依据中国367家高新技术上市公司2007--2014年间的面板数据。建立动态面板Sys_GMM模型,得到企业研发强度与企业收益、风险之间的函数关系式.进而计算出企业研发投入的收益-风险耦合区间和积极退耦、消极退耦区间.研究发现,中国高新技术企业研发投入的收益与风险并不匹配,落人耦合区间的样本企业仅占1.86%.有94.79%的样本企业落入收益-风险积极退耦区间,对于绝大多数中国高新技术企业来说,研发投入是以较小风险博取较大收益的趋利行为.实证结果显示.12.68%可被看作企业研发投入强度的上限.当企业研发强度超过12.68%后.创新活动将进入以较大的风险代价获取较小收益增长的收益-风险消极退耦阶段.研究结论有助于政府主管部门更客观地评价企业研发活动的价值、更全面地掌握企业研发收益的风险代价、更精准地实施企业风险预警,并为企业管理实践摆脱“高收益-高风险”与“低收益-低风险”的两难困境提供启发,为企业确立最佳的研发投入水平提供借鉴.
Based on the panel data of 367 high-tech enterprises in China from 2007 to 2014, it built the dynamic sys- tem GMM estimation model and got the function relation between benefit, risk and R&D intensity. Then it worked out the positive decoupling interval, negative decoupling interval and coupling interval of enterprises' R&D activities. The results show that benefit and risk in Chinese high-tech enterprises' R&D activities do not match. Only 1.86% of the to- tal study sample realize coupling, and 94.79% of the total study sample realize positive decoupling. So R&D invest- ments are profitable to most of Chiese high-tech enterprises, which can gain more interest at less risk. Empirical results show that 12.58% is the upper limit of R&D intensity. If R&D intensity exceeds 12.58% ,the enterprises will fall into negative decoupling interval of gaining less benefit at the cost of larger risk. The result is helpful to the government to assess the value of R&D activities more objectively, to understand the risk cost of R&D benefits more comprehensively, and to implement enterprise risk early-warning more precisely. At the same time, it can provide decision reference for managers to solve the dilemma between "high benefit -- high risk" and "low benefit -- low risk", and help to deter- mine optimal R&D investment level as well.