation was

to develop standards to facilitate internationaltrade in goods and services (Hayes, 1994). Additionally, ISO diffusion reportedly received a greatboost by the EU's use of the standard to accelerateestablishment of a single market (Anderson, Daly, &Johnson, 1999; Conti, 1999). At the firm level, ithas been argued that businesses consider theimprovement ormaintenance of international salesas an adoption motivation (Guler et al., 2002;Simmons & White, 1999). If the above motivationsare founded on accurate assessments of certificationbenefits, then ISO 9000 diffusion should generateenhanced international trade at the macro-level.While our research focus resides in the macro

level effects of ISO 9000 diffusion, it behovesus to consider the micro foundations behind theway in which ISO certification affects relationships between businesses that hail from differentnations. Foss and Pedersen (2004) underscore theimportance in IB of identifying explanatory mechanisms at deeper levels. Additionally, Snowdon andStonehouse (2006) highlight the importance ofgrounding theoretical constructs in actual behaviour. Basing our theoretics on micro foundations ispivotal, as "nations do not trade but firms do"(Murray, Kotabe, & Wildt, 1995: 195). Accordingly,we focus on three particular properties (qualitysignal, common language, and conflict settling) bywhich ISO adoption may enhance trade betweenfirms from different nations and thereby enhancecountry-pair trade.First, ISO 9000 firms can credibly claim that theyhave a documented quality system that is implemented and followed, ensuring that products aremade to exacting specifications (Guerin & Rice,1996). While Rao, Ragu-Nathan, and Solis (1997)find cross-national evidence supporting ISO certification as positively impacting on firm quality,Conti (1999: 458) reports that the quality assuranceelements of ISO 9000 help in business-to-businesscontractual relations "when a supplier's capabilityto design and supply conforming products need tobe demonstrated". Terlaak and King (2006) arguethat ISO can signal a firm's superior but unobservedattributes: in line with this argument, they find ISOcertification to increase facility growth whenbuyers face multiple suppliers, and the attributesof suppliers are intangible. Foreign suppliers ofintermediary products clearly manifest difficultiesin demonstrating product quality: informational,linguistic, cultural, regulatory, and managerialbarriers all conspire to enhance "liability of foreignness" when one considers purchasing from foreign




6/22

_Impact of ISO 9000 on trade and FDI Joseph Cloughertyand Michat Grajek_;_j_ 617

suppliers. In short, ISO 9000 offers a "low-cost"signal of a firm's commitment to quality. Accordingly, the signalling implications of ISO certification help reduce trade barriers between businessesfrom different nations, and thus facilitate trade byreducing the search costs involved with identifyingquality-valuing foreign suppliers.Second, ISO certification has been reported toinvolve common-language properties that facilitatecommunication between businesses (Casper &Hancke, 1999; Dissanayaka, Kumarsawamy, Karim,& Marosszeky, 2001; Grajek, 2004). ISO 9000certification ensures that an organisation has adocumented quality management system that iscodified into a comprehensive manual (Mendel,2002). Consequently, certification establishes anefficient means to both communicate internalsystems to customers and provide a common procedural language to be used across organisations.Not only does ISO 9000 provide purchasers with ameans to decipher a firm's quality commitment,but also the purchasing firm is freed from designingand paying for the system (Hayes, 1994). Thecommon-language properties of ISO 9000 thusrepresent an efficient improvement upon the situation where each business customer specifies uniquerequirements for quality control. For instance, ISOadoption by US auto firms for supplier relationseliminated three other quality standards (Andersonet al., 1999). In this vein, Chen, Otsuki, and Wilson(2006) empirically support that exporters concentrate on fewer foreign markets when they mustcomply with separate country-specific standards.Accordingly, the common-language properties ofISO certification help reduce trade barriers betweenbusinesses from different nations by allowing for theready communication of a firm's production system,and by eliminating costly multiple quality standards.Third, codified procedures a la ISO 9000 may helpsettle - and reduce - organisational disputes, as theauthority of the documented system can "beemployed overtly in everyday battles of organizational contestation" (Mendel, 2002: 414). Lee(1998) reports "better team spirit" and "less staffconflict" as prominently cited benefits of ISOadoption. Casper and Hancke (1999: 968) arguethat ISO "makes links between different productionunits more transparent and, hence, more easilyamenable to improvement". Consider howMatthews (1986) departs from Williamsonianopportunism by emphasising the purely cognitivecosts of organising and monitoring transactionseven when participants are honest. Accordingly,

ISO can help clarify situations where faults are unintended and not due to opportunistic behaviour,thus preventing escalation of inter-organisationalconflict. As evidence, Hayes (1994: 56) reports thatISO firms consider certification to signal that theyare "easier to deal with". The ISO working procedures help clarify the division of labour andresponsibilities involved with complex interorganisational production, and by doing so minimise inter-organisational conflict. Accordingly, theconflict-settling properties of ISO certification helpreduce trade barriers between businesses fromdifferent nations by setting reasonable workingprocedures that smooth inter-firm relations andreduce instances of conflictual hold-up.In sum, ISO 9000 helps standardise practices andterminology, mobilise resources, and structureefforts across organisations. The quality-signal,common-language, and conflict-settling propertieslower the transaction costs and information asymmetries involved with business-to-business relations across borders, thus making arm's lengthtrading relations less costly.

Implicit throughout the above discussion is thatISO's mitigation of the transaction-cost andbounded-rationality problem is most germane forthe home country (what we sometimes refer to asthe "push effect') in country-pair trade. ISO 9000adoption in the home-nation signals the overallquality and reliability of exports (the quality-signal,common-language and conflict-settling effects);hence, we expect home-nation diffusion of ISOcertification to robustly "push" exports in variouscountry-pairs. ISO 9000 adoption in the hostnation (what we sometimes refer to as the "pulleffect') should involve the common-language andconflict-settling properties discussed above; yet thequality-signalling implications are more likely to beabsent for the host effect. In general, home-nationfirmswould be likely to exhibit, at best, moderateconcern as towhether customers were ISO certifiedor not; instead, selling firms are simply generallyopen to sending their products to whoever iswilling to pay the going price. In other words, wehypothesise that ISO adoption matters more onthe selling end than on the buying end of thetransaction. From the micro foundations outlinedabove, we generate three macro-level propositions:

Proposition la: Home-nation diffusion of ISO9000 certification (the selling end) is likely togenerate enhanced exports within country-paireconomic relationships.




7/22

_5__Impactof ISO 9000 on trade and FDI Joseph CloughertyandMichat Grajek_618

Proposition lb: Host-nation diffusion of ISO9000 certification (the buying end) is likely togenerate enhanced exports within country-paireconomic relationships.Proposition lc: The impact of home-nationdiffusion (the selling end) is likelyto be morerobust in magnitude than the impact of hostnation diffusion (the buying end).

COUNTRY-PAIR FDI POSITIONSWhile we have focused to this point on the effectsof diffusion on trade, Neumayer and Perkins (2005:239) note that the goal of ISO 9000 "has been to facilitate international trade and investment by harmonising otherwise diverse and conflicting nationalstandards with international ones". Accordingly, ISO9000 diffusion may also impact country-pair FDIstocks by affecting foreign investment costs. Recallthat FDI consists of both JVs and wholly ownedsubsidiaries - where wholly owned subsidiaries canbe obtained via acquisition or greenfield. We analysethe potential impact of ISO adoption on FDI whileconsidering the same three ISO properties: qualitysignal, common language, and conflict settling.More specifically, we analyse how ISO 9000 certification lowers the transaction costs and information asymmetries involved with firms makingforeign investments. By attempting to explore themicro foundations behind how ISO certificationreduces the costs of engaging in JVs, acquisitionsand greenfields, we base our theoretical propositions on deeper-level explanatory mechanisms.First, the quality-signal properties of ISO 9000

may assist firms inmaking FDI decisions, as highinformation barriers exist regarding the assessmentof foreign firms' quality properties. To the degreethat ISO 9000 represents a low-cost signal of a firm'scommitment to quality, this impacts directly on thecosts of engaging in international JVs and acquisitions. Recall that inadequate target evaluation (andan insufficient pre-acquisition decision process) isoften credited as a factor in failed acquisitions andJVs (Haspeslagh & Jemison, 1991; Hitt, Harrison,Ireland, & Best, 1998; Srivastava & Datta, 2006).For instance, Hitt et al. (1998) make it clear thatEcolab's unsuccessful acquisition of Chemlawn wasdriven by Chemlawn's low managerial and workerproductivity - something discovered only postacquisition. Related to the practice of neglectingtarget evaluation, Singh (2006) recently pointedout that business scholars have failed to delineatethe actual drivers of target selection. Ricart et al.

(2004) implicitly argue that the role of essentialinstitutional intermediaries - which help consummate transactions by evaluating quality and matching potential buyers and sellers - is a place to beginlooking for target selection drivers.We submit that ISO 9000 certification provides aquality signal that reduces the search costs involvedwith identifying "quality valuing" firms to eitherpurchase or JVwith. We do not consider ISO 9000certification to be the ultimate standard of production quality, nor the only indicator that firms use tosearch for JV and acquisition partners: witness, forinstance, the fact that ISO certification is for aplant or premise and not a firm. Hence itwill be -depending on the nature of the transaction - anoisy signal at times. Yet Hudson and Jones (2003)highlight that standards must not generate fullconfidence or information on the part of buyers tobe helpful; instead, as long as they convey someinformation and are cost-efficient, they becomeone signal included in the optimal signalling set.We argue that ISO 9000 represents an importantsignal that reduces the information barriers andtransaction costs with regard to finding foreignfirms (or plants/premises) with which to engage inpartnerships and M&As.While the analysis above highlights the impact ofISO 9000 on the JV and acquisition forms of FDI,ISO 9000 diffusion may also impact on greenfieldtype investments. Almost all forms of FDI rely onhost-environment internal markets to provideproduction process inputs. Yet one of the chiefbarriers to undertaking greenfield FDI is theinability to detect which firms in the host environment can be valuable upstream suppliers byproviding quality inputs: this represents one ofthe principal liabilities of foreignness, and is thereason why many firms seek to acquire localknowledge and contacts through acquisitions andJVs. Akin to our constructs in the trade section,host-nation ISO 9000 diffusion reduces the searchcosts involved with identifying good local suppliersand, by so doing, lowers the costs of engaging ingreenfield-type FDI.Second, the common-language properties of ISO9000 may assist firms interested in making FDIdecisions. Recall that many scholars (e.g., Cartwright & Cooper, 1993; Hunt, 1990; Larsson &Finkelstein, 1999; Pablo, 1994; Shrivastava, 1986)agree that inadequate integration represents thesingle most important factor in explaining failurein acquisitions and JVs.Moreover, the integrationof firms and/or units from different nations




8/22

_Impact of ISO 9000 on trade and FDI Joseph CloughertyandMichat Grajek__^ 619

generates further challenges owing to the distancein culture and managerial practices (Brown,Rugman, & Verbeke, 1989; Fey & Beamish, 2001;Geringer, 1988). Nevertheless, a great deal ofresearch finds similarity between firms to enhanceintegration and thereby enhance success (e.g.,Brown et al., 1989; Datta, 1991; Fey & Beamish,2001; Harrigan, 1988). Coff (2002) sums the aboveup well when he argues that you need some type ofrelated experience to successfully acquire andintegrate strategic assets. Fey and Beamish (2001)go further to uncover the organisational foundations behind the positive relationship between firmsimilarity and successful integration of JVs andacquisitions. In particular, Fey and Beamish finddifferences in managerial practices and values togenerate process losses - losses that derive fromineffective communication, uncertainty over control systems, and difficulties over organising work:hence organisational costs increase substantiallywhen you attempt to integrate dissimilar organisational climates via an M&A or JV.We submit that ISO 9000 certification provides acommon language (i.e., common routine, sharedexperience, similar climate, etc.) that helps firmsfrom different nations consummate (i.e., successfully integrate) international acquisitions and JVs.The common-language properties assist firmswith regard to inter-firm cooperation by enhancingcommunication, reducing uncertainty with regardto control systems, and providing a common approach to organising work. In support of this conjecture, Bergman (1994) provides healthcare industryexamples where similarity in quality managementsystems eased the integration of new JVs andacquisitions. Further, Shani and Rogberg's (1994)case-based research finds ISO 9000 helpful inaddressing the internal change brought on by amerger. In short, ISO 9000 provides a language thatintegrating entities can use to peer inside andcommunicate with each other, and thus aids theall-important integration process.Third, the conflict-settling properties of ISO 9000may assist firms interested inmaking FDI decisions.While we covered above the pivotal need forintegration ifa JVorM&A is to succeed, integrationis often fraught with tensions for the employees ofacquired firms or junior JV partners. Acquired entities are customarily required to alter their organisational style substantially: that is, they bear thebrunt of the change in terms of operations, control,planning systems and procedures, and evenhuman resource issues (Buono & Bowditch, 1989;

Haspeslagh & Jemison, 1991). Accordingly, theformal controls necessary for substantive integration may breed resentments on the part of acquiredfirms' employees- resentments that manifest

themselves in poor morale, employee stress,increased absence, employee turnover, and loweredproductivity (Cartwright & Cooper, 1993). Larssonand Lubatkin (2001) point out that acquiring firmsface a trade-off between "properly integrating" and"harmoniously acculturating" the new acquisition- an acquirer, simply put, cannot have both. In fact,recent research suggests that the greatest employeeresistance is elicited when the success of the interfirm combination depends on gains from integrating similar production and marketing operations(Larsson & Finkelstein, 1999): hence the abovetrade-off ismost present exactly when integration ismost necessary.We submit that ISO certification may mitigate theconflicts (i.e., employee resistance) noted aboveand allow integration to take place withoutsubstantial tensions. This dynamic is due to ISOcertification on the part of target firms, meaningthat less change isnecessary for target-firm employees: hence less conflict results from integration.Related to the common-language properties previously discussed, ISO-certified target firms willhave systems in place that allow the acquiring firmto gain information and take control: thus lessresentment-augmenting change is required duringthe integration process. In fact, Larsson andFinkelstein (1999) empirically support management style similarity to increase both organisational integration and employee cooperation, andfurther find that organisational integration andemployee cooperation generate merger synergies.Larsson and Finkelstein conclude that "when management styles are similar across organizations, thelevel of cooperation is often enhanced and perceptions of the degree of change taking place maybe cushioned" (1999: 8). Thus the cooperation ofthe new employees allows interaction and coordination to take place, which in turn increasesthe likelihood of success. In short, by smoothingrelations between integrating firms from differentnations (attenuating employee resistance yetallowing integration), ISO 9000 certification lowersthe costs of engaging in international JVs andacquisitions.We previously proposed that ISO's mitigationof the transaction costs involved with tradewould be more robust forhome nations (the sellingend) than for host nations (the buying end). The




9/22

*"_Impact of ISO 9000 on trade and FDI Joseph CloughertyandMichat Grajek_620

disproportionate effect of ISO 9000 diffusion on theselling end is likely to be replicated for FDI;however, with FDI the buying end (now the homenation) and the selling-end (now the host nation)are reversed. To illustrate the imbalance betweenthe host-nation and home-nation effects, considerthe quality-signal benefits of ISO diffusion on FDI.ISO diffusion in the host-nation environmentreduces the search costs involved with identifying"quality-valuing" local firms to either purchase, JVwith, or source inputs from (i.e., the costs of foreignacquisitions, JVs, and greenfields are all reduced):thus we would expect ISO adoption in the hostcountry to robustly "pull" FDI via the qualitysignal effect. Yet the diffusion of ISO in the homenation environment should significantly affectonly JV-type FDI, where itwould be in the interestof host-country firms to make efficient pairingswith ISO-certified firms from the home country.Host-nation firms interested in being acquired areunlikely to fully value ISO 9000 certification ofhome-nation suitors; instead, price will be theoverwhelming criterion by which to judge suitors.A target firm will be relatively uninterested inthe degree to which an acquiring firm can signalits quality, and relatively more interested inthe bottom line: that is, no price discount willbe given to ISO-certified bidders. The same typeof dynamic will be present for greenfield-typeFDI, where host-nation suppliers are unlikely todiscriminate between ISO-certified and non-ISOcertified home-nation MNEs interested in buyingtheir products.

Accordingly, host-nation adoption of ISO 9000provides clear benefits for home-nation firms interested in engaging in international JVs, acquisitions,and greenfields: the quality-signal, common-language,and conflict-settling effects help illuminate thetarget environment for FDI. Yet home-nationadoption of ISO 9000 provides relatively fewbenefits for home-nation firms interested inmaking foreign investments; the benefits are mostlyrestricted to home firms seeking international JVs.In other words, we again hypothesise that ISOadoption matters more on the selling end than onthe buying end of the transaction. From the microfoundations outlined above, we generate threemacro-level propositions:

Proposition 2a: Home-nation diffusion of ISO9000 certification (the buying end) is likelytogenerate enhanced FDI within country-pair economic relationships.

Proposition 2b: Host-nation diffusion of ISO9000 certification (the selling end) is likely togenerate enhanced FDI within country-pair economic relationships.Proposition 2c: The impact of host-nation diffusion (the selling end) is likely to be more robust inmagnitude than the impact of home-nationdiffusion (the buying end).

COUNTRY-PAIR HETEROGENEITYAn analysis of how ISO 9000 impacts on FDI andtrade in country-pair economic relationships cannotend with the idea that the effects of ISO diffusionhold universally. We must also factor in the likelihood that ISO's proposed effect on FDI and trademay vary in intensity across different country-pairtypes. In a recent paper, Blonigen and Wang (2004)focus directly on the question of whether it isappropriate to pool data empirically from differentnations; moreover, they find that "the underlyingfactors that determine the level of FDI activity varysystematically across LDCs and DCs" (2004: 17).Thus Blonigen and Wang deliver themessage that itis inappropriate to assume that FDI plays the samerole in developing nations as it does in developednations. While it is not the focus of their research, afew IB researchers (Globerman & Shapiro, 2003;Sethi et al., 2003) have also found the impact of theirexplanatory constructs on FDI to vary substantiallybetween developed and developing nations.There appears, then, to be empirical support forthe need to differentiate between developed anddeveloping nations when one studies FDI drivers.Indicative of the importance of this finding, Kobrinand Wu (2005) observe that the Blonigen-Wangprescription not to pool cross-country data automatically is quickly becoming common knowledgeand practice. The Blonigen-Wang insight is new:hence theoretical justifications for the developed/developing distinction and the potential forother distinction types (e.g., transition nations,newly industrial countries, colonial legacy) havegone unstudied. Mindful that other distinctions arepossible, we consider whether our proposed ISOeffects hold for both developed and developingnations. Furthermore, we proffer two theoreticaljustifications as to what might be behind any

developed/developing distinction when it comesto ISO's influence on trade and FDI.First, the fact that developed and developingnations tend to experience different types ofFDI and trade may contribute to any developed/




10/22

_Impact of ISO 9000 on trade and FDI Joseph CloughertyandMichaJGrajek_"_}j_ 621

developing distinction. Horizontal FDI (foreigninvestment in the same industry abroad as thefirm's domestic industry) is generally found whenthe home and host nations have similar attributes(factor endowments, level of development andsize); in practice, this means that FDI between twodeveloped nations tends to be horizontally based.Further, higher trade costs tend to generate higherlevels of horizontal FDI: hence trade is oftenconsidered a substitute for horizontal FDI. VerticalFDI (investment abroad to source inputs or selloutputs from home production) is generally foundwhen the home and host nations have dissimilarattributes; in practice, this means that FDI from adeveloped home nation to a developing hostnation tends to be vertically based. Further, highertrade costs tend to generate lower levels of verticalFDI: hence trade is often considered a complementforvertical FDI. Accordingly, the fact that developing nations tend to experience more vertical-typerelations in their trade and FDI, while developednations experience more horizontal-type relationsin trade and FDI, might influence the impact of ISOdiffusion. Recall that transaction cost and hold-upproblems tend to be most germane in verticalrelationships where buyers and sellers are firms(Williamson, 1985). We might then expect ISOcertification's impact to be most robust in verticaltype economic relationships (i.e., in developingnations), as this is where transaction cost andinformation asymmetries can be most problematic.Second, the fact that developed and developingnations tend to differwith regard to the presence ofinstitutional intermediaries may also explainany developed/developing distinction. Ricart et al.(2004) note that developing nations oftenexperience institutional voids (a shortage of thespecialised institutional intermediaries needed toconsummate transactions): in particular, the lack ofsoft institutions via which to locate possibletransaction partners undercuts commerce. Consider that government agencies, trade associations,commercial banks, and export management companies often perform information-gathering andmatchmaker services for firms desiring to expandoverseas via exporting and/or investment (Hill,1998). These services play instrumental roles inbridging the financial, logistical, business-practice,and cultural divides endemic to IB. Yet theseservices that help mitigate information asymmetries and transaction costs are less prevalent in thedeveloping world. In this vein, Anderson andMarcouiller (2002) argue that capital-abundant

nations trade disproportionately with each otherbecause they have the strong institutions thatsupport trade security. Simply put, governmentsin

developing nations findit more difficult to

support such public services financially, and themarket options (e.g., commercial banks and exportmanagement companies) are less evolved in thedeveloping world - this is ironic, in that developingnations are the very locations where such serviceswould be most valuable. While Delmas (2002)posits that the developing world's lack of institutions might dampen ISO diffusion, we want topoint out that what diffusion actually taking placewill involve significant effects owing to the dearthof other institutional means via which to reduceinformation asymmetries and transaction costs.Accordingly, we might expect the influence of ISOto be most robust in developing nations, as this iswhere fewer substitutes exist.

In sum, firms from developing nations will facesubstantial challenges in readily conveying productquality, internal production processes and conflictfree relations for two reasons: the vertical nature oftheir international economic relationships, and theshortage of institutions to help mediate transactioncost and information asymmetry problems. ThusISO certification of developing-world firmswill beparticularly helpful in attracting developed-worldcustomers for exports and developed-world buyersfor investment purposes. A simple propositioncaptures the argument above:

Proposition 3: The impactof ISO 9000 diffusionon both trade and FDI is likely to be greater indeveloping nations than in developed nations.

DATA ISSUESData for our empirical analysis are compiled fromdifferent sources, including the UN's Comtrade, theIMF's International Financial Statistics, the WorldBank's World Development Indicators, and theWorldEconomic Forum's Global Competitiveness Report: seeTable 1 for the full list of variable definitions andsources. The primary data consist of bilateral (i.e.,country-pair) trade flows and FDI stocks over the1995-2002 period. Our bilateral FDI data comefrom OECD nations who report bidirectional FDIstocks for their economic relationships with bothnon-OECD and OECD nations. Accordingly, thehome nations and host nations for the variouscountry-pair economic relationships include 52OECD and non-OECD members.




11/22

^_Impact of ISO 9000 on trade and FDI Joseph CloughertyandMichal Grajek_622

Table 1 VariabledefinitionsVariable Source Definition

ExportsN's Comtrade Exportsfromhome to host country inmillions of US dollars)FDI_out OECD StatisticalCompendium Home country's FDI stock inhost country inmillions ofUS dollars)FDI_in OECD StatisticalCompendium Host country's FDI stock inhome country inmillions ofUS dollars)home_GDP IMF'sFS Home country'sGDP (inmillions ofUS dollars)host_GDPMF's IFS Host country'sGDP (inmillions ofUS dollars)homeJSO ISO (2003) Number of home ISO9000 certificationshostJSOSO (2003) Number of host ISO9000 certificationshost_Passengers World Bank'sWDI Number of passengers carriedby aircraftsdeparting fromhost country'sairports per capitahost_Road World Bank'sWDI Total network f paved roadsper km2 inhost countryhost_Electricity World Bank'sWDI Host country'selectricity roduction (in kWhper capita)rel_REERMF's IFS Real effectiveexchange ratebased on relative onsumer prices (highervalueindicatesrealdepreciation of home vs host currency)rel_SMIorld Bank'sWDI Relative stockmarket indexgiven bymarket capitalisationof home relativeto host countryhost_Polcon Henisz (2002) Index of political constraints inhost countrybased on veto points (highervalue indicates tighter control over politicians' decisions)host_TCI World Economic Forum'sCCR Trade cost index reflecting idden barriersto imports f host country(1=highest, 7=lowest)host_Tariffs OECD StatisticalCompendium Custom and other import uties as a fraction f total imports nhostand World Bank's WDI country; OECD data were used ifavailable, otherwise World Bank's data.host_FCM World Economic Forum'sCCR Restrictions o access foreign apitalmarkets inhost country 1=highest,7=lowest)host_HF World Economic Forum'sCCR Hiringand firing ractices inhost country 1=impeded by regulations,7=flexiblydetermined by employers)BIT UN Bilateral investmenttreaty etween home and host nation (0/1)FTA Baier and Bergstrand 2007) Free trade agreement between home and host nation (0/1)

and sources listed hereWe employ the OECD data - not US or Scandinavian - on FDI stocks in order to ensure sufficient

variation in terms of home-nation ISO adoption,and to allow investigation of both the push andpull performance of ISO 9000. The OECD StatisticalCompendium reports both FDI flows and stocks, butwe employ the stock data, as the flow data involverepatriation of profit issues, which can lead toinappropriate measurement of a particular nation'sforeign investment position. Additionally, Hejaziand Safarian (2001) point out that stock data havethe added advantage of mitigating simultaneityissues between FDI and trade.Our explanatory variables aim to control standard factors driving export and FDI activities

beyond the proposed ISO 9000 effects: gravity force(GDP), infrastructure (airline passengers carried,paved road network, and electricity production),financial and wealth conditions (relative exchangerate, and relative stock market index), political risk(Henisz's (2002) veto-points index), trade costs(tariff and non-tariff barriers to trade, and presenceof free trade agreements), labour costs (hiring and

firing practices), and capital costs (presence ofbilateral investment treaties, and restrictions toforeign capital market access). Table 2 reportssummary statistics for our full list of variables. Anarray of other factors is implicitly taken into accountby means of fixed country-pair and period effects.Country-pair specific effects account for commonborder, common language, colonial ties, and otherunobserved factors that do not vary over the sampleperiod. Period-specific effects account for variationinworld GDP and other annual trends.One trade-off involved with the extensive list ofcontrol variables is that each variable containsvarious missing observations. Moreover, the missing observations for each control variable do notline up. Accordingly, the number of feasible observations drops significantly (by over 90%) when thefull set of control variables is included in the esti

mations. We then face a trade-off between having alarge sample and having a rigorous estimationprocedure with multiple controls. Our modelspecification therefore needs to strike a balancebetween the sample selection problem and the




12/22

_Impact of ISO 9000 on trade and FDI Joseph Clouqherty and Michai Grajek_^J_^ 623

Table 2 Summary statisticsLDC-+DC DC^LDC DC^DC

Variablebs. Mean Std. dev Obs. Mean Std. dev Obs. Meantd. dev

Exports68 2038.7 4116.6 1796 1908.4 5924.5 1271 5996.0 14,076.2FDI.out043 170.7 397.9 1858 1492.8 3760.9 1315 7069.7 21,027.4FDI_in50 1423.6 2351.9 1099 194.9 574.3 1204 7399.8 23,227.0home_GDP 1043 224,847 229,507 1858 1,442,356 2,350,017 1315 1,313,508 2257,529host_GDP 1043 1,002,276 1,206,128 1858 214,753 232,848 1315 1,196,615 1,921,242homeJSO043 3302 6827 1858 14,789 16,952 1315 12,054 15,493hostJSO 1043 14,732 17,537 1858 3039 6835 1315 11,182 16,123host.Passengers 875 1.01 0.58 1858 0.45 0.81 1315 1.43.93host_Road 647 167.6 91.8 1348 65.2 111.0 1315 118.05.3host_Electricity 1043 8292.0 4493.8 1858 3373.6 2031.3 1315 9936.6 6056.1rel_REER 584 0.99 0.16 1112 1.03 0.16 1315 1.00.12rel_SMI15 0.99 2.53 1576 139.81 934.99 1042 17.11 62.48host_Polcon 1043 0.77 0.05 1804 0.61 0.23 1315 0.76.04hostJTCI 1043 5.82 0.68 1858 4.69 0.90 997 5.85.79host_Tariffs 741 0.01 0.01 1225 0.03 0.03 1121 0.01.01host_FCM 999 6.48 0.38 1858 4.95 1.19 954 6.37.47host_HF043 3.40 0.92 1838 3.94 0.91 997 3.71.94BIT90 0.42 0.49 1198 0.48 0.50 957 0.00.05FTA90 0.12 0.33 1198 0.10 0.30 957 0.63.48

benefits of including an extensive set of controlvariables.

ESTIMATION ISSUESWe build and estimate a system of two gravityequations for both FDI and exports. In order toestimate the two gravity equations properly, itbehoves us to return to the nature of the FDItrade relationship, as this directly determines theproper estimation technique for the two equations.As already alluded to, it is appealing to consider thetrade-FDI relationship to be substitutable wherehorizontal FDI dominates (e.g., developed-developed country pairs), and complementary wherevertical FDI dominates (e.g., developed-developingcountry pairs). Recall also that empirical scholarship found horizontal FDI to outweigh vertical FDI:hence scholars were surprised when early empiricalwork found a complementary - not a substitutable- relationship between trade and FDI (Hejazi &Safarian, 2001). Grubert and Mutti (1991) began theprocess of reconciling these findings by pointingout thatmuch of that empirical work suffered fromendogeneity by regressing exports on an indicatorof direct investment activity. More specifically, thecomplementarity findings may be driven byunobserved variation in tastes, technology, comparative advantage and government policy, which

could all create a positive correlation between tradeand FDI.A more recent empirical literature takes intoaccount the above endogeneity issues and continues the debate over the complementarity (see

Clausing, 2000; Head & Ries, 2001) or substitutability (see Belderbos & Sleuwaegen, 1998; Blonigen,2001) of tradeand FDI. The Blonigen (2001) studyis particularly important: using product-level data,he identifies substitution (as well as complementarity) between trade and FDI. He suggests that theinability of some researchers to find substitutionrests both with an aggregation bias and withsubstitution effects being a sudden - not gradual -phenomenon. Accordingly, we move forward withthe idea that trade and FDI exhibit a substitutiverelationship that can best be detected via exogenous price indicators - though we are mindful thatcomplementarity will also be at play in ourcountry-pair level data. As seen below, this pullsus in two different econometric directions.The substitution argument roughly correspondsto idealised horizontal FDI - or,more specifically, tothe decision on whether to supply a foreign marketwith arm's length trade or foreign-subsidiary sales.The relative costs associated with the two types offoreign presence mode determine this decision.From the econometric viewpoint, trade flows andFDI are not interdependent - for example, FDI does




13/22

"***_Impact of ISO 9000 on trade and FDI Joseph CloughertyandMichaJGrajek_624

not cause trade - in an idealised horizontal world,but are jointly determined, which leads naturally toZellner's (1962) seemingly unrelated regressions(SUR) model. Moreover, each control variable

-which directly affects the cost of at least one foreignpresence mode - appears in both the FDI and exportequations, since what matters in choosing betweensubstitutes is relative costs. The resulting SUR modelwill then have identical regressors. Error terms arealso likely to be correlated across the two equationsowing to omitted variables (the factors affectingtrade might also affect FDI via the substitutionargument) and common shocks.The complementarity argument corresponds better to a state where cross-border economic activity isvertical in nature, driven by parts of the value chainbeing located in different nations owing to production efficiency. Complementarity between trade andFDI leads to a different econometric specificationfrom that for the substitutability of exports and FDI(i.e., the horizontal world). In this case, the decisiontomove production abroad comes prior to the tradedecision: first,you establish foreign subsidiaries forsourcing home production; second, you source theforeign subsidiaries with inputs and then re-importprocessed goods. In other words, a direct causalrelationship exists between FDI and trade in anidealised vertical world, as FDI outward stocks"cause" both incoming and outgoing trade flows.Yet no reciprocal causality exists in this context, astrade does not cause FDI. This implies a triangularmodel (Lahiri & Schmidt, 1978) in the econometriccontext: FDI variables appear in the export equation,but trade variables do not appear in the FDIequation.As we are unable to separate out the horizontaland vertical-type activities in our country-pair leveldata, our model must simultaneously accountfor both activities. Unfortunately, the two models- SUR with identical regressors, and triangular- cannot be nested and consistently estimatedin a single model without further restrictions.Estimating a triangular model generally requiresthe application of instrumental variable techniquesto account for endogeneity - in the current context,the endogeneity of FDI in the export equation. Thehorizontal nature of the trade/FDI relationship,however, suggests that what impacts on trade willalso impact on FDI, thus rendering identification ofthe export equation parameters impossible.

Ignoring the vertical world (i.e., omitting theinward and outward FDI variables from the export

equation) and estimating the SUR model withexogenous drivers of operating abroad via exportsor FDI represents one possibility. Grubert and Mutti(1991) advocate such an approach, and in lightof our previous discussion, we could interpretthis approach as the estimation of a reducedform model. Although econometrically correct, areduced-form model presents difficulties in inter

preting coefficient estimates. In particular, wewould not be able to distinguish the direct effectof ISO diffusion on exports from the indirect effectsof ISO diffusion on exports via FDI.Another possibility involves assuming uncorrected error terms across equations, thus leading to aspecial type of triangular model known as a"recursive model". With the recursive model, theright-hand-side parameters in the FDI equation areidentified owing to exogeneity of the independentvariables. Hence FDI serves as its own instrument inthe export equation, as it is a linear combination ofexogenous variables, and the FDI equation's errorterm is uncorrected with the export equation'serror term. Yet the uncorrelated-errors assumptionis a strong one, which would be violated in thepresence of omitted variables and common shocksto trade and FDI. An additional benefit of suchan approach is that the complementarity-drivenindirect effects in the export equation should bechannelled through the inward and outward FDIvariables, rather than through the other explanatory variables. We stress "should," however, as thebroad nature of the FDI data (not distinguishingbetween vertical and horizontal FDI) suggests thatthe vertical effectsmay be difficult to control fully.Bearing in mind the potential endogeneity problem, we follow the recursive approach. The FDIand export equations then read

In FDI?UT = X'ijtymi + Kt+ fy + xj,ijt (1)

InEXPORTS^ HSxlnTOlgF+ /J2lnFDlg?X,^EXP (2)+ h + mi etjtwhere the subscripts i, j, and t stand for homenation, host nation and year, respectively. In orderto apply standard estimation techniques, thegravity Equations (1) and (2) are log-linearised.Vector Xyt includes the logged values of allregressors identical across equations. Additionally,we include time effects (Xt nd Kt) and country-pairspecific effects (rjij nd %). Such an approach allows




14/22

_Impact of ISO 9000 on trade and FDI Joseph CloughertyandMichatGrajek_'** 625

for autocorrelated and heteroskedastic error terms,but not for correlation across equations. Underthese assumptions, we can consistently estimateEquations (1) and (2) with equation-by-equationOLS after removing from the data the unobservedheterogeneity embedded in the country-pair specific effects. Primarily, we use a within transformation leading to fixed-effects estimators - though,alternatively, we apply first differencing. Furthermore, all regression estimations are robust toheteroskedasticity and autocorrelation by usingrobust standard errors.We also need to acknowledge that three additionalsources of endogeneity potentially exist beyond theinterdependence of trade and FDI. First, the gravityforce itself is subject to endogeneity problems

asexports necessarily enter GDP in national accountings. Hummels and Levinsohn (1995) report, however, that correcting for GDP endogeneity withinstrumental variables makes very little differenceto coefficient estimates. Second, our research focus ison themacro-level impact of ISO diffusion on tradeand FDI, yet we would like to be able to makeinferences with regard to the micro-level - theagents that make trade and FDI decisions. Forinstance, we would appreciate some degree ofconfidence that ISO-certified firms actually engagein more exports and FDI. Exogeneity of the ISOvariables would allow us to interpret the coefficientestimates as indicating whether ISO-certified firmsexperience increased IB activity at themicro level.Third, endogeneity may owe to reciprocal causation between ISO 9000 diffusion and both FDI andexports. Recall that the dominant literature on ISO9000 attempts not to explain diffusion effects butto explain diffusion drivers. For instance, Guleret al. (2002) find ISO adoption in trading-partnerand trading-competitor nations, and a nation'soverall inward FDI to enhance ISO 9000 diffusionin a particular country. Furthermore, Casella (1996)notes that trade and FDI intensity (both of whichindicate the openness of an economy) mightdetermine the standardisation process. Our country-pair level of analysis may mitigate this problem,however, as the level of trade and FDI in oneparticular country pair is less likely to have a largeeffect on national diffusion of ISO 9000.We apply Sims-type and Hausman-type exogeneitytests in order to support the above intuitions:(1) the validity of a recursive econometric approach;(2) the likely trivial impact of gravity forces on

endogeneity;

(3) the ability tomake inferences with regard to themicro level; and(4) mitigated concern over reciprocal causation.First, we run a Sims test procedure that incorporates lead explanatory variables in the estimated equation and tests their joint significance(Wooldridge, 2002). The underlying logic is thatfuture explanatory variable realisations shouldnot affect present exports and FDI. Finding such arelationship indicates a violation of the strictexogeneity assumption, and renders fixed-effectsestimators inconsistent. Second, we apply aHausman test to a comparison of the fixed-effectscoefficients with the first-differencing coefficients.Under strict exogeneity, both estimators are consistent and have the same probability limit: thusfinding a significant difference between the twoestimates indicates violation of strict exogeneity(Wooldridge, 2002). We apply both tests to allthe explanatory variables as a group (see "overall"in Table 3), and then for the two ISO variables(see "ISO" in Table 3) of primary interest.

EMPIRICAL RESULTSAs already introduced, we follow a panel dataapproach that allows accounting for unobservedheterogeneity pervading cross-sectional studies.Beyond our main fixed-effects estimations (Table 3),we also estimate all regression equations whileusing a first-differencing method (Table 4). Comparing in detail first-differencing with the fixedeffects results allows us to check the robustness ofthe primary fixed-effects estimations. FollowingBlonigen and Wang (2004), we distinguish betweenthree distinct country-pair types: developing todeveloped (LDC->DC); developed to developing(DC->LDC); and developed to developed (DC-?DC). The 20 developed nations in our sampleinclude Australia, Austria, Canada, Denmark,Finland, France, Germany, Iceland, Ireland, Italy,Japan, Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, UK, and the US.Accordingly, each table presents six specifications:one trade and one FDI equation for each countrypair type. Note that we do not have observationnumber symmetry for the trade and FDI equationsin the different country-pair types, as an effectivetrade equation observation requires both inwardand outward FDI measures, and bidirectional FDIdata are not always present. Nevertheless, our mainresults are robust to the imposition of observationnumber symmetry for the trade and FDI equations




15/22

^_Impact of ISO 9000 on trade and FDI Joseph CloughertyandMichat Grajek_626

Table 3 Gravity equation for exports and outward FDI: fixed-effects estimation resultsExportsFDI

LDC->DC DC^LDC DC-* DC LDC->DC DC-+LDC DC-+DCRegressiono.(1) (2) (3) (4) (5)6)home_GDP 0.272** 0.237 0.288** -0.354 -0.360 0.174(0.121) (0.158) (0.125) (0.293) (0.286) (0.311)host_GDP 0.803*** 0.650*** 0.523*** 0.568 0.536*** 0.850***(0.240) (0.076) (0.087) (0.588) (0.176) (0.320)homeJSO 0.116** 0.033 -0.008 -0.079 -0.059 0.064(0.057) (0.050) (0.037) (0.097) (0.126) (0.104)hostJSO 0.085 0.018 0.043 -0.330* 0.148** -0.022(0.087) (0.031) (0.033) (0.190) (0.071) (0.112)FDLout -0.004 0.021 0.005

(0.016) (0.023) (0.011)FDLin 0.043 0.023* 0.016(0.029) (0.013) (0.010)host.Passengers 0.212*** 0.289*** -0.099 -0.363(0.078) (0.100) (0.187) (0.339)

host_Road 0.006* -0.004(0.003) (0.007)host.TCI -0.577 1.210**

(0.393) (0.528)host_FCM 0.263*** 0.406*

(0.076) (0.223)Sims test (overall)3 1.35 (7, 162) 1.30 (8, 216) 1.02 (8, 244) 2.31 ** (5, 224) 0.63 (6, 390) 1.02 (6, 275)Sims test (ISO)9 2.19 (2, 162) 0.48 (2, 216) 0.65 (2, 244) 2.12 (2, 224) 0.05 (2, 390) 1.15 (2, 275)Hausman test (overall)6 22.26*** (7) 31.92*** (8) 39.61 *** (8) 16.06*** (5) 6.44 (6) 3.03 (6)Hausman test (ISO)b 6.74** (2) 1.69(2) 1.63(2) 4.33(2) 2.07(2) 0.19(2)R2 0.227 0.323 0.220 0.070 0.250 0.207Observations 86818 1136 043 1858 1315Clusters (Panels) 21878 29533 410 318*p


16/22

_Impact of ISO 9000 on trade and FDI Joseph Clouqhertyand Michat Grajek_]_}_ 627

Table 4 Gravity equation for exports and outward FDI: first-differencing estimation resultsExportsFDI

LDC-+DC DC-+LDC DC-* DC LDC^DC DC-+LDC DC-* DCRegressiono.1) (2) (3)4)5) (6)

home_GDP.294*** 0.474*** 0.589*** -0.211 -0.377.260(0.064) (0.162) (0.133) (0.221) (0.372) (0.367)host_GDP.879*** 0.559*** 0.253*** 0.802** 0.364* 0.326(0.169) (0.094) (0.080) (0.399) (0.191) (0.427)homeJSO.063** 0.089 -0.044 -0.023 -0.187.108(0.028) (0.057) (0.027) (0.053) (0.120) (0.121)

hostJSO.086 0.004 0.062** -0.045 0.109*0.055(0.059)0.032) (0.027) (0.149)0.061) (0.121)

FDLout0.001 0.002 0.020*(0.010) (0.022) (0.011)FDI_in.028 0.016 -0.008

(0.021) (0.016) (0.012)host_Passengers 0.027 0.0590.184 -0.624(0.052) (0.084) (0.269) (0.383)host_Road 0.008*** -0.005(0.003) (0.011)

host_TCI -0.185 -0.240(0.154) (0.418)host_FCM 0.229*** 0.167(0.056) (0.215)fl2.133 0.216 0.100 0.020 0.067 0.077Observations 613 524 62561 118554Clusters (Panels) 159 217 24582 37572

*p


17/22

***_Impact of ISO 9000 on trade and FDI Joseph CloughertyandMichat Grajek_628

- conform to the pre-existing empirical literatureusing panel data techniques to estimate gravitybased trade equations (e.g., Freund & Weinhold,2004). In terms of the FDI equations, the hostnation GDP pull also seems relatively stable acrossdifferent country-pair types, but is insignificant inthe developing-to-developed country-pairs (regression 4),where the econometrics are weakest and welack precedent theoretics. Additionally, we do notfind the push (home-nation GDP) to be significantfor the FDI equations in Table 3. Furthermore, ourcontrols for FDI-driven trade - outward and inwardFDI stocks in the trade equations - are mostlypositive, as expected, but significant only forinward FDI in regression 2.Host-nation infrastructure, as proxied by thenumber of airline passengers, is positive and significant in the export equation for all country-pairtypes (except LDC-+DC), and negative, thoughgenerally insignificant, in the FDI equations. Similarresults are found for the paved road networkinfrastructure variable, though only in the developed-to-developed country pairs. These findingsconform to the expectation that infrastructureimprovements boost both horizontal-type and vertical-type trade. Furthermore, it appears that thenegative effect of infrastructure on horizontal FDI(via the substitution of trade for FDI when infrastructure costs are low) is potentially countervailedby the positive effect of infrastructure on vertical FDI.Hidden trade barriers - as measured by the WorldEconomic Forum's trade cost index (host_TCI) -turned out to be significant in the developing-todeveloped FDI equation (Table 3, regression 4).Hence fewer hidden trade barriers in developednations generate greater FDI from developingnations. While this estimate is consistent withhow vertical-type FDI will respond positively tothe lowering of trade costs, it should be treated withextreme caution, for a few reasons. As alreadyalluded to, regression 4 indicates the pooresteconometric performance among our set ofestimated equations. Notice that this is the onlyequation for which both the Sims and Hausmantests reject exogeneity of the "overall" explanatoryvariables. The R2 in this equation is also muchlower than in the other Table 3 regression equations. Furthermore, the coefficient estimate forhidden trade barriers is not significant in the firstdifferenced estimations (Table 4, regression 4). Theinconsistent results are not surprising, given therelative lack of theoretical guidelines for FDI fromdeveloping nations to developed nations.

Finally, foreign capital market access (host_FCM)significantly affected cross-border economicactivities for developed-to-developing country pairs.The positive and significant coefficient estimates inTable 3 for the exports and FDI equations (regressions 2 and 5) suggest that lax restrictions regardingthe access to capital markets in developing nationsspurs both imports and FDI from developed nations.This conforms to the expectation that financialmarket openness reduces the cost of doing businessfor both importers and foreign investors.Having discussed the performance of our controlvariables, we can finally turn to the impact of ISO 9000diffusion on trade and FDI - the focus of the paper.Our empirical results yield evidence concerning our

proposed expectations for how ISO diffusion wouldimpact on cross-country economic relationships:According to Propositions la and 2b, we expectedISO diffusion at the selling end (the home nationfor exports and the host nation for FDI) togenerate increased trade and FDI respectively.The empirical results provide some evidence infavour of these propositions in the developing-todeveloped exports equation and the developedto-developing FDI equation.According to Propositions lb and 2a, we expectedISO diffusion at the buying end (the host nationfor exports and the home nation for FDI) togenerate increased trade and FDI respectively.The empirical results reject these propositions, asnowhere does buying-end ISO diffusion enhancetrade or FDI.According to Propositions lc and 2c, we expectedISO diffusion at the selling end to be more robustthan at the buying end. The empirical results doprovide some evidence in favour of the sellingend effects being most robust.According to Proposition 3, we expected the ISOeffects to be most pronounced in developingnations. The empirical results provide strongevidence in favour of this expectation, as thepositive effects of ISO 9000 on the selling end ofthe transaction - home nation for exports andhost nation for FDI outward positions - in thedeveloping world are confirmed.In particular, the empirical results presented inTable 3 indicate three areas where ISO diffusion

significantly impacts on trade and FDI: homenation ISO in LDC->DC exports; host-nationISO in DC-*LDC FDI; and host-nation ISO inLDC->DC FDI. We turn now to specific discussionof these findings.

journal of International Business Studies



18/22

_Impact off ISO 9000 on trade and FDI Joseph Cloughertyand Michai Grajek_^}_^ 629

First, the fixed-effects coefficient estimate onhome-nation ISO adoption is significant and positive in the developing-to-developed export equation (regression 1). The diagnostics point, however,to some concern about strict exogeneity of ourexplanatory variables in general, and the exogeneityof the ISO variables in particular: both Hausmantest statistics are significant. This calls for a detailedexamination of the first-differencing results(Table 4) where the ISO push in the developingto-developed export equation is confirmed by asignificant coefficient estimate. Note that the ISOpush effect s significantlyower inTable 4 (0.063)than inTable 3 (0.116),which might be due to theISO variable accounting for the impact of othertrending omitted variables in the fixed-effectsestimation. First differencing, by eliminating deterministic time trends, is more immune to sucheffects. Moreover, the difference between the fixedeffects and first-differencing coefficient estimatesfor home-nation ISO may also be due to thesamples not fully overlapping: that is, we lose

many observations when first-differencing unbalanced panels (from 868 to 613 here). Wooldridge(2002) notes that positive Hausman tests may bedue to inconsistent sampling between the fixedeffects and first-differencing estimations. The bottom line is that we should treat the magnitude ofthe ISO push effect in Table 3 with caution, yet bereassured as to the significance of this effect.Keeping the above caveats in mind, our coefficient estimate for the ISO push effect in Table 3

suggests that a 10% increase in a developingnation's certificates boosts average exports todeveloped nations by 1.16% in a given year. Thismeans that - at the LDC->DC sample mean - oneadditional ISO 9000 certificate generates an averageUS$71,600 worth of exports in each LDC-DCcountry-pair relationship. To obtain the same effectas a 10% increase in certificates, a developing nation'sGDP would actually need to increase by 4.27%.Second, the fixed-effects coefficient estimate onhost-nation ISO diffusion is significant and positivein the developed-to-developing FDI equation(regression 5). The diagnostics (both the Sims andHausman tests) also indicate exogeneity for theexplanatory variables as a whole and for the ISOvariables in particular; also, the first-differencingcoefficient estimate in Table 4 provides corroboration. Hence ISO diffusion in developing hostnations appears to attract FDI from developednations. In terms of economic significance, thecoefficient estimate (0.148) forhost-nation diffusion

from Table 3 suggests that one additional certificatein a developing nation generates an averageUS$69,600 worth of foreign investment in eachDC - LDC country-pair relationship.Third, the fixed-effect coefficient estimate forhost-nation ISO is significant and negative inthe developing-to-developed FDI equations (regression 4). This result suggests that ISO diffusion indeveloped nations acts as a barrier to foreign investment from developing nations. Recall from ourdiscussion of hidden trade barriers that regression 4is the most troubling estimation in terms both ofeconometrics (the Sims and Hausman tests indicateendogeneity) and of theoretics (FDI from thedeveloping to the developed world is not wellstudied). Moreover, the significance of this effectdoes not hold up in the first-differencing estimation (Table 4, regression 4): hence we treat thisweak and puzzling result as a statistical artefact.In sum, our results support the hypothesisedbenefits of ISO 9000 adoption (the quality-signal,common-language and conflict-settling properties)empirically manifesting on the selling end of bothtrade and FDI transactions. Moreover, splitting thesample into different country-pair types, per Blonigen and Wang (2004), elicits strong support for theheightened importance of ISO diffusion in developing nations. Accordingly, the emergence of theISO institution in developing nations seems to bequite effective in solving the information asymmetry and transaction cost problems involved withFDI and trade.

CONCLUSION AND DISCUSSIONMotivated by the lack of literature analysing themacro-level impact of ISO 9000 diffusion, we setout to study the effects of a nation's ISO certification on country-pair economic relationships. Wesuggest three properties of ISO 9000 (quality signal,common language, and conflict settling) that easethe costs of conducting cross-border trade andinvestment. We also posit that these ISO effectswill be more robust in developing nations, as firmsfrom these nations face particularly high information asymmetries and transaction costs, both whenexporting and when attempting to attract foreigninvestment. Empirical tests based on data coveringthe country-pair trade flows and FDI stocks ofOECD nations over the 1995-2002 period yield twomajor findings:1. ISO 9000 diffusion in developed nations does notappear to enhance trade and FDI between nations.




19/22

**_Impact of ISO 9000 on trade and FDI Joseph Cloughertyand Michai Grajek_630

2. ISO 9000 diffusion in developing nations doesappear to enhance trade (exports to developednations) and FDI (investment from developednations) between nations.

The empirical results yield several implications,four of which we discuss here.First, our results may contribute directly to theNIE scholarship that considers how institutions

play a role that is complementary to that of otherinstitutions. Milgrom, North, and Weingast (1990)found that the medieval institutions of judgesand champagne fairs helped further the effectiveness of the "reputation system" used at that time toreduce the transaction costs of exchange. They state"institutions sometimes arise tomake ... [a]mechanism more effective by communicating information" (Milgrom et al., 1990: 19). The ISO 9000quality system appears to be an example of an institution whose principal aim is to convey information to a community. In the case of ISO, thecommunity consists of worldwide businesses interested in either exporting or investing abroad, andISO appears to complement theWTO system - atleast in the developing world. The ISO 9000 systemmay then assist theWTO in realising its stated goal:to help exporters, importers, and producers ofgoods and services to conduct their business.Second, our results also generate clear policyimplications for developing nations ifwe reconsider our empirical findings from a country's perspective. Recall that ISO diffusion in developingnations was found to have a positive push withregard to exports to developed nations, and not tosignificantly push outward FDI to developednations. ISO diffusion in developing nations wasalso found to have a positive pull with regard toinward FDI from developed nations, and not tosignificantly "pull" in imports from developednations. Taking this from a developing country'sperspective suggests the following: large-scale adoption of ISO generates increased exports to (but notsignificant increased imports from) developednations, and increased inward FDI from (but notsignificant increased outward FDI to) developednations. Hence, from a neo-mercantilist (or developmental) viewpoint, ISO diffusion in developingnations appears to generate the two public policygoals (increased exports and increased inwardinvestment interest) that reside behind the foreigneconomic policies ofmany nations.Third - and related to the above discussion ondeveloping-nation policy implications - our results

provide some evidence as to where the principalbenefits of globalisation amass. IB scholars (e.g.,Buckley& Ghauri, 2004; Eden & Lenway, 2001)haverecently expressed

concern as towhether thebenefits of globalisation accrue to the relatively rich(capital exporting) countries or to the relativelypoor (capital importing) countries. Many punditsand scholars (see Dunning, 2003) have expressedthe opinion that developed nations benefit substantially from globalisation, whereas developingnations experience few - if any - benefits. Ourresults are striking, in that they support the globaldiffusion of ISO 9000 practices to be beneficial fordeveloping nations: hence the global diffusion ofmanagerial practices potentially ameliorates - notexacerbates - global inequalities of wealth.Fourth, our results speak to two additional IBdialogues: the long-lasting dialogue on locationadvantages, and the recent dialogue on institutionalintermediaries. The empirical results suggest thatinstitutional intermediaries can indeed play a fundamental role in supporting commerce. In our case, thediffusion of ISO 9000 managerial practices (aninformal-decentralised international institution) indeveloping nations generates increased inward FDIfrom developed nations and enhanced exports todeveloped nations. These findings conform to theprior wisdom that firms in developing nations facesubstantial barriers to gaining both export marketsand foreign investment interest owing to insufficientgovernment ormarket remedies for transaction-costbased problems. In other words, the impact of ISO9000 is quite robust in developing nations where -because of institutional voids - few substitutes exist.The ISO 9000 system appears, then, to be animportant factor affecting cross-border economicrelationships: thus international institutions thatsignal managerial quality may represent an important location advantage, which helps explain thedestination for FDI and the source of trade flows.

ACKNOWLEDGEMENTSMichat Grajek gratefully acknowledges financial support from the German Federal Ministry of Educationand Research (project 01AK702A). We appreciate theefforts of the special issue co-editors (Witold Heniszand Anand Swaminathan) and substantive conversations with Witold Henisz and Jean-Francois Hennartthat helped in the early development of the work.Comments by Derek Kellenberg, Ken Scheve,Harbir Singh, Maurizio Zanardi, and seminar participants at Tanaka Business School/Imperial College,London School of Economics, Nijmegen School of




20/22

_Impact off ISO 9000 on trade and FDI Joseph CloughertyandMichat Grajek_^ 631

Management, IUI inStockholm and the Conference onthe PE of Regulating MNCs & FDI at Penn State werehelpful in further developing the work. Kemal Azun,

JenniferRontganger, Enno Schroder, and ChristopherXitco provided excellent research assistance. Anyerrors, however, are our own.

REFERENCESAnand,J., Capron, L, & Mitchell, W. 2005. Using acquisitions toaccess multinationaldiversity: hinking beyond thedomesticversus cross-border M&A comparison. Industrial and CorporateChange, 14(2): 191-224.Anand, J.,& Kogut, B. 1997. Technological capabilities ofcountries, firm rivalry and foreign direct investment. Journal ofInternational usinessStudies,28(3): 445-465.Anderson, J. E., & Marcouiller, D. 2002. Insecurity and thepattern of trade: An empirical investigation. Review ofEconomics nd Statistics, 4(2): 342-352.Anderson, S.W., Daly, ]. D., & Johnson, M. F. 1999. Why firmsseek ISO 9000 certification: Regulatory compliance or competitive advantage? Production and Operations Management,8(1): 28-43.Baier, S. L, & Bergstrand, J.H. 2007. Do free trade agreements

actually increase members' international trade? Journal ofInternational Economics, 71(1): 72-95.Belderbos, R.,& Sleuwaegen, L. 1998. TariffjumpingDFI andexport substitution: Japanese electronics firms in Europe. Internation

The Impact of ISO 9000 Diffusion on Trade and FDI: A New Institutional Analysis

Documents

Transcript of The Impact of ISO 9000 Diffusion on Trade and FDI: A New Institutional Analysis