*We thank two anonymous referees, the editor (Anthony Di Benedetto), René Belderbos, and Pierre Mohnen for helpful comments on earlier drafts. The empirical analysis for this article has been performed using the remote access facility at Statistics Netherlands. The views expressed in this article are those of the authors and do not necessarily reflect the policies of Statistics Netherlands.
Determinants of Alliance Portfolio Complexity and Its Effect on Innovative Performance of Companies*
Article first published online: 6 APR 2011
© 2011 Product Development & Management Association
Journal of Product Innovation Management
Volume 28, Issue 4, pages 570–585, July 2011
How to Cite
Duysters, G. and Lokshin, B. (2011), Determinants of Alliance Portfolio Complexity and Its Effect on Innovative Performance of Companies. Journal of Product Innovation Management, 28: 570–585. doi: 10.1111/j.1540-5885.2011.00824.x
- Issue published online: 10 MAY 2011
- Article first published online: 6 APR 2011
- Top of page
- Discussion and Conclusions
- BIOGRAPHICAL SKETCHES
Alliance formation is often described as a mechanism used by firms to increase voluntary knowledge transfers. Access to external knowledge has been increasingly recognized as a main source of a firm's innovativeness. A phenomenon that has recently emerged is alliance portfolio complexity. In line with recent studies this article develops a measure of portfolio complexity in technology partnerships in terms of diversity of elements of the alliance portfolio with which a firm must interact. The analysis considers an alliance portfolio that includes different partnership types (competitor, customer, supplier, and university and research center). So far factors that determine portfolio complexity and its impact on technological performance of firms have remained largely unexplored. This article examines firms' decisions to form alliance portfolios of foreign and domestic partners by two groups of firms: innovators (firms that are successful in introducing new products to the market), and imitators (firms that are successful at introducing products which are not new to the market). This study also assesses a nonlinear impact of the portfolio complexity measure on firms' innovative performance.
The empirical models are estimated using data on more than 1800 firms from two consecutive Community Innovation Surveys conducted in 1998 and 2000 in the Netherlands. The results suggest that alliance portfolios of innovators are broader in terms of the different types of alliance partners as compared to those of imitators. This finding underlines the importance of establishing a “radar function” of links to various different partners in accessing novel information. Specifically, the results indicate that foremost innovators have a strong propensity to form portfolios consisting of international alliances. This underlines the importance of this type of partnership in the face of the growing internationalization of R&D and global technology sourcing. Being an innovator or imitator also increases the propensity to form a portfolio of domestic alliances, relative to non-innovators; but this propensity is not stronger for innovators. Innovators appear to derive benefit from both intensive (exploitative) and broad (explorative) use of external information sources. The former type of sourcing is more important for innovators, while the latter is more important for imitators. Finally, alliance complexity is found to have an inverse U-shape relationship to innovative performance. On the one hand, complexity facilitates learning and innovativeness; on the other hand, each organization has a certain management capacity to deal with complexity which sets limits on the amount of alliance portfolio complexity that can be managed within the firm. This clearly suggests that firms face a certain cognitive limit in terms of the degree of complexity they can handle. Despite the noted advantages of an increasing level of alliance portfolio complexity firms will at a certain stage reach a specific inflection point after which marginal costs of managing complexity are higher than the expected benefits from this increased complexity.
- Top of page
- Discussion and Conclusions
- BIOGRAPHICAL SKETCHES
Alliance formation is often described as a mechanism used by firms to increase voluntary knowledge transfers (Contractor and Lorange, 2002; Kogut, 1988; Mowery, Oxley and Silverman, 1996). Such access to external knowledge can increase a firm's innovativeness by exposing it to novel technologies, increasing its problem solving arsenal, and providing it with new solutions (e.g., Ahuja and Lampert, 2001; Vanhaverbeke, Beerkens, Duysters, and Gilsing, 2006).
Short-term technological alliances and other forms of cooperative R&D links are often described as being ideal sources of such novel incoming knowledge flows. The importance of learning from multiple and diverse ties carrying nonredundant information for innovative performance has been discussed in, e.g., Duysters and Vanhaverbeke (1996) and Belderbos, Carree, and Lokshin (2004). Recent contributions have argued that access to international knowledge flows is especially important for firms aiming to tap into leading-edge knowledge (Griffith, Harrison, and van Reenen, 2006). International R&D alliance formation forms an additional channel allowing access to international knowledge flows along with such well-studied channels as FDI, trade, foreign technology payments (Eaton and Kortum, 1996, 1999; Van Pottelsberghe de la Potterie and Lichtenberg, 2001) and mergers and acquisitions (Bresman, Birkinshaw, and Nobel, 1999; de Man and Duysters, 2005). Collaborating internationally seems to be a less risky alternative to mergers and acquisition as a start-up strategy in new markets. Furthermore, it can be useful for firms that aim to achieve access to local technological expertise and serves as a means to facilitate expansion into these markets. At the same time, it is argued that collaboration reduces risks associated with new product introduction (Barkema and Vermeulen, 1998; Tether, 2002). Recent findings therefore show that in international ventures, firms tend to prefer alliances over mergers and acquisitions (Vanhaverbeke, Duysters, and Noorderhaven, 2002).
A number of recent contributions have argued that by adopting a “portfolio approach” to external innovation information sources, firms put themselves in a better position to achieve and sustain innovation (e.g., Cohen and Malerba, 2001; Faems, Van Looy, and Debackere, 2005; Katila and Ahuja, 2002; Laursen and Salter, 2005; Leiponen and Helfat, 2005). The main contribution of this article to this strand of literature is to focus on an issue relatively unexplored in previous studies: the relationship between alliance portfolio complexity and innovative performance. Although from a practitioner's perspective alliance portfolio complexity has been acknowledged as an important and recurrent issue, there is hardly any academic work that addresses this issue in an empirical setting. Some notable exceptions have made a first attempt to link portfolio issues to alliance formation and performance (Faems et al., 2005; Lavie, 2007; Marino, Strandholm, Steensma, and Weaver, 2002; Powell, Koput, and Smith-Doerr, 1996). However, except for the study of Faems et al. (2005), these studies do not focus on innovativeness in relation to the specific nature of the alliance portfolio.
This research attempts to fill this gap in the literature. Although alliance portfolio complexity can come about in many different forms, this article considers two major aspects that determine alliance complexity: (1) The international and domestic scope of alliance portfolios and (2) the variety of different alliance types in the portfolio. In this sense our analysis takes into account a rich composition of types comprising the alliance portfolio; i.e., links with competitors, customers, suppliers, universities, and research institutes. Our attention is focused on those alliances in which innovative activities are part of the agreement.
This article argues that firms at the frontier of innovativeness seek to establish alliances that are more diverse compared to firms that are further away from the frontier. Furthermore, it is posed that more innovative firms will tend to have alliance portfolios skewed towards foreign partners allowing them to derive disproportionately higher benefits from international knowledge flows and to ensure faster adoption of new products in the foreign markets. This research is the first which deliberately tries to identify the relationship between alliance portfolio complexity and innovativeness. In this paper alliance portfolio complexity is defined as the degree to which alliance portfolios consist of different elements. In line with previous studies (Marino et al., 2002; Powell et al., 1996) our study therefore describes complexity in terms of diversity. In contrast to most previous empirical work relying on alliance press reports, the empirical analysis uses data collected through the Community Innovation Survey (CIS). This harmonized biannual survey is organized by Eurostat and is aimed at collecting information pertaining to both small and large firms' innovation activities.
- Top of page
- Discussion and Conclusions
- BIOGRAPHICAL SKETCHES
The propensity to form alliances can be seen to depend on both “inducements” and “opportunities” to form such linkages (Ahuja, 2000). Heterogeneity of these inducements and opportunities goes some way in explaining why participation in inter-firm collaborations is largely uneven among firms. Firms that have a history of innovativeness can be regarded as technically competent (Stuart, Hoang, and Hybels, 1999) and are desirable potential partners for cooperative innovative activities with other firms (Ahuja, 2000; Singh and Mitchell, 2005). High stocks of technical and commercial knowledge can, conversely, reduce the inducement to enter alliances with certain types of partners, e.g., when the risk of proprietary knowledge leakage is high (Ahuja, 2000). However, these concerns are primarily relevant for horizontal types of cooperation, while the propensity for vertical (customer, supplier) collaboration as well as collaboration with universities may be less affected and have been until now less explored. Following the studies of Stearns, Hoffman, and Heide (1987), Powell et al. (1996), and Marino et al. (2002), it can be said that diversity among the types of alliances increases the complexity of the alliance portfolio (competitors, customers, suppliers, universities, and research institutes). A portfolio approach to alliance strategies can allow a firm to broaden the pool of technological opportunities and to draw on ideas from multiple external sources, providing the firm with information advantages. In fact, Baum, Calabrese, and Silverman (2000) have argued that in terms of innovativeness the diversity of a firm's alliance portfolio is likely to be a more important factor than the sheer number of alliances. This assertion is confirmed in the study of Faems et al. (2005), which established that firms that apply a heterogeneous alliance portfolio are more innovative than other firms. Firms that are able to access a large and diverse stock of knowledge resources are said to be in “the thick of things” (Freeman, 1979, p. 219) and might benefit from information advantages over other companies. Access to a greater diversity of knowledge sources should lead to a higher rate of innovation. Ties to a broad number of different sources may also fulfill a “radar” function in the sense of providing a firm with information on a broad number of relevant technological developments (Ahuja, 2000; Freeman, 1991).
Although beneficial for innovators and imitators, the diversity of ties is expected to be more important for innovators than for imitators. Information from a wide diversity of sources is likely to contain novel information, which is more important for innovators than for imitators. Moreover, the establishment of a “radar” function seems to benefit innovators more than imitators because firms that search for a broader range of novel information and opportunities might be more successful in generating innovations (March, 1991). As a consequence, the benefits of portfolio diversity are expected to be higher for innovators than for imitators. This leads to the following hypothesis:
H1: Alliance portfolios of innovators are broader in terms of the different types of alliance partners as compared to those of imitators, and portfolios of imitators are broader compared to non-innovators.
Since the 1990s a sharp increase in industrial R&D expenditures has been accompanied by two major trends: the growing internationalization of R&D activities, and the spectacular rise in the number of international strategic R&D alliances (e.g., Archibugi and Iammarino, 2002; Doz and Hamel, 1998). These trends gave way to the fact that technological knowledge has become more and more dispersed over the world. The traditional monopoly of the United States in terms of technological know-how has gradually decreased in favor of new hubs of knowledge in Asia and Europe. Global technology sourcing has therefore become a key strategy for firms striving to benefit from geographically dispersed knowledge hubs (Cantwell and Harding, 1998; Florida, 1997; Zander, 1999). Or as stated by Meyer-Krahmer and Reger (1999, p. 33): “International enterprises that are leading performers of R&D are pursuing the strategy of a presence with R&D and product development at precisely those locations where there are the best conditions worldwide for innovation and the generation of knowledge in their product segment or field of technology.”
Access to qualified R&D personnel at foreign locations, adaptation to local needs, lower costs of R&D personnel, and improved access to external knowledge at scientific competence centers located abroad have been mentioned as the most important motives to source technology globally (Brockhoff, 1998; Shrader, 2001). In fact firms are increasingly deliberately seeking partnerships with unique centers of excellence in order to advance their technological knowledge. Empirical research suggests that firms that search for new technologies globally are motivated by the wish to gain highly sophisticated specialized knowledge from international sources (Meyer-Krahmer and Reger, 1999). By linking up with international leading centers of technology, firms get access to novel knowledge that they are unable to find locally. Initial evidence for this is provided by Subramaniam and Venkatraman (2001) who find evidence that international knowledge sourcing enhances firms' capability of producing successful new products for multiple international markets. Furthermore Subramaniam (2006) concludes in his study that cross-national collaboration is an effective channel of international knowledge transfer and integration leading to the development of new products. Innovative firms simply have to go beyond the frontiers of their local networks because searching for information in well-known domains (local search) significantly lowers their chances of finding new information. This creates a need to explore new sources of knowledge outside existing networks through establishing linkages outside their traditional (geographic) regions and networks. There is therefore a growing consensus in the literature that in order to be innovative, firms have to scan for new technologies globally (Narula and Hagedoorn, 1999). Over time international pockets of knowledge have become more distinct instead of similar (Narula, 1996; Narula and Hagedoorn, 1999). Imitators, on the other hand, tend to develop dense networks consisting of strong, local ties, which are characterized by frequent interaction among a select group of local partners. Those firms would rather replicate their existing ties within their local community than search for new ones (Gulati, 1995, 1998; Walker, Kogut, and Shan, 1997). Therefore this pattern of repeated alliance formation is based on a local search process that facilitates exploitation and mimics behavior among firms that eventually develop similar technological skills and know-how (Gilsing, Lemmens, and Duysters, 2007; Knoke and Kuklinski, 1982). Therefore local ties lower the chances of finding new information in the network. In contrast, international ties offer the advantage of providing access to heterogeneous sources of knowledge and information. Links with foreign partners have been shown to benefit especially innovative firms because they often allow access to frontier technological knowledge (e.g., Cincera et al., 2003; Florida, 1997). Neither type is preferred in all cases, as their effect is contingent on the specific innovation strategy that firms pursue (Rowley, Behrens, and Krackhardt, 2000). Innovators are likely to benefit more from international ties because of their ability to provide access to novel and heterogeneous knowledge (Meyer-Krahmer and Reger, 1999), whereas imitators will choose trust and intimacy of local ties over novelty value of these ties. Therefore we hypothesize:
H2: Alliance portfolios of innovators are more internationally oriented as compared to the portfolios of imitators.
Engagement in multiple cooperative agreements allows firms to exploit synergetic effects between these strategies—forming a new alliance in one type of R&D linkage can enhance the effectiveness of other existing R&D collaborations. Such a synergy, or complementarity, has been formally defined by Milgrom and Roberts (1990) and is assumed to exist if the implementation of one practice or strategy increases the marginal return on other practices. Belderbos, Carree, and Lokshin (2006) assess the performance effects of simultaneous engagement in R&D cooperation links with different partners and find that there are benefits of pursuing multiple cooperation strategies simultaneously, especially if the appropriability concerns are low. On the other hand, engaging in different types of R&D alliances, to the extent that they also relate to pursuit of different innovation objectives, will lead to an increase in managerial costs and complexity, resulting in inferior performance, especially in smaller firms. According to the organizational perspective on alliances, firms may find it optimal to separate horizontal and vertical partnerships as long as they entail profoundly different objectives (Gilsing, 2005; Rothaermel and Deeds, 2004). While firms pursue vertical integration in order to achieve a “bullwhip effect” in realizing cost efficiencies, to improve core competences and to ensure successful commercialization of existing products and services (Metters, 1997; Rosenzweig, Roth, and Dean, 2003), they tend to engage in strategic collaboration with competitors to develop new technologies for prospective markets (e.g., Miotti and Sachwald, 2003).
To summarize, engagement in multiple simultaneous cooperation links brings benefits in terms of access to a broader pool of technological opportunities and knowledge acquisition from multiple sources, allowing exploitation of synergetic effects. At the same time it also increases the complexity of the alliance portfolio and the associated management costs and appropriability concerns.
Compared to the management of individual alliances, managing a portfolio of alliances is even more challenging. In our case complexity refers to the international and domestic scope of the portfolio as well as to the variety of different alliance types. The more complex an alliance portfolio, the more management attention is required. Overall too high a degree of alliance portfolio complexity is expected to have a negative effect on portfolio performance because of the increased burden on managing the alliance portfolio (Hoang, 2001). In cases of extreme complexity, the cognitive limits of firms to deal with such a degree of complexity are quickly reached. Therefore rational and optimal decisions regarding the alliance portfolio are often not possible. In this case, bounded rationality (March, 1978; Simon, 1957) limits the ability of companies to come up with optimal solutions to their alliance portfolio challenges. This suggests that each firm has a certain cognitive limit in terms of the degree of complexity it can handle. Therefore this article argues that because of the noted advantages of an increasing level of alliance portfolio complexity firms perform better until a certain inflection point after which marginal costs of managing complexity are higher than the expected benefits from this increased complexity. This suggests the following hypothesis:
H3: Alliance complexity is inverse U-shaped related to innovativeness.
- Top of page
- Discussion and Conclusions
- BIOGRAPHICAL SKETCHES
Sample and Descriptive Statistics
To test the hypotheses we use data from two consecutive Community Innovation Surveys (CIS) conducted in 1998 and 2000 in the Netherlands, as well as information from the production statistics database in the same years. The biannual CIS surveys are conducted by Statistics Netherlands (CBS) on behalf of Eurostat and are aimed at collecting information pertaining to firms' innovative activities. The method and types of questions used in the survey are described in the OECD Oslo Manual (OECD, 1997). An additional advantage of the Dutch CIS surveys is that they have been held every other year rather than in four-year intervals as has been customary in other EU countries. The CIS surveys contain information regarding R&D and innovation activities of the firm, including innovation expenditures, innovation in partnership data, and sources of knowledge used in the innovation process. The CIS and production statistics surveys are sent to all large firms and to a random sample of smaller firms (ten and more employees) in the Netherlands. The questionnaire asks firms directly whether they were able to introduce an innovation and whether the innovation was new to the firm or new to the market. The definition of product innovation is based on the following questionnaire question: “During the period of 1998–2000 (1996–1998 in case of CIS2.5) did your enterprise introduce onto the market any new or significantly improved products?” The questionnaire further details that “Product innovation is a good or service which is either new or significantly improved with respect to its fundamental characteristics, technical specifications, intended uses or user friendliness.” In addition, the following survey question to construct the R&D alliance variables is used: “Did your enterprise have any (if yes, please indicate the type of organization) co-operation arrangements on innovation activities with some other enterprises or institutions in 1998–2000 (in case of CIS2.5 in 1996–1998)?” The reliability of the harmonized questionnaire has been tested by Eurostat and is used to collect official statistics on innovation activities of firms in the EU. To create a longitudinal data set, innovating firms in the 2000 survey are matched to those from the 1998 survey. Each firm is then linked via a unique ID number to the production statistics data.
Table 1 describes the construction of the variables that were used in the statistical analysis and their descriptive statistics. There are 334 firms (17.6% of the sample) with R&D alliance of any type among the firms in the sample. Partnerships with competitors, the focus of much of the alliance literature, are not the most frequently adopted strategy (7.2% of the cases). Alliances with suppliers are most frequent (11.6% of the cases), followed by alliances with customers (11.4%), and university cooperation (6.8%). Some 1561 observations have none of the four types of links. The comparison across industries indicates that the propensity to cooperate is higher in (petro)chemicals and metals. Firms in science-based industries such as electronics and chemicals, but also basic metal industries, report a relatively high share of university cooperation. Collaborating firms are characterized by a significantly higher share of new or improved products introduced to the market.
|All alliances portfolio||Sum of all partnership types a firm reports to be engaged in. The alliance types include international and domestic partnerships||0.57||1.68|
|Foreign alliances portfolio||Sum of partnership types a firm reports. The alliance types include international partnerships only||0.29||1.11|
|R&D intensity||Total innovation expenditures/sales||0.03||0.16|
|R&D intensity squared||Total innovation expenditures/sales squared||0.03||0.71|
|Firm size||Logarithm of number of employees||4.60||1.17|
|Breadth||Constructed as the sum of knowledge sources (customers, suppliers, universities, research centers, etc.) that a firm draws on||2.61||1.87|
|Depth||Constructed as the sum of scores on the knowledge sources that have been rated by firm as important or very important in its innovation activities||0.68||0.93|
|Innovation obstacles||1 if the firm has experienced bottlenecks in its innovation process related to lack of skills, knowledge, rigid organization structure, economic risks, or lack of financing||0.61||0.87|
|Secrecy protection measure||Importance of secrecy as protection of firm's innovations, measured on a 5-point Likert scale||0.14||0.34|
|Complexity protection measure||Importance of product or technological complexity as protection of firm's innovations, measured on a 5-point Likert scale||0.17||0.37|
|Lead time protection measure||Importance of lead time as protection of firm's innovations measured on a 5-point Likert scale||0.35||0.48|
|Legal protection measures||Importance of patents, copyrights, and trademarks as protection of firm's innovations, sum of scores, measured on a 5-point Likert scale||0.22||0.48|
|Foreign multinational||1 if the firm is an affiliate of a foreign multinational, else 0||0.15||0.36|
|Demand pull||Importance of demand-enhancing objectives for the firm's innovations. Constructed as a sum of scores on two categories of objectives, relating to product quality and new products and markets||2.26||0.92|
|Cost push||Importance of cost-saving objectives for the firm's innovations. Constructed as a sum of scores on two categories of objectives, relating to processes, labor, materials, and energy||3.40||2.27|
The analysis uses the variable that indicates the firm's success at introducing products new to the market (and not only to the firm). In our sample there are 534 firms (28% of firms in the sample) that have introduced new-to-the-market products. The average share of sales due to new-to-the-market products is about 5%. In addition, there are 518 firms that introduced new-to-the-firm products but no new-to-the-market products (imitators). Chemicals, machines, electronics, and basic metals industries are characterized by a relatively high share of innovators. Low-technology industries have generally a lower than average share of both innovators and imitators.
Variables and Measures
To test our hypotheses two models are estimated: a model explaining a firm's decisions to form and expand the links in its alliance portfolio and a performance model in which the complexity of its alliance portfolio is allowed to impact the innovative performance. Below it is explained how the variables used in the estimation are measured.
Dependent Variable in the Alliance Portfolio Equation. Our dependent variable in the portfolio formation model is the count of the number of elements in the alliance portfolio. The alliance types include international and domestic cooperation links with competitors, customers, suppliers, universities, and research centers. In each of these types cooperation can take place within national borders (i.e., with domestic partners), and with foreign partners (e.g., in the EU, in the new EU countries, in the United States, Japan, or elsewhere). For each type of alliance firms are asked to indicate whether their partner is located within the national borders or abroad. Adding elements within and across various types of partners produces an overall measure of the alliance portfolio with a maximum of 30 link-types (five types of partners with 6 sub-types for each). These are non-exclusive types of links and firms often report more than one link. In fact, in our sample the majority of firms have fewer than 6 link-types. A large share of firms in the sample report zero links. The CIS surveys do not measure the number of links within each sub-type. It is likely, however, that the number of types of links is highly correlated with the number of agreements within each category.
Independent Variables in the Alliance Portfolio Equation. Based on the review of the literature, this section discusses the specification of the empirical model and describes how the variables are constructed.
Focal variables: Indicators of firm's innovativeness innovators and imitators. In order to test H1 and H2, which predict that a firm's portfolio complexity is partly determined by the firm's position relative to the innovative frontier, our model includes two binary indicators. The CIS survey allows the classification of each firm as either an innovator or imitator. Our binary variables take the value one if a firm is an innovator or imitator, respectively, and else zero. A firm is classified as an imitator if in this period it introduced products new to the firm, but did not introduce products new to the market. A firm is classified as a product innovator if it introduced products new to the market (these firms could also have products new to the firm only). The non-innovator firms are used as a control group; however, none of the parameters are estimated for them. It is possible that in the group of non-innovators there are firms that have introduced process or organizational innovations. In this article the focus is on product innovations only.
Control independent variables. The model further includes a number of control variables, which previous literature on alliances and R&D cooperation found to be important determinants of a firm's decisions to engage in partnerships.
Intellectual property protection. Effectiveness of means available to a firm to protect their innovations and intellectual property from imitators has been found to have substantial influence on its decision to form external cooperative agreements (Ahuja, 2000; Cassiman and Veugelers, 2002). The analysis uses a direct measure of firms' perceptions about the effectiveness of various protection mechanisms it employs. In the survey firms are asked to rate the effectiveness of several methods of protecting own innovations. The analysis distinguishes between the following measures: legal methods of protection (through patents, brand names, or copyright); and strategic protection through (1) secrecy, (2) complexity of existing firm-specific processes, and (3) lead time. The responses pertaining to each of the categories are measured on a Likert scale ranging from not important to very important. For the legal protection methods a summation is performed over the corresponding responses. These responses are measured on the same scale. There is only one survey question for each of the strategic protection measures. Instead of summing them into one measure they are included as separate variables into the model to see the differentiated effects of various strategic measures of IPR protection. Arguably, these appropriability variables are subjective measures based on the management's beliefs about the firm's external environment. Other studies (e.g., Cassiman and Veugelers, 2002; Cohen and Levinthal, 1989; Levin, 1988) have shown that these firm-specific measures capture appropriability effects quite accurately, and influence management's decisions on whether to engage in cooperative R&D.
Foreign multinationals. Experience in establishing and managing partnerships has been shown to be positively correlated with the firms' ability to establish new alliance relationships (Dyer and Singh, 1998; Kale, Singh, and Perlmutter, 2000). Affiliates of multinational enterprises (MNEs) will have an advantage in such collaborative routines and accumulated experience relative to unallied firms. Experience of MNEs in developing foreign links is likely to be most valuable and these firms are expected to have larger portfolios of foreign alliances compared to unallied or domestic firms (Lavie and Miller, 2008). A control dummy for firms that are part of a multinational group is included.
Constraints to innovation. Another control included in the equation is a dummy that aims to capture factors hampering the innovation process. Partnerships can be an effective vehicle to help firms resolve these bottlenecks. A firm's propensity to form alliances has been shown to be stronger when managers anticipate ability to share costs and risks through collaborative relations (e.g., Tyler and Steensma, 1995). Belderbos et al. (2004) find that the existence of organizational rigidities can encourage firms to seek more active collaboration. The empirical specification includes a dummy variable that measures the hampering factors to the firm innovation activities—such as risk, cost, and lack of skills—that can create additional incentives for a firm to seek cooperation partners to overcome existing bottlenecks.
Strategy and firm objectives. The model also includes variables controlling for different objectives of firms' innovation strategies. The analysis distinguishes between a “cost push” objective, that is, capturing the importance of cost-saving objectives of the innovation process, and a “demand pull” objective, that is, capturing the demand-enhancing, product-oriented goals of the innovation.
Firm size. Prior research suggests that the size of companies also plays a role in propensity to form partnerships. Larger firms are more resourceful and may find it less problematic to handle multiple innovation objectives and management of multiple R&D collaborations (e.g., Belderbos et al., 2006; Harrigan, 1988). Therefore, control for firm size, measured as the logarithm of the number of employees, is also included.
Dependent variable in the innovative performance equation. In order to test our third hypothesis pertaining to the effect of portfolio complexity on firms' innovative performance, a model in which the dependent variable is a share of the new products introduced by a firm into the market over the past two years is estimated. Product newness can be defined based on the technological significance of the invention or based on the position in the market or relative to the user (Trajtenberg, 1990). The Community Innovation Survey (CIS) uses the latter approach.
Independent variables in the innovative performance equation. Focal variable: Alliance portfolio complexity. Measuring complexity of alliance portfolios is challenging, because there are many potential sources of such complexity. Our approach focuses on two aspects that determine the complexity of a firm's alliance portfolio: the diversity of types of partners and the number of elements the portfolio contains. For each firm in our sample a score measure is calculated, which gives the proportion of each link-type xij out of the total number of link-types. This score is computed by differentiating between foreign and domestic partners (given by index i=1, 2). The complexity of the alliance portfolio increases along each of the dimensions and is bounded on a unit interval. Rothaermel and Deeds (2006) find evidence that alliance portfolio complexity varies depending on the types of partners in it. Different values of θij can be used to give different weights to rij to allow for such differentiation between portfolios, for example, populated relatively more densely with foreign partners. Our measure of complexity is then expressed as . When θij are all unity, this measure is equivalent to the Blau's index of heterogeneity, which has been used in the alliance literature to measure portfolio diversity (e.g., Powell et al., 1996). The use of weights (which sum to one) increases variation in the data lost due to the dichotomous nature of our alliance variables. Our empirical results are not sensitive to the different values of weights.
Control independent variables in the innovative performance equation. R&D intensity. This equation includes the R&D expenditures variable as a control in the innovation output equation. A positive correlation between technological input and output is predicted in the literature (e.g., Pakes and Griliches, 1984). In line with the previous literature, our model includes the R&D intensity measure (which is preferable to simple R&D expenditure due to scale effects), and its square term. An inverted U-shape relationship between innovation input and output measures is expected, reflecting the decreasing returns on R&D (e.g., Acs and Isberg, 1991). The R&D intensity is measured as the R&D expenditures divided by total sales. These expenditures include, in addition to internal innovation expenditures, outlays on extramural R&D on contracts paid to other parties, such as research centers and expenditures on technology licenses. Hence, the intensity measure also controls for the impact of external technology acquisition. Increasing levels of R&D intensity up to a point will be closely correlated with absorptive capacity. Further increases may be less effective in expanding absorptive capacity due to diminishing scale economies. The R&D intensity is taken from a 1998 survey to allow a two-year lag with which innovation investments affect innovation output.
Knowledge sources. Several recent contributions have argued that firms can better achieve and sustain innovation by adopting a broad sourcing strategy to external innovation information channels (e.g., Faems et al., 2005; Katila and Ahuja, 2002; Laursen and Salter, 2005; Leiponen and Helfat, 2010). A higher degree of openness in external search channels allows firms to broaden the pool of technological opportunities and to draw on ideas from multiple external sources. According to March (1991), using both explorative and exploitative information scanning strategies is important. Access to a greater range of knowledge sources should lead to a higher rate of innovation. In particular, Laursen and Salter (2005) find that both breadth and depth of information sourcing can affect the probability of successful introduction of new products into the market. However, the importance of search depth rather than breadth increases with the degree of novelty of the innovation output of firms. Therefore, equation (2) includes two source-specific incoming spillover variables, “breadth” and “depth.” Following Laursen and Salter (2005), “breadth” is constructed as the number of knowledge sources that a firm draws on. “Breadth” captures the extensive dimension of the use of knowledge sources, while “depth” measures the intensive dimension of their usage. To construct “breadth,” for each firm in each year the sum is computed over all sources of knowledge that a firm reported it had used (the minimum is zero, the maximum is ten). The variables capturing sources of knowledge have a high degree of internal consistency (Cronbach's alpha coefficient around 0.75). The variable “depth” measuring the intensive use of knowledge sources is constructed in a similar fashion. Here the sum is computed over those sources that have been rated by firms as important or very important in their innovation activities.
Firm size and industry controls. Larger firms have more abundant resources and may find it less problematic to handle multiple innovation objectives (e.g., Belderbos et al., 2006; Cohen and Klepper, 1996). The analysis controls for firm size by including a logarithm of the number of employees. Further control variables include a set of two-digit industry dummies. Table 2 and Table 3 report correlations between the variables used in the alliance portfolio and innovation output equations, respectively.
|1||Portfolio, all links||1.00|
|2||Portfolio, foreign links||0.92||1.00|
|8||Protection lead time||0.20||0.17||0.33||0.07||0.09||0.21||0.27||1.00|
|3||Portfolio complexity squared||0.07||0.83||1.00|
The dependent variable in the equation explaining the composition of alliance portfolio is a count variable taking on discrete non-negative integer values, including zero. Use of a Poisson or negative binomial model is standard for such models. One complication is that the explanatory variable that is used as a proxy for the firm's position (innovator or imitator) relative to the innovation frontier may be endogenous. In such a case a use of a standard Poisson or negative binomial model will lead to biased estimates. The main reason for the endogeneity is a bidirectional relationship between alliances and innovation performance of firms. Firms that have a history of innovativeness are more attractive potential partners for other firms compared to less innovative firms. The most innovative firms have the biggest opportunities to form technological alliances compared to firms behind the innovation frontier, while their inducements to inter-firm link formation may vary depending on the type of potential partner (Ahuja, 2000). For this reason the study applies a negative binomial model with endogenous covariates, which in our case is a multinomial variable. This variable is coded to take the values one if a firm is a non-innovator, two if it is an imitator, and three if it is an innovator.
More formally, the model can be described with the following equations (Deb and Trivedi, 2006). Assuming a mixed multinomial logit distribution, the probability of being non-innovator, innovator, or imitator can then be expressed as:
Here z is a vector of explanatory exogenous variables with parameters αj to be estimated, and lij=(li1, li2, li3) is a latent variable incorporating one of the three firm's i innovation states. The expected outcome equation (number of elements in a portfolio) for firm i, i=1,.,N is formulated as:
where xi is a vector of exogenous variables, γi are the coefficients of the endogenous firm innovation type variable relative to the control group, and lij are the latent factors. The maximum likelihood function is given in Deb and Trivedi (2006) and can be estimated with simulated maximum likelihood methods.
The dependent variable in our second model, explaining firms' innovation performance as a function of the alliance complexity measure and other controls, is a censored variable taking on value on a unit interval. Symbolically:
where the dependent variable yi, the share of innovative sales is defined as
where wi is the vector of explanatory variables in the innovation performance equation (they are described in section 3.2), and β is the vector of parameters to be estimated. Prior to the estimation the dependent variable yi is passed through a logistic transformation, allowing us to treat it as a continuous variable in the estimation.
- Top of page
- Discussion and Conclusions
- BIOGRAPHICAL SKETCHES
Table 4 presents the regression results testing H1 and H2. H1 states that alliance portfolios of innovators are broader in terms of the different types of alliance partners as compared to those of imitators, and portfolios of imitators are broader compared to non-innovators. The results obtained in portfolio formation model provide clear support for H1.
|Alliance Portfolio, All Links||Alliance Portfolio, Foreign Links|
|Innovator||2.89*** (0.27)||3.08*** (0.41)|
|Imitator||1.74*** (0.35)||1.33*** (0.44)|
|Legal protection||0.11 (0.13)||0.18 (0.17)|
|Protection through secrecy||0.51*** (0.18)||0.78*** (0.23)|
|Protection through complexity||0.40** (0.17)||0.35* (0.21)|
|Protection through lead time||0.54*** (0.15)||0.60*** (0.20)|
|Firm size||0.31*** (0.06)||0.37*** (0.08)|
|Foreign group||0.37** (0.19)||0.45** (0.22)|
|Obstacles||0.20** (0.08)||0.20** (0.10)|
|Demand-oriented strategy||0.20** (0.09)||0.26** (0.12)|
|Cost-oriented strategy||−0.00 (0.03)||−0.02 (0.04)|
|Number of observations||1895||1895|
|Wald Chi2(64), p-value||657.01 (0.00)||677.08 (0.00)|
|F-test (p-value)||11.78 (0.00)||10.50 (0.00)|
First, the results show that both of the coefficients on the focal innovativeness dummy variables in the “all alliance links” equation are positive and significant for both imitators and innovators, suggesting that their respective portfolios are broader compared to those of non-innovators, which is the base group. Second, the magnitude of the coefficient is statistically significantly higher for innovators than for imitators. To test this formally, an F-test is used to check if this difference is significantly different from zero. The F-test value is 11.78 (p<0.01) and therefore it rejects the null that the difference in coefficients is zero. Additionally, a constrained model is estimated in which the equality between parameters for innovators and imitators is imposed. An LR test value of 81.44, computed as twice the difference between the log-likelihood values of the constrained and unconstrained models, decisively rejects the constrained model. Both the LR test and the F-test lead us to the same conclusion. This result suggests that the alliance portfolios of innovators are broader in range compared to imitators.
Most of our control variables in the alliance equations are significant. A baseline model was also estimated (including focal independent variables only) to check the stability of parameters of interest. While our key conclusions are not affected, the likelihood-ratio test rejects the constrained model in favor of a more comprehensive model. This model was also tested against a parsimonious specification in which only industry dummies and a control for firm size in addition to the focal variables were included. This restricted baseline model is rejected by the likelihood ratio test in favor of an extended model. The null of the equivalence between coefficients on innovator and imitator in the portfolio equation is rejected at 1% level in the restricted baseline model. (These results are available from the authors upon request.)
According to the results, the use of the intellectual property rights protection measures encourages broader alliance formation. The model distinguishes between legal methods of protection using patents, brand names, or copyright and strategic protection through secrecy, complexity of existing firm-specific processes, and the lead time on competitors. All of the strategic protection measures are statistically significant at the 1% level, while the legal measure is not different from zero.
Our overall obstacles to innovation measure is significant at 5%, suggesting that those firms that experience bottlenecks in their innovation processes have a higher propensity for alliance formation. The foreign group dummy is also positive and significant, indicating that foreign firms are more active in forming alliances compared to their domestic counterparts. Firm size is also positive and significant in the alliance equation. Larger firms are more likely to have the critical size and absorptive capacity required to engage in R&D cooperation and forming broader portfolios.
To test H2—whether each type of innovator has different propensities in forming international alliances—the model is reestimated, but now focusing on firms' portfolios including only foreign links, and using it as separate dependent variable. These results are reported in Table 4, column 2. Consistent with H2, the innovator variable is positive and highly statistically significant, while the magnitude of the coefficient on the imitator variable is less than half of that for the innovator in the foreign alliances equation. Both are statistically significant. This result indicates that relative to the base group (non-innovators), innovators tend to have broader foreign alliance portfolios and so do the imitators. Again a test for the difference in the respective coefficients is performed and rejected it at the 1% level (LR test value is 81.19 and F-test is 10.50, p<0.01). To check the robustness of our results, a domestic alliance equation is estimated separately (these results are not tabulated). While these results indicate that both innovator and imitator variables are positive and significant, when tested to see if this difference in coefficients is statistically significant the LR test could not reject that it is not at the conventional level of significance. This result points to the issue that innovators and imitators are different with respect to their international alliances but not to domestic alliances.
The impact of the control variables is similar between the “all” and “foreign alliance” equations. The appropriability protection mechanism through secrecy and lead time seems more important for the propensity to form foreign alliances. Foreign groups have a higher propensity to form foreign alliances compared to domestic firms. The relationship between size and propensity to form alliances is also positive in the “foreign” alliances model. Some authors (e.g., Narula and Hagedoorn, 1999) have argued that there might be a negative correlation between propensity to form foreign alliances and size because smaller firms compensate for their limited resources by actively engaging in foreign cooperation, while larger firms have preference for more permanent forms of entry, such as investment in the wholly owned R&D laboratories. This is not confirmed in our data. Our result is indicative that bigger firms, due to their resources, are able to manage larger and more complex alliance portfolios compared to smaller firms.
H3 predicts that alliance complexity, measured by the aspects pertaining to the diversity of links within the portfolio, has an inverted U-shape relationship to innovative performance. To test this hypothesis, a model of innovative performance is estimated in which the dependent variable is the logistic transformation of the share of sales due to the new products introduced by a firm. This model includes controls for R&D intensity, size of a firm, whether it is foreign or domestic, and the informational sourcing variables as well as industry dummies. Column 1 of Table 5 reports the results of a regression in which a simple count is used (and its square term) across types and links within type to proxy for the portfolio complexity. One advantage of this approach is that it allows computing the inflexion point in terms of the number of elements in a portfolio (cf. Rothaermel and Deeds, 2006). The results reveal that the linear term is positive and the quadratic term is negative and significant, suggesting a curvilinear relationship between complexity and performance. Using the estimated coefficients the inflection point can be computed, which is about 5 linkages.
|Innovative Performance and Alliance Complexity||Innovative Performance and Alliance Complexity|
|Alliance portfolio complexity||0.22*** (0.05)||2.96*** (0.90)|
|Alliance portfolio complexity squared||−0.02*** (0.00)||−2.94** (1.03)|
|Firm size||0.10*** (0.04)||0.10** (0.04)|
|R&D intensity||0.22*** (0.06)||0.23*** (0.06)|
|Foreign group||0.37** (0.15)||0.35** (0.15)|
|Breadth knowledge sourcing||0.09*** (0.03)||0.11*** (0.03)|
|Depth knowledge sourcing||0.17*** (0.06)||0.16*** (0.06)|
|Number of observations||2077||2077|
|Wald Chi2(64), p-value||13.21 (0.00)||12.99 (0.00)|
Column 2 in Table 5 presents the results when our complexity measure (constructed as explained in the section on “measures”) is used. The advantage of this measure over a simple count variable is that by taking both diversity in types and number into account it further allows assignment of different relative weights to different types of links. The disadvantage of the measure is that it does easily produce a prediction on the optimal number of elements in the portfolio and does not specify how to devise the weights for different types of partners. Based on our review of the literature a higher weight is given to foreign linkages, reflecting a situation in which a portfolio with a relatively higher share of foreign links is more difficult to manage than the one populated with predominantly domestic partners. Again, in our results the linear term is positive and the quadratic term is negative and significant, suggesting an inverted U-shape relationship between our complexity measure and performance. Experiments with assigning different weights to foreign and domestic alliances were performed (e.g., θf=0.5; 0.6; 0.7; 0.8), but the outcomes indicate that our empirical results are not very sensitive to the different values of weights. These results are available from the authors upon request.
Discussion and Conclusions
- Top of page
- Discussion and Conclusions
- BIOGRAPHICAL SKETCHES
Despite the vast and still growing attention devoted to strategic alliances in both theory and practice, the study of alliance portfolios is still in its infancy. This article examined the effect of alliance portfolio complexity on innovativeness. More particularly, it studied decisions of firms to form alliance portfolios of foreign and domestic partners by innovators (firms that are successful in introducing new products to the market) and imitators (firms that are successful at introducing new products that are not new to the market). In line with our first hypothesis the results indicate that alliance portfolios of innovators are broader in terms of the different types of alliance partners as compared to those of imitators. This finding can be seen as an extension of the findings of Faems et al. (2005) and points to the importance of establishing a “radar function” of links to various different partners in accessing novel information in a world which is dynamic and not very transparent. In order to be innovative firms should therefore try to develop non-redundant ties that provide access to new information (Burt, 1992).
Non-redundant information can also be found in international alliances. The growing internationalization of R&D activities and the spectacular rise in the number of international strategic R&D alliances has provided us with evidence that global technology sourcing is an important trend. This study found that foremost innovators have a strong propensity to form portfolios consisting of international alliances. Being an innovator or imitator also increases the propensity to form a portfolio of domestic alliances but this propensity is not stronger for innovators. Innovators appear to derive benefit from both intensive (exploitative) and broad (explorative) use of external information sources. The former sourcing is more important for innovators, while the latter for imitators. This points at the importance of access to non-local non-redundant ties to achieve access to novel information.
Finally, alliance complexity is found to have an inverse U-shape relationship to innovative performance. On the one hand, complexity facilitates learning and innovativeness; on the other hand the results suggest that each organization has a certain management capacity to deal with complexity and that this managerial capacity seems to set limits on the amount of alliance portfolio complexity that can be managed within firms. This clearly indicates that firms face a certain cognitive limit in terms of the degree of complexity they can handle. Therefore this article argues that despite the noted advantages of an increasing level of alliance portfolio complexity firms will at a certain stage reach a specific inflection point after which marginal costs of managing complexity are higher than the expected benefits from this increased complexity. Of course, the managerial capability of firms to handle complex alliance portfolios might differ from firm to firm. The nature of the data allowed us to focus on two main aspects that determine complexity: international and domestic scope of alliances and the variety of types in the portfolio. More research is needed to assess the specific alliance capabilities needed to support such complex portfolios (Duysters, Heimeriks, and Jurriens, 2004).
Overall, this study highlighted both the beneficial nature of portfolio diversity as well as its limits from an innovation performance perspective. This can be seen as an important step into expanding our so far limited knowledge of the effects of alliance portfolios on innovative performance.
Limitations and Future Research
Because of the infancy of the alliance portfolio research field there are many limitations that need to be addressed in future research. First of all, the international nature and the type of alliances are not the only dimensions of alliance portfolio complexity, and further study is needed to include additional aspects. In the still infant field of research into portfolio complexity most studies use a very limited conception of complexity. Moreover, studies do not tend to combine size and diversity. In terms of alliance diversity, future work could include portfolio characteristics such as governance structures. A distinction in terms of equity or non-equity could be appropriate. In a similar vein, a characterization of the modes in terms of strategic supplier relationships, cross-licensing, minority stakes, joint ventures, joint marketing, research consortia, etc., could also be very interesting. Furthermore, industry differences between a firm and its partners could add to the complexity of a portfolio, as could degrees of competition between partners. Therefore, the article asserts that portfolio complexity is a multifaceted issue that can be described along a number of dimensions. These dimensions are important candidates for future explorations into the area of alliance portfolio management.
Another interesting avenue for further research is the impact of structural properties of alliance networks on firm innovative output. Recent research has started to uncover the relationship between the structural features (such as embeddedness) of alliance networks firms' innovative performance. Firms that are embedded in networks comprising a wide range of partners will have greater innovative output, to a large extent because such alliance networks allow innovative firms to access diverse, heterogeneous knowledge which further increases innovation (Schilling and Phelps, 2007; Uzzi and Spiro, 2005). Further increases in inter-firm embeddedness will lead to “over-entrenchment,” and associated with it increase in redundancies and decreasing returns to the existing links and decreasing opportunities for new link formation (Burt, 1992; Duysters, Hagedoorn, and Lemmens, 2003; Uzzi, 1997) and eventually to the dissolution of over-embedded partnerships (Hagedoorn and Frankort, 2008). In short, too many alliances in a portfolio might lead to signs of “over-embeddedness.”
In this study a first attempt was made to open up the black box of alliance portfolio complexity by showing that both the international nature and the type of relationships in a portfolio can have a substantial impact on a company's innovative performance.
- Top of page
- Discussion and Conclusions
- BIOGRAPHICAL SKETCHES
- 1991. Innovation, firm size and corporate finances: An initial inquiry. Economics Letters 35 (3): 323–26. , and .
- 2000. The duality of collaboration: Inducements and opportunities in the formation of inter-firm linkages. Strategic Management Journal 21 (3): 317–43.
- 2001. Entrepreneurship in a large corporation: A longitudinal study of how established firms create breakthrough inventions. Strategic Management Journal 22 (6–7): 521–43. , and .
- 2002. The globalization of technological innovation: Definition and evidence. Review of International Political Economy 9 (1): 98–122. , and .
- 1998. International expansion through start-up or acquisition: A learning perspective. Academy of Management Journal 41 (1): 7–26. , and .
- 2000. ‘Don't go it alone: Alliance network composition and startups’ performance in Canadian biotechnology. Strategic Management Journal 21 (3): 267–94. , , and .
- 2004. Cooperative R&D and firm performance. Research Policy 33 (10): 1477–92. , , and .
- 2006. Complementarity in R&D cooperation strategies. Review of Industrial Organization 28 (4): 401–26. , , and .
- 1999. Knowledge transfer in international acquisitions. Journal of International Business Studies 30 (3): 439–62. , , and .
- 1998. Internationalization of research and development. Berlin: Springer.
- 1992. Structural holes: The social structure of competition. Cambridge, MA: Harvard University Press.
- 1998. The internationalization of German companies' R&D. National Institute Economic Review 40 (1): 99–115. , and .
- 2002. R&D cooperation and spillovers: Some empirical evidence from Belgium. American Economic Review 92 (4): 1169–84. , and .
- 2003. Productivity growth, R&D and the role of international collaborative agreements: Some evidence for Belgian manufacturing companies. Brussels Economic Journal 46 (3): 107–40. , , , , and .
- 1996. A reprise of size and R&D. Economic Journal 106 (437): 925–51. , and .
- 1989. Innovation and learning: The two faces of R&D. Economic Journal 99 (397): 569–96. , and .
- 2001. Is the tendency to variation a chief cause of progress? Industrial and Corporate Change 10 (3): 587–608. and .
- 2002. Cooperative strategies and alliances. Amsterdam: Elsevier. , and .
- 2006. Specification and simulated maximum estimation of a non-normal treatment-outcome model with selection: Application to health care utilization. Econometrics Journal 9 (2): 307–31. , and .
- 2005. Collaboration and innovation: A review of the effects of mergers, acquisitions and alliances on innovation. Technovation 25 (12): 1377–87. , and .
- 1998. Alliance advantage: The art of creating value through partnering. Boston: Harvard Business School Press. , and .
- 2003. The effect of alliance block membership on innovative performance. Revue d'Economie Industrielle 103 (1): 59–70. , , and .
- 2004. An integrated perspective on alliance management. Journal of Chain and Network Science 4 (2): 83–94. , , and .
- 1996. Strategic interactions in DRAM and RISC technology: A network approach. Scandinavian Journal of Management 12 (4): 437–61. , and .
- 1998. The relational view: Cooperative strategies and sources of inter-organizational competitive advantage. Academy of Management Review 23 (4): 660–79. , and .
- 1996. Trade in ideas: Patenting and productivity in the OECD. Journal of International Economics 40 (3/4): 251–78. , and .
- 1999. International technology diffusion: Theory and measurement. International Economic Review 40 (3): 537–70. , and .
- 2005. Inter-organizational collaboration and innovation: Towards a portfolio approach. Journal of Product Innovation Management 22 (3): 238–50. , , and .
- 1997. The globalization of R&D: Results of a survey of foreign-affiliated R&D laboratories in the USA. Research Policy 26 (1): 85–103.
- 1979. Centrality in social networks: Conceptual clarification. Social Networks 1 (3): 215–39.
- 1991. Networks of innovators: A synthesis of research issues. Research Policy 20 (5): 499–514.
- 2005. The dynamics of innovation and interfirm networks: Exploration, exploitation and co-evolution. Cheltenham, UK: Edward Elgar.
- 2007. Strategic alliance networks and innovation: A deterministic and voluntaristic view combined. Technology Analysis & Strategic Management 19 (2): 227–49. , , and .
- 2006. How special is the special relationship? Using the impact of US R&D spillovers on UK firms as a test of technology sourcing. American Economic Review 96 (5): 1859–75. , , and
- 1995. Does familiarity breed trust? The implications of repeated ties for contractual choice in alliances. Academy of Management Journal 38 (1): 85–112.
- 1998. Alliances and networks. Strategic Management Journal 19 (4): 293–317.
- 2008. The gloomy side of embeddedness: The effects of overembeddedness on inter-firm partnership formation. Advances in Strategic Management 25: 503–30. , and .
- 1988. Joint ventures and competitive strategy. Strategic Management Journal 9 (2): 141–58.
- 2001. The impact of organizational and alliance based complexity on the development of alliance capacity. INSAD Working paper series. .
- 2000. Learning and protection of proprietary assets in strategic alliances: Building relational capital. Strategic Management Journal 21 (3): 217–37. , , and .
- 2002. Something old, something new: A longitudinal study of search behavior and new product introduction. Academy of Management Journal 45 (6): 1183–94. , and .
- 1982. Network analysis. Beverly Hills, CA: Sage. , and .
- 1988. Joint ventures: Theoretical and empirical perspectives. Strategic Management Journal 9 (4): 319–32.
- 2005. Open for innovation: The role of openness in explaining innovation performance among U.K. manufacturing firms. Strategic Management Journal 27 (2): 131–50. , and .
- 2007. Alliance portfolios and firm performance: A study of value creation and appropriation in the U.S. software industry. Strategic Management Journal 28 (12): 1187–12.
- 2008. Alliance portfolio internationalization and firm performance. Organization Science 19 (4): 623–46. , and .
- 2010. Innovation objectives, knowledge sources and the benefits of breadth. Strategic Management Journal 31 (2): 224–36. , and .
- 1988. Appropriability, R&D spending and technological performance. American Economic Review 78 (2): 424–48.
- 1978. Bounded rationality, ambiguity, and the engineering of choice. The Bell Journal of Economics 9 (2): 587–608.
- 1991. Exploration and exploitation in organizational learning. Organization Science 2 (1): 71–87.
- 2002. Harnessing complexity: The moderating effect of national culture on entrepreneurial orientation and strategic alliance portfolio complexity. Entrepreneurship Theory and Practice 26 (4): 145–61. , , , and .
- 1997. Quantifying the bullwhip effect in supply chains. Journal of Operations Management 15 (2): 89–100.
- 1999. Internationalization of research and technology: Trends, issues and implications for S&T policies in Europe. Research Policy 28 (7): 751–76. , and .
- 1990. The economics of modern manufacturing: Technology, strategy and organization. American Economic Review 80 (3): 511–28. , and .
- 2003. Co-operative R&D: Why and with whom: An integrated framework of analysis. Research Policy 32 (8): 1481–99. , and .
- 1996. Strategic alliances and inter-firm knowledge transfer. Strategic Management Journal 17 (Winter): 77–92. , , and .
- 1996. Multinational investment and economic struture. London: Routledge.
- 1999. Innovating through strategic alliances: Moving towards international partnerships and contractual agreements. Technovation 19 (5): 283–94. , and .
- 1997. Revision of the high technology sector and product classification. Paris: OECD. .
- 1984. Patents and R&D at the firm level: A first look. In R&D, patents and productivity, ed. Z.Griliches, 55–72. Chicago: University of Chicago Press. , and .
- 1996. Inter-organizational collaboration and the locus of innovation: Networks of learning in biotechnology. Administrative Science Quarterly 41 (1): 116–45. , , and .
- 2003. The influence of an integration strategy on competitive capabilities and business performance: An exploratory study of consumer products manufacturers. Journal of Operations Management 21 (4): 437–56. , , and .
- 2004. Exploration and exploitation alliances in biotechnology: A system of new product development. Strategic Management Journal 25 (3): 201–21. , and .
- 2006. Alliance type, alliance experience and alliance management capability in high-technology ventures. Journal of Business Venturing 21 (4): 429–60. , and .
- 2000. Redundant governance structures: An analysis of structural and relational embeddedness in the steel and semiconductor industries. Strategic Management Journal 21 (3): 369–86. , , and .
- 2007. Inter-firm collaboration networks: The impact of large-scale network structure on firm innovation. Management Science 53 (7): 1113–26. , and .
- 2001. Collaboration and performance in foreign markets: The case of young, high-technology manufacturing firms. Academy of Management Journal 44 (1): 45–60.
- 1957. A behavioral model of rational choice. In Models of Man: Social and Rational: Mathematical Essays on Relational Human Behavior in a Social Setting. New York: Wiley.
- 2005. Growth dynamics: The bi-directional relationship between inter-firm collaboration and business sales in entrant and incumbent alliances. Strategic Management Journal 26 (6): 497–521. , and .
- 1987. Performance of commercial television stations as an outcome of interorganizational linkages and environmental conditions. Academy of Management Journal 30 (1): 71–90. , , and .
- 1999. Inter-organizational endorsements and the performance of entrepreneurial ventures. Administrative Science Quarterly 44 (2): 315–49. , , and .
- 2006. Integrating cross-border knowledge for transnational new product development. Journal of Product Innovation Management 23 (6): 541–55.
- 2001. Determinants of transnational new product development capability: Testing the influence of transferring and deploying tacit overseas knowledge. Strategic Management Journal 22 (4): 359–78. , and .
- 2002. Who co-operates on innovation and why: An empirical analysis. Research Policy 31 (6): 947–67.
- 1990. A penny for your quotes: Patent citations and the value of innovations. RAND Journal of Economics 21 (1): 172–87.
- 1995. Evaluating technological collaborative opportunities: A cognitive modeling perspective. Strategic Management Journal 16 (S1): 43–70. , and .
- 1997. Social structure and competition in interfirm networks: The paradox of embeddedness. Administrative Science Quarterly 42 (1): 35–67.
- 2005. Collaboration and creativity: The small world problem. American Journal of Sociology 111 (2): 447–504. , and .
- 2006. Explorative and exploitative learning strategies in technology-based alliance networks. Eindhoven, The Netherlands: Technische Universiteit Eindhoven. , , , and .
- 2002. External technology sourcing through alliances or acquisitions: An analysis of the application-specific integrated circuits industry. Organization Science 13 (6): 714–33. , , and .
- 2001. Does foreign direct investment transfer technology across the borders? Review of Economics and Statistics 83 (3): 490–97. , and
- 1997. Social capital, structural holes and the formation of an industry network. Organization Science 8 (2): 109–25. , , and .
- 1999. How do you mean “global”? An empirical investigation of innovation networks in the multinational corporation. Research Policy 28 (2–3): 195–213.
- Top of page
- Discussion and Conclusions
- BIOGRAPHICAL SKETCHES
Prof. Dr. Geert Duysters is the director of the Brabant Center of Entrepreneurship, a joint Center of Tilburg University and Eindhoven University of Technology. He is also a professor of Entrepreneurship at Eindhoven and Tilburg. His main research interests are in the field of strategic alliances, emerging economies (in particular India and China), innovation strategies, and internationalization strategies. He has published over 80 articles in international refereed journals and books such as the Journal of Product Innovation Management, Organization Science, Research Policy, Organisation Studies, Industrial and Corporate Change, and Journal of International Business Studies.
Dr. Boris Lokshin is assistant professor at Maastricht University, department of Organization and Strategy. His research interests include economics of innovation, inter-firm partnerships, and empirical I.O. His work has appeared in a variety of journals, including Journal of Management, International Journal of Industrial Organization, Research Policy, Review of Industrial Organization, Journal of Product Innovation Management and Oxford Bulletin of Economics and Statistics.