«Reaching for the stars: The Contingent Performance Effects of Basic Research Collaboration Relinde Maria Colen KU Leuven MSI linde.colen ...»
Paper to be presented at the
DRUID Society Conference 2014, CBS, Copenhagen, June 16-18
Reaching for the stars: The Contingent Performance Effects of Basic
Relinde Maria Colen
Managerial Economics, Strategy and Innovation
KU Leuven (HUB)
Onderzoeksgroep globalisering, innovatie en competitie
firstname.lastname@example.org Abstract This paper explores the role of university scientists as firms? key partners in basic research. More specifically, it analyses under which conditions joint basic research with university ?star? scientists improves firms? technological performance. Quantile regression analysis for 153 of the most R&D intensive firms in the biopharmaceutical sector in 1995-2003 shows that the benefits of collaboration with university star scientists are highly skewed and concentrated among the top performing firms. These firms? innovative performance is particularly enhanced if the firm and the star scientist also collaborate on applied research. The benefits of this ?translational? collaboration are reinforced when the firm has secured (temporary)exclusive access to the university star scientist. In contrast, when the firm and star do not perform applied research jointly, exclusive access has a negative effect on the firm?s innovative performance. We discuss the managerial implications of our findings for firms seeking to build productive basic research partnerships with university star scientists.
Jelcodes:D83,L29 Reaching for the stars The Contingent Performance Effects of Basic Research Collaboration between Firms and University Star Scientists Abstract This paper explores the role of university scientists as firms’ key partners in basic research.
More specifically, it analyses under which conditions joint basic research with university ‘star’ scientists improves firms’ technological performance. Quantile regression analysis for 153 of the most R&D intensive firms in the biopharmaceutical sector in 1995-2003 shows that the benefits of collaboration with university star scientists are highly skewed and concentrated among the top performing firms. These firms’ innovative performance is particularly enhanced if the firm and the star scientist also collaborate on applied research. The benefits of this ‘translational’ collaboration are reinforced when the firm has secured (temporary) exclusive access to the university star scientist. In contrast, when the firm and star do not perform applied research jointly, exclusive access has a negative effect on the firm’s innovative performance. We discuss the managerial implications of our findings for firms seeking to build productive basic research partnerships with university star scientists.
Keywords: basic research, innovation, star scientists, pharmaceutical industry
A large body of evidence supports the importance of basic scientific research for advancing economic growth and welfare (Mansfield, 1980; Griliches, 1986; Jaffe, 1989; Adams, 1990;
Salter and Martin, 2001; Toole, 2012). Basic research can be defined as activities that are directed towards the general advancement of men’s knowledge about the physical world, but without specific commercial objectives (Nelson, 1959). Most basic research is sponsored by governments, and conducted at knowledge institutes, most notably universities. Basic research activities expand the knowledge base available for firms on which they can draw in their applied technological activities (Klevorick et al, 1995). Numerous important technical inventions were the direct result of advances in scientific knowledge resulting from basic research at universities. Mansfield (1995) found that around 11% of firms’ new products and around 9% of new processes could not have been developed (or with a substantial delay) in the absence of basic research conducted by universities. Even higher numbers were obtained for the period 1986-1994 (respectively 15% and 11%), suggesting that basic research has increased in importance for industrial R&D over time (Mansfield, 1998).
In general, knowledge interactions between universities and firms have been growing in scale and scope over time (Perkmann and Walsh, 2007; Du et al., 2014), arguably to benefit from each other’s expertise and network. Especially the “star” university scientists, defined as highly productive individuals who are leading researchers in their fields, receive special interest from firms (Zucker and Darby, 1996). By means of collaboration, firms get in close contact with university star scientists, giving them direct access to unique (tacit) knowledge (Rothaermel and Hess, 2007), codified scientific knowledge that is not yet published, such as work in progress (Fabrizio, 2009), and a large social network within the scientific community (Murray, 2004; Subramaniam, 2013).
While most basic research is concentrated at universities, there are also good reasons for firms to conduct in-house basic research. By performing basic research, firms can increase their understanding of the technological landscape in which they search for inventions (Rosenberg, 1990; Fleming and Sorenson, 2004), hire researchers that are reluctant to work for firms in which they are not allowed to do basic research and publish scientific findings (Henderson and Cockburn, 1994; Hicks, 1999), improve their absorptive capacity for external research 2 (Gambardella, 1992; Leten et al., 2011) and get an access ticket to R&D partnerships with universities (Liebeskind et al., 1996).
Prior studies have confirmed that the benefits of in-house basic research are greater when basic research is conducted in collaboration with university scientists (Henderson and Cockburn, 1998; Fabrizio, 2009) and several studies have documented positive performance effects of firm-university collaboration in general (Furman and MacGarvie, 2009; Belderbos et al., 2004 and 2014; Du et al. 2014). In addition, studies have established that the positive impact of research universities on nearby firms is due to specific collaborations with star scientists rather than generalized knowledge spillovers of the university (Zucker et al., 1998).
It has been shown that the number of research collaborations between firms and university star scientists has a positive effect on the number and average quality of firm innovations (Zucker and Darby, 2001; Zucker et al., 2002). While these studies have suggested that collaborating with academic star scientists is likely to result in an improved innovation performance for the firm, little is known about the contingencies for these effects to occur.
Since basic research entails large investments while outcomes are uncertain, it is crucial to understand which modalities of these partnerships are prone to increase the benefits of starfirm collaborations.
This paper contributes to our understanding of how excellence in science strengthens firms’ innovativeness. We investigate two structural characteristics of basic research collaborations with star scientists. First, we examine whether star scientists assist the firm in (follow-up) applied research, taking up a ‘translational’ role between basic research and technological applications. To succeed in the hard translational step from basic research results to commercial development, it is crucial to understand the two worlds and recognize opportunities to link science and technology. Second, we examine whether the firm may benefit from (temporary) exclusivity in the collaboration with star scientist as this mitigates the spillovers of what is arguably unique and highly specialized knowledge to rival firms who may free ride on these research efforts. In sum, we expect that the translational character of the scientist and the exclusiveness of the relationship will act as important moderators for the effectiveness of the star-firm collaboration in basic research, as measured by the innovative performance of the firm.
To test our hypotheses, we make use of a panel dataset on the patent and publication activities of 153 leading innovating pharmaceutical and biotechnology firms. The sample firms have 3 headquarters in the United States, EU and Japan and their technological performance is observed over a period of nine years (1995-2003). Information on scientific publications in the Web of Science database is used to examine firms’ basic research activities and collaborations with university star scientists. Star scientists are identified as individuals leading their field(s) in terms of the number of scientific publications and citations. We use quantile regression analysis to explain firms’ yearly innovative performance, measured by citation-weighted patent counts, as a function of basic research collaboration with university stars, accounting for the aforementioned modalities of these partnerships and controlling for a range of productivity determinants.
We find that basic research collaboration with stars increases inequality in technological performance across firms since the effect of such partnerships only affects the upper quantiles of the innovation performance distribution. Collaborating with ‘translational’ stars – who also engage in joint applied work with the firm – entails an additional performance premium, again at the upper end of the distribution. The firm accumulates further benefits if it manages to temporally secure exclusive access to the translational star. Interestingly, we find that higher rates of exclusivity to non-translational stars – who only do basic research with the firm – suppress the citation-weighted patent output of the most productive firms.
The remainder of this paper starts by grounding our analysis in two related streams of innovation research, i.e. innovation studies on the role of basic research, and the economics of science literature dealing with star scientists. In section three we discuss the rationale behind our hypotheses. Section four introduces the data and explains our empirical methodology. The empirical results are discussed in section five and in the final section we recap the main findings and their implications, summarize work in progress and suggest avenues for further research.
2. Literature and Background Basic Research and Firm Performance Innovation studies have delivered abundant evidence of the important role of basic research in driving innovation and economic growth (Mansfield, 1980; Griliches, 1986; Jaffe, 1989;
Adams, 1990; Salter and Martin, 2001; Toole, 2012). To what extent it is rational for firms to 4 be involved in basic research has however been subject to a long debate among economists.
Nelson (1959) argues that firms are reluctant to invest in basic research due to high degrees of uncertainty, long time frames to bear fruit, and appropriability problems. More recent contributions (Rosenberg, 1990; Gambardella, 1992; Fleming and Sorenson, 2004; Cassiman et al., 2008), on the other hand, argue that there are good reasons for private firms to invest inhouse in basic scientific research.
A first reason for firms to engage in basic scientific research is to develop a deeper understanding of the technological landscape in which they search for inventions (Gambardella 1992). Basic scientific knowledge allows firms to anticipate the results of research experiments without performing them, helping to prioritize research avenues and to avoid costly research trials that lead to low-value outcomes (Rosenberg, 1990; Fabrizio 2009).
In other words, basic scientific knowledge informs firms on the success probabilities of different directions to conduct applied research (Fleming and Sorenson, 2004; Cassiman et al.
2008). Second, it also helps to evaluate the outcomes of applied research and to get a more accurate and encompassing perception of its implications (Rosenberg, 1990). Further in-house basic research helps to recruit scientists that prefer to work in firms that conduct in basic research and where they can publish their research findings (Henderson and Cockburn, 1998;
Hicks, 1999). Finally, in-house basic research generates absorptive capacity: it leads to knowledge and skills required to understand and utilize the findings of basic research conducted elsewhere, most notably at universities (Gambardella 1992, Leten et al. 2011).
Prior empirical studies (e.g. Gambardella, 1992; Cockburn and Henderson, 1998; Fabrizio, 2009; Leten et al., 2011, with the exception of Lim, 2004) have found positive performance effects of conducting in-house basic research. Using samples of US pharmaceutical firms, Gambardella (1992) and Cockburn and Henderson (1998) found that firms which perform more basic research, measured by the number of firm publications, produce a greater number of patented inventions. Similar findings were obtained by Leten et al. (2011), using a global sample of pharmaceutical firms and a more accurate indicator of basic research, i.e. the number of publications in basic research journals. In contrast to these studies, Lim (2004) found no effect of in-house basic research on the patent performance of pharmaceutical firms, and even a negative effect for semiconductor firms. Della Malva et al. (2013) examined whether the own pursuit of basic research puts firms in a better position to generate breakthrough inventions which have a large impact on subsequent inventive activities. They found that the involvement in basic research helps firms to generate breakthrough inventions.
5 Surprisingly, the benefits of basic research are not found in the technology fields that are directly associated with the basic research, but in other areas of the firms’ technology portfolio. This result is consistent with a view of basic research as a “map” that guides firms into new research directions (Rosenberg, 1990; Fleming and Sorenson, 2004).