Master Thesis
The Influence of Entry Modes on
Innovation: firm level evidence from
Central and Eastern Europe
Master thesis International Business and
Management at the University of Groningen
Student:
Johan Marinissen (s1883100): j.marinissen@student.rug.nl
Supervisor:
Dr. Gjalt de Jong
Abstract
By performing logistical regression the relationship of entry modes and innovation is analyzed. In this study entry modes are considered as foreign direct investments in greenfield, joint venture or acquisition. Innovation is measured as new or significantly improved products. In this research I make use of micro enterprise level data of five countries of Central and Eastern Europe. With this I account for firm heterogeneity and agglomeration economies within this region. It was found that acquisition based modes are best for innovation. Joint venture based modes are second best for innovation. Greenfield based modes are on the last place. Most results are insignificant, however one variable stands out as very important and significant, which is co-operation with other enterprises or organizations
Table of Content
Table of Content ...1
Introduction ...5
1. Literature Review ...8
1.1 Origins of Entry Mode Research...8
1.2 Consequences of Entry Modes ...22
2. Hypotheses...25
2.1 Greenfield ...26
2.2 Acquisition...27
2.3 Joint venture ...28
3. Methods...30
3.1 Survey & Sample...30
3.2 Measures of Innovation...31
3.3 Measures of Entry Modes ...32
3.4 Control Variables ...32
3.5 Estimation Method ...33
3.6 Evaluation of Method Assumptions...36
4. Results ...40
4.1Descriptive Statistics ...40
4.2 Logistical Regression...44
4.3 Complementary Log Log Model...46
5. Discussion ...48
6. Conclusion ...50
References ...52
Introduction
By the end of the twentieth century, foreign direct investment (FDI) effectively replaced trade as a driver of economic growth in less developed and emerging economies. A decade later, in 2011, flows to developed countries increased further to $748 billion. By contrast, in developing countries FDI increased to a record of $684 billion. In transition economies FDI increased to $92 billion (UNCTAD, 2012). Accordingly, the determinants and consequences of FDI is a much debated topic in international business research. Since the beginning of this research, the market entry mode choice has been considered one of the most important decisions in the internationalization process (e.g., Wind and Perlmutter, 1977; Anderson and Gatignon, 1986; Hill et al., 1990; Quer et al., 2007; Brouthers and Hennart, 2007).
on the explotation of local differences in autonomous country based operating units, to more differentiated network forms that enable specialization where needed, but also greater integration where possible (Bartlett and Ghoshal, 1993; Nohria and Ghoshal, 1997). The increased complexity of knowledge processes, which are the backbone of new technologies and innovation, leads firms to search beyond their own boundaries for valuable knowledge and skills, in order to complement their own capabilities (Becker and Dietz, 2004). Thus, existing studies do not, at least explicitly, study the relationship between entry mode strategies and subsidiary innovation.
The aim of this study is to understand the relationship between entry modes and subsidiary innovation in a context of CEE and in so doing it offers a theoretical contribution to the existing FDI entry mode literature. This study will make use of firm-level data in order to take the comment of Günther et al. (2011) into account who argues that existing studies are implemented at the country- rather than regional level and, therefore, neglect the role of agglomeration economies in choice of location. They furthermore argue that recent theoretical advances do require micro data sets at the enterprise level in order to take account of firm heterogeneity, a requirement this study fulfills. Furthermore, this study extends the literature on international technology spillovers in three directions. First, this study proposes and empirically tests an integrative framework for a variety of international spillover sources including greenfield FDI, acquisition and joint venture (JV) entry modes. This integrated framework is much needed for theory building and empirical testing in technology spillover research. It has important policy and managerial implications not only for CEE countries, but also for other emerging economies. It may even have implications for western firms with an interest in preventing spillovers that represent leakages of their core competences investing in CEE. Second, within this integrated framework I aim to enhance the knowledge of the dynamic and complex nature of innovation in CEE. Finally, by identifying a broad, integrated set of spillover channels, this study makes a contribution to the extension of the resource-based view (RBV) of the firm to the area of technology spillovers.
The general research question for this study is:
The specific questions are:
1. What are entry modes and how can we measure it? 2. What is innovation and how can we measure it?
3. What is the relationship between entry modes and firm innovation? 4. What are the results of this relationship applied to CEE enterprises?
1. Literature Review
This chapter elaborates on a thorough understanding of the term entry mode in the context of FDI. To provide an answer to this issue, in the first paragraph, I will review the origins of FDI entry mode research. Thereafter, in the second paragraph, this chapter is closed by a discussion about the consequences of entry mode choice, for the reader to have a comprenhensive understanding of entry mode research.
1.1 Origins of Entry Mode Research
Transaction cost economic perspective
control, bounded rationality, asset specifity, firms make a decision to internalize or choose a market form of cooperation in an effort to make the transaction cost perspective as effective and efficient as possible
Institutional perspective
authorities. The institutions of an economy determine the pattern of transaction costs. Businesses thus select, and develop, coordination mechanisms that fit the environment. In addition, Estrin and prevezer (2010) found a critical interaction between formal rules and informal mechanisms. Relatively good formal rules and structures can be undermined by informal mechanisms deterring or blocking entry (for example the case in Russia). An innovative and important strand of the institutional literature has attempted to quantify institutional barriers to new firm entry, using indices about the complexity of starting and doing business in different national contexts as well as about contract enforcement (Djankov et al., 2002). These data, available at the World Bank Doing Business, necessarily represent largely formal measures of institutions because they are constructed in a comparable fashion across 85 countries so as to facilitate the use of cross-country econometric methods. The complexity of institutions in the literature includes a focus on property rights and historical legacies in their establishment (Acemoglu et al., 2002, 2003); on the political power of the elite and the relationship between the state and entrepreneur (Acemoglu and Robinson, 2000); and on the constraints on the elite and safeguards on state intervention and the independence of the judiciary (La Porta et al., 1999; Acemoglu et al., 2001). Thus, institutional theory focuses on the deeper and more resilient aspects of social structure. It considers the processes by which structures, including rules, norms, and routines, become established as authoritative guidelines for social and professional behavior.
OLI perspective
Internalization advantages, associated with reduced transaction and coordination costs, are the result of firms' attempts to create an internal hierarchy of their own in order to attenuate possible failure of the external market and hence to protect their ownership advantages in a foreign market (2009). Ownership and location advantages, specifically, their proxies have been extensively studied by IB scholars. Ownership advantage proxies that have been empirically tested and supported in entry mode literature include multinational experience, country experience, firm size, and firm resources (Agarwal and Ramaswami, 1992; Caves and Mehra,1986; Kogut and Singh, 1988). Location advantage proxies that have been empirically tested and supported include market potential, investment risk, and cultural distance (Agarwal and Ramaswami, 1992; Caves and Mehra, 1986; Chang and Rosenzweig, 2001; Erramilli and Rao, 1993; Kogut and Singh, 1988). According to Dunning (1988), any firm expanding its operations to a foreign market incurs certain disadvantages compared to its local competitors which might range from cultural barriers to limited knowledge about local institutions. The OLI framework posits that, given such challenges, unless a firm has some superiority over local firms in the form of more efficient and effective production process, superior products and/or services, stronger connections with international capital market, that firm has no ownership advantages. Moreover, Agarwal and Ramaswami suggest that to compete with host-country firms in their own markets, firms must possess superior assets and skills that can earn economic rents that are high enough to counter the higher cost of servicing these markets (1992). Thus, based on the OLI framework firms can differentiate themselves from competitors by putting a focus on three significant constructs of FDI in order to develop a synerchy between them and operate on the advantages that drive superior performance in FDI.
Agency theory
profits and working against external threats, but can be opposed when bargaining with each other over the intra firm allocation of resources. Subsidiary managers are both profit seeking and rent seeking, as their actions take place with two different objectives in mind. Agency theory suggests that a firm is a network of contracts. Managers are expected to comply with the interests of external owners of private enterprises. It is, however, a challange for owners to ensure that managers do comply, since it is difficult to specify ex ante contracts that accommodate all possible future contingencies. Following Shleifer and Vishny (1997), asymmetric information between managers and external investors increases monitoring costs and enables managers to pursue their own goals. Thus, information asymmetry is a crucial determining factor influencing the entry mode choice decision. Hennart (1996) argues that information asymmetry is not a problem in itself in the principal agent relationship. It becomes a problem when combined with moral hazard, that is, the potential for agents to operate in their own self-interests against the objectives of the principals. The principals fully recognize that the potential for moral hazard, combined with information asymmetry, can potentially place them at a severe disadvantage. Consequently, to enter into a principal agent relationship, principals require an effective mechanism to control agent behaviour. One effective mechanism involves monitoring agent behaviour. Cantwell (1996) presents another view in arguing that the greater the cultural differences between home and host countries, the higher the costs of monitoring employees. In such instances, agency theory, with its focus on information asymmetry and moral hazard, and the associated internal organization costs of monitoring, provides a valid explanation for using lower-cost/lower-control entry modes such as fran- chising and licensing. Thus the agency theory, in general, deals with two specific problems. First, that the goals of the principal and agent are not in conflict. Second, that the principal and agent reconcile different tolerances of risk.
Resource-Based perspective
Control
Dominant Research Agenda
Determinants of entry modes
In the first paragraph of this section the determinants of FDI in general will be discussed as an introduction to the determinants of the specific entry modes which will be discussed in the next paragraph.
Determinants FDI
(i.e. JV instead of acquisition) can incrementally gain knowledge about the host country with experience that allows them to make a better informed decision in the future without making a large irreversible investment (Ashan and Musteen, 2011).
Parent firm preferences are influenced by the idiosyncrasies of the nations hosting these subsidiaries (Gatignon and Anderson, 1988). Understanding the ownership structure of FDI projects is important because the structure affects the incentives of the investors to apply their resources to the project. Equity shares influence the cost of capital, the level of investment, the degree of technology transfer, the distribution of gains from FDI, the control it can exert on its subsidiary and it protects the integrity of the MNE's assets (Asiedu et al., 2001). Within this ownership decision, the level of firm-specific technology (asset specificity) may also influence mode choice, since firms with greater technology may incur higher transaction costs in safeguarding their technology from misappropriation (Hennart, 1991).
Type of industry is another important determinant of entry mode research. Empirical evidence of Elango and Sambharyab (2004) shows that underlying industries’ structural characteristics influence a firm’s preference for a particular entry mode. In addition, Bhaumik and Gelb (2005) specify that entry mode choice is based on one or more of the following variables: growth of local industry, technology-intensiveness of product, competition in the local market, resource needs of the MNE, local institutions, governance, and business regulations, prior operating experience in developing-country environments, cultural distance between MNE’s home country and host country, extent of liberalization of FDI regulations and industry-specific regulations, perceptions about quality of host country’s managerial labor, and/ or sector of operation of the MNE. They further find that the importance of the choice of the entry mode, from a policy-making perspective, ''is a consequence of the stylized result that the extent of technology transfer by an MNE to a developing or emerging country affiliate depends significantly on the extent to which the parent can exert control over the affiliate, which, in turn, is determined by the mode of entry of the MNE into that country'' (p. 15).
Determinants specific entry modes - acquisition, greenfield, joint venture
Although the growth in global FDI flows in 2011 was driven in large part by cross-border Merger and Acquisitions (M&A’s), the total project value of greenfield investments remains significantly higher than that of cross-border M&A’s, as has been the case since the financial crisis (UNCTAD, 2012). The choice between acquisition and greenfield is represented as a dillema which involves different risks and opportunities. Zekiri and Angelova (2011) argue, acquistition is the appropriate mode of entry when the market is developed for corporate control, the acquirer has high absorptive capacity and is characterized by high synergy, whereas greenfield entry is the appropriate mode when there is lack of a proper acquisition target, in house local expertise, and embedded competitive advantage. According to Harzing, those firms more diverse and with lower R&D intensity are more likely to buy technological capabilities in other countries by acquisitions and this propensity increases where local firms have well established distribution systems and a deeper knowledge of the local market (1999).
Acquisitions are used primarily to secure brands and technology quickly and pre-empt similar moves by competitors. Thus acquisition adds innovation, differentiation, and brand advantages to the existing cost advantage (Luo and Tung, 2007). The model of Yokota and Chen predicts that, with relatively high R&D spillovers and relatively small difference in cost between M&A and greenfield FDI, an R&D intensive firm is likely to choose greenfield investment over M&A. On the other hand, with relatively small R&D spillovers, a MNE is more likely to prefer M&A rather then greenfield investment. They further find in the situation with high R&D spillovers, foreign firms with strong technological advantages are likely implement greenfield investment ventures to avoid issues associated with the leakage of their technology (2012).
Filatotchev find that the availability of scarce resources affects the location of FDI and the likelihood of greenfield entry (2005). For example, according to Norris (2011) entering China's complex landscape with a greenfield investment is the path with the greatest risks and rewards. The manufacturing efficiencies that come with a newly designed and constructed facility are the primary drivers for such an investment option.
entry if the local market is very competitive, such that, despite the technological/competitive edge, the magnitude of returns on the investment made in the host country is uncertain (Hennart and Park 1993).
If a rapidly growing industry also promises high rates of return on investment into the future, it may be reasonable for the MNE to minimize its agency and restructuring costs by opting for a greenfield project, even though such a move would increase the transactions cost in the short run (Cheng 2006). In concentrated and high-growth industries, foreign firms prefer entry by setting up greenfield operations rather than pursuing acquisitions or JV (Elango and Sambharyab, 2004). Furthermore they argue that growing markets seem to encourage entry of foreign firms by greenfield operations and foreign firms also have greater motivation in growing industries to set up greenfield operations and gain the potential long run benefits of independent operations.
In the study by Liu and Zou (2008), with Chinese hightech industry data, they examine differences in the impact of greenfield FDI and M&A on the innovation performance in a single framework and show that both greenfield FDI and M&A have a positive association with the R&D activities of foreign firms.
Firms typically acquire related firms based on the assumed benefits from reducing costs, increasing product value, avoiding price wars, increasing bargaining power, and enhancing organisational effectiveness (Prajogo and Sohal, 2006). The other motivation for acquisitions is that the acquired firms have unrelated but complementary strategic assets, such as customer and supplier relationships, distribution systems, and technological know-how. Moreover, managers believe that they can increase their efficiency and decrease production cost of target firms by transferring technology and learning management skills from other organisations (DahlgaardPark and Dahlgaard, 2010). The acquirer has to deal with the problem of restructuring the acquired organization (such that its business culture is similar to that of the parent MNE) and the cost associated with such restructuring (Liu and Zou, 2008).
durable goods to local conditions requires skills better captured through acquisition than through greenfield investment.
The study of Cheng (2006) finds that the higher the level of industrial concentration in the foreign market, the higher the likelihood of a firm's FDI entry through acquisition or brownfield over greenfield. The choice of the entry mode would depend on the rate of growth of the local industry, and on whether, broadly speaking, the MNE operates in a manufacturing or a services-sector industry. If an industry is fast growing, and therefore fast changing, it may be essential for an MNE to quickly have a stake in it, so as not to lose its first-mover advantage to other MNE's or local firms. In such an event, an entry by way of acquisition, for example, may be more suitable.
Acquired subsidiaries evolve to become more embedded in the internal MNE network at the expense of their external network ties. Greenfield subsidiaries do not exhibit any such internal-external trade-off because they do not have a history of ties that are dependent of the parent MNE. Zhao (2009) furthermore finds that the acquisition decision seems to be based on innovation output quality, rather than innovation input or output quantity. However, right from the beginning on acquisitions require some level of adaptation since two companies, the acquirer and the acquiree, have to become one. Zhao (2009) finds that firms engaging in acquisition activities are less innovative and have often experienced declines in technological innovation during the years prior to the bid.
1.2 Consequences of Entry Modes
Most of the literature focuses on the relationship between entry modes and performance. In addition, the main theoretical approach to the analysis of spillovers from MNE's has been industrial organization economics and new growth theory, e.g. the literatures on FDI spillovers and R&D spillovers (Grossman and Helpman, 1991). Few studies consider, however, more recent developments in strategic management research, such as the RBV (Barney, 1991). Given the importance of technological knowledge as a contribution by MNE's to host economies, Liu and Zou (2008) made the first attempt to extend the RBV to analyze technology spillovers. They find that foreign greenfield R&D activities by MNE's in a host country significantly affect the innovation performance of domestic firms and there exist both intra-industry and inter-industry spillovers from foreign greenfield R&D, There are, however, only inter industry spillovers in M&A. They further find that importing foreign technology and investing in domestic R&D have positive impacts on domestic innovation. Extant literature prevails to use performance as a rather broad measure of financial and non financial profits of the overall company/ subsidiary, wheras a more precise definition of the performance measure would yield much more.
affect the extent of knowledge transfer since the investment size and the subsidiary's role vary with it, as is confirmed in the evidence of Yang et al. (2008) about determinants of conventional and reverse knowledge transfers in three transition economies in Central and Eastern Europe. Nevertheless, there is general evidence of the relationship between entry modes and innovation, as Yokota and Chen (2012: p. 269) argue, ‘’although there are many studies that examine the effects of total FDI on host firms’ productivity, the relationship between the various channels of technology transfer and the firm’s performance in host countries has received little attention’’.
According to Liu and Zou (2008) international technology spillovers have become a central theme in the study of the impact of trade and FDI on host countries. Knowledge originating in one country increasingly transcends national boundaries, and contributes to technological progress in other countries. Innovation is no longer purely based on domestic resources. FDI and trade are viewed as the main channels for technology spillovers. In addition, Tether (2002) observed that some reviewers argue innovation is no longer the driver of individual firms, but is a matter of collective action, with firms acting together and participating with suppliers, customers, competitors, consultants and/or universities in co-operative arrangements for innovation.
Yokota and Chen suggest that the choice of entry mode of the MNE crucially depends on the magnitude of R&D spillovers in entry modes (2012: p. 277). Aghion et al. suggested that ''entry can induce reallocation of inputs and outputs, trigger knowledge spillovers, and affect innovation incentives in incumbent firms'' (2009: p. 20). They describe a firm’s motivation to innovate as an escape-entry strategy, that is, in order to discourage leading-edge competitive entry, an incumbent at the technology frontier will innovate in order to maintain its market position.
2. Hypotheses
I use the RBV as explaining theory in bringing the two concepts of interest in this study - entry modes and innovation, together. The RBV is particularly suitable for explaining international entry strategies because its dynamic foundations provide a basis to analyze the dynamics of firm growth (Penrose, 1995), in particular the direction and mode of international growth (Mahoney and Pandian, 1992; Meyer, 2006). The RBV looks at the firm's strategies from its resource endowment and deployment (i.e., inside-out view of the firm), this will throw new light on the firm's strategies as explained in Wemerfelt (1984). According to Barney, sustained competitive advantage derives from the resources and capabilities a firm controls that are valuable, rare, imperfectly imitable, and not sustainable (1991). In addition, the literature on FDI spillovers can be seen to be highly relevant to the RBV. Foreign entry introduces the risk of losing control of key knowledge resources. It is based on the notion that MNE's must possess superior technology, management skills and intangible assets in order to overcome the difficulties of doing business abroad. Technological superiority, or possession of some intangible assets, is believed to provide advantages for MNE's over indigenous firms. The firms investing in foreign countries therefore have distinctive characteristics which may differ from those in host countries (Sharma and Erramilli, 2004). The data used for this study fits the characteristics of the RBV as well.
Entry modes and innovation
entry, each requiring a different form of co operation, different magnitudes of innovation will be evoked. (1). Greenfield, (2). Acquisition, and (3). Joint venture. Due to a lack of research taking these entry mode variables sperately into consideration, there is no reference point as to what type of entry will induce the most innovation. This study fills this gap by taking the entry modes separately into consideration.
2.1 Greenfield
weak (Peng, 2000). The probability that this leads leads to a loss of focus in innovation on additional/ complementary processes is therefore high and innovation activities may be hampered. Therefore the following hypothesis is proposed:
Hypothesis 1: Greenfield based entry modes will limit subsidiary innovation
2.2 Acquisition
innovation. Innovation outcomes of acquisitions are driven by the pre acquisition knowledge of the acquirer and its similarity with the targets’ knowledge. They find a strong interaction effect of the depth of the acquirer’s existing knowledge and that of its acquisitions on innovation output. They also find that moderate similarity leads to greater postacquisition innovation than does either low or a high similarity between the acquirer’s and the target’s preacquisition knowledge (p. 125). Furthermore Lee and Lieberman (2010) suggest that an acquisition is the most efficient way to obtain tacit experiential knowledge and thus giving access to unique information, providing important input for innovation processes. Therefore the following hypothesis is proposed:
Hypothesis 2: Acquisition based entry modes will limit subsidiary innovation
2.3 Joint venture
organizational learning and the creation of new knowledge in the investing firm resulting from its combinative capability (Kogut and Zander, 1996), and by interacting with organizations holding complementary knowledge, notably JV's. Thus, when new knowledge development is more important, firms more likely choose collaborative entry modes (Madhok, 1997). Therefore the following hypothesis is proposed:
3. Methods
In this section the data used in this study will be discussed. In addition, the dependent, independent, and control variables will be measured. For all categorical variables or variables measured at the nominal levels, k-1 dummies were created before they were used in the analysis. No dummies were, however, created for ordinal level variables.
3.1 Survey & Sample
I test the hypotheses using data from a unique multi-country, multi-industry database. The source is a 2011 subsidiary-level survey conducted in the Czech Republic, Hungary, Romania, Poland, and the Slovak Republic by Institute for Economic Research Halle (IWH). This 2011 IWH survey database offers the opportunity to measure the dependent variable (i.e., subsidiary innovation), the key independent variable (the specific entry modes) as well as headquarters, subsidiary and industry characteristics.
The 2011 IWH survey database is part of a larger project aimed to systematically collect information about innovation activities and the role of foreign investors in former Eastern and Central European (CEE) countries. The overall population of subsidiaries from which the IWH sample is taken from Orbis (broken down per ownership for each country) and consists of foreign-owned manufacturing and service subsidiaries located in the five CEE countries. The selection of these countries in economic transition balances country size, geographic location, and the level of economic development. The population includes different types of foreign investors such as pension funds, banks, foundations, individuals, families, or any combination of these different types of owners (Dut, 2013).
request. The questionnaire was the same in all countries. It was first tested for coherency to at least four pre-tests per country before being submitted to the subsidiaries between 6 August and 3 September 2009. The final questionnaire required 15 minutes on average for completion. The IFAS interviewers received intensive training by IWH regional experts concerning innovation and business activities in CEE countries. The interviews were conducted by native speakers from each of the countries under observation. The 2011 IWH survey used selection questions for various parts of the survey implying that very few companies answered all questions in the survey. Between 21 September and 16 December 2011 IWH completed the required interviews in line with the sample stratification (Dut, 2013).
3.2 Measures of Innovation
as a dichotomous variable. This is binary coded with 1, denoting respondents that answered yes to product innovation. No answers are coded 0.
3.3 Measures of Entry Modes
I limit the entry choice to ownership-based entry modes (OBE) because OBEs (joint ventures, acquisitions, and greenfields) often represent more complex investment forms and structures than contract-based entry modes (e.g., exporting, subcontracting, licensing, and franchising), and these high order entry modes have generated more research attention. I measure entry mode with the question ''what descrtibes best the initial entry mode of your foreign investor?'' Each entry mode was coded 1 if respondents answered yes to and 0 if they answered no to whether or not they belong to any of the entry modes.
3.4 Control Variables
To make the findings as robust as possible I will include a set of control variables. To first discuss the actual measurements, for all control variables such as country and sector of business and participation in R&D cooperation measured at the nominal levels, K-1 dummies were created. For example, in the case of country of origin which has five categories, Hungary was used as the reference category in the analysis. On the other, no dummies were created for variables measured at the ordinal level such as quality of skilled labour, quality of unskilled, quality of apprentices and trainees and quality of junior employees with university degrees. Accordingly, such variables were treated as continuous variables.
First, results of Winne and Sels (2010) indicate that owners/managers, human capital, employees and HR practices significantly contribute to innovation in start-ups. It thus can be assumed that human capital is an additional driver of innovation, therefore I will include the human capital of the subsidiary as a control variable. I measure human capital of the parent company by asking ''How do you evaluate the quantitative labour supply at your enterprise’s location currently. The respondents can choose between very good, good, poor, very poor.
beneficial to a firm’s innovation performance (e.g. Cincera et al., 2003; Belderbos et al., 2004; Lööf and Broström, 2008). Therefore I will include potential and actual R&D corporation of the subsidiary as control variable. I measure potential cooperation by asking ''how do you evaluate the potential for technological cooperation with 1. universities and other public research institutes, 2. with other enterprises (customers, suppliers, competitors) at your enterprise’s location currently?''
The respondents can choose between very good, good, poor, very poor. I measure actual cooperation by asking ''Did your enterprise participate in any R&D co-operation with other enterprises or organizations?'' This measure ranges from 0 - No to 1 - Yes.
Third, Inkpen and Beamish (1997) suggest that the advantages and drawbacks of R&D collaborations may be heightened when the partner is foreign. Indeed, some foreign companies or research institutions in a given industry may possess specific more advanced, knowledge of interest for domestic firms, either because it can be transferred into products for the domestic market, or because it is needed to penetrate the foreign market. This indicates that the level of innovation is dependent on either local or foreign partners. Therefore I include local and foreign embeddedness as the third control variable. I measure local and foreign embeddedness by asking the question ''With which of the following partners did your enterprise co-operate in the area of R&D?''
For details on the operationalization of the different measures I refer to appendix 1 in which in an overview will be given of the actual questions respondents had to answer and out of which the data was developed.
3.5 Estimation Method
Descriptive statistics
To evaluate the formulated hypothesis, the analysis begins with a presentation of descriptive statistics of the dependent, independent and control variables used in the study. For nominal level variables, percent distributions are provided while means are estimated for ordinal and interval level variables (See Table 1).
Given that the dependent variable is dichotomous, a logit or logistic regression model is specified instead of the conventional OLS model which is built on the assumption the dependent variable of interest is a continuous variable. By applying the logistic regression model in the current context, I am able to predict whether or not an enterprise will implement any product innovation. In STATA, the logistic regression model would predict the probability of initiating a product innovation as Y=1. The logistic regression model is thus specified as follows (Long, 1997):
ln(P/1-P)=α+βx
where P is the probability of innovating and α is constant term as in an OLS model. βis the parameter coefficient associated with the predictor(s) of product innovation. Ln is a natural logarithm.
The estimated coefficients can then be exponentiated and interpreted as odds ratios. Variables with ratios higher than 1 are interpreted as being more likely to implement a product innovation while ratios that are less than 1 are considered as being less likely to implement a product innovation. Also, for a better understanding, these ratios can be converted into percentages. To effectively test the three hypotheses postulated, five models were run. Consistent with prior research on the subject, the first model (Model 1) is a multivariate model that examines the effect of all theoretically relevant control variables on product innovation. Since it was of interest to compare the effect of each entry mode on product innovation, three sets of bivariate models examining separately, the effects of greenfield, joint venture and acquisition entry modes on product innovation. Thus, respectively, Models 2, 3 and 4 are bivariate models showing the effect each entry mode has on the likelihood of initiating a product innovation. Each of these models has no other covariate other than the respective entry modes. However, in order to examine the independent effect of greenfield, joint venture and acquisition entry modes on product innovation, a final model (Model 5) was estimated. This is a full model that includes all the entry modes variables and controls for all the controls variables outlined in the study.
log-log regression model is premised on the understanding that although the dependent variable is dichotomous, the cases are unevenly distributed among categories and thus, the logit link
functions which assume a symmetrical distribution might produce biased estimates. Given this, I used the complementary log-log function which is better suited for asymmetrical distributions (see Long, 1997). It has the form:
ln(-ln[1-Pr(Yi=1/xi])=xiβi
where Y is the likelihood of implementing a product innovation. While xi denotes each predictor
variable, βi represents the coefficient associated with each predictor variable. For a more intuitive
understanding, the estimated coefficients are exponentiated through the function: Pr(Yi=1/ xi)
=1-exp [-exp (xiβi)]. Just like the odds ratios in the logistic regression model, the results are then
interpreted as incidence ratios. For categorical variables, a risk ratio higher than one indicates that enterprises with that attribute have higher likelihood of implementing a product innovation than those in the reference category. The reverse is true if the ratio is less than one. The analysis in presented in Table 3.
Why logistical regression?
should be no outliers in data. *Large sample: Uses the maximum likelihood method, so a large sample size is required for logistic regression (Greenland et al., 2000).
3.6 Evaluation of Method Assumptions
To ensure that the regression estimates presented in the current study are fairly robust, diagnostics of the various assumptions underlying the standard logistic regression method were carried out. Overall, these diagnostics suggest that these assumptions were not violated and that the resulting estimates can be reasonable relied upon. The diagnostics of four of such assumptions are briefly discussed below.
Assumption of no outliers or influential data
1 2 34 5 6 7 8 9 10 11 12 13 14 15 16 17 18 1920 21 22232425 2627 28 29 303132 33 34 35 36 37 383940 41 42 43 44 4546 47 48 49 50 5152 5354 5556 57 58 59 6061 62 63 64 65 6667 68 69 7071 72 7374 75 76 77 78 79 8081 82 838485 86 87 88 89 90 91 92939495 96 97 98 99 100 101102 103 104 105 106 107 108 109 110 111 112 113114115 116 117 118 119120 121 122 123 124 125 126 127128129 130 131132133134135 136 137 138139 140 141142 143144 145 146147148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164165 166 167 168 169 170 171172 173 174 175 176 177 178 179 180 181 182 183 184185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203204205 206 207 208 209210 211 212213 214 215 216 217 218 219 220 221 222 223224 225 226 227 228 229 230 231 232 233 234 235 236 237 238239 240 241 242243244 245 246 247 248 249 250 251 252 253 254 255256 257 258 259 260 261 262 263264 265 266267268 269 270 271 272 273274 275 276 277 278 279 280 281 282 283284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320321 322323 324 325326 327 328 329 330 331 332 333 334 335 336 337 338 339340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369370 371 372 373 374375 376 377378 379 380 381 382383 384 385386 387 388 389390 391392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410411 412 413 414 415416417418 419 420 421 422 423 424 425 426427 428 429 430 431 432 433 434435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454455 456 457 458 459 460461 462463 464 465466 467 468469470 471 472 473 474475 476 477 478479 480 481 482 483 484 485 486 487 488 489 490 491492 493 494 495 496 497 498 499 500 501502503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521522523 524 525 526 527 528529530531 532 533 534535 536 537 538539 540541542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558559560 561 562 563 564 565 566 567 568 569570571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587588 589 590 591 592593 594 595 596597 598 599 600 601 602603 604 605 606 607608 609 610611 612 613 614 615 616 -4 -2 0 2 P e a rs o n r e s id u a l 0 200 400 600 Ids of Respondents
Figure 1: An Index Plot of Regression Residuals
The assumption of linearity in logistic Regression
Figure 2: Log odds of product innovation and quality of skilled labour Pro b a b ili ty i n n o v a ti o n 2 = 1 recode of q02_2
Observed P(innovation2=1) Predicted P(innovation2=1)
DK very goo
.5 .596774
Figure 2 suggests a positive linear relationship between the log odds of product innovation and quality of skilled labour. Despite the fact that a few of the observations are clearly outside of the regression line, it is fairly reasonable to suggest that the assumption of linearity has not been violated.
Multicollinearity
Normality
Unlike in the OLS regression model, satisfying the normality of errors assumption is not a necessary condition in logistic regression. In fact, in logistic regression, the distribution of the errors may be normal, poison or binomial. To check this assumption in the current study, the residuals were examined using a histogram and normal quantile plots (See Figures 3 and 4). These plots suggest that the residuals are fairly normal.
0 .2 .4 .6 .8 D e n s it y -4 -2 0 2 Pearson residual
-4 -2 0 2 4 P e a rs o n r e s id u a ls -4 -2 0 2 4 Normal Quantile
Figure 4: Normal Quantile Plot
4. Results
Table 1: Descriptive Statistics and Correlation Matrix of variables used in the analysis Mean SD 1 2 3 4 5 6 7 8 1 Innovation 0,5665584 0,4959529 1,0000 2 Green field (%) 0,5097406 0,5003114 -0,0845 1,0000 3 Joint Venture (%) 0,2613636 0,4397347 0,0282 -0.4957* 1,0000 4 Acquisition (%) 0,2564935 0,4370523 0,0861 -0.5022* -0,1209 1,0000 5 Hungary (%) 0,0925325 0,2900116 0,0306 -0.0230 -0,1389 0,1716 1,0000 6 Czech Republic(%) 0,3003247 0,4587717 -0,0058 0.0120 0,0133 -0,1172 -0,2092 1,0000 7 Poland(%) 0,3506494 0,4775611 0,0729 0.0605 0,2055* -0,1512 -0,2347 -0,4814 1,0000 8 Romania(%) 0,2077922 0,4060570 -0,1172 -0.0420 -0,1316 0,1481 -0,1635 -0,3355 -0,3764 1,0000 9 Slovakia(%) 0,0484013 0,2154178 0,0305 -0.0497 -0,0488 0,0744 -0,0723 -0,1482 -0,1663 -0,1159 10 Manufacturing (%) 0,5470779 0,4981833 0,0860 -0.0703 0,0217 0,1536 0,0092 -0,0940 -0,0217 0,1248 11 Service(%) 0,4529221 0,4981833 -0,0860 0.0703 -0,0217 -0,1536 -0,0092 0,0940 0,0217 -0,1248 12 Unskilled labour 3,7305190 1,0844550 0,0092 -0,0311 0,0081 0,0912 -0,0136 -0,1475 0,0948 0,0978 13 Skilled labour 3,6250000 0,7864451 0,0203 0,0238 -0,0123 -0,0130 -0,0116 -0,1065 0,0909 0,0000 14 Apprentices and Trainees 3,0584420 1,1934070 0,0319 -0,0037 -0,0323 0,0336 0,0783 -0,1301 0,0981 0,0085 15 Junior emp. With univ. degree 3,7759740 0,9753752 -0,0309 -0,0155 0,0230 0,0511 -0,0013 -0,0674 -0,0056 0,0849
Potential cooperation:
16 With univ. and public institution 3,4788960 1,1474440 0,0739 -0,0634 0,0287 0,0886 0,0279 -0,0482 0,0818 -0,0534 17 With other enterprises 3,8717530 0,8879212 0,0213 -0,0393 0,0568 0,0137 -0,0422 -0,0171 0,0871 -0,0703 18 Actual R&D cooperation(%) 0,2110390 0,4083776 0,3721* -0,0419 -0,0270 0,0880 0,1369 -0,0004 0,0118 -0,1276
Type of partner in R &D Coop.:
Table 1: Descriptive Statistics and Correlation Matrix of variables used in the analysis 9 10 11 12 13 14 15 16 17 18 19 1 Innovation 2 Green field (%) 3 Joint Venture (%) 4 Acquisition (%) 5 Hungary (%) 6 Czech Republic(%) 7 Poland(%) 8 Romania(%) 9 Slovakia(%) 1,0000 10 Manufacturing (%) -0,0062 1,0000 11 Service(%) 0,0062 -1,0000 1,0000 12 Unskilled labour -0,0621 0,1800 -0,1800 1,0000 13 Skilled labour 0,0408 -0,0067 0,0067 0,1406 1,0000
14 Apprentices and Trainees -0,0317 0,1102 -0,1102 0,3765 0,1291 1,0000
15 Junior emp. With univ. degree -0,0022 -0,0318 0,0318 0,2241 0,2973 0,2529 1,0000
Potential cooperation:
16 With univ. and public institution -0,0156 -0,0125 0,0125 0,1509 0,1651 0,2479 0,2762 1,0000
17 With other enterprises 0,0327 0,0302 -0,0302 0,0958 0,0870 0,1421 0,1245 0,2567 1,0000
18 Actual R&D cooperation(%) 0,0308 0,0630 -0,0630 0,0332 0,0240 0,0914 0,0168 0,1796 -0,0015 1,0000 Type of partner in R &D Coop.:
Table 1: Descriptive Statistics and Correlation Matrix of variables used in the analysis 20 21 22 23 24 25 1 Innovation 2 Green field (%) 3 Joint Venture (%) 4 Acquisition (%) 5 Hungary (%) 6 Czech Republic(%) 7 Poland(%) 8 Romania(%) 9 Slovakia(%) 10 Manufacturing (%) 11 Service(%) 12 Unskilled labour 13 Skilled labour
14 Apprentices and Trainees
15 Junior emp. With univ. degree
Potential cooperation:
16 With univ. and public institution
17 With other enterprises
18 Actual R&D cooperation(%)
Type of partner in R &D Coop.:
19 Headquarter/own enterprise(%) 20 Not with local suppliers(%) 1,0000 21 Not with foreign suppliers(%) 0,5647 1,0000 22 Not with local customers(%) 0,5101 0,4251 1,0000 23 Not with foreign customers(%) 0,3232 0,4366 0,4863 1,0000 24 With local research inst.(%) 0,4752 0,3474 0,3819 0,2357 1,0000
4.2 Logistical Regression
Table 2: Logistic regression models of product innovation Model 1 Model 2 Model 3 Model 4 Model 5
Variable β OR β OR β OR β OR β OR
Entry mode
Green field (No is reference) ---- ---- -0,342 0.710* ---- ---- ---- ---- -0,298 0,742 Joint Venture (No is reference) ---- ---- ---- ---- 0,130 1,139 ---- ---- -0,088 0,915 Acquisition (No is reference) ---- ---- ---- ---- ---- ---- 0.406* 1.500* 0,177 1,193
Country of origin Hungry (reference) ---- 1,000 ---- 1,000 Czech Republic 0,193 1,213 0,274 1,315 Poland 0,431 1,529 0,530 1,699 Romania -0,171 0,843 -0,151 0,860 Slovakia 0,399 1,490 0,386 1,472 Sector
Manufacturing (Service is reference) 0.368* 1.445* 0.340! 1.404!
Quality of human capital
Unskilled labour -0,026 0,974 -0,038 0,963
Skilled labour 0,006 1,006 0,012 1,012
Apprentices and Trainees -0,031 0,97 -0,025 0,975 Junior emp. With univ. degree 0,092 1,097 0,092 1,097
Potential cooperation
With univ. and public institution -0,022 0,978 -0,034 0,966
With other enterprises 0,035 1,036 0,029 1,029
Actual cooperation
Yes, in R &D cooperation 2.416** 11.203** 2.376** 10.764**
No (reference) ---- 1,000 ---- 1,000
No response/Don't know 0,693 1,999 0,577 1,780
Headquarter/own enterprise
No(reference) ---- 1,000 ---- 1,000
Yes -0,244 0,783 -0,203 0,816
Not with local suppliers
No(reference) ---- 1,000 ---- 1,000
Yes 1,101 3,007 1,078 2,94
Not with foreign suppliers
No(reference) ---- 1,000 ---- 1,000
Yes -0,673 0,51 -0,758 0,468
Not with local customers
No(reference) ---- 1,000 ---- 1,000
Yes 0,786 2,195 0,842 2,321
Not with foreign customers
No(reference) ---- 1,000 ---- 1,000
Yes 0,591 1,806 0,652 1,919
With local research inst.
No(reference) ---- 1,000 ---- 1,000
Yes 0,168 1,183 0,194 1,214
With foreign research inst.
No(reference) ---- 1,000 ---- 1,000
Yes -0,585 0,557 -0,594 0,552
Constant -0,757 ---- 0.444*** ---- 0.234* ---- 0,166 ---- -0,597 ----
Model fitness parameters
Log likelihood -363,014 -419,300 -421,259 -419,192 -360,964 Pseudo R-squared 0,1388 0,005 0,006 0,006 0,1436 Notes:
4.3 Complementary Log Log Model
Table 3: Commplementary log-log regression models of product innovation Model 1 Model 2 Model 3 Model 4 Model 5 Variable β Exp(β) β Exp(β) β Exp(β) β Exp(β) β Exp(β)
Entry mode
Green field (No is reference) ---- ----
-0,232 0.793* ---- ---- ---- ---- -0,228 0,796 Joint Venture (No is reference) ---- ---- ---- ---- 0,088 1,091 ---- ---- -0,044 0,957 Acquisition (No is reference) ---- ---- ---- ---- ---- ---- 0.268* 1.308* 0,094 1,099
Country of origin Hungry (reference) ---- 1,000 ---- 1,000 Czech Republic 0,111 1,117 0,158 1,171 Poland 0,277 1,320 0,340 1,405 Romania -0,130 0,878 -0,116 0,891 Slovakia 0,249 1,283 0,233 1,262 Sector
Manufacturing (Service is reference) 0.238! 1.268! 0.227! 1.255!
Quality of human capital
Unskilled labour -0,049 0,952 -0,054 0,947
Skilled labour 0,007 1,008 0,011 1,011
Apprentices and Trainees -0,023 0,977 -0,018 0,982 Junior emp. With univ. degree 0,053 1,055 0,055 1,056
Potential cooperation
With univ. and public institution -0,005 0,995 -0,016 0,984
With other enterprises -0,006 0,994 -0,008 0,992
Actual cooperation
Yes, in R &D cooperation 1.377*** 3.964*** 1.410*** 4.095***
No (reference) ---- 1,000 ---- 1,000
No response/Don't know 0,437 1,548 0,344 1,411
Headquarter/own enterprise
No(reference) ---- 1,000 ---- 1,000
Yes -0,194 0,823 -0,211 0,810
Not with local suppliers
No(reference) ---- 1,000 ---- 1,000
Yes 0,516 1,675 0,468 1,598
Not with foreign suppliers
No(reference) ---- 1,000 ---- 1,000
Yes -0,339 0,713 -0,427 0,653
Not with local customers
No(reference) ---- 1,000 ---- 1,000
Yes 0,169 1,184 0,246 1,279
Not with foreign customers
No(reference) ---- 1,000 ---- 1,000
Yes 0,429 1,536 0,464 1,590
With local research inst.
No(reference) ---- 1,000 ---- 1,000
Yes 0,027 1,028 0,013 1,013
With foreign research inst.
No(reference) ---- 1,000 ---- 1,000 Yes -0,226 0,797 -0,189 0,827 Constant -0,642 ---- -0,062 ---- -0.202** ---- -0.249*** ---- -0,533 ----
Model fitness parameters
Log likelihood -363,059 -419,300 -421,259 -419,192 -360,869 Notes: 1. Statistical significance: *p<0.05; **p<0.01; ***p<0.001 ! Significant at p<0.1
5. Discussion
The results presented in Model 1 suggest that the manufacturing industry and enterprises that participated in any R&D co-operation with other enterprises or organizations from 2007 to 2009 are significantly associated with initiating product innovation. Compared with the service sector, the manufacturing industry sector is 44% (OR=1.44) more likely to induce product innovation. More dramatically, Model 1 indicates that participation in R&D co-operation significantly increases product innovation by more than 10 times (OR=11.20) than non-participation. The effects of the other control variables on product innovation are only substantive. Although substantively important, these variables are not statistically significant predictors of product innovation. For example, Model 1 suggests that R&D co-operation with others other than local suppliers, having highly skilled labour and junior employees with university degrees are positively associated with product innovation but not statistically significant. Also, compared to Hungary, product innovation is substantively more likely to occur in countries such as the Czech Republic, Poland and Slovakia.
Other than the effect of the control variables, it was hypothesized that Greenfield based entry modes would limit subsidiary innovation. At the bivariate level, the logistic regression results in Model 2 (see Table 2) suggest that compared with other entry modes of foreign investment, a greenfield entry mode is 0.71 times or 29% less likely to implement any product innovation. The effect is statistically significant at the conventional alpha level of 0.05. The current study also hypothesized that joint venture based entry modes will foster subsidiary innovation. As shown in Table 2, the bivariate results in Model 3 suggest that joint venture based entry modes may, indeed, lead to product innovation but the effect is not statistically significant. Specifically, compared to other entry modes, joint venture based entry mode is about 1.14 times or 14% more likely to bring about product innovation.
relationship between an acquisition entry mode and product innovation. Thus, the hypothesis that acquisition based entry modes would limit subsidiary innovation cannot be validated with the current evidence. Rather, as Models 4 and 5 suggest, an acquisition entry mode induces product innovation although the statistical significance attenuated in the final model.
Substantively, the effect of joint venture entry mode in Model 5 is now inconsistent (OR=0.92) with the positive effect observed in Model 3 (OR=1.14). Statistically, we might describe this observation as spurious since the original relationship between a joint venture entry mode and product innovation completely disappeared with the introduction of the control variables in Model 5. On the other hand, although statistically not significant, the effect of greenfield entry mode in Model 5 is consistent with the bivariate results in Model 2. This suggests that the control variables moderate the effect of greenfield on product acquisition. Furthermore, consistent with the observation in Model 1, Model 5 indicates that participation in R&D co-operation with other enterprises and being in the manufacturing sector increase product innovation by 10.76 and 1.40 times respectively. As with the case in Model 1, all other control variables were not statistically significant in final model.
6. Conclusion
In answer to my research question I can say that contrary to expectations, an acquisition entry mode is associated with a higher likelihood of product innovation. Although this effect attenuated in the multivariate context, it is consistent with the bivariate results. Thus, there is some evidence to suggest that acquisition entry mode fosters product innovation and not otherwise. This leads me to the conclusion that in a Central and Eastern European context, the acquisition entry mode is better for innovation then argued in extant literature which is mostly based on Western economies or other emerging giants, like China. Furtermore this study did not give any conclusive evidence concerning the other investigated entry modes because statistical significance is not present. However, by increasing the sample size this evidence will most probably be met because LR is build upon the assumption of large sample sizes (Greenland, et al., 2000). This study, therefore serves as guiding framework for future research the measurement of entry modes and subsidiary innovation.
Though the effect sizes of the bivariate analysis decrease in their prediciting power in the multivariate analysis, I can conclude from this evidence that the control variables do matter and could be used in future research. Participation in R&D cooperation (significantly) appears as the most important variable affecting the entry mode - innovation relationship. Thereby in contradiction to the evidence developed by Lhuillerya and Pfister (2009), who argue that R&D collaborating firms had to abondan or delay their innovation projects due to difficulties in their partnerships. Their evidence was based on French CIS data, whereas this studies evidence is based on Central and Eastern European CIS data, suggesting that, although Central and Eastern European countries are part of the European Union, they are differences respect to their innovation inducement.
Furthermore, being in the manufacturing industry appeared as a significant predictor of innovation in a Central and Eastern European context. This suggests the focus of these countries is on this industry and acts as driving force of attracting manufacturing companies towards them.
Limitations and future research
studies focusing on CEE by adding a larger sample, future studies focusing on western economies interested in individually examining ownership based entry modes and innovation and future studies focusing on other emerging economies. By doing this a global picture on the innovation literature emerges and can be acted upon by multinationals and policy makers.
Focusing on the dependent variable, this study is furthermore purely based on hard defined data – yes or no with regards to innovation. A rather broad measure of performance. However, the variable does not measure the level or extent of innovation. The only knowledge is that it is the introduction of a good or service that is new or significantly improved with respect to its characteristics or intended uses. Future research would yield promising results when the dependent variable is measured at the level of innovation, i.e. intensity of innovation, so that more concise conclusions can be drawn from the sample.
Focusing on the independent variable, this study does not take regard of the cultural and contextual differences within the context of this study. Cultural characteristicsof the home country or home based organization are not included with the scope of this study. The measurement of the thechnology gap between home base and foreign entry mode is not included which could be a predictor as well, because the larger the technology gap, the larger the potential for technology spillovers.
References
Acemoglu, D., Johnson, S. and Robinson, J. (2001). 'The Colonial Origins of Comparative Development: an Empirical Investigation'. American Economic Review, 91, 1369– 1401.
Acemoglu, D., Johnson, S. and Robinson, J. (2002). 'Reversal of Fortune: Geography and Institutions in the Making of the Modern World Income Distribution'. Quarterly Journal of
Economics, 117, 1231–1294.
Acemoglu, D., Johnson, S., Robinson, J. and Yunyong, T. (2003). 'Institutional Causes, Macroeconomic Symptoms: Volatility, Crises and Growth'. Journal of Monetary Economics, 50, 49–123.
Aghion, P. and Griffith, R. (2005). 'Competition and Growth: Reconciling Theory and Evidence', Cambridge, MA: MIT Press, 2005.
Aghion, P., Blundell, R., Griffifth, R., Howitt, P. and Prantl, S. (2009). 'The Effects of Entry on Incumbent Innovation and Productivity'. Review of Economics & Statistics, 91, 20-32.
Ahsan, M. and Musteen, M. (2011). 'Multinational enterprises’ Entry Mode Strategies and Uncertainty: A Reviewn and Extension'. International Journal of Management Reviews, 13, 376–392.
Almeida, P. and Phene, A. (2004). 'Subsidiaries and Knowledge Creation: the Influence of the MNC and Host Country on Innovation'. Strategic Management Journal, 25, 847-864.
Anand, J. and Delios, A. (1997). 'Location Specificity and the Transfer of Downstream Assets to Foreign Subsidiaries'. Journal of International Business Studies, 28,579-604.
Anand, J. and Delios A. (2002). 'Absolute and Relative Resources as Determinants of International Acquisitions'. Strategic Management Journal, 23, 119–134.
Anderson, J. A. (1982). 'Logistic Discrimination'. In P. R. Krishnaiah and L. N. Kanal (Eds.), Classification, Pattern Recognition and Reduction of Dimensionality, Handbook of
Statistics, 2, 169-191. Amsterdam: North Holland.
Andersson, U., Forsgren, M. and Holm, U. (2007). 'Balancing Subsidiary Influence in the Federative MNC: A Business Network View'. Journal of International Business Studies, 38, 802–818.
Andersson, E. and Gatignon, H. (1986). 'Modes of Entry: a Transactions Cost Analysis and Propositions'. Journal of International Business Studies, 17, 1–26.
Agarwal, S. and Ramaswami, S. N. (1992). 'Choice of Foreign Market Entry Mode: Impact of Ownership, Location and Internalization Factors. Journal of International Business
Studies, 23, 1−27.
Asiedu, E., Esfahani, H. and Hadi, S. (2001). 'Ownership Structure in Foreign Direct Investment Projects'. Review of Economics & Statistics, 83, 647-662.
Bagozzi, R.P. and Yi, Y. (1991). ’Multitrait-Multimethod Matrices in consumer research. Journal of consumer research, 17, 426-439 - estimation methods
Barney, J. B. (1986). 'Organizational culture: Can it be a Source of Sustained Competitive Advantage?', Academy of Management Review, 11, 656-665.
Barney, J. (1991) Firm Resources and Sustained Competitive Advantage. Journal of
Management, 17, 99 -121.
Barkema, H.G. and Vermeulen, F. (1997). 'What Differences in the Cultural Backgrounds of Partners are Detrimental for International Joint Ventures?' Journal of International Business
Studies, 28, 845-864.
Barkema, H.G. and Vermeulen, F. (1998). 'International Expansion Through Start-up or Acquisition: a Learning Perspective. Academy of Management Journal, 4. 7–26.
Bartlett, C. A. (2006). 'Philips versus Matsushita: A New entury, a New Round, Harvard
Business Case, 9.
Bartlett, C. and Ghoshal, S. (1993). 'Managing Innovation in the Transnational Corporation'. In C. Bartlett, Y. Doz & G. Hedlund (Eds.) Managing the Global Firm, 8, 215-270.
Becker, W. and Dietz, J. (2004). 'R&D Cooperation and Innovation Activities of Frms - Evidence for the German Manufacturing Industry. Research Policy, 33 (2), 209–223.
Bhuamik, S.K. and Gelb, S. (2005). 'Determinants of Entry Mode Choice of MNCs in Emerging Markets Evidence from South Africa and Egypt'. Emerging Markets Finance and
Birkinshaw, J. and Hood, N. (1998). 'Multinational Corporate Evolution and Subsidiary Development'. London: Macmillan.
Bleeke, J. and Ernst, D. (1991). 'The Way to Win in Cross-border Alliances'. Harvard
Business Review, 69, 127- 35.
Bouquet, C. and Birkinshaw, J. (2008a). 'Weight Versus Voice: How Foreign Subsidiaries Gain Attention from Corporate Headquarters'. Academy of Management Journal, 51, 577–601.
Bouquet, C. and Birkinshaw, J. (2008b). 'Managing Power in the Multinational Corporation: How Low-power Actors Gain Influence. Journal of Management, 34, 477–508.
Brouthers, K. D. (2002). 'Institutional, Cultural and Transaction Cost Influences on Entry Mode Choice and Performance'. Journal of International Business Studies, 33, 203–221.
Brouthers, K.D. (2013a). 'Institutional, Cultural and Transaction Cost Influences on Entry Mode Choice and Performance'. Journal of International Business Studies, 44, 1-13.
Brouthers, K.D. (2013b). 'A Retrospective on: Institutional, Cultural and Transaction Cost Influences on Entry Mode Choice and Performance'. Journal of International Business
Studies, 44, 14–22.
Belderbos, R., Carree, M., Diederen, B., Lokshin, B. and Veugelers, R. (2004). 'Heterogeneity in R&D cCooperation Strategies'. International Journal of Industrial
Organization, in press.
Brouthers, K. D. and Hennart, J.F. (2007). Boundaries of the Frm: Insights From International Entry Mode Research. Journal of Management 33, 395–425.
Brouthers, K. D. and Brouthers, L.E. (2001) 'Explaining the National Cultural Distance Paradox' Journal of International Business Studies, 32, 177-189.
Brouthers, K. D. and Brouthers, L. E. (2000). 'Acquisition or Greenfield Start-up? Institutional, Cultural and Transaction Cost Influences'. Strategic Management Journal, 21, 89– 97.
Cantwell, J. (1989). 'Technological Innovation and Multinational Corporations'. Oxford:
