18
Ž . Research Policy 27 1998 541–558 Patterns of internationalization of Spanish innovatory firms Jose Molero Faculty of Economics, Department of Economic Structure and Industrial Economics, UniÕersidad Complutense, 28023 Madrid, Spain Abstract The aim of this paper is to contribute to the debate about the factors determining the growing internationalization of small and medium firms originating from intermediate countries. The reason for selecting the Spanish innovating firms consists of the interest they have for analysing the interaction between the characteristics they have as innovatory agents and their behaviour regarding the global economy. The paper is developed from the following hypothesis: first, in the case of firms of countries with a more recent history of opening up to the international economy, the patterns of overseas expansion combine different kinds of elements such as factors included in the eclectic theory and tacit components of the company advantages. Second, in the trajectories followed by those firms some accumulative learning patterns can be found. The starting point of this research is a previous one carried out on the patterns of technological innovation of Spanish innovatory firms. On this basis, I have combined two sources for obtaining relevant data. The first consists of the creation of a data bank of innovatory firms with 814 innovating companies.The second source is a survey carried out with those firms. The statistical analysis is made through the application of discriminant analysis and logit models. They are good tools for systematic comparison between clusters and allow me to follow a stepwise method for a close study of increasing complex international activity. q 1998 Elsevier Science B.V. All rights reserved. Keywords: Internationalization; Innovatory agent; Discriminant analysis; Logit model 1. Introduction The recent evolution of the international economy has revealed important changes regarding the struc- ture of the relationships among economic agents and in the variables determining the conditions of com- petitiveness. Two factors stand out over many others: the first is the growing number of elements of eco- nomic organization affected by internationalization; the second refers to the increasing complexity of the innovatory process. These two features reinforce each other to the extent that today’s economic analysis has to consider both of them simultaneously when trying to account for the new dynamic of the firms operating at the international level. The aim of this paper is to contribute to the debate about the factors determining the growing internationalization of small and medium firms, es- pecially those originating from intermediate coun- tries. The reason for selecting the Spanish innovating firms instead of all Spanish firms with activity abroad consists precisely of the interest they have for analysing the interaction between the characteristics they have as innovatory agents and their behaviour regarding the global economy. An important point to underline is the concept of internationalization used in this research; unlike other works, I do not limit the study to Spanish firms investing in other countries. It is, of course, a very important element, but the international expansion of firms include, together 0048-7333r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved. Ž . PII: S0048-7333 98 00064-X

Patterns of internationalization of Spanish innovatory firms

Embed Size (px)

Citation preview

Page 1: Patterns of internationalization of Spanish innovatory firms

Ž .Research Policy 27 1998 541–558

Patterns of internationalization of Spanish innovatory firms

Jose MoleroFaculty of Economics, Department of Economic Structure and Industrial Economics, UniÕersidad Complutense, 28023 Madrid, Spain

Abstract

The aim of this paper is to contribute to the debate about the factors determining the growing internationalization of smalland medium firms originating from intermediate countries. The reason for selecting the Spanish innovating firms consists ofthe interest they have for analysing the interaction between the characteristics they have as innovatory agents and theirbehaviour regarding the global economy. The paper is developed from the following hypothesis: first, in the case of firms ofcountries with a more recent history of opening up to the international economy, the patterns of overseas expansion combinedifferent kinds of elements such as factors included in the eclectic theory and tacit components of the company advantages.Second, in the trajectories followed by those firms some accumulative learning patterns can be found. The starting point ofthis research is a previous one carried out on the patterns of technological innovation of Spanish innovatory firms. On thisbasis, I have combined two sources for obtaining relevant data. The first consists of the creation of a data bank of innovatoryfirms with 814 innovating companies.The second source is a survey carried out with those firms. The statistical analysis ismade through the application of discriminant analysis and logit models. They are good tools for systematic comparisonbetween clusters and allow me to follow a stepwise method for a close study of increasing complex international activity.q 1998 Elsevier Science B.V. All rights reserved.

Keywords: Internationalization; Innovatory agent; Discriminant analysis; Logit model

1. Introduction

The recent evolution of the international economyhas revealed important changes regarding the struc-ture of the relationships among economic agents andin the variables determining the conditions of com-petitiveness. Two factors stand out over many others:the first is the growing number of elements of eco-nomic organization affected by internationalization;the second refers to the increasing complexity of theinnovatory process. These two features reinforce eachother to the extent that today’s economic analysishas to consider both of them simultaneously whentrying to account for the new dynamic of the firmsoperating at the international level.

The aim of this paper is to contribute to thedebate about the factors determining the growinginternationalization of small and medium firms, es-pecially those originating from intermediate coun-tries. The reason for selecting the Spanish innovatingfirms instead of all Spanish firms with activity abroadconsists precisely of the interest they have foranalysing the interaction between the characteristicsthey have as innovatory agents and their behaviourregarding the global economy. An important point tounderline is the concept of internationalization usedin this research; unlike other works, I do not limit thestudy to Spanish firms investing in other countries. Itis, of course, a very important element, but theinternational expansion of firms include, together

0048-7333r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved.Ž .PII: S0048-7333 98 00064-X

Page 2: Patterns of internationalization of Spanish innovatory firms

( )J. MolerorResearch Policy 27 1998 541–558542

with the abovementioned form, others which rangefrom the selling of a significant portion of theirproduction in foreign markets to a variety of techno-logical activities. This enlarged vision complicatesthe analysis but also makes it more realistic.

The theoretical perspective combines different ap-proaches, trying to situate the internationalization ofsmall and medium firms as a process with differentstages and alternatives. Recent evolutionary theorieshave enriched the theoretical debate through theconsideration of technology and innovation as anendogenous part of the economic dynamics. Amongthe features of that approach, we have to underlinethe specific character of technological know-how,

Žthe importance of its tacit components embodied in.persons and organizations , the accumulativeness of

its results and the variety and diversity of the learn-ing processes. All these are elements to bear in mindwhen analysing the international economy’s competi-tiveness. In this sense, the expansion of the firms is acentral field to study the explanatory power of thosefactors.

However, more classic factors—especially thoseincluded in the ‘eclectic’ vision of Dunning—remainas important in understanding international strategiesof firms. In fact, what can be considered as explicitand tradeable components of the firm can be ana-lyzed in the light of variables related to advantagesof location or to the transaction cost. The point isthey do not explain the overall behaviour if the tacitfactors are excluded.

More specifically, the paper is developed from theŽ .following hypotheses: 1 In the case of firms be-

longing to countries with a short history of openingup to the international economy, the patterns ofoverseas expansion combine three kinds of elements.First, factors included in the classic OLIargumentation 1; second, tacit components of thecompany advantages and, finally, elements arisingfrom recent historical changes in the macroeconomicframework, especially the incorporation of Spain to

1 Ž .OLI ownerships–location–internalization is the eclecticparadigm elaborated by Dunning in which he synthesizes thedifferent available arguments for explaining the international ex-pansion of firms. A recent and dynamic version can be seen in

Ž .Dunning 1993 .

Ž .the EEC. 2 In the trajectories followed by thosefirms some accumulative learning patterns can befound. They are not mechanical and fixed for everyparticipant—in fact there are many dissimilarities—but constitute ‘guide-paths’ for most of them. More-over, in the various stages of the process of interna-tionalization, there are different variables which bet-ter express the success strategies of the firms.

In Section 2, I shall explain briefly the fundamen-tal theoretical considerations supporting the analysis,with special attention devoted to the necessity ofcombining more than one isolated theory. In Section3, I shall detail the research methodology. Section 4deals with the results emerging from the research,from the more descriptive to the statistical andeconometric ones. Finally, in the conclusions, I willdiscuss both theoretical and empirical implications ofthe findings.

2. Some theoretical considerations

In coherence with the internationalization conceptguiding this research, the background theory includesthree main fields: the theory of innovation, the tech-nological foundation of foreign trade and the re-viewed versions of international investment.

As far as the theory of innovation is concerned, Ijust want to give a reminder of some of the mostremarkable features of what is known as the evolu-

Ž .tionary or, as I prefer it, structural conceptualiza-tion, which starts from realizing the concrete andspecific features of technology and on the accumula-tive and endogenous character of technological inno-vation. From here, several characteristics of techno-logical innovation are underlined.

The first one has to do with tacit vs. explicitŽcomponents of technological knowledge Atkinson

and Stiglitz, 1969; Sahal, 1981; Freeman, 1982; Nel-son and Winter, 1982; Rosenberg, 1982; Dosi, 1984;

.Pavitt, 1984 . Whereas the latter can be discussed inthe light of theories like transaction costs or publicgoods, the tacit part of technology means that mostof conventional analysis cannot be applied becausethere is no option of transacting them through themarkets. Moreover, most of those tacit componentsare embodied in persons and organizations. Thismeans the strategies of firms for valuing this advan-

Page 3: Patterns of internationalization of Spanish innovatory firms

( )J. MolerorResearch Policy 27 1998 541–558 543

tage necessarily are different and they develop meth-ods for a joint valuation together with other organi-zational elements of the company.

A second important consideration arises from theŽevolutionary character of innovation Dosi et al.,

.1988; Freeman, 1990; Dosi, 1991 . The interactionbetween the firm and the environment contains fun-damental characteristics among which there are thelearning through different mechanisms and the accu-mulativeness of the process. To learn means to rec-ognize mistakes and to adapt to new circumstances,it is against any linear conceptualization of innova-tion and allows for differences in personal and orga-nizational capabilities in the face of similar chal-lenges. The accumulative character of innovationlead us to the necessity of taking into account histor-ical paths and dependencies on analysing technicalchange.

A third very relevant feature consists of the grow-ing international character of technological develop-ment. Although I do not support ‘technoglobalism’in its radical form, what is important is the increas-ing internationalization of technological markets;thus, the profitability and reproduction of technologi-cal innovations have to be achieved in the interna-tional framework instead of in national markets,which introduces an additional stimulus for business

Žtransnationalization Casson, 1991; Patel and Pavitt,.1991, 1995; Archibugi and Michie, 1995 .

The technological foundations of foreign trade areone of the most promising research fields of therecent economic literature. Since the ‘paradox’ ofLeontieff onwards, there have been a growing num-ber of studies and models showing the significanceof ‘knowledge-based’ factors in explaining the inter-national specialization of countries and the revealedsectoral advantages. It is not my intention to explorethat rich and complex debate, suffice it to remarkthere are two different—albeit complementary—ap-proaches.

On one hand, there is the macro perspective,which from general or sectoral perspectives has of-fered vast empirical evidence of the relevance oftechnological component for understanding interna-

Žtional specialization Fagerberg, 1988; Dosi et al.,.1990 . On the other, studies more related to business

consideration also have demonstrated the fundamen-tal role played by technological capabilities of the

firm in providing their international competi-tiveness 2.

Bearing those elements in mind, the question hereis to relate the behaviour of already confirmed inno-vatory firms to their international expansion towardsforeign markets in an economy in which the openingup of economic relations is a relatively recent fact. Inother words, our problem consists not of looking fortechnological vs. non-technological features with re-gard to foreign trade, but of finding out whetherthere are different variables explaining separate de-velopment of companies with demonstrated techno-logical activity.

Despite some controversies, there is a rather gen-eral acceptance of the decisive influence of techno-logically-related variables on exports. It is not thesame situation as far as internationalization of pro-duction is concerned. The two most classical ap-

Žproaches transaction costs and Dunning’s eclectic.synthesis are now being enriched by other theories

which are based on a great extent on new empiricalfindings of small and medium multinational compa-nies.

The so called ‘Uppsala’ school bases its argu-ments on the experience of Swedish companies andplaces as a central element of the debate an evolu-tionary concept of international expansion of firmsŽJohanson and Wiedersheim-Paul, 1975; Johanson

.and Vahle, 1977, 1990 : firms rise to higher levels ofinternational commitments after establishing and ob-taining experience in previous and simpler steps. Thefirm is viewed as an active agent adapting its strate-gies to uncertain environments depending upon itsown capabilities and the opportunities offered by thecontext. Thus, there is a gradual process derivedfrom the learning and the reduction of the uncer-tainty of the environment.

There are two fundamental consequences arisingfrom that view: on one hand, the learning process

2 Ž .The very much quoted work of Porter 1990 has to beŽ .complemented by others like Cantwell Cantwell, 1992, 1995 and

Ž .his notion of ‘technological competitiveness’ or Chesnais 1994and the concept of ‘structural competitiveness’. Additional exam-ples arise from many other fields such as the analysis of ‘strategic

Ž .agreements’ Michalet, 1991; Mytelka, 1991 . A recent veryŽ .interesting contribution is that of Pavitt 1994 .

Page 4: Patterns of internationalization of Spanish innovatory firms

( )J. MolerorResearch Policy 27 1998 541–558544

implies a parallel changing organizational path. Onthe other, international expansion usually starts fromnear geographical or cultural contexts; the reason isthe aim to reduce risks and uncertainties as much aspossible.

Two main limits can be found in that approachŽ .Alonso, 1995 . First, the rather unclear nature of theproposed relationship which hinders its transfer tooperative variables, in spite of the effort made by

Ž .Johanson and Vahle 1977 to clarify the model. Forthe empirical work, the biggest weakness rests on thefact there are no criteria to explain the transitionfrom one phase to another.

The second limitation derives from its excessivelylinear view of the international process, which ishardly compatible with the complexity of the pre-sent-day world and with the diversity of opinionsfaced by the company. ‘‘In fact, when the prescrip-tions of the theory are taken to extremes, the pro-posed stages come to be seen as a sort of obligatory

Ž .sequence’’ Alonso, 1995, p.15 . In the present re-search, a more flexible view is taken into account,which implies the possibility of different existingtrajectories of internationalization of firms. It is dueboth to the more complex international scenario—which in different places and sectors can offer differ-ent opportunities and limitations—and to the exis-tence of richer organizational possibilities for facinginternational competitiveness, depending on thefirm’s skills and learning capabilities.

A final comment on recent contributions to thetheory of new multinational corporations regards oneof the essays which gives most significance to thetechnological component. In several recent works,

Ž .Cantwell Cantwell, 1992, 1995 has emphasized theexplanatory power of the technological foundationsof multinational activities. The starting point is thenew conceptualization of technological developmentarising from the evolutionary theory of innovationand technology previously mentioned.

Cantwell wonders why the majority of theoreticalapproaches to multinational corporations continue tobase the discussion on an old conceptualization oftechnology and innovation. The central point refersto the tacit vs. public elements of technology. Thus,the more widespread justification of MNCS as agentswhich internationalized technological advantages ofthe firm, needs to be completed with the tacit capa-

bilities obtained through a continuous and collectivelearning. Nevertheless, he takes a further step; theclassical argument which highlights the advantage ofMNCs for internalizing the technological creationcan only be applied to the public component of thetechnology; in principle it is not possible to discussthe tacit component in the light of transaction coststheory.

Why can this approach be especially adapted tothe case of intermediate countries or to small andmedium firms? In my opinion, the answer rests onthe lower public component of the technologicalcompetence of those firms. As a matter of fact, thereis empirical evidence from technology exports fromthird world countries which shows how in thesecases the codified way of technological exportationthrough patents and production licenses is muchlower than in the case of most developed countries

Žor firms Dahlman and Sercovich, 1984; Katz, 1984;.Lall, 1987 . In the Spanish case, we have observed

similar behaviour among the majority of our firmsŽwhich export technology Sanchez and Vicens, 1991;´

.Molero, 1996 .Having shown the basic theoretical background of

this research, the task of translating those considera-tion into empirical hypothesis is by no means asimple one. In fact, the reflection on the Spanishcase is guided by two considerations which consti-tute the two core hypotheses already expressed in thefirst pages of this paper.

First of all, peculiarities qualifying the Spanishindustrial system allow me to recommend an ap-proach to firms’ behaviour which is very open tocontributions from different theoretical perspectives.Thus, the framework of the Uppsala proposal needssome qualifications in order to make it more suitableto our reality. In this sense, I understand that first-level variables like the size of the firm and how longit has been in operation are fundamental factors inthe process of interacting with international competi-tiveness and in the establishment of particular strate-gies vis-a-vis the learning for outside expansion.

In a similar way, classic factors included in thebasic ILO argument are still crucial for firms whichtake their first steps in international expansion. Whenconsidering the Spanish industrial structure it is notdifficult to agree that most firms do not have condi-tions for expansion abroad. As I said, they are

Page 5: Patterns of internationalization of Spanish innovatory firms

( )J. MolerorResearch Policy 27 1998 541–558 545

usually too small to allocate enough resources to thatadventure, full of uncertainties and risks. But even ifsize is not an absolute barrier there are other ele-ments which make the creation of firm’s advantagescapable of being exploited in a foreign economymore difficult. All this is clearly confirmed by therelatively new and limited process of outside interna-tionalization of Spanish firms.

The justification for a mixed approach can also beused to understand my position with regard to thenecessity of making more flexible the concept ofsequential trajectories of firms when they follow astrategy of internationalization. In fact, I should liketo show empirical evidence to support or reject thatthesis, but from the very beginning I want to under-line that I do not believe in either linear or one-waysequences. On the contrary, my thesis is that avariety of trajectories can coexist. In practical terms,it means there can be firms following the ‘Uppsala’

Ž .predicted pattern perhaps older and bigger firmstogether with other strategies which go directlythrough technological expansion and others usingbasically foreign direct investment.

3. Methodology: sources and means

The starting point of this research is a previousone carried out on the patterns of technologicalinnovation of Spanish innovatory firms 3. From itsresults, we know some central characteristics of thisgroup of firms whose international trajectory I ana-lyze here. On this basis, I have combined two sourcesfor obtaining relevant data. The first consists of theenlargement of the previous data bank of innovatoryfirms up to 814 innovating companies. In this databank, the available information about their interna-tional activity coming from existing statistics anddata bases is included. Although that information isvery simple it allows us to characterize some aspectsof the international position they occupy and theextension of their foreign activity.

The second source consists of a survey carried outwith those firms. After eliminating firms with differ-

3 A full presentation of it can be found in Buesa and MoleroŽ .1992 . A reviewed version of main results, regarding this paper,

Ž .is included in Molero and Buesa 1996 .

Žent identification problems changes in address or.legal status , we sent 719 questionnaires from which

we got 241 answers. For our analysis, we had toremove some incomplete answers or cases which donot enter clearly in some of the groups we estab-lished. So, we arrive at 205 valid cases.

In the corresponding questionnaire, a long list ofpoints is included trying to clarify the strategies andresults of the internationalization process. A centralplace is occupied by questions intending to discoversome crucial elements regarding the organization ofthe firm. Another important group of questions refersto the relationships between the firms’ internationalposition and their results.

The variables included in the analysis come fromthe two sources and belong to economic, organiza-tional and technological spheres. Table 8 includesthe description of the most representative variablesand their content.

The analysis has three complementary parts. Thefirst one has to do with a description of the coreaspects of the international activity of the firms,covering trade, technology and investment. The sec-ond deals with the analysis and evaluation of theanswers with regard to the motives the firms havefor developing those international tasks. The thirdconsists of a systematic comparison of differentgroups of firms according to the complexity of theirinternational trajectory. Four clusters have been es-tablished on the basis of the information emerging

Ž .from information contained in the two sources: afirms without any sort of international activityŽ . Ž .NIFIRMS ; b firms whose international presence

Ž .consists only of exports from Spain EXPFIRMS ;Ž .c firms exporting goods and developing technologi-

Ž . Ž .cal tasks abroad EXTECFIRMS and d firms witha full range of international activity: exports of goods,technological activities and direct investmentsŽ .COMIFIRMS .

The statistical analysis is made through the appli-cation of discriminant analysis and logit models.They are good tools for systematic comparison be-tween clusters and allow me to follow a stepwisemethod for a close study of increasingly complexinternational activity.

The conclusions—both theoretical and applied—are established on the basis of combined results ofthe three steps. Thus, although it makes the study

Page 6: Patterns of internationalization of Spanish innovatory firms

( )J. MolerorResearch Policy 27 1998 541–558546

Table 1Basic data of different groups of innovatory firms according to their level of internationalization

Kind Kind of sector Percentageof firms on total firmsIndustrial high Industrial low Services Total

technology technology

Number % Number % Number % Number %

NIFIRMS 12 17.91 30 44.78 25 37.31 67 100 32.7EXPFIRMS 33 47.83 33 47.83 3 4.34 69 100 33.6EXTECHFIRMS 15 48.39 8 25.81 8 25.81 31 100 15.2COMIFIRMS 24 63.16 10 26.32 4 10.53 38 100 18.5Total 84 40.97 81 39.51 40 19.51 205 100 100

NIFIRMS: firms without internationalization.EXPFIRMS: firms only exporting.EXTECHFIRMS: firms exporting and with technological activity abroad.

Ž .COMIFIRMS: firms exporting, with technological activity abroad and with foreign direct investment complete internationalization . TheŽ .classification of sectors has taken into consideration the composition of the labour force; see Buesa and Molero 1995 .

Source: own elaboration.

more complex, the richness of the findings is greaterand brings us more possibilities.

4. The internationalization of the innovatory firm

4.1. Basic description

Our previous research on innovatory firms locatedŽ .in the Madrid region Molero and Buesa, 1996 was

enlarged for this specific topic of internationaliza-tion. With the information arriving from that database and the survey, it is possible to describe somebasic features of the sample 4.

The first data refers to the fact that nearly onethird of firms do not show any kind of internationalprojection. Another third just go outside throughexports and only the remaining third has started amore sophisticated way of internationalizing by in-vesting or developing technological activities in othercountries. Taking into account that we are dealingonly with firms which show positive technological

4 From now onwards, all data and tables refer to 205 firmswhich have complete and well controlled information. Because ofthe origin of the data bank, our sample is not a random one and isbiased in favour of the Madrid region. Nevertheless, I wish tohighlight the fact that Madrid concentrates a large part of totalSpanish innovatory firms. For example, Madrid gathers around40% of total Spanish firms’ R&D expenditures.

activity, these results are a clear demonstration of therelative backwardness in the opening up of the Span-ish economy.

In many respects, these results are comparable tothe Italian ones. In fact, a recent study has provedthat from the total of Italian innovatory firms, 25.42%do not export and 86.45% do not control any firm in

Ž .a different country Archibugi and Ceccagnoli, 1994 .In both cases, the presence abroad of non-innovatoryfirms suggests the existence of a relative divorcebetween the processes of innovation and internation-alization, although the process seems to be moreaccentuated. 5

The majority presence of industrial firms in com-parison with other sorts of economic activity, is aconsequence of a higher propensity to the localmarket by service companies. Finally, the type ofagents controlling the firms affects their externalactivity. In fact, foreign capital firms show a moreintensive international activity.

Continuing with this presentation, Table 1 showsbasic data distributed by type of sectors. I havedivided the sample into three main clusters: two

5 In Spain, a considerable number of exporting firms do notcarry out any sort of research or technological development and

Žbase their exportation on more traditional factors Alonso and.Donoso, 1994 . In Italy, only 12.18% of exporting firms can be

Ž .catalogued as non-innovatory Archibugi and Ceccagnoli, 1994 .

Page 7: Patterns of internationalization of Spanish innovatory firms

( )J. MolerorResearch Policy 27 1998 541–558 547

industrial, within which I distinguished between lowand high technological content—and one groupingservice activities. Figures clearly say there is a higherlevel of internationalization in branches with highertechnological level; only 14.28% of these firmsŽ .against 32.7% in the global sample do not developinternational activities. On the other hand, 28.56% of

Ž .them against 18.5% have a complete process ofinternationalization, including commercial, techno-logical and investment tasks.

The less internationalized firms belong to serviceswhere 62.5% of firms only operate in the domesticmarket, whereas just 10% of firms have completedtheir process of internationalization. Apart from otherreasons, these results are coherent with the locallyoriented tendency services usually have all over theworld, at least until recent times.

The final descriptive approach to the topic hasbeen made by the figures included in Table 2. Here Isummarize the export propensity of each group, mea-sured through the percentage of sales devoted toforeign markets. As can be seen, outside commercialorientation is higher in firms with more complexinternational projection; While for firms with onlycommercial expansion 72.5% of them export lessthan 25% of sales, in other more internationalizedgroups, that percentage comes down to 50 to 60%. Itseems to show the existence of a reinforcing influ-ence from higher internationalization towards astronger position in foreign markets, although wemust be very cautious due to the descriptive natureof the data.

4.2. Factors in international expansion

Answers coming from the survey give us thepossibility of knowing central features of the interna-tional expansion of these firms in each of the threeareas included as forming part of the international-ization strategy.

Regarding exportation, two groups of factors wereincorporated into the survey. On one hand, a numberof them concerning firm specific elements: level ofprices, product characteristics, organizational capa-bilities and after-sales services. On the other, therewas a group of factors related to the national orinternational context: rate of exchange, market pro-tection, responses to domestic difficulties and institu-tional support.

Table 3 summarizes the results and for a bettercomprehension of their meaning, I have made a

Ž .valuation index CVI which gives a simple view ofŽthe predominant orientation. It ranks from 0 no

. Ž .importance to 3 high importance . Thus, the resultsclearly show the most important factors for explain-ing export activity are product characteristics, ex-change rate and price leÕel. They are very consistentwith other findings which have proved the behaviourof our exporting firms tends to be ‘passive’; they sellbasically because somebody asks for their products,

Žnot because they promote them Alonso and Donoso,.1994 .

The presence of factors more dependent on ex-plicit strategies has a lesser importance. Organiza-tional capabilities have only a medium significance

Table 2Export propensity of different groups according to their level of internationalization

Kind of firm Exports as percentage of sales Percentage on totalinternational firmsWithout data Up to 25% More than 25% Total

Number % Number % Number % Number %

Ž .YIFIRMS Of which 3 2.17 89 64.5 46 33.3 138 100 100EXPFIRMS 1 1.45 50 72.5 18 26.1 69 100 50EXTECHFIRMS 2 6.45 16 51.6 13 41.9 31 100 22.5COMIFIRMS 0 0 23 60.5 15 39.5 38 100 27.5

YIFIRMS: total firms with international activity.EXPFIRMS: firms only exporting.EXTECHFIRMS: firms exporting and with technological activity abroad.

Ž .COMIFIRMS: firms exporting, with technological activity abroad and with foreign direct investment complete internationalization .Source: own elaboration.

Page 8: Patterns of internationalization of Spanish innovatory firms

( )J. MolerorResearch Policy 27 1998 541–558548

Table 3Exporting explanatory factors

a bType of factors Degree of importance CVI

High Medium Low Null

Firms’ specificPrice level 55.10 28.60 6.60 9.60 2.29Product characteristics 71.40 17.30 6.80 4.50 2.55Organizational capabilities 34.80 39.40 18.20 7.60 2.01After-sale services 25.20 30.50 18.30 25.90 1.55

( )Related to the context national or internationalŽ .Rate of change pesetas 56.40 24.80 12.00 6.80 2.31

Domestic market difficulties 16.10 36.90 20.00 26.90 1.42Protectionist structures 31.50 26.20 22.30 20.00 1.64Institutional support 21.50 23.80 21.50 33.10 1.34

a Measured by the percentage of firms responding to each cate-gory. The first four columns add up to 100%.bCVI: Each percentage of answers is weighted by the following

Ž . Ž . Ž . Ž .scale. High 3 , Medium 2 , Low 1 , Null 0 . The resultingvalue is divided by the number of valid answers in each category.Source: own elaboration.

—although far from the levels of previous factors—and services related to sales are important only for25% of the firms. Again, I have to underline thecoherence of these findings with other available stud-ies recently made for explaining export activities of

Ž .Spanish firms Alonso and Donoso, 1994 .It is rather interesting to highlight the slight im-

portance of institutional factors related to domesticand foreign markets or to official aid. In this sense,these findings differ from others which have under-lined the significant role played by those factorsŽ .Alonso and Donoso, 1994 . In my opinion, it is notcontradictory because of the kind of firms I deal withhere. In fact, the extreme importance given to prod-uct characteristic is strongly linked with the innova-tive nature of this sample, which can be an easyexplanation of their lesser dependency upon local orinternational difficulties.

The analysis of explanatory investment factorsfollows a similar procedure. I have used two clustersof questions. The first refers again to firms’ particu-larities: technological capabilities, intangible assetsand organizational capabilities. The second includesfeatures of the country of investment destination.Some of them belong to classic elements like thesize of the market, production costs, the existence ofprotectionist structures, the possibilities of using the

country as an exporting platform and local State aids.I have added another factor more related to recenttendencies as is the case of the existence of localtechnological resources useful for the investor. Table4 gathers the results and includes the valuation indexŽ .CVI to facilitate their interpretation.

As far as country characteristics are concerned,the most remarkable places are occupied by whatcould be called ‘classic factors’. The size of themarket ranks in first place, having a high importancefor 57.10% of the firms. The second place is occu-pied by production costs, although its significance isnot so important. The rest of local features play alimited role in explaining the investment. An espe-cial significance is shown by the low level reachedby technological resources in the strategy of Span-ish innovatory investing firms; only 27.3% of themconsider those resources very important. Therefore,in considering location factors, the traditional onesstill have a very important role.

As on average, factors referring to firm’s charac-teristics present a higher importance, with the excep-tion of the size of the market. Among them, techno-logical capabilities show a clear first place, which isconsistent with the characteristics of the firms; 55.6%of them claim to give a high importance to thatfactor. Organizational capabilities reach a level closeto medium importance, as well as exporting activi-

Table 4Investment explanatory factors

a bType of factors Degree of importance CVI

High Medium Low Null

Firms’ specificTechnological capabilities 55.60 30.60 8.30 5.60 2.36Intangible assets 37.10 40.00 14.30 8.60 2.06Organizational capabilities 30.60 61.10 5.60 2.80 2.19

Country of destinationSize of the market 57.10 37.10 2.80 2.80 2.49Costs of production factors 42.90 31.40 14.30 11.40 2.06Technological resources 27.30 30.30 21.20 21.20 1.64Exporting platform 33.30 27.80 13.90 25.00 1.69Protectionist structures 20.60 26.50 32.40 20.60 1.47State support 14.70 23.50 26.50 35.30 1.17

a Measured by the percentage of firms responding to each cate-gory. The first four columns add up to 100%.bCVI: as in Table 3.

Page 9: Patterns of internationalization of Spanish innovatory firms

( )J. MolerorResearch Policy 27 1998 541–558 549

Table 5Firms with technological activity abroad

Frequency Type of activity

Technology transfer Patenting R&D

Number Percentage Number Percentage Number Percentage

Regularly 18 8.6 14 6.7 13 6.2Rarely 18 8.6 21 10 15 7.2Never 173 82.8 174 83.3 181 86.6Total 209 100 209 100 209 100

Source: own elaboration.

ties. Even closer to the medium is the significance ofintangible assets, which recent literature tends toplace close to technological capabilities in order to

Ž .make them fully operative Cantwell, 1995 . Proba-bly, it is a result of the ignorance of the real impor-tance of this complementarity which leads the firmsto answer in that way instead of the true role theyplay.

The analysis of technological activity abroad hasmore difficulties because it can be present throughmore highly diversified mechanisms. Thus, the sur-vey included three different possibilities: technologytransfer, patenting and R&D. Table 5 shows a globalview of the situation. Regularly only 8.6% of thefirms are engaged with transferring of technology,6.7% patents and 6.2% carry out R&D tasks inforeign countries. They are poor results whateverconsiderations we may make. Roughly figures aredoubled when we include non-regular activities.Nevertheless, more than 80% of the firms have neverbeen involved in technological operations outsideSpain. 6 Among the three, R&D appears as the leastfrequent way of internationalizing technologicaltasks.

Technology transfer is made more frequently un-der the form of technical assistance than throughlicensing. It corresponds well to the general be-

6 Of course, this is leaving aside the activity of importingtechnology. Although it is very relevant in the Spanish caseŽ .Molero, 1995 , it has to do with a different part of firm’soperations. In fact, considering data from the balance of paymentswe know that 179 of the 814 firms included in our database paidfor foreign technology in 1990–1992. For the same sample,between 10% and 15% received technological incomes fromabroad.

Ž .haviour of Spanish firms Casado, 1995 , albeit inour sample the difference is less important, which inmy opinion reflects the nature of the firms. If weconsider the size of the firm, licensing is even rareramong firms with fewer than 50 employees. Nationalcapital controlled firms are less active in licensingthan those with foreign share.

Generally speaking, the transference is a littlemore frequent to firms with which they do not haveany other link. Not surprisingly, firms with foreignequity are more active in transferring technology tolinked firms. According to the size of the firms, thesmallest ones, rarely belonging to groups, transfertechnology more often to unrelated firms.

The rare patenting activity is carried out morefrequently with the purpose of protecting technologythan for developing commercial linkages. As Table 6

Table 6Ž .Motives for patenting abroad CVI

Exports Protection ofsupport technology

Size of firmsUp to 50 employees 2.75 2.8051 to 200 2.08 2.60201 to 500 1.66 2.25More than 500 1.60 3.00Total 2.13 2.28

Type of capitalMixed 2.00 3.00Foreign 2.25 2.80Public 2.00 3.00Private national 2.17 2.55Total 2.13 2.28

Ž .CVI defined like in Table 3 .Source: own elaboration.

Page 10: Patterns of internationalization of Spanish innovatory firms

( )J. MolerorResearch Policy 27 1998 541–558550

shows, this is especially clear for largest firms; onlyŽ .small ones fewer than 50 employees give both

objectives the same level of importance.The consideration of the kind of capital control

confirms the former conclusion in all cases. How-ever, national private firms show the more balancedsituation, probably due to the smaller size they usu-ally have. The most marked difference is betweennational public and mixed firms. In both cases, theprotection of technology reaches the highest possibleimportance.

The final point regarding factor analysis has to dowith the motives for carrying out R&D activities inother countries. Two kinds of motives have beentaken into account in the survey. One refers tosupply side elements and includes questions aboutlocal human and financial resources; the other con-sists of questions regarding demand side elementslike technology adaptation and the creations of spe-cific products for differentiated markets.

Table 7 is a synthesis of the findings and allowsus to underline the following aspects. First of all, onemust consider the outstanding position occupied bydemand factors, especially own technology adapta-tion. The possibility of creating new products isclose to medium importance, as well as the availabil-ity of local human resources. Financial support is theleast important motive, at least as far as the averagebehaviour is concerned.

The size of the firm introduces interesting qualifi-cations. Thus, largest companies place more trust infinancial resources coming from foreign economieswhile medium firms are stimulated more strongly byhuman resources availabilities. On the other hand,public companies are more dependent on foreignresources. Firms with foreign capital participationare more oriented to demand conditions than nationalprivate ones, reflecting the different source of theirrespective technologies.

4.3. Factors influencing different leÕels of interna-tionalization

The aim of this section is to acquire a deeperknowledge of the strategies of internationalizationthrough another method. Starting from the informa-tion gathered in Section 4.1, my intention is tocompare the characteristics of firms belonging to thedifferent groups, according to their level of interna-tionalization. So, three tasks need to be fulfilled: theconstruction of the groups, the elaboration of vari-ables for the analysis and the selection of the tech-nique to analyze those variables.

With regard to the construction of the groups offirms to be studied, I have taken the following steps.In the beginning, the comparison will be made be-

Ž .tween firms with no international activity NIFIRMSand others with any sort of internationalization,

Ž .whatever the level reached YIFIMRS . The vari-

Table 7Ž .Motives for R&D activities abroad CVI

To make To make To adapt own’s To createuse of foreign use of foreign technology specific productsfinancial resources human resources

SizeUp to 50 employees 1.00 1.50 1.50 1.5050 to 200 1.33 2.50 2.00 2.25201 to 500 1.00 2.66 2.33 2.33More than 500 2.50 1.66 2.40 2.00

Type of capitalMixed n.a. n.a. n.a. n.a.Foreign 1.25 2.33 2.66 3.00Public 3.00 0.00 0.00 0.00Private national 1.22 2.12 2.40 1.88Total 1.43 2.00 2.28 2.07

Ž .CVI defined like in Table 3 .Source: own elaboration.

Page 11: Patterns of internationalization of Spanish innovatory firms

( )J. MolerorResearch Policy 27 1998 541–558 551

ables used in this phase are the first two groupsincluded in Table 8; first, variables representing thestructure of the firms and, second, variables measur-ing the technological activity. The two are commonto the groups under comparison.

The second stage analyses the differences be-tween two groups of already internationalized firms.

ŽOn one hand, companies which only export EX-.PFIRMS and, on the other, firms which in addition

to exporting carry out other types of internationalŽ .activity OIFIRMS . The variables for this analysis

include the two former groups plus the third clusterreferring to foreign trade.

The third comparative exercise is made betweenthe two groups with greater international activity. Onone side, there are firms which export goods and alsocarry out technological activities in foreign countriesŽ .EXTECFIRMS . On the other side, there are firmswhich, besides doing those tasks, have direct invest-ments abroad, that is, firms with a complete process

Ž .of internationalization COMIFIRMS . The list of

variables includes the three mentioned groups plusthe fourth one, corresponding to levels of technologi-cal exportation.

The analysis in the three stages will be donecombining two methods. The first consists of theapplication of ‘discriminant analysis’ in order to findout which variables discriminate the best betweendifferent clusters. The lists of variables to use in theanalysis are those already described, taking in eachcase only those common to the respective groups. Inaddition to that, I shall estimate Logistic regressionstaking as dependent variables the groups establishedand as explanatory ones those which present goodbehaviour within discriminant analysis.

The combination of these statistical–econometricmethods will provide more explanatory power to ourconclusions. In fact, the sequential findings willallow me to discover the similarities and differencesaffecting different levels of firms’ internationaliza-tion and will complete the panorama shown in theprevious section.

Table 8Variables included in the analysis

Content

( )Structure of the firms DB,SSECTOR Number of economic sector, according to Spanish classification.SIZE Number of employees.FORCAP Firm controlled by foreign capital.NATCAP Firm controlled by national capital.YEBIRTH Year of firm’s birth.YECONTROL Year of firm’s control by present owners.

( )Technological actiÕity DBPATENTS Number of patents and utility models in Spanish office 1988–1992.PATUSA Number of patents in USA.TECHPAY Payments for technology imports, 1990–1992.TECHREV Revenues for technology exports, 1990–1992.TECHBAL Rate of balance between payments and revenues in percentages.

( )Foreign trade SREGEXP Number of periods in which the firm exported.PROPEXP Percentage of sales allocated in foreign markets.DIFEXP Difference between year of birth and year of export starting.

( )Technological actiÕity abroad SMINEXTECH Rare technology activity and only of one sort.LOEXTECH Rare technology activity of two or three sorts.MEEXTECH One regular technology activity and another rare one.HIEXTECH At least two regular sorts of technological activity.

Ž . Ž .Source: own elaboration from the data base DB and Survey S :.

Page 12: Patterns of internationalization of Spanish innovatory firms

( )J. MolerorResearch Policy 27 1998 541–558552

Table 9Discriminant function between NIFIRMS and YIFIRMS

Ž . Ž . Ž . Ž . Ž .Dsy0.56 SECTOR ))) q0.20 SIZE ))) y0.35 YEBIRTH ))) q0.15 PATENTS ))) q0.22 TECHREV )))

Ž .q0.37 TECHBAL )))

Chi squareds55.18.Significances0.000.Percent correctly classified casess71.57.)))Significance levels0.01.

Having explained the analytical procedure, thecomments will follow the same path. Thus, the firstones correspond to the possibility of discriminatingbetween internationalized and non-internationalizedfirms. The application of discriminant technique leadsto the results included in Table 9.

According to these results, the basic process for afirm to enter or not into any kind of internationalprojection is associated with three variables of thestructural composition of the firms and another threeof their technological activity. With regard to thefirst group, SECTOR is the variable with greatestinfluence; the way in which SECTOR has beenconstructed and the sign it has, tells us that the typeof sectors least oriented to international activities areincluded in the groups starting with statistical num-bers over four; in other words, they belong morefrequently to services, construction and light manu-facturing branches like textiles, clothes, furniture,food and so on. The contrary tendency works forpower, electricity, oil refining and chemicals, and toa lesser extent, machinery, electronics, etc. The sec-ond variable within this group is the antiquity of the

Ž .firms YEBIRTH . The sign of its parameter is veryimportant because it indicates a positive relationshipbetween the age of the firm and a more frequentinternational activity, whatever the way in which it iscarried out. SIZE has a direct and positive influencewhich means the larger the firm the more oriented itis to the YIFIRM group; in other words, smallercompanies are more biased to the group withoutinternational activity.

The three technological variables PATENTS,TECHREV and TECHBAL point in the same direc-tion. The more intense the three variables, the moreinternational is the projection of the firm. It is animportant result to analyse more deeply.

The application of logistic regressions allows usto progress in the explanatory meaning of the rela-

tionships. Thus, I have estimated a logit model inwhich the belonging to any of the groups is relatedin probabilistic terms to the variables which signifi-cantly acted in the Discriminant. The result is in-cluded in Table 10.

Two variables lose the capability of explainingthe probability of a firm belonging to the predefinedclusters: the size of the firm and the rate of balanceof technological flows. Four factors still continue tohave a significant role: technological activity, de-fined by the number of patents and related industrialutility models, and returns from technological exter-nal payments. According to its coefficient, the roleof PATENTS is to increase the probability of be-longing to the group of firms with international

Ž .expansion YIFIRMS .Two arguments can help to understand that out-

standing result. On the one hand, there is a directrelationship between the technological activity of thefirms measured through patents and their possibili-ties of starting the internationalization process; onthe other, the slight tendency of Spanish firms to

Žpatent, whatever their technological level Molero.and Molas, 1992 . Therefore, this indicator is a very

qualified one for selecting firms with internationalpotential.

Table 10Results of logistic regression NIFIRMS and YIFIRMS

Variables B Significance

SECTOR y0.23 0.001YEBIRTH y0.30 0.009PATENTS 0.50 0.09TECHREV 0.49 0.007CONSTANT 2.60 0.0000

Chi squareds54.72.Significances0.0000.Percent correctly classified casess72.06.

Page 13: Patterns of internationalization of Spanish innovatory firms

( )J. MolerorResearch Policy 27 1998 541–558 553

Table 11Discriminant function between EXPFIRMS and OIFIRMS

Ž . Ž . Ž . Ž . Ž .Ds0.34 SECTOR ))) q0.65 SIZE ))) q0.53 PATENTS ))) q0.28 PATUSA ))) y0.34 REGEXP )))

Chi squareds19.40.Significances0.0000.Percent correctly classified casess67.39.)))Significances0.01.

Additionally, two structural variables still have apowerful presence: the sector to which firms belongand the year they were created. In both cases the signis negative, which confirms the interpretation I of-fered for discriminant coefficients.

To summarize the findings: firms of sectors suchas services, construction and light industries, rela-tively new, small and with a less intense technologi-cal performance have much smaller probabilities ofinternationalizing their activities than technologicallymore intensive, large, older ones, and those belong-ing to the rest of the sectors.

The second step consists of the comparison be-tween firms whose internationalization is made only

Ž .through exports EXPFIRMS and others with aŽ .more complex process OIFIRMS . The explanatory

Žvariables of the first step structure of firms and.technological activity are now enlarged with those

which are common to the new groups: the onesreferring to foreign trade, REGEXP, PROEXP andDIFEXP. The results of the discriminant analysis aresynthesized in Table 11.

There are similarities and differences with regardto the first discriminant. Three variables maintaintheir importance: two representing firms’ structureŽ .SECTOR and SIZE and one of technological activ-

Ž .ity PATENTS . PATUSA arises as a new discrimi-nant variable with effect like PATENTS. From thegroup of variables included in foreign trade, the onlyone capable of explaining group differences is REG-EXP, which refers to regular exporting activity. Thesign it has means the more regular exports are, theless prone is a firm to diversify its forms of interna-tional expansion. Very briefly, the results allow us tostate that firms with more complex processes ofinternationalization belong to services, constructionand traditional industries; moreover, they have alarge size and carry out a more intense technologicaleffort. Nevertheless, the more regular exports are the

less likely is a firm to diversify its ways of interna-tional expansion.

The application of a logistic regression using thesignificant variables of the discriminant functionleads to the results of Table 12. Logit analysisremoves REGEXP from the model and maintains theother four; two with lower levels of significanceŽ .SECTOR and SIZE and the other two, PATENTSand PATUSA, with higher levels, albeit not negligi-ble ones.

Summing up the findings of this step, it is possi-ble to affirm that the basic distinction among firms

Ž .with the lowest international level EXPFIRMS andŽ .with any kind of higher level OIFIRMS is to be

found, first, in the sector of activity and in the size ofthe companies. The larger the size, the greater theprobability that a firm belongs to OIFIRMS; in otherwords, smaller firms diversify their ways of interna-tionalization less. As far as SECTOR is concerned,

Žfirms of branches already defined services, light.industries, etc. have more probabilities of having a

more complex process of international expansion.In complementary fashion, we have to consider

domestic technological activity and the capability ofpatenting abroad. The first seems to maintain thekind of effect described before; more technologicallyintensive firms tend to be more oriented to comple-

Table 12Results of the logistic regression EXPFIRMS and OIFIRMS

Variables B Significance

SIZE 0.49 0.02PATENTS 0.29 0.11SECTOR 0.18 0.02PATUSA 0.59 0.16CONSTANT y2.15 0.002

Chi squareds21.49.Significances0.0003.Percent correctly classified casess66.67.

Page 14: Patterns of internationalization of Spanish innovatory firms

( )J. MolerorResearch Policy 27 1998 541–558554

menting exports with other international activities.The same effect is seen in the possession of patentsin the USA; they support more diverse ways ofexternal expansion.

The last step deals with the difference betweenfirms which, in addition to exporting, develop tech-

Ž .nological activities abroad EXTECHFIRMS andothers which have a complete process of internation-alization, including exports, technological activity

Ž .and direct investments COMIFIRMS . As in theprevious analysis, now I have added to the list ofvariables those common to the new clusters; thereare several indicators of the intensity and complexityof their technological presence abroad MINEX-TECH, LOEXTECH, MEEXTECH and HIEX-TECH, as defined in Table 8.

In this step, I have modified the analysis for thefollowing reasons. The discriminant function worksreasonably well, with 72.13% of cases correctlyclassified. However, when I estimated the logit re-gression, most of the variables were removed fromthe model, only two remaining. Moreover, the per-centage of group 1 firms correctly classified was33.33. This fact suggested attempting another methodfor a better explanation of the clusters. Thus, I havedirectly estimated a logistic regression including allvariables used in Discriminant analysis, not onlythose which had there a positive explanatory rolethere. 7 The results are included in Table 13.

As in the first comparison, firms belonging toservices and more traditional industries, have lessprobability of being classified within the groups offirms with the most advanced level of international-ization. Together with this, two other variables arisefor the first time: the presence of foreign capital

Ž .among shareholders FORCAP and the year in which

7 I have also estimated this kind of logit models in the twoprevious cases. There, unlike now, the results were very similar tothose presented in the text. Thus, in the first case—the compari-son between NIFIRMS and YIFIRMS, the variables included inthe model were the same and the percentage of cases correctlyclassified, practically coincided with the one obtained using onlysignificant discriminant variables. In the second case—compari-son between EXPFIRMS and OIFIRMS—the new logit includesthe same variables as the first one, except PATENTS, the levels ofsignificance being very similar in the two models. Thus, in thesesteps there are no practical differences.

Table 13Logistic regression EXTECHFIRMS and COMIFIRMS

Variables B Significance

SECTOR y0.28 0.07FORCAP y0.91 0.05YECONTROL 0.48 0.05TECHREV y0.34 0.08TECHPAY 0.75 0.00LOEXTECH y4.71 0.00MEEXTECH 1.30 0.10CONSTANT y0.22 0.84

Chi squareds23.82.Significances0.001.Percent correctly classified casess74.14.

the present control of the firm took placeŽ .YECONTROL . The first indicate that national firmshave more frequently completed international expan-sion than those with foreign capital. The secondrefers to the importance of firm’s restructuring for itsdirect international investment. The sign of this lastvariable shows firms which did that reorganizationmost recently have more probabilities of belongingto the groups with foreign investment; in other words,it seems necessary to carry out some organizationalchanges to facilitate the access to direct internationalinvestment.

Variables measuring the level of internationalŽ .technological payments TECHREV, TECHPAY

also have significant coefficients, higher in the caseof payments. The contrary signs they have allow usto understand the different directions of the influ-ence. While the level of incomes decreases, theprobability of a firm being part of the group with acomplete international presence, the level of pay-ments increases the relative probability of belongingto the group which enters into international invest-ment. In terms of the Spanish structure, these resultsare very important because they reflect the consider-able level of external dependency of our system of

Žinnovation Sanchez and Vicens, 1991; Casado, 1995;´.Molero, 1996 . Thus, firms with higher levels of

technological dependency have more possibilities todevelop the complete process of internationalization.

The last variables for comment belong to thecomplexity of technological activity abroadŽ .LOEXTECH, MEEXTECH . The results show that

Page 15: Patterns of internationalization of Spanish innovatory firms

( )J. MolerorResearch Policy 27 1998 541–558 555

a low level of that complexity is significant fordetermining the probability of a firm not investing

Žabroad, while the most complex activities HIEX-.TECH do not affect the probability of belonging to

any of the groups because that variable is not in-cluded in the model. Interestingly enough, the sign

Žof medium level of technological activity MEEX-.TECH tells us that this variable positively affects

the probability of a firm belonging to the groupswith foreign investment. Although with some ele-ment of risk, we can understand from this result thattechnological internationalization is a precondition tosubsequent international activity, but only if it iscarried out at a medium level of intensity.

5. Conclusions

My first comment is to state the usefulness of theapproach adopted here. It has proven to be efficientin revealing important elements of the strategy ofinternationalization of the companies. In this respect,I think the first of the explicit hypotheses is fullyproved because the type of variables explaining thebehaviour of the firms includes a great variety ofelements belonging to different theoretical ap-proaches.

The results obtained from direct questions to thefirms basically confirm the validity of most argu-ments included in the eclectic paradigm. Thus, loca-tion aspects like the size of the markets, the costs ofproduction, the level of prices and so on are impor-tant factors for explaining substantial parts of theinternational strategies of the companies, whateverthe level we take into consideration. They are presentindeed in exporting, investing and developing tech-nological tasks. Other factors referring to firms’organizational and technological capabilities are alsovery significant for the firms. In fact, they are veryclose to the size of the market in the global evalua-tion of explanatory investment factors.

More difficult is to find in the answers signs ofelements belonging to new theoretical developmentssuch as technological competitiveness. However,given the complementarity existing among many ofthe isolated answers, especially those related to tech-nological and organizational capabilities, I think theimportance they acquired in the evaluation can be

considered as an indication supporting those theories.The same can be done when interpreting the motivesfor developing R&D activities in foreign countries.The first place reached by the need for adapting theirown technology derives, to a great extent, from thespecific nature of technological knowledge and itstacit features.

The second method enriches the conclusions withmore evidence. For a synthetic discussion, I haveprepared Table 14 which synthesizes the variablesthat proved significant in each of the three steps ofthe final part of the analysis. Regarding the firsthypothesis, I will make the following comments.

From the group of structural variables, the sectorof activity is a significant factor in all the steps,which allows us to suggest the existence of different‘opportunities of internationalization’ within sectors.It is partially due to the international situation of thebranches, which is revealed by the importance ofbelonging to more complex technological sectors intwo of the three stages we made. It also has to dowith the particular process of recent Spanish indus-trialization. This is expressed by the fact that firmsentering first and third stages belong more frequentlyto branches with most tradition in the Spanish econ-omy, such as construction, services or light indus-tries.

The size of the firm clearly separates cases in thetwo first levels while it is not an important factor forselecting firms of an internationally complete group.In those two, larger firms show greater probabilitiesfor taking more complete steps and reaching higherlevels of internationalization.

The presence of national vs. foreign capital ac-quires an important presence in the latter differentia-tion when considering direct foreign investments,although it has no importance in lower levels. Apartfrom being coherent with other findings, in my viewit confirms the relatively poor interest of foreigncompanies located in Spain in developing interna-tional strategies from this country.

The two dates included have proved to be impor-tant. One, the year of creation, is relevant for thefirst and most simple process of internationalizationin the sense that the more recent the firms are theless probable it is that they have started international-izing. The year of the last control of the companyhas a significant role in the most complex level of

Page 16: Patterns of internationalization of Spanish innovatory firms

( )J. MolerorResearch Policy 27 1998 541–558556

Table 14Significant variables in statistical–econometric analysis

Variables NIFIRMS–YIFIRMS EXPFIRMS–OIFIRMS EXTECHFIRMS–COMIFIRMS

D1 L1 D2 L2 L3

SECTOR )) )) )) )) )

SIZE )) )) ))

FORCAP ))

YEBIRTH )) ))

YECONTROL ))

PATENTS )) ) )) )

PATUSA )) )

TECHPAY ))

TECHREV )) )) )

TECHBAL ))

REGEXP ))

LOEXTECH ))

MEEXTECH )

))sSignificance level 0.05 or lower; )ssignificance level higher than 0.05.Dsdiscriminant analysis; Ls logistic regression.Source: own elaboration.

transnational operations; in this case, recent changesin the control of the firms push them to complete theinternationalization process.

Indicators of technological dynamism also getpositive results. Thus, domestic patents and to alesser extent external ones, play a significant role inpushing firms to increase their level of international-ization. This means that very active technologicalfirms carry out international activities more fre-quently.

The fact of being more active in receiving techno-logical payments from outside—and its correspond-ingly more balanced situation—is a significant factorin two moments: the initial one of having any kindof international presence and the final one of reach-ing the complete level. In this latter moment, themost interesting role is played by payments fortechnology importation because they are relevant toachieve a complete pattern of international expan-sion; the joint consideration of both effects leads usto the need to take into account the technologicaldependency of the firms as a critical variable.

Of considerable importance is the relative absenceof trade variables among explanatory factors. Onlyregularity in exporting seems to have an influenceupon establishing different strategies; most regularexporting firms usually limit their strategy to foreigntrade actions.

The second general hypothesis of the existence ofa learning process cannot be confirmed if we con-sider it only in a restricted perspective. However, ifwe have a more flexible approach in which theexistence of trajectories of internationalization doesnot imply the following of particular paths by everyfirm, the results confirm, at least partially, that the-sis. From our results, two findings will throw lighton this dimension.

First, the low presence of variables related toexports gives support to the idea that learning throughtrade is neither a determinant nor a precondition forfirms to develop further more complex strategies.Rather, it seems to demonstrate the existence ofcompanies whose international strategy is carried outjust by exporting. Moreover, the only moment inwhich regular exports arise is for discriminatingagainst higher levels of internationalization.

Equally important is the fact that technologicaltasks developed abroad have some effects on acced-ing to the complete form of internationalization. Ashas been explained, as the complexity level of thosetasks rises their influence on a firm entering a pro-cess of investment disappears. The consideration tomake is parallel to the former one: although thereseems to be some influence—as the significance ofintermediate levels of technological presence proves—the lack of impact of higher levels of technologi-

Page 17: Patterns of internationalization of Spanish innovatory firms

( )J. MolerorResearch Policy 27 1998 541–558 557

cal tasks allows us to maintain the thesis of coexist-ing different trajectories. Sometimes, it is a learningprocess from simple international presence to moresophisticated ones but, on other occasions, firmsdevelop alternative routes without having passedthrough previous ones.

The hypotheses presented in this paper are diffi-cult to prove or reject in a radical way because theyinvolve a lot of implicit conditions and refer to aheterogeneous group of elements. Moreover, thecombination of theoretical elements included in newapproaches make it complicated to isolate testablequestions. Nevertheless, I think the findings offeredin previous sections are enough to affirm with empir-ical support some basic questions affecting the inter-nationalization of Spanish innovatory firms, whichcan be added to other evidence available for small

Žcountries both from the North Hircsh and Thomsen,. Ž1993 and the South of Europe Archibugi and

Ceccagnoli, 1994; Schiattarella, 1994; Malerba and.Orsenigo, 1995 . In all these cases, the application of

more open conceptualizations is very helpful forunderstanding a new reality not always resemblingprevious patterns of internationalization.

Acknowledgements

I wish to express my gratitude to Antonio Fonfriafor his fruitful help in dealing with statistical analy-sis and to Barry Readman for his patience in readingmy poor English; also my gratitude to the anony-mous referees whose suggestions helped me to im-prove and clarify the text. The defects still presentare of my exclusive responsibility.

References

Alonso, J.A., 1995. Internationalization process and forms ofŽ .market penetration: a dynamic proposal. In: Molero, J. Ed. ,

Technological Innovation, Multinational Corporations and theNew International Competitiveness. The Case of IntermediateCountries. Hardwood Academic Publishers, Reading.

Alonso, J.A., Donoso, V., 1994. Capacidades competitivas de laempresa exportadora espanola. ICEX, Madrid.˜

Archibugi, D., Ceccagnoli, M., 1994. Innovazione e internazional-izzazione nelle imprese manifatturiere Italiane. Mimeo. CNDR,Rome.

Archibugi, D., Michie, J., 1995. The globalization of technology:a new taxonomy. Cambridge Journal of Economics 19.

Atkinson, A.B., Stiglitz, J.E., 1969. A new view of technologicalchange. The Economic Journal, September.

Buesa, M., Molero, J., 1992. Patrones del cambio tecnico y´polıtica industrial. Eudema, Madrid.´

Buesa, M., Molero, J., 1995. Configuracion productiva y capaci-´dad de innovacion en la industria espanola. Informacion Com-´ ˜ ´ercial Espanola, no. 743, July.˜

Cantwell, J., 1992. Innovation and technological competitiveness.Ž .In: Buckley, P.J., Casson, M. Eds. , Multinational Enterprises

in the World Economy: Essays in Honour of John Dunning.Edward Elgar, Aldershot.

Ž .Cantwell, J., 1995. In: Molero Ed. , Multinational corporationsand innovatory activities: towards a new evolutionary ap-proach.

Casado, M., 1995. La capacidad tecnologica de la economıa´ ´espanola: un balance de la transferencia de tecnologıa. Infor-˜ ´macion Comercial Espanola, no. 740, April.˜

Ž .Casson, M.C. Ed. , 1991. Global Research Strategy and Interna-tional Competitiveness. Blackwell, Oxford.

Chesnais, F., 1994. La mondialisation du capital. Syros, Paris.Dahlman, C.J., Sercovich, C., 1984. Exports from semi-industrial

economies and local technological developments: lessons fromnewly industrializing countries. Journal of Development Stud-ies 16, 169–170.

Dosi, G., 1984. Technical Change and Industrial Transformation.Pinter Publishers, London.

Dosi, G., 1991. Perspectives on evolutionary theory. Science andPublic Policy, December.

Dosi, G., et al., 1988. Technical Change and Economic Theory.Pinter Publishers, London.

Dosi, G., Pavitt, K., Soete, L., 1990. The Economics of TechnicalChange and International Trade. Harvester Wheatsheaf, HemelHempstead.

Dunning, J., 1993. The Globalization of Business. Routledge,London.

Fagerberg, J., 1988. International competitiveness. The EconomicJournal, 98 June.

Freeman, Ch., 1982. The Economics of Industrial Innovation.Pinter Publishers, London.

Ž .Freeman, Ch. Ed. , 1990. The Economics of Innovation. EdwardElgar, Aldershot.

Hirsch, S., Thomsen, S., 1993. Internationalization of small coun-try industrial firms. A comparative study of Denmark andIsrael. Paper to the 19th Annu. Conf. EIBA, Lisbon, Decem-ber.

Johanson, J., Vahle, J.E., 1977. The internationalization process ofthe firm: a model of knowledge development and increasingforeign market commitments. Journal of International Business

Ž .Studies 8 1 .Johanson, J., Vahle, J.E., 1990. The mechanism of international-

Ž .ization. International Marketing Review 7 4 .Johanson, J., Wiedersheim-Paul, F., 1975. The internationalization

of the firms: four Swedish cases. Journal of ManagementStudies, October.

Katz, J., 1984. Domestic technological innovations and dynamic

Page 18: Patterns of internationalization of Spanish innovatory firms

( )J. MolerorResearch Policy 27 1998 541–558558

comparative advantage. Journal of Development Studies 16,196.

Lall, S., 1987. Learning to Industrialize: The Acquisition ofTechnological Capabilities by India. Macmillan, London.

Malerba, F., Orsenigo, L., 1995. Technological innovation andŽ .international competitiveness in Italy. In: Molero, J. Ed. ,

Technological Innovation, Multinational Corporations and theNew International Competitiveness. The Case of IntermediateCountries. Hardwood Academic Publishers, Reading.

Michalet, Ch.A., 1991. Strategic partnerships and the changingŽ .internationalization. In: Mytelka, L. Ed. , Strategic Partner-

ships and the World Economy. Pinter Publishers, London.Ž .Molero, J. Ed. , 1995. Technological Innovation, Multinational

Corporations and the New International Competitiveness. TheCase of Intermediate Countries. Hardwood Academic Publish-ers, Reading.

Molero, J., 1996. La exportacion de tecnologıa como factor´ ´estrategico del desarrollo industrial: un analisis sectorial. In-´ ´formacion Comercial Espanola, no. 752, April.´ ˜

Molero, J., Buesa, M., 1996. Patterns of technological changeamong Spanish innovative firms: the case of Madrid region.Research Policy, 25.

Molero, J., Molas, J., 1992. Spanish innovative performance.Evidence from U.S. patenting. Documento de trabajo no.9214. Facultad de Economicas y Empresariales, UniversidadComplutense.

Ž .Mytelka, L. Ed. , 1991. Strategic Partnerships and the WorldEconomy. Pinter Publishers, London.

Nelson, R., Winter, S., 1982. An Evolutionary Theory of Eco-nomic Change. Harvard University Press, Cambridge, MA.

Patel, P., Pavitt, K., 1991. Large firms in the production of theworld’s technology: an important case of non-globalisation.Journal of International Business Studies. First Quarter.

Patel, P., Pavitt, K., 1995. The localised creation of global techno-Ž .logical advantage. In: Molero, J. Ed. , Technological Innova-

tion, Multinational Corporations and the New InternationalCompetitiveness. The Case of Intermediate Countries. Hard-wood Academic Publishers, Reading.

Pavitt, K., 1984. Sectoral patterns of technological change: to-wards a taxonomy and theory. Research Policy 6, 196.

Pavitt, K., 1994. Technological competencies in the world’s largestfirms: characteristics, constraints and scope for managerialchoice. Workshop on Innovation, Patents and TechnologicalStrategies. OECD, Paris, 8–9 December.

Porter, M., 1990. The Competitive Advantage of Nations.Macmillan, London.

Rosenberg, N., 1982. Inside the Black Box. Technology andEconomics. Cambridge Univ. Press, Cambridge.

Sahal, D., 1981. Patterns of Technological Innovation. Addison-Wesley Publishing, Reading, MA.

Sanchez, P., Vicens, J., 1991. Exporting technology. Recent de-´velopments in the export of technology by Spanish companies.

Ž .Science and Public Policy 18 5 , 139.Schiattarella, R., 1994. Considerations on the determinants and

effects of multinational activity. Paper to the 20th Ann. EIBAConf.