Introduction

Technology is a core imperative for modern organizations. Nowadays firms seeking sustained competitiveness need to generate and exploit technological capabilities on a worldwide scale. Early conceptualization of the multinational enterprise (MNE) linked its expansion into foreign markets to the exploitation of MNE proprietary advantages (technology and associated knowledge) generated in the home base of the headquarters (Frost 2001). Two factors seem to have accounted for the initial rejection of the foreign subsidiaries’ creative technological activities. First, the only motivation that could encourage such activities would be the need to adapt the MNE centrally-derived capacities to local conditions that were considered of secondary importance (Pearce 1999). Second, factors related to the nature of technology provided an overwhelming “gravitational” force pulling such activities towards a focused centralized facility (Manolopoulos et al. 2007). These factors had to do with the need for precise coordination, control and security of technology inasmuch as interdependent operations were geographically dispersed.

Generally the viewpoint that the sources of technological ideas and projects for MNE subsidiaries originate at home reflects the “head office assignment” perspective. This approach posits that the foreign subsidiary is an instrument of the MNE and that management at the headquarters is responsible for defining the imperatives of subsidiaries abroad (Vernon 1966). The significance of MNE subsidiaries can be decisive if they provide key resources in the form of technology, knowledge, resources etc. to other MNE actors (Andersson et al. 2001, Gupta/Govindarajan 2000); but still the role of headquarters is of primary importance in the “head office assignment” view.

However, in recent years the benefits of a more decentralized technological approach have gained growing recognition (Hedlund/Rolander 1990). Following Pearce (1999), we can summarize our understanding of MNE decentralized technology operations in two main conclusions. First, MNE competitiveness is increasingly associated with overseas innovative activities and original product development rather than mere technological adaptation. Second, the accelerating impact of local environmental forces can often determine the location, extent and scope of decentralized technological operations of the subsidiary. Such forces may derive from the level and distinctiveness of a host country’s scientific background and infrastructure.

In the main, this decentralized approach signals the advent of a liberalism era, whereby innovative technological ideas and projects can emerge from subsidiaries around the world rather than exclusively MNE headquarters (Birkinshaw 2000, Birkinshaw/Hood 2001). This “subsidiary-focused” perspective stresses the fact that embeddedness of the subsidiary within its local context of suppliers, customers, distributors, customers, competitors etc. provides the impetus to creative technological ideas (Birkinshaw/Hood 2000, Paterson/Brock 2002). This is because the foreign subsidiary in its local market possesses the specific knowledge to compete in that market and can be the generator of significant technological outflows to other MNE actors worldwide.

Despite the abundance of research seeking to capture the determinants of de/centralized MNE knowledge-related competitiveness, the majority of studies deal with the ability of subsidiaries to generate distinctive technologies (Birkinshaw/Hood 1998a, Kogut/Zander 1996); or, the organizational designs to accommodate this goal (Gupta/Govindarajan 1994, Nahapiet/Ghoshal 1998). Relatively little research seems to exist on the direct impact of de/centralized technology sources on foreign subsidiary performance. In this paper, we use the term “technology sourcing” to refer to those sources in the MNE or local contexts, which pertain to technology acquisition, development, application and diffusion in the foreign subsidiary. Our research objective is to explore how internal MNE (centralized) and local environmental (decentralized) technology sources affect foreign subsidiary performance. In other words, how do the “head office assignment” and “subsidiary- focused” perspectives influence MNE subsidiary performance in its local market as regards technology sourcing? And, which of the two approaches is more valuable in explaining subsidiary performance?

Apart from its research interest, such a study has apparent managerial and public policy importance since it may inform corporate and governmental decision-makers about the implementation of appropriate strategies. Particularly the public policy significance is key because enhanced subsidiary performance is likely to generate positive externalities in terms of development of local employee skills, resistance to downsizing, and spillovers of technological and management know-how to the local economy (Birkinshaw/Hood 1998b, Blomstrom/Kokko 1998, Tavares/Young 2005).

An additional valuable characteristic of this research pertains to the fact that most studies that investigate MNE technology-related effects examine local (host) markets of advanced economies (mainly the US or the UK); or, large emerging markets, such as China. In the current study, we seek to investigate the impact of technology sourcing on MNE subsidiary performance in a small advancing economy, notably Greece. This country, which has a population of just above ten million inhabitants, is an interesting case to examine because of its dualism. It is a geographically peripheral EU country with an advancing economy that could either restrict MNE subsidiary technological activities to routine adaptation, reflecting a primarily market-seeking rationale for investments; or, provide a valuable and creative positioning for subsidiaries in a broad MNE strategic programme aiming at possible expansion from Greece into the emerging Southeast European economies.

The remainder of this paper is organized as follows. The second section explores the state-of-art on de/centralized technology sourcing and advances the research hypotheses to be tested. In the third section, we present the methodological details linked to the research design and operationalization of variables. In the fourth section, we discuss the results of the statistical analysis. The concluding section explores the implications of the study and future research directions.

Research Background and Hypotheses

Researchers have pointed out that the performance of MNEs is positively associated with their ability to generate and utilize effectively technological resources. According to Caves (1996) and Dunning (1988), the creation and exploitation of technology is the main reason explaining the success and growth of MNEs over time. Dunning (1993) notes that technology embraces all forms of the firm’s physical assets, human learning and capabilities, which lead to efficient production of goods and services. Technology is viewed as a major proprietary asset that allows firms to develop unique capabilities, which, in turn, may induce and facilitate expansion into foreign markets (Buckley/Casson 1976, Håkanson 1981). Both internal MNE and local environmental technology sources can affect performance of the foreign subsidiary in the market in which it operates.

MNE Technology Sourcing

Early writings argued that technology should be diffused internally within the firm so that management can exploit the advantages related to cost minimization and hierarchical control systems (Barney 1991). Vernon (1966) argued that the exploitation of MNE competitive advantage is mostly derived from technological knowledge transfers from headquarters to subsidiaries. This is a perspective largely associated with transaction cost economics (Williamson 1985). MNEs exploit technological asset interdependencies internally in order to develop unique organizational competencies through accumulating proprietary knowledge rather than outsourcing (Cantwell 1989, Pavitt 1990). This viewpoint is consistent with the “head office assignment” approach and the traditional roles played by the overseas MNE subsidiary.

These roles of dispersed affiliates are determined by the MNE headquarters and pertain to factors linked to the availability and cost of physical inputs in the production process. Once a role is allocated to the MNE subsidiary, its technology is sourced from the MNE group’s established knowledge resources (Ivarsson/Jonsson 2003, Papanastassiou/Pearce 1994). Although the ability of MNEs to engage in centralized technology sourcing “does not imply that such knowledge transfers are necessarily easy, successful, or routine” (Allred/Swan 2004, p. 261), the MNE has to find the best processes and mechanisms to disseminate these technology transfers between its subsidiaries (Zander 1997).

In support of this centralized view, there is recent evidence suggesting a strong link between the involvement of subsidiaries in the innovation network of the MNE system and the knowledge management mechanisms they develop (Johnston/Paladino 2007). Similarly, the technological richness of the MNE is connected with the subsidiary’s level of technological innovation (Almeida/Phene 2004). These arguments are additionally related to the finding that when headquarters provide an abundant level of resources to its affiliates, the performance of the subsidiary in the local market is high (Luo 2003). The implicit assumption in this line of thinking is that MNE technological knowledge flows that end up in the subsidiary are conducive to enhanced performance in the market it operates. Hence, this discussion leads us to the first hypothesis pertaining to the association between internal technology sourcing and performance of the MNE foreign subsidiary:

Hypothesis 1. Internal MNE sources of technology will be positively associated with subsidiary performance in the foreign country.

Local Environmental Technology Sourcing

Subsequent research developments on technology sourcing of the subsidiary highlight a growing involvement of MNE affiliates in technological operations. According to this research stream, subsidiaries are increasingly engaged in the innovation of products, contributing to the expansion of MNE core technologies. On this issue, Papanastassiou and Pearce (1999, p. 90) note that “the accessing and creative use of technology in a decentralized fashion should enrich both the knowledge available to the group and the effectiveness with which it is used commercially”. Thus, the technology strategy of the headquarters has to monitor and evaluate all MNE decentralized initiatives that emerge from overseas subsidiary operations embedded in localized knowledge systems. There exist numerous studies (e.g., Cantwell/Hodson 1991, Fors/Zejan 1996) that show that sophisticated technology acquisition is associated with establishment of subsidiaries in close proximity with technological “centres of excellence”.

Viewed in the light of this “subsidiary-focused” perspective, the roles of foreign affiliates are to a considerable degree influenced by the level of their local embeddedness. These roles largely depend on the incorporation of technological and idiosyncratic factors arising from the environment of the host country and the subsidiary itself (Andersson/Forsgren 2000, Andersson et al. 2005). This argument is linked to the finding that the more the subsidiary is focused on its local market (compared to its MNE system), the greater its degree of entrepreneurship (Birkinshaw et al. 2005). Not only that but also headquarters may benefit significantly from reverse knowledge transfers by subsidiaries located in highly competitive countries (Ambos et al. 2006).

In favour of the importance of decentralized technology sourcing, Almeida and Phene (2004) support the notion that technological innovation of the subsidiary is connected with its knowledge linkages to local organizations. Technological reliance on the local environment as a source of subsidiary competitive advantage is well justified in the works of Birkinshaw (1996) as well as Kotabe and Murray (1990). Additionally, in seemingly the sole study that directly tests the relationship between technology-related aspects and performance of the foreign subsidiary, Andersson et al. (2001) find that the affiliate’s technological embeddedness in the host country enhances its market performance. Consequently, these arguments lead us to the next hypothesis regarding the link between local technology sourcing and subsidiary performance:

Hypothesis 2. Local environmental sources of technology in the foreign country will be positively associated with subsidiary performance in that country.

MNE vs. Local Environmental Technology Sourcing

The literature provides ambiguous evidence concerning the relative mix of centralized vs. decentralized technology generation and diffusion. Yamin and Otto (2004) find that both internal and environmental knowledge sharing are crucial for MNE innovative performance. At the foreign subsidiary level, Granstrand et al. (1993) posit that apart from MNE factors, subsidiaries in their technology strategies should take into consideration also local environmental aspects. In addition, involvement of the subsidiary in the MNE innovation network is linked to the level of engagement with its local organizations (Johnston/Paladino 2007), a fact that underscores the interconnectedness between the internal and local contexts for the subsidiary. Moreover, both internal and local environmental contexts are significant for technological innovations of the subsidiary (Almeida/Phene 2004).

Therefore, although we expect that there will be a positive influence on subsidiary performance of both MNE and local technology sources, their relative impact is unknown. In other words, it is plausible that both of them will have a stronger impact on performance than the other. Thus, we propose the following null hypothesis to guide our empirical analysis:

Hypothesis 3. Internal MNE sources of technology will have a stronger (weaker) effect on subsidiary performance in the foreign country than local environmental sources of technology.

Sampling and Measures

Foreign subsidiaries located in Greece were examined in this study. ICAP Greek Financial Directory, a standard source of financial data for firms based in Greece, was the sampling frame employed. The ICAP database included 317 foreign subsidiaries, out of which 88 usable questionnaires were collected through a nationwide postal survey. The response rate of 28 percent is very satisfactory and perfectly acceptable when compared with similar postal surveys (Agarwal/Ramaswami 1992, Harzing 1997). The industries in which these firms operated included a wide range of manufacturing and service sectors, with food and beverages, pharmaceuticals, and textiles standing out as the most prevalent. Major countries of origin of these subsidiaries were the US, the EU, Japan, and other European countries. Appendix A summarizes the data on respondents and response rates by industry and country of origin for the investigated subsidiaries.

The structured questionnaire was pretested during a three-stage process. First, it was scrutinized by two knowledgeable academics and a professional consultant, who provided improvements on the wording and layout. This resulted in a major revision of the questionnaire. Second, it was sent to five chief executive officers (CEOs) of subsidiaries operating in different industrial sectors. In most cases their recommended amendments were similar and gave rise to a second revision. Third, the questionnaire was posted to ten randomly selected firms, chosen by their country of origin, for the final testing. Afterwards the survey questionnaire was posted to all 317 CEOs, followed by a second mailing three weeks after the initial one. In order to examine potential non-response bias we compared respondents and non-respondents against three variables: number of employees, level of sales and age of the subsidiary. None of the respective t-tests for differences of means was statistically significant.

The independent variables in the current study are technology sources of MNE and local environmental contexts. These sources were captured through a scale asking respondents to evaluate the importance of twelve technology aspects in their operations (see Appendix B for the operationalization of the variables). CEOs were asked to evaluate the prevalence of the sources under investigation using a four-point scale. The first six sources (C1-C6) represented centralized technology sourcing linked to the MNE internal context (Cronbach α is 0.76); while the remaining six (D1-D6) referred to the decentralized local Greek environmental context in which the subsidiary is located (Cronbach α is 0.73). All twelve items were drawn from Papanastassiou and Pearce (1999). In the MNE category of questions (centralized sources), the items referred to the application (or adaptation) of headquarters’ (or MNE system) pre-existing technologies in the products marketed in Greece. In the local environment category of questions (decentralized sources), the items concerned technology sources (or inputs) derived from the interaction with indigenous actors involving Greek firms, academic communities, research centres, employees etc.

Additionally, three control variables are employed in this study: age of the subsidiary, as assessed through the number of years the firm had been in operation; size of the subsidiary, as assessed through the number of employees in the firm; and, export intensity, as measured through the ratio of exports to total sales of the firm. Export intensity is included because of the potential influence of sales in markets outside Greece on foreign subsidiaries’ operations and performance (e.g., Dunning 1993, Hogenbirk/van Kranenburg 2006). This may be especially true for foreign subsidiaries based in Greece, given its location on the EU periphery.

The dependent variable in the current study is performance of the foreign subsidiary in the Greek market. Subsidiary performance is a complex construct to assess mainly due to the fact that it is influenced by different MNE objectives in the host market (Birkinshaw/Morrison 1995). However, according to Keating (1997), performance of a subsidiary should be evaluated against criteria that maximize its profits and return. In accord with this, Gray et al. (2001) state that return on investment (ROI) is the most utilized measure of performance, and is particularly convenient in measuring the success of subsidiary activities. Indeed, 65 percent of US MNEs reported by Reece and Cool (1978) employed ROI as the sole criterion for assessing their performance. ROI has been used on its own or in combination with other performance indicators in numerous studies examining foreign subsidiary performance (e.g., Birkinshaw/Morrison 1995, Borkowski 1999, Demirag 1987, Luo 2003).

In the present research, we employ ROI as the independent variable. We had also used return on equity and return on capital as indicators of performance in initial drafts of the questionnaire. However, we realized that ROI was the only financial indicator calculated by all investigated subsidiaries on an annual basis. Due to pragmatic reasons, for some subsidiaries ROI was likely to refer to a range of values rather than a specific figure. Therefore, in our research we classified reported ROI figures in to a 4-point Likert scale ranging from 4 (very high) to 1 (negative).

Results and Discussion

Table 1 provides the means, standard deviations, and zero-order correlations for the variables of the study. We checked for multicollinearity in the models through the examination of variance inflation factors (VIF) for each independent variable. The VIF values range below the upper limit of 10, which is typically suggested as the highest acceptable value (Hair et al. 1995, Netter et al. 1989). This indicates that multicollinearity does not pose a problem for the results of this study.

Table 1 Descriptive Statistics and Zero-Order Correlations
Full size table

Next, we performed an ordered probit regression analysis. We used the ordered probit regression model as this is the appropriate technique when an ordinal polychotomous dependent variable (ROI) exists. We estimated four probit models to investigate the relationships between independent and dependent variables. Three of these models are baseline or restricted models whereby either the control variables, internal MNE (centralized), or local environmental (decentralized) technology sources are regressed independently on the dependent variable. The other one is the full model that includes control variables, centralized and decentralized technology sources. Table 2 presents the results of the four regression models.

Table 2 Ordered Probit Regression Results
Full size table

Hypothesis 1 predicted that internal MNE sources of technology would be positively associated with subsidiary performance in the foreign country. The results confirm this hypothesis. Among the centralized technology sources, only ‘core application of parent existing technologies embodied in established products’ (C4) is not significantly associated with performance. In general, the MNE technology sources are positively and significantly linked to performance in the restricted model, the full model, or both. This result supports those by Manolopoulos et al. (2005), who found that in the case of foreign subsidiaries based in Greece, MNE-related factors predominantly determine the roles allocated to subsidiary R&D departments. Thus, it appears that in line with the centralized view, when the subsidiary receives technological knowledge, ideas and inputs from the MNE system, its performance in the market it operates is enhanced. This is seemingly the first study providing empirical support for this link.

On the other hand, the results of the probit analysis do not appear to support Hypothesis 2, which stated that local environmental sources of technology in the foreign country would be positively associated with subsidiary performance in that country. Table 2 reports that only in one instance was a decentralized technology source positively connected with performance (at the 10 percent level): this was the case in the restricted model for technological inputs derived from joint projects with the local academic community (D3). This finding is in accord with results in Germany (Terpstra 1977) and the UK (Lawson 1999), which show advantageous synergies concerning technology and R&D spillovers. In these studies, the positive spillovers derive from close links between foreign subsidiaries and universities located in these (larger) countries. This significant and positive influence on performance has, nonetheless, disappeared in the full model of our research when all variables were included.

Hence, the marginally positive result should not mask the overall picture that indicates that two local environmental technology sources (D1 and D2) in both the restricted and the full model consistently have a significant and negative effect on the performance of the subsidiary. In other words, technological collaboration and joint research activities with Greek scientific institutions, research centres and local firms have a negative impact on subsidiary performance. This may be because technological collaboration with these firms is likely to be exploitative in nature, unlike that with academic universities that can be largely exploratory.

To elaborate, universities largely pursue basic technological research that can provide a win-win situation for the parties involved; whereas, the other local organizations are likely to seek to promote and take advantage of applied research that may not necessarily be of considerable interest to subsidiary managers. This technological know-how of an exploitative (applied research) type is likely to have already been supplied to the Greek subsidiary by the MNE group. Consequently, subsidiary management can view collaborative activities with local organizations other than universities as a misuse of time and resources, which would adversely affect subsidiary profitability.

The unexpected finding pertaining to the primarily negative impact of decentralized technology sourcing on performance is, however, close to that of Davis and Meyer (2004). These authors found that subsidiaries operating in competitive environments of small countries with advanced supplier conditions engage in few (rather than many) R&D activities. Again, the MNE group is likely to be the main provider of R&D know-how to these affiliates. Thus, Davis and Meyer (2004) empirically support the notion that “small countries cannot like large countries rely on the strength of the endowment of [their] market and supply conditions” (p. 376), an argument that the evidence of the present study seems to endorse. We will return to this point in the concluding section of this paper.

In addition, in order to investigate the relative importance of centralized vs. decentralized technology sourcing we compared the full model with the two restricted models. To illustrate, we determined the comparative significance of a set of variables over another by performing χ2-tests involving both the full and the restricted models. Such an examination attempting to explore the relative importance of one set of variables over another is often encountered in prior studies (Hansen/Wernerfelt 1989, Kotha/Nair 1995). Figure 1 shows the results of this analysis.

Fig. 1
figure 1

Variance Decomposition Model

Full size image

The restricted model containing the centralized technology sources is a significant improvement over the base model (Δχ2 = 9.27, p < 0.01). Likewise, the restricted model containing the decentralized technology sources is also a significant improvement over the base model (Δχ2 = 17.03, p < 0.05). The inclusion of both centralized and decentralized technology sourcing improves the full model significantly over the centralized model (Δχ2 = 4.28, p < 0.05) and the decentralized model (Δχ2 = 16.09, p < 0.05). Further, the relative effect of MNE sources is stronger than that of local environmental sources, as measured by the adjusted pseudo R2 and comparative significance test (Δχ2 and p) levels. Thus, with regard to Hypothesis 3 the evidence validates the view that internal MNE sources of technology have a stronger effect on subsidiary performance in the foreign country than local environmental sources of technology. It may be that such a result is valid for Greece and generally other smaller countries since the strategic significance of these host countries for MNE technology development is often not very high. This is a view inferred in subsidiary studies undertaken in such economies (Davis/Meyer 2004, Manolopoulos et al. 2005, 2007).

In a nutshell, this study provides evidence suggesting that internal MNE technology sourcing has a positive effect on subsidiary performance; in contrast to the effect of local environmental technology sourcing, which is primarily negative. Also, the impact of local environmental technology sourcing is of weaker magnitude than that of MNE technology sourcing.

Conclusions

Implications

There are valuable insights for research, management, and public policy that can be drawn from these findings. With regard to the implications for research, this present work is to our knowledge the sole empirical study examining the connection between technology sourcing and performance. But, even more important, the present research empirically shows that MNE factors are more significant than local environmental aspects for subsidiary technology sourcing. Thus, the “head office assignment” view in relation to technology sourcing is confirmed from the evidence of this paper. These results sound a cautionary note to researchers in their examination of subsidiary technology strategies. While the decentralized “subsidiary-focused” perspective appears to have gained considerable momentum in the literature recently, this study highlights the key significance and primary importance of the centralized “head office assignment” view.

In an attempt to explain the finding regarding the predominance of the centralized viewpoint, we highlight implications that are appropriate for subsidiary research in small countries. As outlined above, one could assert that the strategic significance of Greece is not that strong for MNEs that come to invest in that country. Thus, MNEs would rely on their internal systems and networks to provide technological ideas and know-how to their Greek affiliates. Associated with this argument is the suggestion that the local context is not so relevant in explaining the investigated link between technology sourcing and performance for subsidiaries located in Greece. The evidence from Manolopoulos et al. (2005) indicates that some foreign subsidiaries in Greece have evolved into regional hubs servicing the larger Southeast European market. This argument confirms similar conclusions provided by recent studies investigating subsidiaries in other small countries (Benito et al. 2003, Davis/Meyer 2004, Hogenbirk/van Kranenburg 2006). Consequently, the fact that the “subsidiary-focused” perspective is not validated in the present research may imply that investigation is required to explore which is the relevant market for foreign subsidiaries in small countries: is it the host market in which they are based; or, neighbouring markets into which subsidiaries have expanded and may serve as terrains for their technology strategies. Export intensity was included as a control variable in this research, but, while the results were significant, the sign was both positive (base model) and negative (full model) in the regressions that were calculated (Tab. 2). Therefore, there are no meaningful conclusions that can be inferred and further study is clearly required.

With regard to implications for MNE managers, the evidence of our work indicates that the technological resources and know-how generated by the MNE headquarters and group appear to matter a lot when it comes to subsidiary performance in small countries like Greece. Management at headquarters have to ensure that they generate and source to their small country subsidiaries technological knowledge flows and outputs on a constant basis. Similarly, MNE subsidiary managers need to closely monitor technological developments at headquarters; and, encourage technology transfers to their small country operations through building strong network relationships with their parents and MNE actors. It might also be suggested that MNE headquarters managers should be cautious in supporting subsidiary management for collaborative technological projects with local partners (other than universities) as this cooperation could have detrimental consequences on subsidiary performance. Certainly more investigation is required into the reasons of these adverse effects of collaboration with host small country scientific institutions, research centres and local firms on performance.

The study provides also challenging insights for public policy. Our finding that the role of local technology sourcing for subsidiary performance is very possibly negative clearly comes as a disappointment to public policy decision-makers in small countries. Such a challenge for public policy occurs at a point in which other recent studies question the relevance of policy initiatives seeking to support local company interaction and collaboration (e.g., Alecke et al. 2006, Bode 2004). Nevertheless, two related considerations stem from the findings of this study and are likely to support useful policy intervention.

First, our results imply that a fruitful strategy that public policy makers in small countries can implement is to attract to their territories foreign MNEs that have a solid record of technological innovation and output. This appears to be a realistic policy to pursue inasmuch as the MNE context is very important for technology sourcing and subsequent performance. Therefore, targeting should focus upon firms with an acknowledged strong technological background and record, with inducements such as tax exemption schemes. Second, our findings suggest that the influence of joint projects with the local academic community might have a positive impact on subsidiary performance. Further research is needed to reach a firm conclusion on the validity of this statement. In spite of this, if indeed it is the case, the specific linkages with academic institutions should be encouraged and cultivated. Such a recommendation paves the way for a targeted policy to support collaborative arrangements between local universities and foreign subsidiaries. In general, our two suggestions linked to a targeted (rather than an “all-inclusive”) course of action on subsidiary attraction and subsequent after-investment incentives are in accord with similar recommendations in the literature for “tailor-made” inward foreign direct investment policies (e.g., Young et al. 1994, Tavares/Young 2005).

Future Research Directions

The limitations of this study guide suggestions for future research. The generalizability of the findings of the current research would be strengthened if studies were carried out in other small countries with dissimilar levels of economic and technological development; and, different geographic locations within their continents. Further, comparison of datasets involving subsidiary activities in small and large countries would illuminate whether the two country categories indeed produce different results. Similarly, extensions of the research reported in this paper could include examination of other variables likely to affect the link between technology sourcing and performance such as the roles of foreign subsidiaries, including market servicing roles (local and/or foreign markets, such as Southeast Europe in the present paper).

Future research can also assess performance indicators other than ROI. While the use of ROI is well-documented and widely used in the MNE subsidiary literature, other variables capturing the whole range of foreign subsidiary objectives in the foreign country would be helpful. Such variables could be market share, R&D output, or even indicators attempting to capture more societal aspects of the subsidiary’s activity. These indicators could be the increase of technology-related employment in the subsidiary; or, the number of spin-offs generated from the subsidiary.