External R&D: exploring the functions and qualifications of R&D personnel
André Spithoven*
Belgian Science Policy Office, Avenue Louise 231, 1050 Brussels, Belgium.
Vrije Universiteit Brussel, Pleinlaan 2, 1050 Brussels, Belgium.
E-mail:
Peter Teirlinck
Belgian Science Policy Office, Avenue Louise 231, 1050 Brussels, Belgium.
Hogeschool-Universiteit Brussel, Stormstraat 2, 1000 Brussels, Belgium.
Vrije Universiteit Brussel, Pleinlaan 2, 1050 Brussels, Belgium.
E-mail:
* Corresponding author
Abstract: Knowledge has become an important production factor. Tacit knowledge is embodied in the minds of people, and is part and parcel of their skills. Since businesses concentrate on strategic tasks, many others are outsourced. This also applies to R&D.External knowledge sources help firms to leverage its internal R&D efforts. Internal and external knowledge have to be knitted together on the work floor giving rise to innovative products and processes. Since tacit knowledge is embodied in personnel, it is interesting to look at the role of the functions and the qualifications of R&D personnel in relation to activities developed in the framework of technical collaboration agreements. Using the R&D OECD survey of firms in Belgium, insights are offered in the way that external knowledge impacts on the organisation of the internal division of labour. We demonstrate that absorptive capacity, embodied in people, is a key element in using external knowledge.
Keywords:R&D personnel; Absorptive capacity; Collaboration; Knowledge exchange; Function; Qualification.
1 Introduction
Due to the complexity and knowledge intensity of developing new products and processes, the R&D active company has to look elsewhere to complement its internal R&D efforts (Cassiman and Veugelers, 2006). To this end a company is able to engage in collaborative agreements that can have a formal or informal character (Bonte and Keilbach, 2005). At least two issues can be addressed by such an agreement. First, a collaboration agreement can spell out the exact arrangement between the partners involved concerning the costs and organisational engagements. Second, the agreement might include appropriability issues if the aim of collaboration is to device a new product or set up a new process (Gulati and Singh, 1998; Helper et al., 2000). Hence the IP management of companies becomes a vital issue. This is especially the case in the early stages of formal R&D collaborations (Bader, 2006).
Knowledge stocks and flows are recognised as a crucial element of value creation in firms (Kang et al., 2007). The paper adds to the knowledge based view of the firm which sees them as repositories of competences and knowledge (Kogut and Zander, 1996; Grant, 1996) by devoting more attention to the embodiment of these competences with respect to the ability to cooperate and to share knowledge. Knowledge generation often results from new combinations of existing knowledge (Cohen and Levinthal, 1990). This is not restricted within the boundaries of the firm as R&D increasingly necessitates the use of external sources (Chesbrough, 2003; Coombs et al., 2003; Howells et al., 2003).
The paper contributes to the literature on absorptive capacity. Absorptive capacity has been coined by Cohen and Levinthal (1990) to indicate the need for firms to identify, assimilate and commercialise external knowledge. They point to the fact that absorptive capacity is path dependent and cumulative, and is build through investments in its labour force. Usually, R&D investments are seen as an indicator for firms to invest in absorptive capacity. But the receptivity of the labour force to the R&D investments remains understudied (see, however, Roper et al. 2008).
This paper intends to fill this gap by looking closer to the internal knowledge base of firms. The focus is on the relation between the function and qualification of R&D personnel and external knowledge relations in terms of the development and exchange of knowledge in formal research cooperation. When investigating these relationships, attention is paid to the use of intellectual property and product innovation.
2 Literature review and hypothesis development
The necessity of firms to rely on external knowledge has brought the notion of absorptive capacity centre stage. That is why the debate has attracted considerable attention from scholars (Lane and Lubatkin, 1998; Lane et al, 2006; Nieto and Quevedo, 2005; Todorova and Durisin, 2007; Zahra and George, 2002).
Human resources are a crucial element in the research and development (R&D) activities in firms (Jain and Murray, 1984; Allen and Katz, 1992). This paper deals with the role of two key dimensions in gaining from knowledge spillovers related to human resources – the function or skills and formal qualification or educational level of the labour force – when they engage in formal collaborative agreements on behalf of the firm. Skills thus become a qualitative aspect of absorptive capacity because firm’s employees embody tacit knowledge which is considered to be a main source of spillovers (Kaiser, 2002).
Various authors pointed to the fact that competitive pressures and increased complexity force firms to look for external knowledge and collaborative actions to innovate (Powell et al., 1996; Chesbrough, 2003; Coombs et al., 2003; Howells et al., 2003). Collaborative agreements are a means to gain access to external knowledge (Colombo and Garrone, 1996; Lazaric and Marengo, 2000).
2.1 Function of R&D personnel and external knowledge relations
The literature on open innovation conceptualises firms as organisations in which continuously and simultaneously ‘make’ or ‘buy’ decisions are being taken by individuals (Chesbrough 2003). In this way, the employees of the firm are making the boundaries permeable in order to attract the necessary knowledge and technology; while at the same time prevent unintended spillovers by scrutinizing alternatives to commercialisation and other appropriability mechanisms.
Jacobides and Billinger (2006) and Rothaermel et al. (2006) have systematically analysed boundary construction. They introduced the concept of organisational absorptive capacity by acknowledging the roles of internal functions and competences which serve as antennas directed at detecting external knowledge. They posit that these internal functions are a prerequisite to successfully interact and outsource.
Lazonick (2005) builds on the resource-based view of the firm to demonstrate how innovation depends on the development and accumulation of specialised internal capabilities. To stimulate the development of internal capabilities the firm needs organisational integration: a set of relations that creates incentives for employees who participate in hierarchical and functional divisions of labour to apply their skills and efforts to the innovation process (Helper et al., 2000). To absorb knowledge from the external environment, firms need organisational integration in which employees function as interfaces with the environment. These employees have to possess the skills to screen, interpret and assimilate the knowledge and transfer it through internal communication and diffusion on the work floor. Lam (2000) finds that the narrower the interfaces to the external environment, the less knowledge and ideas are absorbed. The less internal employees learn about external ideas, the smaller the chance that they will succeed in their innovative efforts.
Hence, absorptive capacity is more than stimulating R&D investment in the vain of Cohen and Levinthal (1990). Absorptive capacity has to be enlarged by acknowledging its interface role towards the external environment while acting as internal evaluator as to the value of the external knowledge.
Hypothesis 1 The higher the share of researchers and R&D managers; the higher the probability to cooperate to create new knowledge or exchange knowledge developed in the firm.
2.2Qualification of R&D personnel and external knowledge relations
Next to the skills, which can be developed at the working place, skills are also obtained through education which results in formal qualifications in terms of degrees or diplomas (OECD, 1995). The relation between R&D and education has been established in the past (Scherer and Huh, 1992) and was reiterated in recent research (Roach and Sauermann, 2010). High qualified employees are associated with higher R&D investment levels. Education and training are, moreover, crucial to innovation (Lundvall and Nielsen, 1999, Lam, 2005). Similar to the skill-levels, educational levels facilitate the detection and management of relevant external knowledge flows (OECD, 2008), which is a key ingredient in absorptive capacity (Roach and Sauermann, 2010).
Especially in the case of doctorate degrees the body of research is considerable (Stuart and Ding, 2006; Bercowitz and Feldman, 2008). These category of R&D personnel, at least the most recently graduated of them, possess knowledge that borders at the frontier of science and technology. They are especially receptive to so-called upstream research such as universities and public research centres that provide basic research and, also, their work is related to the patent activity of the firm (Sauermann and Cohen, 2008).
Roach and Sauermann (2008) raised the question if corporate scientists (R&D personnel having a PhD) have a weaker inclination to engage in collaborative relations than academic scientists. They found that corporate scientists are more inclined to be engaged in research involving downstream R&D than academic scientists. They also value the salaries and resources for research more than academic scientists.
The level of qualification is also thought relevant in the concept of absorptive capacity; where a positive relation between the two is assumed (Azagra-Caro et al., 2006; Roper et al., 2008). Based on the previous discussion the second hypothesis can be formulated.
Hypothesis 2 The higher the share of highly qualified employees; the higher the probability to cooperate to create new knowledge or exchange knowledge developed in the firm.
2.3Using intellectual property in external knowledge relations
Intellectual property (IP) catches all protection mechanisms that enhance a willingness of firms to undertake risky R&D activities through cooperation (Gulati and Singh, 1998). By protecting the innovation firms hope to capture the benefits attached to it (Chesbrough, 2003). This might be through preventing the innovation from being imitated and functioning as a unique selling point for a while; or from licensing the innovation to other firms and creating additional revenue through this channel (Arora et al., 2001).
The returns from R&D activities can be negatively influenced by intellectual property because it reduces competitive elements, augments market prices and prevents a smooth technology transfer of the most recent technologies. Intellectual property is also highly industry dependent (Granstrand, 2003).
The use of IP mechanisms, such as cross licensing, also serves to have an impact on industrial standards; gaining access to technical portfolios of other firms; increasing innovation speed and freedom of operation (Lichtenthaler, 2005). From these elements a third set of hypotheses can be derived.
Hypothesis 3a The expectation to register or obtain patents stimulates the probability to develop knowledge in formal cooperation agreements
Hypothesis 3b The expectation to gain revenues from inventions stimulates the exchange of knowledge to external parties.
2.4Conceptual model
All variables in the preceding paragraphs, together with the control variable are described in Table A1 in the Appendix. Figure 1 recapitalises the conceptual model and the hypothesised impacts of the independent variables.
Figure 1 Conceptual model on external knowledge relations
3 Data description and variables
3.1Data and sample
The analysis makes use of the biannual OECD R&D business survey organised in Belgium in 2008. This survey starts from a population of 5,270 firms of which 2,083 stating that in 2006 and 2007 they performed R&D activities occasionaly (302) or in a permanent (1,781) fashion. Not all of these firms engage in external knowledge relations. Deleting the missing values reduces the total observations for which we have information to a sample of 645.
Table 1 gives an overview of the information on all variables for the entire sample and the reduced sample we are about to use for the analysis.
Table 1 Descriptive statistics: mean and standard deviation in the sample of R&D active firms and the sample without missing values (N=645)
Variable / Obs. / Mean / Standard deviation / Obs. / Mean / Standard deviation / Difference in mean, unequal variances (p-value)Dependent variables
COOP / 737 / 0.385 / 0.487 / 645 / 0.391 / 0.488 / 0.84
EXCH / 736 / 0.363 / 0.481 / 645 / 0.366 / 0.482 / 0.90
Independent variables
a. Function of R&D personnel
FRESMAN / 712 / 0.555 / 0.354 / 645 / 0.566 / 0.348 / 0.57
FTECHNIC / 711 / 0.356 / 0.337 / 645 / 0.350 / 0.332 / 0.73
FSUPPORT / 711 / 0.088 / 0.179 / 645 / 0.084 / 0.172 / 0.68
b. Qualification of R&D personnel
QPHD / 678 / 0.113 / 0.213 / 645 / 0.116 / 0.216 / 0.79
QMASTER / 678 / 0.507 / 0.347 / 645 / 0.504 / 0.346 / 0.87
QBACHEL / 678 / 0.225 / 0.272 / 645 / 0.224 / 0.273 / 0.98
QOTHER / 678 / 0.155 / 0.257 / 645 / 0.155 / 0.258 / 0.99
c. Output variables
INNOPROD / 765 / 0.827 / 0.378 / 645 / 0.839 / 0.368 / 0.57
IPACQ / 777 / 0.147 / 0.354 / 645 / 0.155 / 0.362 / 0.66
IPREG / 775 / 0.268 / 0.443 / 645 / 0.273 / 0.446 / 0.85
IPEARN / 774 / 0.124 / 0.330 / 645 / 0.136 / 0.344 / 0.49
Control variables
LEMP / 2002 / 3.749 / 1.727 / 645 / 3.761 / 1.811 / 0.89
GP / 1186 / 0.557 / 0.497 / 645 / 0.550 / 0.498 / 0.76
IDCP / 2064 / 0.224 / 0.417 / 645 / 0.222 / 0.416 / 0.91
IDCS / 2064 / 0.017 / 0.129 / 645 / 0.019 / 0.135 / 0.79
IDFP / 2064 / 0.068 / 0.252 / 645 / 0.078 / 0.268 / 0.44
IDPE / 2064 / 0.223 / 0.417 / 645 / 0.217 / 0.413 / 0.74
IDSB / 2064 / 0.092 / 0.289 / 645 / 0.112 / 0.315 / 0.16
IDKI / 2064 / 0.305 / 0.461 / 645 / 0.310 / 0.463 / 0.82
IDOS / 2064 / 0.070 / 0.256 / 645 / 0.043 / 0.204 / 0.006***
RDOUT / 2050 / 0.584 / 0.376 / 645 / 0.482 / 0.500 / 0.000***
RDINT / 2002 / 0.202 / 0.493 / 645 / 0.231 / 0.301 / 0.045**
RSHR / 743 / 0.551 / 0.362 / 645 / 0.555 / 0.356 / 0.83
Table 1 indicates that only in the case of three control variables the two samples differ markedly. First, in the case of the industry dummy on other services, this category of firms is underrepresented in the reduced sample since the share in the full sample is 7% whereas this is 4.3% inthe case of the reduced sample.
3.2Variables
Three groups of variables are discussed: dependent, independent and control variables. The dependent variables cover the external knowledge relations firms engage in: formal collaborative agreements (COOP) and the exchange of knowledge that was generated in the framework of such an agreement (EXCH).
The independent variables can be grouped into three aspects that have an impact on the dependent variables. First, the OECD (2002) acknowledges different functions that employees classified as R&D personnel can have: researchers and R&D managers (FRESMAN); technicians (FTECHNIC) and other supporting staff (FSUPPORT). Researchers and R&D managers are involved in the conception and generation of new knowledge, products, processes, methods and systems. Also the management of the R&D projects falls under this heading. Technicians and equivalent staff perform tasks that demand technical knowledge and experience. These employees are involved in the application of the new knowledge and are usually under the supervision of R&D managers. Other supporting staff captures the secretarial, clerical employees and craftsman that participate in R&D projects or are closely related to these projects.
Second, the OECD (2002) classifies the R&D personnel using the formal qualification. Four academic qualifications are used: doctorate holders (QPHD); holders of basic university degrees at masters (QMASTER) and bachelor (QBACHEL) level; and other qualifications below university level (QOTHER). Abstraction is made of the field of science in which these degrees have been obtained.
Both the functions and qualifications of R&D personnel have been measured in terms of full time equivalents.
The third group of independent variables are related to the output of the R&D activities. Under normal circumstances the most obvious output are innovative products that the firms aim to bring to market (INNOPROD). Other types of output are more related to enhance the knowledge used within the company. In this case firms can acquire patents, licences or other intellectual property rights (IPACQ). Firms can also register patents (IPREG). Or firms can sell patents, licences and other intellectual property rights to third parties and earn money from them (IPEARN). This last activity is found to be popular in the context of open innovation (Chesbrough, 2003) and requires good functioning technology markets (Arora et al., 2001)
Six control variables are used in the analysis because they, too, have potential impacts on the external knowledge relations firms’ maintain. First, the size of the firm is often used. The size is measured by the natural log of the total employment of the firm. The effect is somewhat unclear, since on the one hand smaller firms are assumed to lack the human resources to engage in external knowledge relations. On the other hand, the need for additional information is making itself more felt in the case of small firms.
Second, the firm that is part of a group might rely on an internal division of labour that is directed to source relevant knowledge in host countries and apply this knowledge in their home countries. Also, subsidiaries might be more inclined to look for external information which helps tem to enter or integrate in local markets.
Third, a set of industry dummies accounts for the fact that R&D activities are known to differ according to the industry firms are operating in (Hatzigronouglou, 1997). Likewise, aspects such as patenting are also known to differ across industries (Griliches, 1990). This paper relies on the industry classification by Marsili and Verspagen (2002). Drawing on Pavitt’s (1984) taxonomy they use empirical data (such as patents, R&D statistics, scientific inputs, innovation surveys) to develop a typology of five regimes. These regimes are based on a set of criteria: the nature of the knowledge base; technological opportunity conditions; and technological entry barriers. Marsili and Verspagen (2002), like Pavitt (1984) only consider manufacturing, which is inadequate in the Belgian context for it is highly specialised in services. Therefore two additional regimes are added, the knowledge intensive services and other services, in order to cover the services industry.
The R&D intensity, which is an indicator that reflects the technological needs of the firm, is an important element in the context of open innovation. Cassiman and Veugelers (2006) already pointed to the complementarity between internal R&D efforts and the use that can be made from external knowledge relations. Cohen and Levinthal (1990) further highlighted the use of R&D intensity in the development of absorptive capacity by internal employees in screening and using external knowledge flows. Following Caloghirou et al. (2004), R&D intensity is measured in terms of R&D personnel.
The fifth control variable, on R&D outsourcing, captures the fact if firms engage in subcontracting R&D to third parties. This activity also indicates if a firm is open to external knowledge flows, but buying R&D results is obviously requiring less intensive involvement of the internal R&D personnel than the other external knowledge relations that have been implied by a formal collaborative agreement.
Finally, a sixth control variable is on the ratio of fundamental and applied research to development. This distinction is less relevant in recent times. It might be, however, that external knowledge relations are more directed to the development of a specific product or targeted at a well defined result. All these variables are succinctly defined in Table A1 in the Appendix.
4 Analysis
The undertakings of collaboration and the exchange of knowledge devised in the framework of the technological collaboration agreement are not independent to each other, these actions are, indeed, correlated. The binary dependent variables, COOP and EXCH, appear to have four mutually exclusive cases: (COOP, EXCH) = (0,0), (1,0), (0,1), and (1,1) respectively.