Endogenous R&D networks when labour unions have preferences over wages and employment

Abstract

We develop a model to analyse the pattern of R&D network formation when unions have relative preferences over wages and employment. Within a three-firm industry, we show that when the unions place a low weight on wages and technological spillovers are low, a partial R&D network that includes two firms but excludes the third emerges in equilibrium. In contrast, when the unions care a lot about wages, a complete R&D network that includes all firms emerges. For all other intermediate levels of union preferences over wages, there is no strong stable equilibrium network. Empirical implications emerge from these findings, which are also discussed.

Keywords:

• labour unions
• R&D collaboration
• wages
• innovation

JEL Classification:

• L13
• J50

Acknowledgements

We would like to thank Cristiano Antonelli (the Editor) and two anonymous referees. An earlier version of the paper was presented at the 2013 Royal Economic Society Conference (Royal Holloway, UK). The usual disclaimer applies.

Disclosure statement

No potential conflict of interest was reported by the authors.

Funding

Zikos acknowledges financial support from the University of the Thai Chamber of Commerce under its grant scheme.

Notes

1. For more details see, for example, http://www.birminghammail.co.uk/news/midlands-news/jaguar-land-rover-union-bosses-7812147http://www.birminghammail.co.uk/news/midlands-news/jaguar-land-rover-union-bosses-7812147.

2. Of course, research networks are an increasingly common academic phenomenon too. See Paier and Scherngell (2011) and Scherngell and Barber (2009) who analyse European framework initiatives.

3. Issues related to R&D cooperation within an oligopoly have been a central issue in the literature on Research Joint Ventures (RJVs); see for example, d' Aspremont and Jacquemin (1988), Poyago-Theotoky (1995 1999) and Kamien, Muller, and Zang (1992). In this paper we adopt instead the ‘network approach’, which treats network formation as an endogenous decision, allowing us to accommodate networks that include only a subset of firms in the industry. Most importantly, our focus on R&D networks is supported by empirical evidence documenting a substantial increase of non-equity alliances (such as R&D networks) relative to equity forms (such as RJVs), reaching an overall share of more than 85% by the mid-1990s (Caloghirou, Ioannides, and Vonortas 2003).

4. Goyal (2009) provides an excellent survey of R&D network models (see Chapters 3 and 10).

5. A notable exception is Manasakis and Petrakis (2009). They develop a model of a duopoly market and show that under different unionisation structures firms' incentives to form RJVs are sensitive to the extent of (public) spillover effects.

6. Union/firm bargaining may actually take place at a number of levels. Our assumption, however, reflects the empirical observation that such negotiations are often formally conducted at the level of each individual enterprise: Du Caju et al. (2008) note that ‘company level’ bargaining is ‘very usual’ in 23 European countries, the US and Japan (p. 15). Indeed, even where a firm is not fully unionised, the wages negotiated with its unionised workforce will often determine those for their non-unionised counterparts (Visser 2006).

7. The monopoly union model is adopted for simplicity. We expect our results would be qualitatively similar in the alternative framework where each union-firm pair negotiates over its wage. This is because high (low) bargaining power of the unions increases (reduces) the wage effects within a given network, suggesting a qualitatively similar mechanism to a high (low) level of union preferences over wages, β, in our model.

8. This corresponds to a reduction of the capital/labour ratio within a Leontief production function.

9. Our results would be qualitatively similar in the alternative framework where R&D is aimed at enhancing productivity of the labour input, given that the production technology exhibits constant returns to scale (so that ).

10. Linear product demand is typical in existing R&D network models. In this vein, we also retain the assumption of homogeneous goods as it allows us to focus sharply on the trade-off between inter-firm ‘cooperation’ (at the network formation stage) and ‘competition’ (in the product market). This trade-off, which is consistent with evidence from OECD (2001), drives the firms' network formation decisions in the present setting.

11. Because strong stability is a refinement of pairwise stability, a strongly stable network is necessarily pairwise stable.

12. We use ‘Mathematica 8’ (see e.g. Wolfram 1999) for the figures, and set and , which is inconsequential in a qualitative sense.

13. The hold-up argument suggests that cost-reducing investments have the effect of lowering marginal costs, raising the demand for labour and thus increasing wages as unions attempt to capture a share of the quasi-rents generated by R&D activity.

14. The analysis for imperfect private spillovers, given that public spillovers are absent, is analogous.

15. Goyal and Moraga-González (2001) focus on pairwise stable networks.

16. In spite of a relatively high failure rate, as Narula and Duysters (2004, p. 202) point out, lasting collaborative agreements in R&D ‘stand out in terms of their effectiveness’.