It is well-established that labor market features, such as the level,1 the agenda and the pattern of employers-employees negotiations, the bargaining power distribution between firms and unions, and the labor market institutions (minimum wages, unemployment benefits, etc.), are amongst the crucial determinants of firms’ innovation activities (see e.g. Hirsch 2004).
In his seminal paper, Grout (1984) firstly introduced the “hold-up” argument in the litera- ture. In a “one firm-one union” framework, Grout argues that in the absence of legally binding contracts, once the firm has incurred the sunk costs of investment, its union has incentives to extract a portion of the quasi-rents created by the firm’s investment through higher wage demands. The union’s hold-up behavior, in turn, leads the firm to underinvest. The higher the union’s power is, the lower are the firm’s incentives to innovate. However, unionization is not always associated with underinvestment. In a patent race for a labor savings innovation, Tauman and Weiss (1987) show that a unionized duopolist has stronger incentives to adopt the new technology than its non-unionized counterpart. Ulph and Ulph (1994, 1998), in a duopoly where firms bargain with their firm-level unions over employment and wages (“Efficient Bar- gaining”), show that a more powerful risk-averse union may encourage its firm to overinvest in order to win the patent race for a cost-reducing innovation. More recently, Calabuig and Gonzalez-Maestre (2002) show that if the market size is small, an industry-wide union provides stronger incentives for a firm to win a patent race than a firm-level union. Moreover, Haucap and Wey (2004) show that innovation incentives are not monotonic in the degree of centralization of wage-setting. Innovation incentives are the strongest when an industry-wide union sets a uniform wage rate, they are the weakest under an industry-wide union coordinating, via wage discrimination, its wage demands, while they lie in-between under perfectly decentralized firm-level union wage setting. Finally, these theoretical findings are, to a major extent, supported by the inconclusive empirical evidence on the impact of the unionization on the firms’ incentives to innovate. Menezes-Filho and Van Reenen (2003), surveying the bulk of the empirical literature, conclude that, although there are consistently strong and negative impacts of unions on R&D expenditures in North America, this is not the case for Europe where no such clear pattern can be reached. In addition, Hirsch (2004) concludes that the existing empirical evidence does not allow us to establish, or reject, causal union effects on R&D investments.
Although the impact of labor market features on firms’ R&D investments have been extensively addressed in the literature, some key features regarding the nature and the organization of R&D activities, along with their interplay with the labor market features, have been ignored so far. Firstly, the spillover effects of a firm’s R&D activities that may cause underinvestment problems. And secondly, the organizational mode of R&D investments, i.e. whether firms invest non-cooperatively, or cooperatively by forming a Research Joint Venture (RJV). As Vonortas (1997) notes, RJVs are regarded as “the cure for a number of failures in innovation markets” as far as spillovers are internalized and thus incentives for R&D investments are restored.
This paper aims to fill this gap by reconsidering the role of unions for the firms’ incentives to invest in cost-reducing R&D activities when R&D spillovers are present and firms have the option to form an RJV. Our envisaged model is a homogeneous unionized Cournot duopoly, where firms can invest in cost-reducing R&D activities before adjusting their quantities in the market. Each firm’s R&D output partially flows to its rival, contributing to the latter’s unit cost reduction. Workers are organized either in two firm-level unions (decentralized unionization structure) or in an industry-wide union (centralized unionization structure). After firms have chosen their R&D expenditures and before the market competition stage, firm-level unions set their firm-specific wage rates, or else the industry-wide union sets the uniform wage. At an initial stage, firms have the option to form an RJV in order to invest cooperatively in the next stage, or to stay separately. We thus extend d’Aspremont and Jacquemin (1988) by adding alternative unionization structures to their non-union model.
In the above setup we address the following three questions. First, how does the presence of R&D spillovers affect firms’ R&D investments, employment and output levels, firms’ profits and social welfare? Second, how does the unionization structure affect equilibrium market outcomes and welfare under alternative organizational forms of the R&D activity (RJVs or non-cooperative R&D spending)? Third, how does the unionization structure affect the firms’ incentives to form an RJV, and in particular when the formation of the RJV is costly?