همکاری در فعالیت های R & D: تصمیم گیری شرکت های خاص
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|10269||2008||20 صفحه PDF||سفارش دهید||12610 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : European Journal of Operational Research, Volume 185, Issue 2, 1 March 2008, Pages 864–883
In this paper, we study the strategic R&D collaboration by introducing a virtual player to reveal cooperative incentives and keeping investment share and market share independent of each other. Not consistently with the traditional opinions, we show that the superiority of the R&D cartel is due to the coexistence of cooperation and competition when spillovers are exogenous. Moreover, we conclude that high R&D input share must be reflected implicitly by high market share, and that firms’ R&D decisions vary with firms’ specific characteristics when spillovers is endogenous.
Collaboration plays an important role in research and development (R&D) activities, and has spurred research into their effects on firms’ strategic decisions. The advantages of collaboration in R&D are outstanding, especially in high-tech firms: the collaboration helps firms to crack new markets, gain skills and technologies, realize economies through reorganization and exploitation of complementarities, share costs and risks, and control competitive forces (Veugelers, 1998). A famous model for R&D collaboration mixing with competition in the literature is the two-stage game presented by D’Aspremont and Jacquemin (1988). In their model, firms cooperate in the first stage (R&D collaboration) while compete in the second stage (market competition), and the outputs of firms’ post-innovation products in the second stage only depend on the R&D investment (expenditure) decisions in the first stage. By extending this model, the mainstream of the literature in Industrial Organization (IO) find that spillovers increase the relative profitability of the R&D collaboration once spillovers are sufficiently high (D’Aspremont and Jacquemin, 1988, De Bondt, 1996 and Kamien et al., 1992; etc.). Furthermore, the cartelized RJV (research joint venture) is the best R&D scenario; at least for industries with nearly firms with symmetric constrains (Amir et al., 2003). The two-stage game models bring convenience to compare the equilibria or the social welfare among different R&D scenarios defined as Kamien et al. (1992). But they suppose that firms in the alliance contract on the future market. Indeed, they ignore the fact that such formal and complete interfirm contracts are extremely difficult to write down and enforce, especially for those trying to identify the output share with the input share. There are two main difficulties. First, the R&D collaboration normally progresses over a long period while the market share relies on the protean market circumstances, as well as firms’ internal characteristics. It is difficult to choose the quantity of post-innovation products only according to R&D investment. On the contrary, many R&D strategic decisions rely partly on firm’s current market share. Second, what actions can be taken in the second stage is unbelievable, at least imperfectly unbelievable. There is no sound reason to suppose that such complete binding contracts or agreements are at work or that they are even possible. Before the R&D collaboration, the agreement of the market share of post-innovation products must suffer from the uncertainty in the production process and that in the R&D process. The risk is so high that firms have no incentives to contract on the output sharing of post-innovation products, since the R&D collaboration has to suffer inherent uncertainty during a long period. As shown in Negassi (2004)’s empirical study, the market variable which indicates firms’ market share is positive but not significant in explaining collaboration. One plausible interpretation is that firms invest in R&D and estimate approximate prospective quantity independently and simultaneously. Based on the two-stage game models, the theoretical studies mainly focus on the role of spillovers. According to Katz (1986), spillovers refer to the research done by one firm which can be used by another firm even though the latter does not receive permission (i.e., purchases a license) to use the inventive output. Besides spillovers, there are various determinants presented by other studies. According to the theoretical studies, the determinants of the R&D collaboration generally include spillovers, absorptive capacity,1 complementarity,2 heterogeneity,3 firm size, market share, R&D intensity,4 human capital, technological transaction, appropriability, and public subsidies (see Hernán et al., 2003, Negassi, 2004 and Belderbos et al., 2004; etc.). Recently, empirical interpretations are paid more attentions (e.g., see Cassiman and Veugelers, 2002, Becker and Dietz, 2004, Confessore and Mancuso, 2002 and Negassi, 2004). By focusing on the frequency of occurrence of the R&D collaboration or the R&D expenditures, most empirical studies discuss which determinants are more beneficial to the R&D collaboration. Although most determinants have been considered in the literature, their effects on firms’ decisions are discussed separately. Indeed, there exist interrelations among them. Up to now, there is no research into the significant determinants holistically. Moreover, the coexistence of cooperation and competition in R&D activities make non-cooperative game theory partly inefficient in anticipating firms’ decisions. Traditional non-cooperative game theory assumes that all people are exclusively pursuing their material self-interest and do not care about “social” goals per se. While in an actual competition players care about their own profits as well as the profits of their opponents in order to keep or improve their competitiveness. They have incentives to cooperate each other since their cooperation may benefit their total profits. Certainly, the cooperators maybe have adverse incentives to increase their opponents’ profits. The relations among firms are seldom of a wholly cooperative type or a wholly competitive one: firms sometimes cooperate with each other and do not cooperate at other times, or they cooperate in some areas as well as compete in others. The two-stage game models resolve the R&D collaboration5 through maximizing the total profits prior to the individual profits. This implies that the cooperative incentives are more important than the competitive ones. Indeed, the cooperative incentives and the competitive incentives always coexist, and their importance in the alliance cannot be distinguished. Thus, a more general framework, maybe a thoroughly different one, is needed for analyzing these players’ rational behavior. Fortunately, the rationalization has been studied a few years ago, by Fehr and Schmidt, 1999 and Xu, 1999 and so on. They measure the cooperative incentives by introducing proxies or parameters into player’s individual utility function. The common drawback of their measurements is that it is difficult to operate in reality, e.g., how to choose the values of the parameters for a given economic situation. Thus, there are two problems for further study. First, which determinants are holistically crucial and how they affect firm’s decisions in the R&D collaboration? Second, how to reveal firms’s cooperative incentives in a rational mode? In this paper, we integrate four crucial determinants. First, two determinants are crucial in the process of knowledge transfer: spillover level and the absorptive capacity.6 A firm’s spillover level indicates the ratio of spillovers to its total knowledge base. The spillover level describes outgoing transfer while the absorptive capacity determines the level of incoming transfer. Both the absorptive capacity and the spillover level jointly determine the effective transference of knowledge from one firm to another. Although spillovers play a minor role after taking account of the absorptive capacity, as discussed in the literature (e.g., see Cohen and Levinthal, 1989 and Negassi, 2004), this paper will show that spillovers are still very important to firm’s cooperative strategies after distinguishing the incoming transfer from the outgoing one. A sound explain is that it is difficult to empirically measure the spillovers which occur through the channel characterized by theoretical studies. Second, we introduce the conception of cost-reducing ability, which is exactly the ability to reduce unit cost through R&D activities. It may be a crucial reason why R&D decisions diverge across firms, since every firm tries to reduce its production costs as possible as he can. Finally, uncertainty indicates the failure probability of the R&D collaboration, which describes the opposing aspect of the probability of success in the R&D collaboration. On the other hand, differently from the literature, we introduce a virtual player whose utility is the total profits of all firms to reveal cooperative incentives, and consider the input (R&D investment) share and the output (quantity of new products) share simultaneously. The factors such as agreement or people’s conception (e.g., long-term reputation, altruism, collectivistic consciousness, etc.) can be viewed as a virtual player, who coordinates its actions with other real players. The coexistence of cooperation and competition is exactly the situation where real players (firms) and the virtual player play one game. We will obtain the following findings on firms’ decisions in R&D activities. First, firm’s high input share generally is reflected by high market share, and firm’s R&D investment increases with the probability of success in the R&D collaboration. Second, from the viewpoint of both R&D investment and social welfare, the R&D cartel is superior to the other R&D scenarios mentioned in Section 3 when spillovers are exogenous. Differently from the literature, we show that this superiority comes from the coexistence of the concern about total profit and that about individual profits. Third, unlike the results in some empirical papers, we find that spillovers play a dual but still significant role in changing R&D investment. When spillovers are endogenous, firms take different actions on the R&D investment according to their own situations. Whether or not they increase their investment with spillovers is determined by their own characteristics. Finally, we analyze firms’ interactions. This paper is organized as follows. Four crucial determinants are described in Section 2. Our models corresponding to different R&D scenarios are presented in Section 3. In Section 4 different R&D scenarios are compared when spillovers are exogenous. In Section 5 we analyze the shift of firm’s optimal actions under the condition that firms can internalize spillovers, and extend the R&D collaboration to a non-formal situation in Section 6. Finally, we conclude this paper in Section 7.
نتیجه گیری انگلیسی
The models discussed in this paper differ from the two-stage game models presented in the literature: • several determinants are integrated here; • a virtual player is introduced here to reveal the cooperative incentives, whose utility function is the total profit of the alliance; • the investment share and the market share are considered simultaneously, in consideration of the difficulty to agree on the market share of the post-innovation products. Under the framework with the above aspects, we analyzed the R&D collaboration, by comparing with the non-cooperative R&D and the R&D monopoly. This analysis results in the following four conclusions. First, all firms are convinced of the positive correlation between the R&D expenditure and the output of the post-innovation products. Namely, their enlargement in the R&D investment must be reflected by a larger output of the future products. Moreover, the uncertainty in the R&D collaboration negatively affects the R&D expenditure. Second, with the exogenous spillovers, both the outputs and the R&D expenditures in the R&D cartel are larger than those in other R&D scenarios. The R&D cartel has to coordinate members’ market share to maximize the total profit, without loss of the members’ individual profits. Therefore, the R&D cartel is superior to other scenarios, even though the producer surplus reaches the maximum in the monopolistic R&D. Differently from that in the literature, we show that this superiority comes from the coexistence of the concern about the total profit and that about the individual profits. Third, after introducing the firm-specific absorptive capacity and the cost-reducing function, we claim that the spillover level plays a dual but still significant role in deciding the R&D investment. Not consistently with the prevailing viewpoint, firms take different actions on the R&D investment when the spillovers are endogenous. Namely, when firms in a RJV or a cartelized RJV increase the spillover level, whether or not they increase their investment with spillovers are determined by the firm-specific determinants such as the absorptive capacity, the market competition, the degree of risk-aversion, etc. It is a balanced result among the R&D determinants. Finally, the interactions among members in the R&D cartel are determined by the combined cost externality. A firm’s reaction to other also varies with its characteristics. In this paper, we do not consider the stability of the R&D collaboration, and the problems which may appear in management and reorganization. Overall, both the firm-specific R&D resources and the coexistence of competition and collaboration make it more complex to anticipate firms’ decisions.