不确定性条件下研发联盟的风险分担设计  

Design on Risk Sharing of R&D Alliance on the Condition of Uncertainty

作　　者：汪翔[1,2] 孟卫东[1] 吴国东[2] 

机构地区：[1]重庆大学经济与工商管理学院,重庆400044 [2]重庆师范大学经济与管理学院,重庆400047

出　　处：《管理工程学报》2013年第3期189-195,222,共8页Journal of Industrial Engineering and Engineering Management

基　　金：国家社科基金重点资助项目(08AJY026);教育部新世纪优秀人才支持计划资助项目(NCEF08-0609);重庆师范大学青年基金资助项目(2011XWQ08)

摘　　要：Shapley值考虑了研发收益如何在企业间进行公平的分配,却没有考虑研发风险如何在风险态度不同的企业之间进行有效的分担。本文在Shapley值的基础上设计了一个直接分配机制,该机制不仅实现了Shapley意义上的公平分配,而且在一定条件下,还能实现风险在不同企业之间的有效分担,提高整个联盟的确定性等价收益。The Shapley Value is often used to fairly distribute the revenue of an R＆D alliance among its members, but not to efficiently allocate the risk among members with different risk attitudes. For example, if an alliance is composed of a large-scale enterprise and a small-scale enterprise, the large-scale enterprise can endure external risks and is more likely neutral to risks. However the small-scale enterprise can not bear the same risk and is more likely risk-averted. The Pareto efficient curve is used to understand how an alliance allocate risks by assigning the small enterprise fixed revenue and assigning the big enterprise residual revenue. The large-scale enterprise endures the entire risk while the small one bears no risk. We aim to design a distributing mechanism for an R＆D alliance that not only achieves to allot income fairly, but also to improve the risk allocation among members. Therefore, this distributing mechanism must have two properties： （ I ） each enterprise＇s expectation revenue of this mechanism should be equal to that of the Shapley Value, and （2） certainty equivalence of the alliance should exceed that of the Shapley Value. We consider the direct distributing mechanism in which the members＇ revenue is only the function of R＆D Alliance＇s total revenue, not the function of marginal revenue just like Shapley Value. In this paper we design a direct mechanism, which can not only distribute the revenue fairly, but also allocate risk among different members efficiently, and increase the certainty of equivalent revenue of R＆D alliance on the given condition. We find that it is efficient for an alliance to assign risks if all members are risk-averted, and the risk （ measured by deviation of revenue） of enterprise is inversely proportional to the square of its Arrow-Pratt coefficient of absolute risk aversion. If an alliance consists of risk-aversion enterprises and risk-neutrality enterprises, it is efficient to assign risk-neutralityenterprises all the risk. If there is at

关 键 词：风险分担 不确定性 研发联盟 SHAPLEY值 

分 类 号：F270.5[经济管理—企业管理]