The themes of technological innovation, entrepreneurship, and organizing

Determinants and Consequences of R&D Strategy Selection

Diana A. Filipescu

Autonomous University of Barcelona, Spain

Claudio Cruz Cazares

Autonomous University of Barcelona, Spain

ABSTRACT

Nowadays firms are not able to achieve all innovation in-house due to the specific set of technologies required by most products and processes, obliging firms to access external knowledge. In this context, the aim of this chapter is two-fold with the final goal of increasing our knowledge on firm innovating behavior. First, this chapter analyzes the determinants of the R&D strategy (RDS) selection posting the make, buy and make-buy as the three RDSs. Second, this chapter analyzes the consequences that each of the RDSs has on firm innovativeness. Results show that commercial and organizational resources, jointly with the information sources, influence the selection of the strategy. As for the second part of the analysis, we see that all RDSs have positive effects on firm innovative performance but these effects are not straightforward and simple since they vary depending on firm's type and on the radicalness of the innovation achieved.

Copyright © 2012, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.

INTRODUCTION

In order to survive in the competitive scene that companies have faced in recent years and which is characterized by a high level of dynamism (Teece, 1998; Lopez & Garcia, 2005; Diaz et al., 2008),

DOI: 10.4018/978-1-61350-165-8.ch023 the continual renewal of competitive advantages through innovation (Cho and Pucik, 2005) and the development of new capabilities (Grant, 1996) has become necessary (Danneels, 2002; Branzei & Vertinsky, 2006). In this context, technology represents one of the most important factors in increasing the national and international competi­

tiveness of firms, while successful technological innovation in new products and processes is gradu­ally more regarded as being the central issue in economic development (Porter, 1998). Moreover, as highlighted by Shrivastava and Souder (1987) and Bone and Saxon (2000), a key element in the competitive business strategy is the combination of technological innovation and R&D activities.

Since the objective of R&D strategies (RDSs) is to guide the firm in acquiring, developing and applying technology in order to generate sustain­able competitive advantages (Swan & Allred,

2003) , it is extremely important for the firm to select the best way of achieving the technology needed (Clarke et al., 1995). It is emphasized in the literature that firms establish their boundaries based on the decision regarding the type of R&D activities - whether they should be integrated within the company or not. In fact, Williamson (1975) identified a dichotonomous decision be­tween make (internal) and buy (external) RDS; later on, Veugelers and Cassiman (2006) added a third, complementary one, make-buy RDS1.

The effect of these RDSs over firms’ innova­tive results has been thoroughly studied; however, there is no general conclusion except that all of them are, in a way, highly significant in a firm’s innovative impetus. Explicitly, Diaz et al. (2008) find that the three RDSs have a positive impact, while Veugelers and Cassiman (2006) suggest that only make-buy yields the best results, whereas buy has the lowest ones. Even more, most of the research studies carried out in this field of investigation focus on the choice between make and buy strategies, the attention towards a make- buy one or to the reasons behind the selection of one strategy as opposed to another being almost inexistent (e. g., Veugelers & Cassiman, 1999).

When referring to the make RDS, firms un­derstand a sole source of knowledge and, thus, important sources of competitive advantages achieved with high costs whose results cannot be clearly foreseen. The buy strategy, on the other hand, is a relatively low-cost one with more predictable results, offering solutions to some problems related to a lack of capacity. However, it does not stand for competitive advantages since there is a high probability that competitors attain it as well. As for the combination between them, the make-buy strategy, it enhances both the advantages and disadvantages of make and buy RDSs, being extremely complex to manage it.

Taking this into consideration, this study has a two-fold focus. Firstly, it aims at finding the determinants of the innovation strategy selection and, secondly, it has the objective of understand­ing which of the RDSs produce the best results in term of firm’s innovative performance. In order to reach this objective, data from the Technological Innovation Panel (PITEC) provided by the Na­tional Statistics Institute from Spain are analyzed, precisely the period 2004-2007.

As for the main contributions of this study, they are as follows: Firstly, we look at RDSs as a whole process, considering the determinants of selecting one strategy over another, and next their consequences over a firm’s innovative perfor­mance. Secondly, we consider the R&D Capital Stock Model developed by Grilliches (1979), which emphasizes the relation between RDSs and a firm’s innovative performance, employing lagged values of RDSs in order to improve pros­pects of valid causal inference (Baum, 2006) and to reduce possible endogeneity problems (Bernard and Jensen, 1999; Salomon and Shaver, 2005). Thirdly, we will look at both manufacturing and service industries, aiming at offering a better understanding of the latter, which has not been analyzed in the sense that our investigation does.

The study is organized as follows. First, the advantages and disadvantages of each of the RDSs are presented. Next, the determinants and consequences of RDSs are described, with their respective hypotheses, based on the absorptive capacity and open innovation approaches as well as the resource-based view (RBV) theory. The conceptual model is then presented, followed by the description of the methodology employed in this analysis. Results and discussions are offered in a late section, finalizing the study with the conclusions, contributions, limitations and future research lines of investigation.

R&D STRATEGIES: ADVANTAGES AND DISADVANTAGES

The make strategy is defined as the internal devel­opment of R&D activities, being such a complex one that it requires the creation of internal depart­ments in order to develop it (Dosi, 1988). The information flow between the R&D department of the firm and those which will use the resultant technology could considerably increase as a result of the integration of R&D activities (Fernandez, 2005). At the same time, in-house R&D consti­tutes a unique source of knowledge and allows for an objective assessment of real innovation needs (West, 2002), with economies of scale being enhanced, transaction costs avoided and barriers against imitation constructed (Contractor & Lorange, 1988).

However, choosing to achieve R&D internally has some disadvantages: It is less cost-effective, riskier and less predictable, it takes longer to commercialize the new product (West, 2002), and the firm might remain isolated in only one technology if the R&D department is not flexible (Perrons & Platts, 2004).

The speed of development in new technologies is increasing, and due to this some firms prefer to externalize R&D activities since it is not feasible for them to develop such specific technologies internally (Quinn, 2000). Moreover, as stated in the RBV, it is not necessary for firms to own all the resources and capacities when they could access them externally (Barney, 1999). The buy strategy is valuable because it is more reliable and firms’ results are more predictable, since the technology has been already developed and tested (Kessler & Bierly, 2002). In addition, it allows risk calculation a priori, offers solutions to some problems related to a lack of capacity, increases the speed of access to new technology, and reduces risk (West, 2002). When firms decide to externalize the R&D, they acquire access to new knowledge areas (Haour, 1992) through the productive networks that are created (Nishiguchi, 1994). Value creation, agil­ity, quality of technology commercialization, and transfer cost of technology are some factors which condition the buy RDS (Tsai & Hsieh, 2006).

Nevertheless, acquiring external technology is not a competitive advantage per se, since it is available for competitors as well (Barney, 1991), providing a short-term strategy to the firm (Ku- rokawa, 1997). External dependence, functional inequalities, and coordination problems are other disadvantages of the buy strategy (Kotabe & Helsen, 1999). In addition, if a firm has a learn­ing gap it will be unable to take advantage of the technology acquired (Steensma, 1996).

As for the complementarity between make and buy RDSs, it is based on the assumption that existing products are more complex, as they must be technologically feasible and economically vi­able, this complexity requiring multidisciplinary knowledge that may sometimes be exclusively found beyond the firms’ boundaries, involving the combination of internal and external R&D (Chesbrough & Crowther, 2006). Two different theoretical approaches hold that the make and buy strategies may be regarded as complements rather than as alternatives. In this way, the make-buy strategy would be characterized by the advantages and disadvantages for the make and buy strategies but with the additionality that it is more complex to manage and more expensive to achieve.

The first approach highlighting the comple­mentarity of the make and buy strategies is the absorptive capacity (Cohen & Levinthal, 1990), being defined as the firm’s ability to recognize the value of external knowledge and to assimilate and apply it for commercial ends (Abecassis - Moedas & Mahmoud-Jouini, 2008). Acquisition, assimilation, transformation and exploitation are the four organizational capabilities constituting the firm’s absorptive capacity (Zahra & George, 2002). Complementarity between make and buy should be emphasized since firms must develop in-house R&D in order to generate or increase their capabilities to scan (acquisition-assimilation) and to integrate (transformation-exploitation) the external knowledge acquired through the buy strategy (Arora & Gambardella, 1990). In other words, a firm will not make the most of the buy strategy efficiently if it does not develop R&D activities internally (Colombo & Garrone, 1996). Furthermore, greater knowledge gained through in-house R&D may serve to modify or improve external technological acquisitions (Veugelers & Cassiman, 1999).

The second theoretical background underlying such complementarity is the open innovation ap­proach (Chesbrough, 2003). It argues that firms have changed from the closed-innovation process to a more open process of innovation, where knowledge and technology flow are two-fold: inside-out and outside-in. The former implies that the internal inventions not used by the firm should be commercialized in order to profit from them. The latter, more important in this research, refers to the fact that not even large firms can afford to solely rely on their internal research and might acquire external knowledge though buying patents, licenses, or establishing cooperation agreements, or licensing processes or R&D activities. That is, open innovation is the combination of internal and external ideas and technologies in order to achieve new products, processes and technologies and to reduce time to market (Enkel et al., 2009). As a consequence, those firms acting within a closed innovation perspective will reduce their knowledge-base over the long term (Koschatzky, 2001).

DETERMINANTS OF R&D STRATEGIES

Based on the RBV, we consider that the main determinants of RDS selection are the firm’s resources. Amit and Schoemaker (1993) charac­terized firm resources as the stock of available factors owned and controlled by the firm. Tan­gible factors, like financial and physical assets, and intangible factors, such as human capital and technological know-how, are the components of firm resources (Grant, 1991). Firm capability is the ability to use and transform the owned resources into a desired end. Without these capabilities, the mere possession of a large amount of resources does not guarantee the creation of a competitive advantage (Song et al., 2007). Thus, according to Penrose (1959), Barney (1991) and Grant (1991), a firm should possess certain intangible assets that competitors cannot copy easily, in this way gaining a sustainable competitive advantage in the market.

A firm’s organizational resources are viewed as determinants of RDS selection since they reflect the efficiency and synergy between the depart­ments involved in the R&D process (e. g., R&D, production, and marketing). These resources also embody management and organizational excel­lence and enable the integration of internal and external knowledge (Bughin & Jacques, 1994; Dy - erson & Mueller, 1999), enhancing the absorptive capacity. However, these internal resources have been analyzed as a fostering factor in innovation processes (Bughin & Jacques, 1994; Galende & De la Fuente, 2003), but not as a determinant of RDS selection, a gap in the research literature that this contribution aims at filling.

Theory considers a firm’s experience as an intangible asset which represents the basis for obtaining a sustainable competitive advantage (Nonaka et al., 2000; Barney et al., 2001). A firm’s experience is also understood as attainment of knowledge. It is worth underline that a firm’s experience is related to a better management of communication and of necessary creativity to innovate, and to a more effective capacity for absorption (Rothwell, 1986; Cohen & Levinthal, 1990). Hence, it is more likely that experienced firms develop the make RDS since it requires organizational capabilities to control the complex process of R&D, and due to the fact that it is usu­ally more risky and expensive (West, 2002). On the contrary, firms which lack experience are more prone to select the buy strategy so as to externalize risk and overcome environmental uncertainties (Poon & McPherson, 2005). These arguments lead us to pose the following hypothesis:

H1: The greater the firm’s organizational resources, the lower the probability of selecting the buy strategy and the higher the probability of selecting the make-buy strategy.

A firm’s relations with foreign clients (Galende & Suarez, 1999), regularly measured by its inter­national achievements, are important intangible assets as well. International achievements are often considered useful for properly exploiting technological innovations (Teece, 1986). As ar­gued by Prashantham (2005), internationalization often leads to crucial growth and useful learning outcomes; it also increases a firm’s market size, thus favoring innovation activity (Galende & Suarez, 1999). Some studies have analyzed the firm’s internationalization as a determinant of R&D activities and found a positive relationship (Kumar & Saqib, 1996; Salomon & Shaver, 2005; Vila & Kuster, 2007; Filipescu et al., 2009). In the study carried out by Veugelers and Cassiman (1999), it was observed that export intensity re­duces external R&D selection.

One of our arguments is that international firms are more likely to combine make and buy strategies since technological advances are achieved through the development of internal R&D, increasing competitiveness and enacting disruptive innova­tions which are usually needed for entry into new international markets (Etflie et al., 1984; Galende and De la Fuente, 2003). When a firm becomes international, it creates new networks and, at the same time, gains access to foreign information and communication technologies, as well as produc­tion methods, transportation, and international logistics, which could reduce business transaction costs with potential suppliers, facilitating the buy strategy. Considering this, we state the second hypothesis of this study:

H2: The greater a firm’s commercial resources, the greater the probability of selecting the make- buy strategy.

We consider that the technological/knowledge intensity might also affect the RDSs selection. Industries experiencing a large number of tech­nological changes estimate R&D externalization to be the best option because it does not seem right to rely on internal R&D when the market is changing substantially (Noori, 1990). Similarly, when there is great technological diversity in the market, firms are likely to externalize R&D (Cesaroni, 2004).

On the other hand, it is suggested that when technological changes are unpredictable, R&D integration becomes necessary (Shrivastava & Souder, 1987) so as to prevent technological innovations that dramatically threaten market stability (Cooper and Schendel, 1976). In the end, a balance between these two poles might be found in the absorptive capacity and the open innovation approaches; that is, firms need to be aware of shifts in technology and gain greater flexibility through the buy strategy. At the same time, however, they might also develop in-house R&D in order to integrate the acquired technol­ogy efficiently and gain a competitive advantage.

Finally, information sources are also impor­tant in the selection of a specific type of RDS. Explicitly, firms which positively assess internal information tend to look for the complementar­ity between the two RDSs (Veugelers & Cassi - man, 1999). Thus, in order to take advantage of external knowledge, firms develop their internal capabilities. When the external knowledge is highly valued, firms will avoid solely developing the make or buy strategies, as they will prefer to combine them in order to make the most of external knowledge through internal R&D development.

CONSEQUENCES OF R&D STRATEGIES

The R&D Capital Stock Model developed by Grilliches (1979) emphasizes the positive rela­tion between RDSs and the firm’s innovative performance. This model stresses that R&D activities enhance innovations and these foster firm performance. Furthermore, Grilliches (1979) also argues that the R&D’s effects on innovations are lagged since R&D projects take a minimum of one year to be completed. However, there is no evidence which accounts for lagged effects of make, buy and make-buy RDSs, and this is one of the contributions of this study.

The empirical evidence of the effects of the make and buy strategies on firm innovativeness is insufficient and somehow controversial. Interest­ingly, the buy strategy has always been associated with negative effects when it was the only RDS evaluated. Kessler et al. (2000) analyzed how RDSs affect new product development and found that the buy strategy during the generation of the idea was negatively related to product competitive­ness and that externalization during the techno­logical development decreased innovation speed. Lanctot and Swan (2000) developed a scale for measuring a firm’s tendency to externalize tech­nology development and discovered a negative effect of externalization of product and process technology on the firm’s success. Finally, Fey and Birkinshaw (2005) found a negative relationship between contracting R&D and the creation of new products and technologies. These results are understandable if we consider that those firms solely externalizing R&D activities will not take advantage of it since they have not previously generated the absorptive capacity.

There is empirical evidence showing that inter­nal R&D produces better results than on external one. For example, Beneito (2006) found that the buy strategy had positive effects on incremental innovations, while the make strategy had positive effects on both incremental and radical innova­tions. Haro-Dominguez et al. (2007) and Chen and Yuan (2007) observed positive effects of external and internal R&D on new product development, although the effects were greater for internal R&D.

Results of studies where both the make and buy strategies were analyzed are also quite diverse. Jones et al. (2001) observed that external R&D significantly detracted from firm performance in terms of product, market and financial measures, while the make strategy had a positive effect on new product development. Diaz et al. (2008) concluded that both internal and external R&D increased the probability to achieve innovations. Santamaria et al. (2009) considered low-, medium - and high-technology industries and found that the make strategy was significant for both groups, while buy was positive for process innovations in the former and product ones in the latter.

Scarce is also the empirical evidence of the complementarity between make and buy strategies on firm innovativeness. Beneito (2006) observed that external R&D had no effect on innovation outputper se; however, when it was combined with internal R&D, positive effects arose. Adopting the supermodularity and productivity approaches, Veugelers and Cassiman (2006) found support for this complementarity. They observed that the make-buy strategy had the highest impact on sales due to new products. Nevertheless, following the same methodology, Schmiedeberg (2008) did not observe any trace of complementarity for a sample of German firms. Tsai and Wang (2007) concluded that external R&D had no effect on firm performance when used in isolation; rather, its ef­
fect depended on internal R&D efforts. Hence, the level of knowledge positively influenced inward technology for improving firm performance. In 2009, the authors analyzed inward technology for low - and medium-technology firms and found contradictory results. Internal R&D negatively moderated the role of external R&D on sales due to new products.

Drawing on the discussion above, the follow­ing hypotheses are formulated:

H3: All RDSs will have a positive effect on the firm’s innovative performance.

However, we do not expect all RDSs to have the same impact on firm innovativeness. Based on the open innovation (Chesbrough, 2003), innovation network (Pyka & Kuppers, 2003) and absorptive capacity (Cohen & Levinthal, 1990) approaches and on previous empirical research (i. e., Veugelers & Cassiman, 2006; Cruz-Cazares et al., 2010), we believe that the make-buy strategy will pro­duce the greatest effect on the firm’s innovation performance because innovations occur mainly through the combination of ideas, resources and technologies (Fey & Birkinshaw, 2005).

H3a: The make-buy strategy will have the high­est impact on the firm’s innovative performance.

Finally, despite the flexibility gained through the externalization ofthe R&D activities, coordina­tion problems, transactional costs and functional inequalities emerge when externalizing R&D (Kotable & Helsen, 1999), and researchers have found empirical evidence of these limitations (i. e.,Kessler et al., 2000; Fey & Birkinshaw,

2005) . Thus, based on theoretical and empirical evidence, we would argue that the buy strategy will have the lowest impact on the firm’s innova­tive performance.

H3b: The buy strategy will have the lowest impact on the firm’s innovative performance.

MODEL

The conceptual framework developed in this study is presented in Figure 1. As mentioned, this study is composed of a two-step analysis. In the first part, aiming to explain the determinants of the RDS, we set Hypotheses One and Two, arguing that the organizational resources and commercial resources will condition the firm’s selection of a certain RDS. Furthermore, we consider that the firm’s information sources will also influence firm’s selection. Additionally, at this first stage we control for the technological/ knowledge intensity, the fact that a firm belongs to a holding group and the firm’s size. In the second stage ofthe analysis, following the R&D Capital Approach, we assume that these strategies will have a positive and lagged effect on firm’s innovative performance. The firm’s size and the technological/ knowledge intensity are also used as control variables in the model estimation. We do not consider that the fact that a firm belongs to a group might affect the firm’s innovative performance. In the Methodol­ogy section, the variables operationalization and the modes estimates to test the model presented in Figure 1 are described.

METHODOLOGY Data

Figure 1. Conceptual framework

The empirical analysis carried out in this paper draws on the Technological Innovation Panel (PITEC) dataset (2003-2007). The PITEC dataset is conducted by the Spanish National Institute of Statistics (INE) in collaboration with the Spanish Science and Technology Foundation (FECYT) and the Foundation for Technological Innova­tion (COTEC). Particularly, PITEC compiles the information provided by the Spanish Com­munity Innovation Survey (CIS), which follows the guidelines ofthe Oslo Manual (OECD, 1997). It includes data on the technological innovation
activities like the main obstacles for achieving innovations, the main technological information sources, innovation and R&D expenditures, quali­fications of the R&D personnel, outsourcing R&D activities classified by origin and type of partners, and the effects the innovation achievements. This dataset is available in a set of files for each year (i. e., a file for each year 2003, 2004 ... 2007). However, due to sampling procedure limitations for the year 2003, we specifically used data for the period 2004-2007, with a sectoral coverage of manufacturing and service firms2.

In order to have a consistent longitudinal dataset, we excluded those firms that have ex­perienced any important contingency during the period 2004-2007 from the sample (i. e., mergers and acquisitions, fusions, changes in the industrial activity and missing values in some of the year’s corresponding to the period under analysis). In addition, we excluded those firms belonging to the primary sector (agriculture and mining) and those involved in construction and energy pro­duction and/or distribution activities. In order to observe the differences between manufacturing and service firms, the panel dataset was divided into two samples, the first one containing 6,776 observations of1,694 service firms, and the second one 14,188 observations of 3,547 manufacturing firms. As observed, both samples are balanced panels. Following Miotti and Sachwald (2003), innovative and non-innovative firms are included in the panel in order to prevent bias in the sample.

The main advantage of estimating the model in a panel dataset is that it allows for solving the endogeneity problem by including the independent variables lagged, thus improving the inference of causal effects (Baum, 2006).

Variables Definitions

Dependent Variables

As commented, this study is composed of a two-stage analysis. In the first stage we evaluate the determinants of RDS selection, leading us to consider the different strategies as dependent variables. This variable was constructed by the two original variables of internal and external

R&D expenditures. Firms were asked to report the percentage of R&D expenditures for each of the internal and external activities over the total innovation expenditures. Thus, if a firm reported zero expenses on both activities, the resultant variable has the value of 1 (no R&D), if the firm reported positive internal expenditures and zero external, the dependent variable has the value of 2 (make), for firms with zero internal expenditures and positive external expenditures, the variable has the value of 3 (buy), finally, if the firm reported positive expenditures for internal and external R&D, the composed variable has the value of 4 (make-buy). As observed, the resultant variable has four levels, each for one strategy, and these levels are mutually exclusive.

For the second stage ofthis study, when evaluat­ing the consequences ofthe RDSs on firm innova­tive performance, we consider the percentage of sales due to new products - new to the firm-, as the first dependent variable. Additionally, we also attempt to measure the novelty of the innovations achieved, thus, a second dependent variable is considered which measures the percentage of sales due to new products that are new to the market. Measuring the effect of the innovation activities as the percentage of sales due to new products is accepted in the literature (Veugelers & Cassiman,

2006) . The main advantage of considering this indicator of innovative performance is that it does not merely measure the innovations achieved but also the successful innovations that reached the market (Cano & Cano, 2006).

Independent Variables

As argued in the theoretical background, for the first stage ofthe analysis, we consider that organi­zational and commercial resources will influence the selection of the RDSs. The age of a firm is considered a proxy for measuring the organiza­tion resources since it represents experience and learning capacity (Galende & De la Fuente, 2003). This variable has the value of 1 if the firm is of recent creation and 0 otherwise. The degree of firm internationalization is considered as a proxy for measuring the organizational resources (Filipescu et al., 2009). This variable has the value of 1 if the firm commercializes its products in a local and national market, 2 if the firm commercializes in other EU countries and 3 if the firm sells its products in any other country. The last independent variables for the determinants of the RDS selec­tion are the firm’s information sources. With the PITEC sample, there are eleven different items that capture the importance of each ofthe information sources. Firms are asked to mark, from 1 to 5, with the latter being the highest degree, the importance given to that specific information source. Using a factorial analysis those eleven items were re­duced into three different factors. The first one captures the information coming from the firm’s market, the second one gathers the information from public organisms and the latter represents the information from conferences and publications3. As observed, all of these information sources are beyond firm boundaries and, following the open innovation approach, we expect that the higher their assessment, the more prone firms will be to externalize R&D activities.

The dichotomous variable of belonging to a group is included as a control variable since firms with access to group technology will have incen­tives to achieve the buy strategy due to a reduction in the transaction costs. For controlling the industry effect, a dummy variable was included which has the value of 1 if the service firm belongs to one of the industries considered with a high knowledge intensity according to Felix’s (2000) classification. The Oslo Manual (OECD, 1997) classification was used to categorize manufacturing firms de­pending on their technology intensity being low, medium or high. The logarithm of the number of employees is included to account for firm size.

The main explanatory variables of firm innova­tive performance are the RDSs which, contrary to the first stage ofthe analysis, are separately intro­duced into the model as dichotomous variables,
with four resulting variables: no R&D, make, buy and make-buy. To assert the lagged effect of the RDSs, we included these variables at t-1 and t-2. The commercial resources, measured as the degree of firm internationalization, are considered inde­pendent variables so as to explain firm’s innovative performance. The firm’s technological resources are also considered to be an explanatory variable ofthe firm’s innovative performance, and they are measured as the percentage of R&D expenditures over total innovation expenses. In order to solve the endogeneity problem, both technological/ knowledge and commercial resources at t-1 were introduced into the model. Finally, the knowledge and technological intensity levels and firm size are included in the model as control variables.

RESULTS Descriptive Analysis

Next, Table 1 shows the distribution ofRDS selec­tion among service and manufacturing firms. As observed, 38.90% of those service firms in low- knowledge sectors do not achieve any RDS. This percentage is reduced by more than half for those firms in high-knowledge sectors. We can observe the same behavior for manufacturing firms, since the percentage of firms with no R&D decreases as the technological intensity increases. There is a clear tendency for all sectors: the make strategy is the most selected, followed by the make-buy and finally the buy strategies. Interesting is the fact that the highest percentage of firms developing the buy strategy is represented by the low-knowledge - intensity service firms.

Table 2 shows the relationship between the RDSs achieved and firm innovative performance. As commented, we consider the percentage of sales due to new products for the firm and for the market as two variables accounting for the firm’s innovative performance. For service and manu­facturing firms, these descriptive results highlight the importance of achieving R&D activities in order to obtain radical innovations. Observe that for those firms without R&D activities the per­centage of sales for products new to the firm is quite similar to those firms achieving any RDSs; however, for the percentage of sales for products new to the market, the difference between those firms achieving the make or make-buy strategies is considerably higher than in those firms without R&D activities. Additionally, we could assume that the buy strategy does not influence the prod­ucts new to the market since its percentage is very low.

Determinants of RDS Selection

Since the dependent variable of the first stage of the analysis is a categorical one, we estimate a multinomial logit model with random intercepts that accounts for the individual, unobservable heterogeneity4. In Table 3, the model estimates for the determinants of RDSs selection for service and manufacturing firms are seen. Two models were estimated for each sample; in the first one, the make strategy is the reference variable which is compared with buy and make-buy strategies. In the second model, the buy strategy is the reference and it is compared against the make-buy strategy. By doing so, we are able to compare the prefer­ence for selecting one strategy against the others5, something that the previous literature failed to do (e. g., Veugelers & Cassiman, 1999).

Innovation

Strategies

a. Service firms

b. Manufacturing firms

Low-knowledge

intensity

High-knowledge

intensity

Low-Tech.

intensity

Medium-Tech.

intensity

High-Tech.

intensity

no R&D

38.92

14.64

20.12

19.09

9.49

make

28.77

44.59

42.95

39.81

47.82

buy

11.03

5.72

6.41

8.33

4.07

make-buy

21.27

35.05

30.53

32.78

38.62

Table 2. Percentage of RDSs by sector

Innovation Strategies

a. Service firms

b. Manufacturing firms

New to the firm

New to the market

New to the firm

New to the market

no R&D

15.53

5.88

15.06

6.57

make

16.14

16.13

17.00

10.81

buy

14.86

7.07

15.51

6.44

make-buy

16.11

18.63

16.44

12.65

As observed in Table 3, organizational re­sources determine RDS selection. For service and manufacturing firms, the negative and significant effects of buy and make-buy strategies in Models A and C indicate that when organizational re­sources are high firms tend to select the make strategy. From Models B and D we see that there is no significant difference between buy and make - buy. These results lead us to reject Hypothesis 1 and contradict previous studies that highlight that young firms lacking organizational resources avoid the make strategy since they are likely to

select the buy strategy in order to externalize risk and overcome uncertainties (Poon & McPherson, 2005).

Hypothesis 2 states that the higher commercial resources, the higher the probability for selecting the make-buy strategy; however, this hypothesis cannot be supported. Again, for Models A and C we can observe that the make strategy is preferred over the other two, while make-buy is preferred over the buy strategy for Models B and D. These results are in line with those of Veugelers and Cassiman (1999), who stated that the higher the commercial resources, the lower the probability to select the buy strategy.

The information sources for innovations also determine RDSs selection although not in the expected direction. For both service and manu­facturing firms, the information for innovation coming from the firm’s market reduces the pos­sibility to solely externalize the R&D activities (Models A and C); however, - due to the positive sign of make-buy in all models-, it does influence the selection of the make-buy strategy. This fact highlights the open innovation approach since those firms aware of changes, tendencies and facilities of competitors, clients and equipment suppliers look forward to gaining the flexibility and speed through externalizing while, simultaneously, developing internal R&D to create knowledge and barriers to imitations.

Against what was expected, the negative and significant sings of the buy and make-buy strate­gies in Models A and C show that when firms highly value the importance of information com­ing from public institutes, they tend to select the make strategy. As for the information coming from conferences and publications, it does not seem not so clear to determine RDS selection since it is not significant for any case in the service sample, and for the manufacturing firms results merely show that the make strategy is preferred over buy.

Table 3. Determinants of RDS selection

Variables

Service firms

Manufacturing firms

A. make as ref.

B. buy as ref.

C. make as ref.

D. buy as ref.

buy

make-buy

make-buy

buy

make-buy

make-buy

Organizational

Resources

-1.9380*

(1.0599)

-0.7352**

(0.3627)

0. 9709 (1.5614)

-2.2619**

(1.1165)

-0.9865**

(0.4960)

1.0978 (1.5691)

Commercial Re­sources

-0.8915***

(0.1158)

-0.2472***

(0.0745)

0.3944* (0.2147)

-0.3342***

(0.0805)

-0.1728***

(0.0619)

0.2366* (0.1266)

Info. firm’s market

-0.3773***

(0.1107)

0.0255

(0.0706)

0.4344**

(0.1782)

-0.1848**

(0.0774)

0.2224***

(0.0542)

0.4022*** (0.1095)

Info. public organ­ism

-0.4706***

(0.1119)

-0.1638**

(0.0730)

0.3678**

(0.1853)

-0.5679***

(0.0788)

-0.1399**

(0.0540)

0.4780*** (0. 1220)

Info. conf. & publi­cations

-0.0620

(0.1220)

0.0241

(0.0814)

0.2451 (0.2248)

-0.3914***

(0.0843)

-0.0230

(0.0606)

0.1879 (0.1281)

Group

0.6768***

(0.2135)

0.4419***

(0.1610)

0.0451 (0.4025)

0.1243

(0.1478)

-0.0540 (0.1114)

-0.1557 (0.2312)

High knowledge Int.

-2.8135***

(0.2796)

-1.3856***

(0.2499)

1.9523***

(0.4235)

Medium Tech.

0.3122*

(0.1834)

0.1581 (0.2499)

0.1278 (0.2767)

High Tech.

-1.0814***

(0.1576)

-0.5303***

(0.1203)

0.6147** (0.2463)

Size

0.3581***

(0.0548)

0.2147***

(0.0422)

-0.1901**

(0.0944)

-0. 1469*** (0.0506)

-0.0162

(0.0418)

0.1736* (0.0915)

Constant

-0.7542*

(0.3949)

-0.3465

(0.3112)

5.0410***

(0.7447)

-1.4963***

(0.2978)

-0.6798***

(0.2347)

4.3382*** (0.5221)

log-likelihood: -5022.4754 N. observations: 5082 N. firms: 1694

log-likelihood: -10316.92 N. observations: 10641 N. firms: 3547

* p <.1; ** p <.05; *** p <.01; standard errors in brackets

As for the belonging to a group variable, ob­serve in ModelA that buy and make-buy strategies have a positive and significant coefficient showing that the make strategy is the less preferred. These results indicate that belonging to a group, due to a reduction in the transaction cost, stimulates the externalization of R&D, being in isolation or in
combination with internal R&D. Nevertheless, these results are not extendable for manufactur­ing firms.

Service firms in high-knowledge industries will be likely to select the make strategy, while the buy one is the least preferred. Regarding manufacturing firms, we can see that the technological intensity sector to which the firm belongs also conditions RDS selection. For medium-technology firms, the buy strategy seems to be preferred over make, while for high-technology firms, the buy strategy is the least selected.

Contrary to the regular effects on service and manufacturing firms ofthe determinants described above, firm size presents almost an inverse effect for the two samples. On one hand, for the service sample, the bigger the firm is, the higher the probability to select the buy strategy, while the lowest probability accounts for the make strategy. On the other hand, bigger manufacturing firms will prefer the make or make-buy strategies over exclusively buying.

Consequences of RDSs

The second-stage analysis aims at evaluating the effect of RDSs on firm innovative performance. To do so, the effects ofthe RDSs are evaluated on the percentage of sales due to new products for the firm and for the market. Due to the truncated nature of the dependent variables, we estimated a

Tobit model with random effects for service and manufacturing firms.

As commented, in order to empirically demon­strate the lagged effect of R&D activities on firm innovativeness (Griliches, 1979), and looking to solve any endogeneity problem, we include the RDSs in the model lagged at t-1 and t-2. Addition­ally, this double-lagging allows us to capture the short - and long-term effects of the RDSs on firm innovativeness.

First of all, the results of the new-to-the-firm model corresponding to service firms presented in Table 4 are surprising. As observed, none of the RDSs has any short-term (t-1) or long-term (t-2) effect on firm innovativeness. This might indicate that service firms do not need to attain R&D activities in order to obtain non-radical in­novations. Continuing with service firms, we can see that all RDSs have positive and significant effects over the percentage of sales due to new products for the market, confirming Hypothesis

Table 4. Consequences of RDS selection

Service firms

Manufacturing firms

New to firm

New to market

New to firm

New to market

Make t-1

5.3661 (5.6233)

30.2971*** (6.4906)

11.9663*** (2.8946)

20.5586*** (3.1499)

Buy t-1

2.4156 (6.5951)

24.2054*** (7.8600)

9.3077** (3.5386)

3.9463 (4.0303)

Make-buy

3.6505 (5.8399)

36.1143*** (6.7580)

10.7482*** (3.0190)

26.0626*** (3.2936)

Make t-2

4.5290 (3.2322)

17.7258*** (3.9480)

4.1783** (2.1354)

9.4276*** (2.4053)

Buy t-2

-3.2977 (4.7804)

-0. 2255 (6.2348)

2.4960 (2.8437)

5.6346* (3.2673)

Make-buy

5.7134 (3.5605)

17.9095*** (4.2867)

4.4934** (2.2567)

12.3009*** (2.5100)

Commercial Resources

0.3342 (1.3543)

6.9480*** (1.4511)

0.9060 (0.9486)

1.0844 (0.9949)

Technological Resources

-0. 0458 (0. 0554)

-0.1091* (0.0624)

-0.0747** (0. 0258)

-0.0778** (0.0277)

High - knowledge Int.

-4.5117 (3.8399)

13.9377*** (4.2633)

Medium-Tech.

-4.2021* (2.2428)

-1.1146 (2.3086)

High-Tech.

1.9892 (1.9301)

5.3593** (1.9791)

Size

-0.7496 (0.7361)

-1.5860** (0.7895)

0. 2306 (0.6254)

-0.7416 (0.6366)

Constant

-6.9791 (6.1713)

-64.6966 *** (6.9983)

-9.5435** (4.0823)

-38.4766*** (4.3749)

* p <.1; ** p <.05; *** p <.01; standard errors in brackets

3. However, as predicted in Hypotheses 3a and 3b, not all strategies have the same impact. Due to the complementarity stressed in the open in­novation (Chesbrough, 2003) and absorptive capacity (Cohen and Levinthal, 1990) approaches, the make-buy strategy produces a greater impact for achieving successful radical innovations, confirming Hypothesis 3a. Despite the difficulties that arise when R&D activities are externalized (Narula, 2001), the buy strategy positively impacts the achievement and commercialization of radical innovations for service firms.

When increasing the spectrum, we can clear­ly detect the so-called short-term effect ofthe buy strategy (Kurokawa, 1997) since it is no longer significant at t-2. As for the make and make-buy strategies, they still impact on firm innovativeness two years after they were achieved; nevertheless, their effect is reduced by half. Interestingly, the greatest effect of the make-buy strategy loses strength at t-2, and its effect is practically the same as the make strategy.

Important results emerge when paying atten­tion to manufacturing firms. First, considering the positive and significant effects of RDSs on sales due to new products for the firm, it could be as­sumed that R&D activities are needed to achieve
non-radical innovations. Here, Hypothesis 3 is again verified. However, due to the coefficient magnitude, we are not able to support Hypothesis 3a since the make-buy strategy does not produce the greatest impact. It seems that for achieving non-radical innovation, the make strategy is the best option for manufacturing firms. In line with Chen and Yuan (2007), the estimates show that the buy strategy is the one with the least significant effect, confirming Hypothesis 3b. When focusing on the long-term effect, the make and make-buy strategies still positively influence the innovative performance but, as service firms, its effect is re­duced by more than half of that of the year before. The buy strategy is no longer significant at t-2.

We are not able to support Hypothesis 3 for the last model since the buy strategy does not seem to influence the sales due to new products for the market. Based on the coefficient magnitude, Hypothesis 3a is supported, in which we indicated that the make-buy strategy would produce the greater impact (Veugelers & Cassiman, 2006). The same pattern of the long-term effect-reduction is also detected in this model. Nevertheless, the buy strategy increased when referring to the coefficient size and significance, indicating that this strategy influences radical innovations, but exclusively in a long-term. The absorptive capacity approach is useful to clarify this behavior. Those firms that achieve the make and buy strategies simultane­ously can profit from the complementarity in a short-term since they can easily integrate the external knowledge into the firm’s routines. On the contrary, those firms which exclusively buy technology and knowledge can profit from it after two years, since the integration of knowledge takes longer.

Concerning the commercial resources, we can observe in Table 4 that they slightly affect firm innovativeness. The only effect registered in the estimates is positive for the sales due to new products for the market for the service firms, meaning that firm internationalization requires highly innovative products and allows access to knowledge that stimulates these innovations (Fili­pescu et al., 2009). Contrary to what was expected, the firm’s technological resources, measured as R&D expenditures, negatively influence firm in­novativeness for service and manufacturing firms. A possible explanation would be that the innova­tion process is inefficient for most of the firms in the sample since those R&D expenditures that are not successfully transformed into innovations are sunken costs that negatively affect firm per­formance (Koellinger, 2008). Finally, knowledge/ technological intensity and firm size little affect firm innovativeness. This might indicate that the knowledge and technology required for achiev­ing this type of innovation is less complex, and a single firm can manage it successfully.

CONCLUSION

Why do firms select one RDS over another? What are the effects of the RDSs on firm innovative performance? Are these effects similar for all RDSs? Responding to these questions marked the aim of this research. We selected a sample of service and one of manufacturing firms in order to empirically answer these questions, thus far inconclusive.

Initially, two hypotheses were stated for the first part of the analysis - determinants of RDS selection. In the first hypothesis it was argued that firms with higher commercial resources would tend to select the make-buy strategy since it could give the firm the capacity for going beyond the local market and, at the same time, a firm with activities abroad could access new information and networks, thus facilitating the externalization of R&D activities. However, neither for the service nor for the manufacturing sample were we able to support our hypotheses.

Our second hypothesis stated that the higher the organizational resources, the higher the prob­ability for selecting the make-buy strategy and the lower the probability for selecting the buy strategy. The results obtained did not give support for this statement either. For high commercial resources, the make strategy is preferred over the others for service and manufacturing firms. This might indicate that when firms are aware of their available resources, they feel confident to achieve R&D activities on their own without taking into account the technologies and knowledge available in the market. Nevertheless, as mentioned next, this excess of self-confidence might be negative.

As for the second part of the analysis, three hypotheses were proposed. In the first one, op­posing some previous studies (Fey & Birkinshaw, 2005), we stated that the make, buy and make-buy strategies will produce positive effects on firm in­novative performance and re sults gave support for it, for both service and manufacturing firms. Based on the absorptive capacity and open innovation approaches, we hypothesized that the make-buy strategy would produce the better results and the buy strategy the worse results. These hypotheses were confirmed but with some nuances. We ob­served that the effects of all of the strategies are not straightforward and simple since they vary depending on firm type and on the radicalness of innovation achieved. For manufacturing firms, the make-buy strategy produces the higher results in firm innovativeness as long as the innovation achieved is radical. Nevertheless, when the in­novation achievement is non-radical, the make strategy is the best option. For service firms there is not a significant difference between choosing one strategy or another if the firm aims to achieve non-radical innovations, that is, innovations new to the firm, but when the innovation is radical, the make-buy strategy is the one producing the better results.

Contrary to some studies which stated that the buy strategy did not affect the innovation output (Schmiendeberg, 2008), or affected it negatively (Fey & Birkinshaw, 2005), we found that this strategy positively affects firm innovation perfor­mance for service firms when achieving radical innovations and for manufacturing firms achiev­ing non-radical innovations. Additionally, this strategy has a slight effect on radical innovations for manufacturing firms, but only two years later. Drawing on the absorptive capacity, we could assume that those firms which exclusively buy technology and knowledge can profit from it after two years, since the integration of knowledge takes longer in the absence of an efficient absorptive capacity. On the contrary, those firms that achieve the make and buy strategies simultaneously can profit from the complementarity in a short-term since they can easily integrate external knowledge into firm routines.

This study has academic and managerial impli­cations. Firstly, as far as the authors’ knowledge goes, this is the first study that empirically looks at RDSs as a whole process, considering the de­terminants of selecting one strategy over another and next their consequences over firms’ innova­tive performances. Second, contrary to previous research, we consider and use the R&D Capital Stock Model (Grilliches, 1979) and observe the lagged effect of the RDSs on firm performance, improving the prospects of valid causal inference (Baum, 2006) and reducing possible simultaneity problems (Bernard and Jensen, 1999; Salomon and Shaver, 2005). Thirdly, we look at both manufac­turing and service industries, aiming at offering a better understanding of the latter which has not been analyzed in the sense that our investigation has. Finally, for managers this study can be useful to understand the characteristics ofthe RDSs and their potential impact on innovative performance to a greater extent.

This study is not free of limitations, which come especially from the fact that we dealt with a longitudinal sample which, according to Baltagi (2007), includes problems in the design, data collection, and data management of panel surveys. It is also possible that panel data show bias due to sample selection problems and attrition (Wooldridge, 1995). Other limitations are related to the introduction and measurement of some other variables in the analyses, thus conferring a more complete image of R&D activities. Moreover, the approach used to measure some of the factors may be less precise than desired.

As for future research lines, it would be inter­esting to be able to realize comparisons between similar studies. By replicating this investigation in distinct geographical contexts, results could be generalized to larger populations. In this way, it would reveal if institutional factors play a role in influencing the relation (Kogut et al., 2002; Peng et al., 2005; Kumar, 2009). Moreover, comparisons among firms with different types of ownership could be also acknowledged, as well as the employment of other criteria to separate the database (e. g. size, sector).

The themes of technological innovation, entrepreneurship, and organizing

About the Contributors

Farley S. Nobre (PhD, MSc, BSc) is Professor at the School of Management of Federal University of Parana, Brazil. His research interests include organizations, knowledge management systems, innova­tion and sustainability. …

The Roles of Cognitive Machines in Customer — Centric Organizations: Towards Innovations in Computational Organizational Management Networks

Farley Simon Nobre Federal University of Parana, Brazil ABSTRACT This chapter proposes innovative features of future industrial organizations in order to provide them with the capabilities to manage high levels …

Tools That Drive Innovation: The Role of Information Systems in Innovative Organizations

Jason G. Caudill Carson-Newman College, USA ABSTRACT The purpose of this chapter is to examine computer technology as a tool to support innovation and innovative processes. The primary problem that …

Как с нами связаться:

Украина:
г.Александрия
тел./факс +38 05235  77193 Бухгалтерия
+38 050 512 11 94 — гл. инженер-менеджер (продажи всего оборудования)

+38 050 457 13 30 — Рашид - продажи новинок
e-mail: msd@msd.com.ua
Схема проезда к производственному офису:
Схема проезда к МСД

Оперативная связь

Укажите свой телефон или адрес эл. почты — наш менеджер перезвонит Вам в удобное для Вас время.