Unionization structure, licensing and innovation

Abstract

We show the effects of the unionization structure (viz., decentralized and centralized unions) on a firm's incentive for technology licensing and innovation. The incentive for technology licensing is stronger under decentralized unions. We identify circumstances under which the benefit from licensing creates a stronger incentive for innovation under decentralized unions. If the union's preference for employment is high, the benefit from licensing may create higher incentive for innovation under decentralized unions. However, if the union's preference for wage is high enough, the incentive for innovation is higher under a centralized union irrespective of licensing ex-post innovation. If the centralized union decides whether or not to supply workers to all firms, the possibility of higher innovation under decentralized unions increases. We further show that perfectly substitutable workers can be better off under decentralized unions if the labor productivity depends on the unionization structure, which occurs in our analysis when, e.g., licensing after innovation occurs only under decentralized unions or innovation (with no licensing) occurs only under a centralized union.

Introduction

Labor unions differ substantially between countries with respect to the degree of wage setting centralization (Calmfors and Drifill, 1988, Moene and Wallerstein, 1997, Flanagan, 1999, Wallerstein, 1999). Decentralized wage setting is often contrasted with centralized wage setting. Under a decentralized wage setting, wages are set between an employer and a firm-level union, while under a centralized wage setting, an industry-wide union negotiates wages with all firms (Haucap and Wey, 2004). While the centralized argument is egalitarian in nature and generally makes the workers better off if the workers are sufficiently substitutable (Horn and Wolinsky, 1988, Davidson, 1988), the rigidity associated with this system is generally bad for overall economic performance (Nickell, 1997, Siebert, 1997). The rigidity associated with a centralized wage setting has been attacked in recent years and tendencies towards a more decentralized wage setting is gaining popularity in the policy circle. For example, countries such as Sweden, Australia, the former West Germany, Italy, the UK and the USA move towards a more decentralized unionization structure, as documented in Katz (1993). The OECD Jobs Study recommends making the wages and labor costs more flexible to reflect local conditions (OECD, 1996, p. 15). The trend over the past decades towards more decentralized unions can also be found in OECD (2004). Haucap and Wey (2004) provide evidence for a move towards a more decentralized wage setting.

Given the diversity in labor market unionization and the popularity towards a more decentralized wage setting regimes, the purpose of this paper is to examine the effects of the unionization structure on unions' utility highlighting the effect of the unionization structure on innovation and technology licensing.

We consider an economy where a firm innovates to reduce its labor requirement, and has the option to license¹ the invented technology to a non-innovating firm.² Assuming a given technology difference between the firms, we first show the incentives for technology licensing under decentralized and centralized unions. This setting is relevant if the technologically advanced firm innovates irrespective of the unionization structure, while its incentive for technology licensing may depend on the unionization structure. We show that the technologically superior firm has a stronger incentive for technology licensing under decentralized unions compared with a centralized union.

Next, we examine the incentives for innovation under different unionization structures. First, note that the labor unions use wage mark-ups to extract rents from the firms. Ex-post innovation and/or licensing, the unions adjust their wages to extract the gains from innovation and licensing. On one hand, the higher wage charged by the unions encourages the firms to engage in more labor saving activities through innovation and licensing. On the other hand, the rent seeking behavior of the unions reduces the firm's incentive for investing in innovation or engaging in technology licensing. This rent extracting behavior of the unions acts as a tax on the firms and increases their costs. Thus, it holds up the firms from investing in innovation or engaging in technology licensing by reducing returns from these activities. However, the strengths of the hold-up problems depend on the unionization structure and on the unions' preferences for employment and wage.

The hold-up problem under decentralized unions is constrained by competition between the unions, while the hold-up problem under a centralized union is often constrained by the uniformity rule of wage determination. While determining the incentive for licensing ex-post innovation, we find that competition between the unions under decentralized unions is more effective in softening the hold-up problem, thus creating a stronger incentive for licensing under decentralized unions. Hence, there are situations where licensing occurs only under decentralized unions. The gain from licensing tends to increase the incentive for innovation under decentralized unions by reducing the negative effects of the hold-up problem under decentralized unions.

If the unions' preferences for wages are not very strong, hold-up problems are mild. In this case, the gain from licensing may be strong enough to compensate the hold-up problem under decentralized unions to an extent that the incentive for innovation can be higher under decentralized unions. If either the unions' preference for wage are strong or there is no licensing, the uniformity rule under a centralized union is more effective than competition between the unions to soften the hold-up problems. As a result, the incentive for innovation is higher under a centralized union in these situations. However, if a centralized union decides whether or not to supply workers to all firms, it increases the hold-up problem under this type of unionization structure. In this situation, the possibility of higher innovation under decentralized unions increases.

Combining the results on innovation and licensing, we show that even the perfectly substitutable workers can be better off under decentralized unions if the firms' labor productivities depend on the unionization structure (thus internalizing the effects of innovation and licensing), which occurs in our analysis when, e.g., licensing after innovation occurs only under decentralized unions or innovation (with no licensing) occurs only under a centralized union.³

The contributions of our paper are then threefold. First, it contributes to the vast literature on technology licensing, which has generally ignored the impacts of input markets on technology licensing, thus implicitly assuming perfectly competitive input markets. We show that the incentive for licensing is significantly affected if the input supplier (which is labor union in this paper) has market power. Moreover, the structure of the input market (viz., firm-specific or an industry-wide) plays an important role.⁴

Second, it contributes to the literature showing the effects of the unionization structure on innovation. While the earlier works⁵ have shown the impact of union bargaining power on the incentives for innovation under decentralized unions, recently, Calabuig and Gonzalez-Maestre, 2002, Haucap and Wey, 2004 show the (ambiguous) effects of different unionization structure on innovation in R&D competition and patent race models. The main intuition behind their results is related to the different types of constraints under different unionization structures, which affect the hold-up problems. Haucap and Wey (2004) show that the uniformity rule under a centralized union is more effective in constraining the unions' hold-up potential and leads to higher incentives for innovation.⁶ Calabuig and Gonzalez-Maestre (2002) show how the hold-up problems are affected by the market size, which may make production by the non-innovating firm unprofitable. In our analysis, hold-up problems not only affect the incentives for innovation, but also affect the incentives for licensing. The uniformity rule under a centralized union is more restrictive for technology licensing, and creates higher incentive for licensing under decentralized unions. Whether this gain from licensing under decentralized unions is strong enough to create stronger innovation incentives under decentralized unions depends on the unions' preferences for employment and wage, which affect the equilibrium wage and the extent of the hold-up problems.

Third, our analysis sheds light on the literature examining the preference for a particular unionization structure. In contrast to the earlier works (e.g., Horn and Wolinsky, 1988, Davidson, 1988), we show that homogeneous workers can be better off under decentralized unions than under a centralized union if the unionization structure affects the incentive for innovation and licensing, thus affecting the cost structures, which affect the intensity of product market competition.

Though our paper is theoretical in nature, it is worth discussing the empirical evidences on the effects of union on innovation. There is no doubt that the level of wage bargaining, the bargaining agenda and the bargaining power distribution between firms and unions are among the important determinants of innovation (Bassanini and Ernst, 2002, Hirsch, 2004), yet there is controversy about their exact effects on firms' performance, innovation and labor productivity (see, Flanagan, 1999, for a survey on this topic).

The existing works examining the effects of union on innovation mainly show the effects of union power on the incentives for innovation. Freeman and Medoff (1984) show that the effect of unionization is ambiguous on innovation. Using COMPUSTAT data, Bronas and Deere (1993) show that there is a significant negative relationship between firm-specific unionization rate and innovation. Using mainly aggregative industry level data, Ulph and Ulph (1989) find a negative relation for the high-tech industries in England, while Addison and Wagner (1994) find a positive but insignificant relation. It is documented in Menezes-Filho et al. (1998) that most U.S. studies show a negative effect between union power and innovation, while the evidence from some European studies is less compelling.⁷ Menezes-Filho and Van Reenen (2003) also show strong and negative effects of unions on innovation in North America, while that is generally not the case in the UK.

Although the existing empirical works mainly show the effects of stronger unions, instead of different unionization structures, the implied hold-up problems under different union powers may have similar implications for different hold-up problems under different unionization structures. In line with the empirical evidence, our results show that the effects of different hold-up problems under different unionization structures are not straightforward in determining their effects on innovation. Moreover, as we show, the relative effects of different unionization structures on innovation may vary in the presence of other business strategies such as technology licensing. Hence, it will be worth investigating the effects of the unionization structure on innovation while controlling for technology licensing.

The remainder of the paper is organized as follows. 2 Decentralized and centralized unions: licensing, 3 Decentralized and centralized unions: innovation respectively consider the incentives for licensing and innovation under decentralized and centralized unions. Section 4 looks at the utility of unions under both unionization structures. Section 5 considers two extensions of the basic model. Section 6 concludes.