Master thesis MSC Business Administration
Strategic Innovation Management
The effects of R&D alliance portfolio characteristics on firms’ benefits from
external knowledge
Anja K. Everts
S3015289
Word count: 10028
University of Groningen
Faculty of Economics and Business
2
Abstract
Alliances have become important mechanisms to gain access to external knowledge that is used to develop capabilities for introducing new products. Alliances can thereby enhance firms’ financial performance. This study explores how characteristics of a firm’s R&D alliance portfolio influence its financial performance. We do so by aggregating R&D alliance portfolios into two sub-components: the portfolio’s internationalization and its concentration at corporate level. International alliances in the portfolio can enable the acquisition of more novel and non-redundant knowledge compared to domestic alliances. Moreover, the central position of a headquarter enables better distribution of external knowledge acquired from alliances across the firm,
compared to subunits. Therefore, internationalization and concentration of alliances at headquarter level are proposed to facilitate the appropriation of knowledge from alliances. Working with a panel dataset of 47 firms in the electronics industry over a time window of 15 years, evidence is found for the positive influence of R&D alliances on financial performance. Furthermore, results indicate that international alliances within a firm’s R&D portfolio have a positive and significant impact on firm performance, as well as alliances managed at corporate level. The implications for managers and future research are discussed.
3
Introduction
Firms obtain ideas for innovations from a wide range of sources. The extent to which firms are able to develop products from ideas depend on their ability to appropriate knowledge from their sources (Von Hippel, 1998; Cohen & Levinthal, 1990). A growing body of literature is focusing on alliances as mechanism for acquiring new external knowledge (Kim & Inkpen, 2005). However, it is not inevitable that forming alliances enables firms to internalize and use new knowledge. Recent literature suggests that a firm’s alliance capability and its absorptive capacity determine the extent to which firms differ in terms of learning from alliance partners (Lane & Lubatkin, 1998; Simonin, 1999). Firms’ reliance on basic R&D and their alliance experience also influence the learning effect (Cassiman & Veugelers, 2006; Rothaermel & Deeds, 2006). Moreover, variety in alliance performance is partially explained by differences between firm capabilities, interdependencies with internal technology investments and with the presence of an alliance department (Heimeriks, Duysters & Vanhaverbeke, 2007; Kale & Singh, 2007; Noseleit & De Faria, 2013; Wassmer, 2010; Wuyts & Dutta, 2014; Antonelli & Colombelli, 2015). Although there is much more known since the past few years, there remain unexplored alliance characteristics that can contribute to financial performance (George et al., 2001).
Alliances are often studied as individual events or transactions, with the consequence that synergistic effects of the portfolio are not taken into account (George et al., 2001). In order to refine our understanding of alliance portfolio characteristics that contribute to performance, this study explores coherent alliance
portfolios and suggests that two portfolio characteristics will have an influence on financial performance. The first portfolio characteristic is the internationalization of alliances in the portfolio. International alliances are agreements among firms in different country’s that enable the sharing or exchange of resources and hence, engage in co-development or provision of product, services and technologies (Gulati, 1998). Forming an international alliance increases coordination costs and complexity compared to domestic alliances (Hitt, Hoskisson & Kim, 1997). However, in terms of innovation, more differential contacts lead to lower levels of knowledge redundancy and more access to novel information which provides more entrepreneurial
opportunities (Gilsing et al., 2008; Hansen, 1999). Internationalization opens doors to novel knowledge of other environments that can be in short supply in a firm’s home country which can be useful to develop new products (Eisenhardt & Schoonhoven, 1996). As technology cycles become shorter and technological developments are rapidly evolving, high technology firms need to keep pace of the fast going developments (Oxley & Sampson, 2004). This partially explains the ongoing increase in international alliances as well as in R&D internationalization (Mowery, Oxley & Silverman, 1996; Oxley & Sampson, 2004), and raises the importance to study the implications of international alliances. The second portfolio characteristic is the concentration of alliances that are coordinated at headquarter or subunit level. Managing knowledge is an issue of coordination (Grant, 1996). The increase in internationalization also raises the complexity and costs of coordination due to communication difficulties (Chiesa, 1996). Studying alliance portfolios can provide insights in the coordination of alliances. When a firm manages alliances mostly at corporate level, this indicates that a firm’s technological knowledge base is more concentrated. Centralized governance can facilitate the creation and transfer of knowledge across units (Teece, 1982). However, it is not clear whether this advantage holds for alliances since subunit alliances can also benefit from being in closer proximity to their partners, thereby overcoming constraints from diversity.
To be able to answer a part of the question of why some firms benefit more from alliances than others, this study is deconstructing alliance portfolios from several firms in the electronics industry over a time window of ten years with additional financial information for 15 years. Building on previous literature and theories, mainly from the knowledge-based view, the characteristics of the R&D alliance portfolio in terms of
4
alliances are often formed to supply a firm with new technological knowledge. These alliances are particularly useful for shortening the learning cycle, accelerating product development and reducing R&D costs and risks (Hagedoorn, 1993). First, this study creates an extensive panel-database of a sample from firms in the electronics industry. Second, information on firm (R&D) alliances is extracted from the SDC database and complemented with financial firm information from Orbis. Then the firms’ R&D portfolios are decomposed in order to study the portfolio characteristics for all firms in the sample. Finally, fixed-effects regression analyses are performed to test the proposed relationships.
As managers increasingly tap into external knowledge sources for innovation activities, it is important to understand how internal factors and choices influence the process of transferring this knowledge to commercial ends (Monteiro & Birkinshaw, 2017; Szulanski, 1996). For this reason, it is important to understand which conditions must apply to gain benefits from external R&D sources. Furthermore, existing literature found inconsistent results for the assumption that access to external technological knowledge increases firm performance (Tsai & Wang, 2008). This study aims to fill in this gap by providing empirical evidence on the influence of R&D alliance portfolio characteristics on financial performance. By doing so, this research mainly contributes to the knowledge-based view. Theory from the knowledge-based view is addressed in order to explain the implications of knowledge transfer from access of external knowledge through alliances. Besides, this study contributes to literature on dynamic capabilities, an extension of the resource based view, as this theory is concerned with explaining firm heterogeneity. Moreover, the empirical perspective on this topic is important for managers of innovating firms in guiding them on how to organize R&D for efficient external knowledge appropriation and it is relevant in guiding managers on how to organize alliance portfolios.
5
Theoretical background and hypotheses development
This section describes theory of the knowledge based view that is concerned with knowledge and the context in which it is transferred. The theory is adopted to address the benefits and challenges of transferring
knowledge. Furthermore, the elements that are related to differences in performance are discussed, leading to development of the hypotheses.
Knowledge acquisition through alliances
Rapidly evolving business environments have focused the attention of literature towards resources and organizational capabilities as sources of sustainable competitive advantage (Grant, 1996b). Alliances can give firms access to different resources and especially to new knowledge (George et al., 2001). Teece (1998) was among the first researchers who mentioned that a firm’s ability to create, integrate, transfer and use
knowledge on an ongoing basis underpins the firms’ capabilities and competitive advantage (Easterby-Smith & Prieto, 2008). Theory about the knowledge-based view proposes that knowledge processing capabilities differ among firms. Since knowledge is created within individuals in the firm, it is considered to be a firm specific resource rather than a generic resource (Grant, 1996). When knowledge is transferred between entities, as in alliances, complexity arises from two properties of knowledge that are addressed in literature of the knowledge-based view (Nonaka & Takeuchi, 1995). First, knowledge resides in individuals as well as in systems, procedures or tools and therefore it can be difficult to integrate. Second, knowledge that is
considered to be tacit is hard to articulate, which makes it also difficult to share. Szulanski (1996) found that knowledge transfer within an organization can be impeded by the nature of knowledge, the sender and recipient - their experience, motivation etc.- and the organizational context (Szulanski, 1996). When knowledge is transferred between companies, the inter-firm context creates another dimension that can hamper knowledge exchange. Moreover, explicit knowledge is traditionally viewed as public good; once it is created, it can be used at zero costs by other users (Grant, 1996). Furthermore, knowledge is idiosyncratic as it is often restricted to a certain time and place. This all imposes several challenges and benefits related to knowledge transfer and integration. The challenges prohibit inimitability and increase complexity, therefore knowledge can lead to sustainable competitive advantage. When knowledge is sticky (e.g. tacit), or when capabilities based on sticky knowledge accumulate slowly over time, the acquisition of this knowledge is subject to path dependency (Mesquita, Anand & Brush, 2008). In this case, the acquired knowledge is more likely to result into sustainable advantages (Mesquita, Anand & Brush, 2008) as it is more difficult to imitate. Tacit knowledge, which is most difficult to exchange, is best revealed through its application (Kogut and Zander, 1992). Benefits of knowledge transfer are that knowledge accumulated in an alliance can also be useful input to other related projects, with little additional costs. Firms are effective institutions to capture benefits from knowledge spillovers by internalizing these spillovers across their divisions (Grant, 1996). However, this ability to assimilate and appropriate external knowledge to valuable innovations differs across firms.
6
Eisenhardt & Schoonhoven, 1996). Alliances have become important mechanisms to provide external knowledge that can be used to develop capabilities for introducing new products, thereby enhancing firms’ financial performance.
Prior studies have found that the ability to recognize new knowledge is related to a firms existing fields of expertise, also referred to as prior knowledge stock (Cohen & Levinthal, 1990; Lane, Salk & Lyles, 2001). The concept of absorptive capacity assumes that technological learning contributes to internal learning effects if the firm has a relevant knowledge base (Cohen & Levinthal, 1990). Learning involves the transfer and absorption from knowledge of partner firms in order to explore new knowledge areas (Dyer & Nobeoka 2000; Kale, Singh & Perlmutter, 2000). In this connection, Lane & Lubatkin (1998) argued that the ability of firms to learn from their partners is determined by the relative characteristics of both firms (Land & Lubatkin, 1998). Their findings show that not only a firms absorptive capacity determine firms’ ability to learn, but also the system in which the knowledge is processed, in other words, the organizational form (Lane & Lubatkin, 1998). Alliances are expected to enhance firm performance despite the difficulties due to properties of knowledge, and literature explains that these benefits depends on the fit between partners, firms ability to learn and their capabilities to organize their alliances (Argyres & Silverman, 2004; Cohen & Levinthal, 1990; Lane & Lubatkin, 1998). The following section will continue on these fundaments by reviewing the literature on R&D alliances, alliance organization and firm performance.
Challenges for R&D Alliances
Not surprisingly, in today’s fast-paced and knowledge-intensive environment, R&D alliances are a popular means to acquire and leverage technological capabilities (Oxley & Sampson, 2004). For a long time, the popularity of R&D alliances has increased (Lavie & Miller, 2008; Pearce & Singh, 1992). Many firms are establishing alliances to obtain new skills, enter new markets and share risks and resources (Inkpen &
Beamish, 1997). In technology-based industries, firms tend to form alliances with the motivation to learn from a partner and to gain access to assets (Zhang, Baden-Fuller & Mangematin, 2007). Technological alliances, also referred to as R&D alliances, are characterized by risk, uncertainty and a potentially high payoff (Dunning, 1995). R&D alliances enable the sharing and transfer of technological capabilities among organizations (Luo & Deng, 2009; Sampson, 2007; Lin et al., 2012). Furthermore, R&D alliances have the characteristic that technologies or knowhow can be shared in order to develop and co-create new insights from combined knowledge (Lin et al., 2012). Hence, R&D alliances are seen as learning platforms to acquire external knowledge (Lin et al., 2012).
7
the more investment in internal R&D is needed in order to reap the benefits (Lin et al., 2012). Lavie & Miller (2008) focused on alliance portfolio composition. The study shows effects of the whole alliance portfolio composition within one industry and found no effects for technology alliances. This would imply that technology alliances are not having an impact on firm performance. Although firms expect positive effects from alliances on firm performance, empirical research found mixed or non-significant results for this relationship with respect to R&D alliances.
Since it is known that diverse sources of knowledge are needed in order to innovate and remain competitive, it is argued that having more access to external knowledge sources is beneficial for firm performance (Lavie & Miller, 2008; Pearce & Singh, 1992). Therefore, firms that are having a larger R&D alliance portfolio have more access to external knowledge that can be integrated and transformed into valuable knowledge, which in turn could increase the ability to gain competitive advantage which increases the financial performance over time. Hence, the first hypothesis states that:
H1: There is a positive relation between R&D alliances and firms financial performance.
Over the past few years, the internationalization of R&D and amount of inter-organizational linkages in research have increased, together with the growth of computer-communication networks (Asakawa 2001; Chiesa, 1996). R&D internationalization has complications for management since it offers new challenges with respect to the coordination of activities (Von Zedtwitz, Gassmann & Boutellier, 2004). Besides, from a learning perspective, partners need to adjust to both organizational and cultural differences of their
international alliance partners (Barkema, Bell & Pennings, 1996). Proximity to other members is viewed as facilitating condition in knowledge transfer (Kagami, 2006). On the cognitive dimension, similar cultures, values and common goals are also seen as facilitators, meaning that differences instead might hamper
knowledge transfer between partners (Kagami, 2006). Besides, it is argued that the transfer of tacit knowledge is more complex if partners differ in cultural context (Kagami, 2006). Costs for coordination are assumed to increase for firms with an international alliance portfolio (Hitt, Hoskison & Kim, 1997). Firms’ R&D investments can increase further during the alliance with a foreign partner, since it can require customization of technologies in accordance with local preferences or standards (Lavie & Miller, 2008).
Despite these complexities and additional coordination costs, international alliances can open the doors to resources that are in short supply in a firm’s home country (Eisenhardt and Schoonhoven, 1996). As Cantwell (1994) mentioned, skills and technology that are underlying global competitiveness are often embedded in different countries (Cantwell, 1994). The established R&D structure of an international company is considered to be related to the geographical dispersion of key knowledge sources (Chiesa, 1996). Since national
environments have idiosyncratic knowledge bases, the diversity in dispersed knowledge bases provides broader learning opportunities (Kim & Inkpen, 2005). Establishing cross-border alliances (e.g. international alliances) is a means to gain access to technological capabilities of a partner and to its technological
environment (Kim & Inkpen, 2005) which enhances the focal firms’ existing knowledge stock (Faems et al., 2010; Wassmer, 2010). Besides, it is argued in literature that cross-border R&D alliances are often established with a long-term strategic intent, because of the tradeoff of complexity and high potential learning results (Hagedoorn & Narula, 1996).
8
by suggesting that cross-border alliances yield to higher benefits than domestic alliances. Building on this knowledge, this study tests if learning effects from international R&D alliances also occur in another
innovative industry and attempts to enlarge our understanding of international R&D alliances, by testing if the learning effect is large enough to have an influence on firm performance.
In short, it is argued that knowledge from different countries provide broader learning opportunities and thus can enhance a firm’s existing knowledge base. Since knowledge can be in short supply in a specific country, it is expected that international R&D alliances add valuable knowledge to a firm’s portfolio on the longer term. Therefore, this study suggests a positive influence from international R&D alliances on firm performance in terms of financial results.
H2: International R&D alliances in the portfolio positively influence financial performance.
Since organizational context is viewed as determinant of a firms’ ability to learn (Lane & Lubatkin, 1998), the extent to which the R&D organization is centralized or decentralized is expected to influence the benefits from alliances (Zhang et al., 2007). Referring to the capabilities to organize alliances, there are new challenges in organizing R&D since internationalization of R&D has increased. Traditionally, structuring R&D was viewed as a trade-off between factors to determine whether the coordination of R&D activities as a whole could better be centralized or decentralized (Chiesa, 1996). Multinational companies nowadays cannot simply choose between centralized or decentralized R&D anymore. They have to create a structure which can encompass both the benefits from centralization on generic research and from decentralized research on specific research (Von Zedtwitz et al., 2004). For each alliance, a firm can choose whether to coordinate this alliance on the corporate level or at divisional level. The centrality of R&D organization structure determines to which extend the technological knowledge base is concentrated at the corporate level or dispersed at divisional level (Argyres, 1996; Argyres & Silverman, 2004; Chacar & Lieberman, 2003; Zhang et al., 2007). Therefore, the share of alliances that are coordinated at the corporate level indicates to which extend the technological knowledge base is concentrated in the firm.
Argyres and Silverman (2004) found that firms with a centralized R&D organization structure generate innovations with a higher level of impact on a broader range of technological areas, compared to firms with decentralized R&D (Argyres & Silverman, 2004). This study classified centralized R&D organization as structure where large R&D labs were present at corporate level located in headquarters, while a decentralized organization has a relatively small central lab and more R&D labs in different units. As a headquarters function is described as one of control and coordination, it is often pointed out that headquarters are better able to coordinate projects that involve more than one business unit (Goerzen, 2005). Furthermore,
headquarters have a higher level of point centrality (Ghoshal & Barlett, 1990; Asakawa, 2001). Teece (1982) already argued that centralized governance can facilitate the creation and transfer of knowledge across units (Teece, 1982). Centralizing research can increase a firm’s architectural knowledge to understand how
components in the system interact (Henderson & Clark, 1990). A decentralized knowledge base is extensively developed within business units or divisions, while a centralized knowledge base is characterized by strong planning and control on corporate level (Zhang et al., 2007).
9
centralization than from decentralization, while specific research on incremental technologies benefit more from decentralized coordination. A firm will benefit more from economies of scope if broad R&D projects are managed at corporate level, since more resources need to be allocated and risks are higher, but potential benefits are also expected to be more significant and can be useful to other projects. Economies of scope can be acquired if the entity involved in transfer can oversee the other areas to which new knowledge can contribute to. As economies of scope are related to research expenditures (Henderson & Cockburn, 1996), a firm would also benefit in financial terms from broad corporate R&D projects on the longer term, since research expenditures can be allocated to other research fields. Therefore, it is expected that a firm with more R&D alliances managed at corporate level, will have more broad and ambitious R&D projects compared to firms that manage most R&D alliances at subunit level. Besides, a higher share of alliances managed at corporate level indicates that a firms technological knowledge base is more concentrated, indicating that the firm is better able to reap benefits from centralization, hence they will benefit more in terms of financial firm performance. Hence the third hypothesis proposes that:
H3: Concentration of the alliance portfolio at corporate level positively influences financial performance.
Methodology
This section explains how data is collected and which measures are used. Furthermore, this section provides information on the gathered data and sample, as well as on the performed statistical analyses.
Data and sample
This research paper attempts to address a business phenomenon faced by many companies, which has not been fully addressed by literature, albeit in general literature streams are already quite elaborated. In this case, theory testing is the most appropriate approach for the methodology according to Van Aken, Berends & Van der Bij (2012). When firms establish alliances, the most critical decisions are made at firm-level while other day-to-day decisions have to be taken at the alliance level. The differences in performance from these established alliances depend on the whole portfolio of alliances that a firm has as well as on firms ability to manage these (Inkpen, 1998). Studying only firm-level or dyad-level factors does not provide insights into the whole context that coordinate communication at these levels (Kim & Inkpen, 2005). Several researchers have therefore argued that portfolio analysis might provide more insights into firm capabilities and performance since it can take path dependency into account and the accumulation of skills and resources (Christensen & Snyder, 1997; Lin et al., 2012).
To develop a sample with alliances from comparable industries, all companies with US SIC code 36XX are isolated. This code includes all companies that have their core business in electronics and other electrical equipment and components. The industry is appropriate for this research because it is identified as an industry with a high likelihood to engage in cooperation (Noseleit & de Faria, 2013), which is important when studying commercial results from external knowledge sharing. Furthermore, this industry is characterized by its
continuously large R&D investments as firms in this industry have to stay at the frontier (Grindley & Teece, 1997). To gain a sample of firms that are likely to have a large alliance portfolio, a selection is made of the hundred largest companies. This is preferred because these firms are more likely to have multiple geographic units.
In order to construct a database for testing the hypotheses, data is collected from multiple sources. For
10
gathered, data on alliances are collected via the firm-level Securities Data Corporation’s (SDC) Platinum database. This database contains information about announced formal collaborations between companies that are established within a certain year. It is the most comprehensive database containing alliance information (Sampson, 2007). For every year from 2002 until 2012, all alliances are selected for the companies from the industry sample to construct alliance portfolio’s. The SDC database also offers insights in the type of alliance that is established and details about the partner firm(s). This information is used to create firm-level variables, amongst which alliance portfolio size and the share of joint ventures formed. Alliances are only labelled as R&D alliance when firms explicitly mention that joint research is going to be conducted, technology is going to be exchanged, or if an agreement is established between the parties to use each other’s knowledge or license. Therefore, conducting this research only on R&D alliances increases the likelihood to measure external knowledge exchange. As this study implies to gain better understanding of firm-level characteristics that influence the extent to which knowledge can be absorbed, several firm-level indicators of R&D alliances from the SDC database are collected and added to the database. This has resulted in extended information about the R&D alliances; the ratio of R&D alliances carried out by headquarters versus subunits and the number of alliances that are established abroad. If firm-level information is not published either in Orbis or the SDC database, then extensive research through annual reports is carried out to complement the database. The financial information is collected from 2002 to 2016 to be able to measure the longer term effects from the alliances. This effort is made because measuring more results increases the reliability of results (Van Aken et al. 2012; Lavie & Miller, 2008). The initial sample contains alliance portfolios over a ten year time
window, for the 60 largest companies in the electronics industry with financial information for fifteen years. The sample contains firms from 11 different countries and results in information about 415 R&D alliances.
Measures
Dependent variables
Firm financial performance will be measured by computing each firm’s return on assets (ROAt+1), following previous studies in this research field (Hitt, Hoskisson & Kim, 1997; Lavie & Miller, 2008). The ROA measures firm profitability in terms of generating revenues from their assets and is, in this case, measured with firm’s net income per period. The ROA is found in Orbis or extracted from annual reports or SEC filings for each firm over fifteen years, to complete the database with historical financial data as well as with most recent data.
Independent variables
The R&D alliance portfolio will be measured by all R&D alliances that each firm in the sample has formed during a ten year period. Alliances in the SDC database are only classified as R&D alliance when this is explicitly stated. However, if alliance announcements do not specifically describe to conduct joint R&D, there can be still an agreement to exchange technology. In the database, this is shown with a label ‘technology transfer’. Besides, when a firm announced in its agreement that licenses are exchanged and they are not labelled as manufacturing or marketing license, it is considered to exchange knowledge related to technology. These criteria are taken into account to calculate the amount of R&D alliances per year for each firm in the sample.
To test the second hypotheses and assess the internationalization of the portfolio, the variable international R&D alliances is constructed. This information is extracted from the SDC database, where alliances are labelled as same nation alliance or cross-border alliance. The same nation alliances are classified as domestic alliances. The alliance description given in the SDC database offers information on an additional
11
the concentration of alliances managed at the headquarter or subunit. It is expected that the effects of the studied hypotheses can be seen on the longer term. An appropriate time frame to measure results from alliances is five years (Lavie, 2007; Noseleit & de Faria, 2013). The independent variables are therefore measured over a five year time window.
Control variables
Control variables are constructed on firm- and alliance portfolio-level. First of all, larger firms spend proportional more on R&D and have more resources to divide. Therefore, most studies control for firm size (Cohen & Klepper, 1996). The log number of employees is used as proxy for firm size. Second, R&D expenses differ among firms in a way that higher R&D investments are related to higher innovation performance (Lane & Lubatkin, 1998). Information from Orbis is used to calculate R&D intensity for each year by looking at R&D expenses divided by firm revenue. The number of non-R&D alliances that a firm is holding in its portfolio might influence the amount of resources that a firm has available to govern all its alliances and hence it is a characteristic of the total alliance portfolio. Therefore, the total number of alliances established every year is also measured. The results of this variable are taken over a five year time window to compare it with the other alliance portfolio characteristics. Furthermore, it is known that joint ventures are an in-between governance mode that increase the momentum for closer collaboration. Therefore, this study controls for the share of joint ventures within the R&D alliance portfolio, in order to control for governance structure (Jiang, Tao & Santoro, 2010). Lastly, year dummies are included for each year to control for unexpected variations from events related to time that might influence the results, also called time-series trends (Sampson, 2007).
Analysis
The created dataset enables analyzing changes in firm performance over time, which is suitable for finding differences between companies since it accounts for individual heterogeneity. It is expected that variation in the return on assets is caused by firm-level characteristics and alliance portfolio characteristics. In order to assess the effect of these characteristics on the financial performance, a fixed effects linear regression analysis is applied to the analyses. This model accounts for firm specific, time-invariant characteristics that remain unobserved.
Results
This section presents the results from the performed analyses. First, the descriptive statistics are discussed and then the regression results.
Descriptive statistics
12
alliances will be kept in the equation, although one has to bear in mind that this might have consequences for interpretation of the results. This will be further discussed in the discussion section. The four independent variables, the subcomponents of the R&D portfolio, also show high correlations with the variable R&D alliances. This might be because of the same reasons and therefore, the variable R&D alliances will not be included in the regressions to test the second and third hypotheses to avoid the risk of multicollinearity. Similarly, in the regression for hypothesis 2 the subcomponents to test hypothesis 3 will not be included and vice versa.
The control variable total alliances turns out to have a correlation with the variable international alliances and domestic alliances. This can be expected with respect to the previous result, as both alliances are constructed from the R&D alliance portfolio. However, the mean VIF value for hypothesis 2 is 2.26 and the highest VIF value is 4.82 as can be seen in table 3, therefore the VIF tests suggests that multicollinearity is not an issue. Furthermore, total alliances and subunit alliances show a correlation of 0.823. The highest VIF value for the model for hypothesis 3, as shown in table 4, is 5.73 and the mean VIF value is 2.61, also indicating that multicollinearity is not an issue here.
Table 1. Descriptive statistics and correlations
Variables Mean S.D. 1 2 3 4 5 6 7 8 9 10
ROA(t+1) 7.167 14.07 1.000
R&D alliances 3.476 5.64 0.101 1.000
Firm size (log) 10.993 0.94 -0.072 0.409 1.000
R&D intensity 6.944 6.52 0.087 0.126 -0.191 1.000 Total alliances 12.936 23.01 0.028 0.883 0.428 0.031 1.000 Share of JV’s 0.338 0.763 -0.027 0.420 0.175 0.093 0.427 1.000 HQ-alliances 2.317 3.733 0.192 0.872 0.259 0.289 0.673 0.017 1.000 Sub. alliances 1.160 2.999 -0.054 0.793 0.438 -0.127 0.823 -0.021 0.394 1.000 Int. alliances 2.533 4.419 0.043 0.950 0.387 0.109 0.795 0.417 0.817 0.768 1.000 Dom. alliances 0.911 1.94 0.197 0.710 0.280 0.128 0.718 0.271 0.670 0.509 0.451 1.000
Table 2. VIF values for hypothesis 1
Variable VIF 1/VIF
R&D alliances 4.90 0.204 Firm size (log) 1.38 0.723 R&D Intensity 1.11 0.898 Total alliances 5.06 0.197
Share of JV’s 1.20 0.835
Mean VIF 2.73
Table 3. VIF values for hypothesis 2
Variable VIF 1/VIF
Cross-border alliances 3.05 0.328 Domestic alliances 2.14 0.468
Firm size (log) 1.35 0.739
R&D Intensity 1.16 0.862
Total alliances 4.82 0.208
Share of JV’s 1.06 0.948
13 Table 4. VIF values for hypothesis 3
Variable VIF 1/VIF
Headquarter alliances 2.39 0.419 Sub-unit alliances 3.84 0.261
Firm size (log) 1.42 0.703
R&D Intensity 1.25 0.799
Total alliances 5.73 0.175
Share of JV’s 1.04 0.961
Mean VIF 2.61
The final sample consists of 286 observations for 47 firms. Data is collected for 60 firms from the top 100 largest firms, however, since there are a few firms without data on alliances, these firms are automatically not included in the regression models. In the whole sample, the total alliance portfolio size ranges from 0 to 53 alliances and the amount of R&D alliances within the total alliance portfolio is 7 at the maximum. The firms within the sample are headquartered in diverse locations; 15 firms are headquartered in the US, 15 in Japan, 13 in China and the remaining 16 in the European Union, showing that the industry is quite dispersed. When looking at the 47 firms that formed alliances during the researched time window, there are 14 firms left from the United States, 13 from japan, 9 from China and only 9 from the European Union. This information shows that most of the largest firms in this industry are forming R&D alliances, with an exception of the European Union; here almost half of the firms are not accessing external knowledge through R&D alliances.
Regression results
Table 5 represents the results from the fixed effects regression analyses. A robustness check is performed within the regressions to control for multicollinearity. Model 1 includes only control variables. Within this model, the results show no significant values for the control variables. Model 2 represents all control variables and adds the independent variable R&D alliances portfolio, to test if the expected relationship exists between having an R&D alliance portfolio and higher ROA (hypothesis 1). The coefficient of R&D alliance portfolio is positive and highly significant with b= 1.303 and p= 0.013. Furthermore, the total number of alliances also shows a significant effect on ROA. However, this effect is slightly negative, as b= -0.255 and p= 0.080. This negative coefficient is notable, since prior research indicated a positive relation between alliance portfolio size and firm performance (Baum, Calabrese & Silverman, 2000). Besides, the F-test for this model shows
significance, therefore, hypothesis 1 is confirmed. Model 3 introduces the variables international alliances and domestic alliances within the R&D alliance portfolio, to be able to see the differences from
14
research on R&D intensity and performance, as it is assumed that R&D intensity is not solely enough to enhance firm performance (Lin, Lee & Hung, 2006). The coefficient of control variable total alliances differs, this will be discussed further in the discussion section.
Table 5. Fixed effects regression with robustness check results Dependent variable: ROA t+1
Variables Model 1 (Only controls) Model 2 (Hypothesis 1) Model 3 (Hypothesis 2) Model 4 (Hypothesis 3)
R&D alliances portfolio 1.303** (0.506) International alliances 1.353** (0.598) Domestic alliances 0.952 (0.652) Corporate level (HQ) alliances 1.567** (0.585) Sub-unit alliances 0.696 (0.542) Firm size (log) -5.972
(6.526) -7.069 (6.273) -7.330 (6.209) -6.708 (6.350) R&D intensity -0.529 (0.399) -0.490 (0.376) -0.494 (0.376) -0.488 (0.376) Total alliance portfolio 0.140
(0.131) -0.255* (0.143) -0.227 (0.148) -0.263* (0.132) Share of JV’s 7.774 (7.271) 6.916 (7.478) 6.843 (7.566) 6.683 (7.414)
Year dummies included included included included
Constant 77.986 (71.301) 91.099 (68.447) 93.664 (67.445) 87.075 (69.392) N 286 286 284 286 R² 0.1624 0.1856 0.1862 0.1902 F 4.04*** 5.01*** 4.76*** 4.85*** *< 0.10, ** < 0.05, *** < 0.01
Discussion
The discussion section presents a brief summary of the findings as well as the theoretical and managerial implications. Moreover, the limitations of this research are discussed to direct future research.
Research implications
15
it is widely assumed in literature and in practice that R&D alliances yield to positive results on firm performance, empirical evidence was incongruent. Finding this result points to the importance of R&D alliances as a means to access and apply knowledge from external sources. Additionally, two R&D portfolio characteristics are studied. An ongoing increase of international alliances and internationalization of R&D increases the importance to study the implications for managing knowledge, as this also increases complexity in managing and coordinating alliance activities. The second hypothesis of this study found that international alliances within the portfolio are positive and significant related to financial performance. Domestic alliances are also positively impacting financial performance, but this effect is not significant. This result indicates that within the portfolio, the international R&D alliances are primarily responsible for the increase in financial performance. Furthermore, the result of the third hypothesis suggests that the concentration of alliances within the firm matter, as it is found that concentration of alliances at headquarter level contributes significant to financial performance. These findings contribute to refinement of our understanding of alliance portfolio characteristics and their contribution to firm performance.
Establishing an R&D alliance is subject to challenges and might therefore not always be considered to be an appropriate choice. Not only because of uncertainty and risks, but also because of negative knowledge spillovers and difficulties of knowledge transfer due to properties of knowledge and context. Managers need to balance the risks against the potential returns of R&D alliances, and therefore it is important to find empirical evidence in order to guide these practices. This study found that R&D alliances are correlated with higher results in return on assets in the studied sample, suggesting that the benefits of R&D alliances do outweigh certain risks, at least in the high-technology electronics industry. This study furthermore suggests a portfolio approach to manage alliance activities. As firms increasingly tap into external knowledge sources (Lundvall, 1992), it is important to understand which efforts yield to the expected advantages. By showing that internationalization and concentration of alliances managed at the corporate level have a significant impact on financial performance, this study emphasizes the importance of alliance portfolio characteristics for alliance portfolio management. As the relations between these characteristics and financial performance were not well understood, this study is contributing to the existing literature with empirical evidence of the
relations.
Notably, prior research found a positive impact from alliance portfolio size on firm performance (Baum, Calabrese & Silverman, 2000), while this study found a slightly negative effect from total alliance portfolio size. Although this effect was not found in all the models, it might indicate that in high-technology industries, such as the electronics industry, increasing the alliance portfolio size might not always yield to better firm performance. Further research might examine this further. Additionally, in order to increase our understanding of alliance portfolio characteristics, future research might elaborate on these findings by examining the
interaction results of internationalization and concentration of the alliance portfolio. So far, theories are not extensive in explaining appropriate context conditions. However, these results provide arguments to suggest that there could be a combination between portfolio characteristics that yields the best outcome under certain environmental conditions. Managerial implications of this study can be extended by searching for an optimal combination between these and other portfolio characteristics and by finding out under which conditions advantages of such organization holds.
Limitations
16
other aspects, further research should examine the effects on different industries to enable a more reliable interpretation of the results. Furthermore, the variables for the alliance portfolio characteristics are constructed from the dependent variable. Therefore, including all the variables in one model would lead to unreliable effects due to multicollinearity. With this in mind, the variables for the models are carefully selected in this study. The tests for multicollinearity indicate that the control variable total alliance portfolio size is correlated with a few variables for R&D portfolio characteristics. Although the VIF values indicate no issues, the results should be taken with caution due to the indication of multicollinearity from the Pearson’s correlation. With regard to the alliance data, there is no alliance data found for 13 firms in the sample. Due to time limitation, not all alliance announcements are cross-validated with other sources. Albeit the SDC database is considered to be one of the most comprehensive sources that provide alliance information, it might not provide complete coverage because firms are not required to disclose their alliance activities (Anand & Khanna, 2000;
Sampson, 2007). Finally, the R-squared (within) of this study indicates a value of 0.1902 for the fourth model. Further research could include more control variables or portfolio characteristics in order to explain more of the heterogeneity between firms and use a larger sample size for larger reliability.
Conclusion
It is long argued that adapting and upgrading firm capabilities is necessary to sustain competitive advantages and therefore, firms need continuous acquirement of knowledge (Mesquita, Anand & Brush, 2008). Empirical findings of this study suggest that R&D alliances have a positive impact on firms’ financial performance. This result is in line with expectations from prior literature, suggesting that R&D alliances are characterized by a high potential payoff (Dunning, 1995) as these alliances enable the firm to directly apply new external knowledge (Kogut & Zander, 1992). Prior studies also emphasizes the downsides of R&D alliances as risks and uncertainty tend to be very high (Dunning, 1995). Despite these challenges, R&D alliances are
increasingly used as means to access external knowledge. Researchers found prove for the positive relation of R&D alliances on firm innovation, while results on firm performance are remaining mixed or insignificant. Contributing to the current research stream on R&D alliances, this study shows the positive relation on financial performance. In addition, this study highlights that portfolio characteristics can partially explain the heterogeneity among firms. Albeit prior research has widely studied similarity between alliance partners and firm level absorptive capacity, our understanding about alliance characteristics still calls for refinement. From a sample of 47 firms in the electronics industry, it is found that international R&D alliances within the
portfolio have a positive and significant impact on financial performance. Logically this effect might be different in other industries where knowledge is more or less clustered. However, for this industry, establishing international R&D alliances contributes more to firm financial performance compared to domestic R&D alliances. Moreover, the influence of concentration of R&D alliances formed by headquarter of subunit locations is studied. As a positive significant effect is found between concentration of R&D
17
References
Anand, B. N., & Khanna, T. (2000). Do firms learn to create value? The case of alliances. Strategic management journal, 295-315.
Antonelli, C., & Colombelli, A. (2015). External and internal knowledge in the knowledge generation function. Industry and Innovation, 22(4), 273-298.
Argyres, N. (1996). Evidence on the role of firm capabilities in vertical integration decisions. Strategic Management Journal, 129-150.
Argyres, N. S., & Silverman, B. S. (2004). R&D, organization structure, and the development of corporate technological knowledge. Strategic Management Journal, 25(8‐9), 929-958.
Asakawa, K. (2001). Organizational tension in international R&D management: the case of Japanese firms. Research Policy, 30(5), 735-757.
Barkema, H. G., Bell, J. H., & Pennings, J. M. (1996). Foreign entry, cultural barriers, and learning. Strategic management journal, 151-166.
Baum, J. A., & Oliver, C. (1991). Institutional linkages and organizational mortality. Administrative science quarterly, 187-218.
Baum, J. A., Calabrese, T., & Silverman, B. S. (2000). Don't go it alone: Alliance network composition and startups' performance in Canadian biotechnology. Strategic management journal, 267-294.
Cantwell, J. (1994). The theory of technological competence and its application to international production. Transnational Corporations and Innovatory Activities, 17, 107.
Cassiman, B., & Veugelers, R. (2006). In search of complementarity in innovation strategy: Internal R&D and external knowledge acquisition. Management science, 52(1), 68-82.
Chacar, A. S., & Lieberman, M. B. (2003). Organizing for technological innovation in the US pharmaceutical industry. In Geography and Strategy (pp. 317-340), Emerald Group Publishing Limited.
Chiesa, V. (1996). Managing the internationalization of R&D activities. IEEE Transactions on Engineering Management, 43(1), 7-23.
Christensen, T. J., & Snyder, J. (1997). Progressive research on degenerate alliances. American Political Science Review, 91(4), 919-922.
Cohen, W. M., & Klepper, S. (1996). A reprise of size and R & D. The Economic Journal, 925-951.
Cohen, W. M., & Levinthal, D. A. (1990). Absorptive capacity: A new perspective on learning and innovation. Administrative science quarterly, 128-152.
Dunning, J.H. (1995). Reappraising the eclectic paradigm in an age of alliance capitalism. Journal of International Business Studies, 26, 461–492.
18
Easterby‐Smith, M., & Prieto, I. M. (2008). Dynamic capabilities and knowledge management: an integrative role for learning? British Journal of Management, 19(3), 235-249.
Eisenhardt, K. M., & Santos, F. M. (2000). Knowledge-based view: a new theory of strategy? Forthcoming in Handbook of Strategy and Management. Pettigrew A, Thomas H, Whittington R.
Eisenhardt, K. M., & Schoonhoven, C. B. (1996). Resource-based view of strategic alliance formation: Strategic and social effects in entrepreneurial firms. organization Science, 7(2), 136-150.
Faems, D., De Visser, M., Andries, P., & Van Looy, B. (2010). Technology alliance portfolios and financial performance: value‐enhancing and cost‐increasing effects of open innovation. Journal of Product Innovation Management, 27(6), 785-796.
George, G., Zahra, S. A., Wheatley, K. K., & Khan, R. (2001). The effects of alliance portfolio characteristics and absorptive capacity on performance: A study of biotechnology firms. The Journal of High Technology Management Research, 12(2), 205-226.
Ghoshal, S., & Bartlett, C. A. (1990). The multinational corporation as an interorganizational network. Academy of management review, 15(4), 603-626.
Gilsing, V., Nooteboom, B., Vanhaverbeke, W., Duysters, G., & van den Oord, A. (2008). Network
embeddedness and the exploration of novel technologies: Technological distance, betweenness centrality and density. Research policy, 37(10), 1717-1731.
Goerzen, A. (2005). Managing alliance networks: Emerging practices of multinational corporations. The Academy of Management Executive, 19(2), 94-107.
Grant, R. M. (1996). Toward a knowledge‐based theory of the firm. Strategic management journal, 17(S2), 109-122.
Grant, R. M. (1996b). Prospering in dynamically-competitive environments: Organizational capability as knowledge integration. Organization science, 7(4), 375-387.
Grindley, P. C., & Teece, D. J. (1997). Managing intellectual capital: licensing and cross-licensing in semiconductors and electronics. California management review, 39(2), 8-41.
Gulati, R. (1998). Alliances and networks. Strategic management journal, 19(4), 293-317.
Hagedoorn, J. (1993). Understanding the rationale of strategic technology partnering: Nterorganizational modes of cooperation and sectoral differences. Strategic management journal, 14(5), 371-385.
Hagedoorn, J., & Narula, R. (1996). Choosing organizational modes of strategic technology partnering: international and sectoral differences. Journal of international business studies, 27(2), 265-284.
Hagedoorn, J., & Schakenraad, J. (1994). The effect of strategic technology alliances on company performance. Strategic management journal, 15(4), 291-309.
19
Heimeriks, K. H., Duysters, G., & Vanhaverbeke, W. (2007). Learning mechanisms and differential performance in alliance portfolios. Strategic organization, 5(4), 373-408.
Henderson, R. M., & Clark, K. B. (1990). Architectural innovation: The reconfiguration of existing product technologies and the failure of established firms. Administrative science quarterly, 9-30.
Henderson, R., & Cockburn, I. (1996). Scale, scope, and spillovers: the determinants of research productivity in drug discovery. The Rand journal of economics, 32-59.
Hitt, M. A., Hoskisson, R. E., & Kim, H. (1997). International diversification: Effects on innovation and firm performance in product-diversified firms. Academy of Management journal, 40(4), 767-798.
Inkpen, A. C. (1998). Learning and knowledge acquisition through international strategic alliances. The Academy of Management Executive, 12(4), 69-80.
Inkpen, A. C., & Beamish, P. W. (1997). Knowledge, bargaining power, and the instability of international joint ventures. Academy of management review, 22(1), 177-202.
Jiang, R. J., Tao, Q. T., & Santoro, M. D. (2010). Alliance portfolio diversity and firm performance. Strategic Management Journal, 31(10), 1136-1144.
Kagami, Y. (2006). In search of alliance capability-as conditions facilitating knowledge creation in creative industry. Keio Communication Review, 28, 87-105.
Kale, P., & Singh, H. (2007). Building firm capabilities through learning: the role of the alliance learning process in alliance capability and firm‐level alliance success. Strategic Management Journal, 28(10), 981-1000.
Kale, P., Singh, H., & Perlmutter, H. (2000). Learning and protection of proprietary assets in strategic alliances: Building relational capital. Strategic management journal, 217-237.
Kim, C. S., & Inkpen, A. C. (2005). Cross-border R&D alliances, absorptive capacity and technology learning. Journal of International Management, 11(3), 313-329.
Kogut, B., & Zander, U. (1992). Knowledge of the firm, combinative capabilities, and the replication of technology. Organization science, 3(3), 383-397.
Lane, P.J., Lubatkin, M. (1998). Relative absorptive capacity and interorganizational learning, Strategic Management Journal 19, 461–477.
Lane, P. J., Salk, J. E., & Lyles, M. A. (2001). Absorptive capacity, learning, and performance in international joint ventures. Strategic management journal, 22(12), 1139-1161.
Lavie, D. (2007). Alliance portfolios and firm performance: A study of value creation and appropriation in the US software industry. Strategic management journal, 28(12), 1187-1212.
Lavie, D., & Miller, S. R. (2008). Alliance portfolio internationalization and firm performance. Organization science, 19(4), 623-646.
20
Lin, C., Wu, Y. J., Chang, C., Wang, W., & Lee, C. Y. (2012). The alliance innovation performance of R&D alliances—the absorptive capacity perspective. Technovation, 32(5), 282-292.
Lundvall, B.A., 1992. National Systems of Innovation: Towards a Theory of Innovation and Interactive Learning, Pinter, London.
Luo, X., & Deng, L. (2009). Do birds of a feather flock higher? The effects of partner similarity on innovation in strategic alliances in knowledge‐intensive industries. Journal of Management Studies, 46(6), 1005-1030. Mesquita, L. F., Anand, J., & Brush, T. H. (2008). Comparing the resource‐based and relational views: knowledge transfer and spillover in vertical alliances. Strategic Management Journal, 29(9), 913-941.
Monteiro, M., Birkinshaw, J. (2017) The external knowledge sourcing process in multinational corporations. Strategic Management Journal, 38, 342-326.
Mowery, D. C., Oxley, J. E., & Silverman, B. S. (1996). Strategic alliances and interfirm knowledge transfer. Strategic management journal, 17(S2), 77-91.
Nonaka, I., & Takeuchi, H. (1995). The knowledge-creating company: How Japanese companies create the dynamics of innovation. Oxford university press.
Noseleit, F., & de Faria, P. (2013). Complementarities of internal R&D and alliances with different partner types. Journal of Business Research, 66(10), 2000-2006.
Oxley, J. E., & Sampson, R. C. (2004). The scope and governance of international R&D alliances. Strategic Management Journal, 25(8‐9), 723-749.
Pearce, R., & Singh, S. (1992). Globalizing research and development. Springer.
Robinson, C., & Schumacker, R. E. (2009). Interaction effects: centering, variance inflation factor, and interpretation issues. Multiple Linear Regression Viewpoints, 35(1), 6-11.
Rothaermel, F. T., & Deeds, D. L. (2006). Alliance type, alliance experience and alliance management capability in high-technology ventures. Journal of business venturing, 21(4), 429-460.
Sampson, R. C. (2007). R&D alliances and firm performance: The impact of technological diversity and alliance organization on innovation. Academy of Management Journal, 50(2), 364-386.
Simonin, B. L. (1999). Ambiguity and the process of knowledge transfer in strategic alliances. Strategic management journal, 595-623.
Szulanski, G. (1996). Exploring internal stickiness: Impediments to the transfer of best practice within the firm. Strategic management journal, 17(S2), 27-43.
Teece, D. J. (1982). Towards an economic theory of the multiproduct firm. Journal of Economic Behavior & Organization, 3(1), 39-63.
Teece, D. J. (1992). Competition, cooperation, and innovation: Organizational arrangements for regimes of rapid technological progress. Journal of economic behavior & organization, 18(1), 1-25.
21
Tsai, K. H., & Wang, J. C. (2008). External technology acquisition and firm performance: A longitudinal study. Journal of Business Venturing, 23(1), 91-112.
Van Aken, J., Berends, H., & van der Bij, H. (2012). Student Projects. In Problem solving in organizations: A methodological handbook for business and management students. Cambridge University Press.
Von Hippel, E. (1998). Economics of product development by users: The impact of “sticky” local information. Management science, 44(5), 629-644.
Von Zedtwitz, M., Gassmann, O., & Boutellier, R. (2004). Organizing global R&D: challenges and dilemmas. Journal of International Management, 10(1), 21-49.
Wassmer, U. (2010). Alliance portfolios: A review and research agenda. Journal of management, 36(1), 141-171.
Wuyts, S., & Dutta, S. (2014). Benefiting From Alliance Portfolio Diversity The Role of Past Internal Knowledge Creation Strategy. Journal of Management, 40(6), 1653-1674.
