创新要素投入产出效率随机变动测度与结果分析

创新要素配置在投入产出两个环节中生成大量不完全技术信息，投入产出效率随机变动更加明显。从投入产出数据集合中提炼共同因子，与不可观测变量构成预测器，建立因子增广向量自回归模型(FAVAR)，分析随机变动方差构成，测度投入产出效率随机变动程度。随机变动效应包括水平效应、稳定性效应和规模效应。投入产出效率自回归扰动项表示随机变动程度，将其细分为共同因子，计算预测器方差及其与投入产出效率的协方差。基于669家上市公司月度、季度、半年度和年度技术研发数据，比较分析投入产出效率随机变动程度及形成原因，引入脉冲响应法分析变量方差构成与变动特征，提出相关政策建议。

the Measuring of the Innovation Factors Input-Output Productivity Stochastic Change and the Result Analysis

The innovation factors allocation has much imperfect technology information during the input-output process, the input-output productivity (IOP) random changes prominently.Extracting the common factors from the input-output data set, consisting the predictor with the unobserved variables together, construct the factor augmented vertical auto-regression model (FAVAR).Analyzing the components of the stochastic change variance, can measure the IOP stochastic change degree accurately.The effect of the stochastic change includes the level effect, the scale effect and the sustained effect.The IOP stochastic change degree is indicated by its auto-regression perturbation, and divided into variances of the common factors and the predictor, and the co-variance of the unobserved variable and the IOP.It is to comparatively analyze the IOP stochastic change degree and its cause, based on the 669 quoted companies month, quarter, half year and whole year R&D data through the measuring thought, analyze the variables variances and change character with impulse responsible approach, then put forward the policy proposals.