时变弹性生产函数与全要素生产率

Production Functions with Time-Varying Elasticities and under the Catch-up Consensus:Total Factor Productivity

由于缺乏资本价格等国民核算资料，收入份额法在我国的使用受到一定限制。传统Cobb-Douglas生产函数估计的资本和劳动力产出弹性，反映的只是整个研究时期的一个平均水平，适用于估计整个研究时期的平均贡献率，但用于估计逐期贡献率很可能是有偏的甚至是有误的。本文给出了一个改进生产率增长测算的新思路，系统地提出利用非参数模型、变系数模型、可变参数模型和面板数据模型替代收入份额法，估计不同时期资本和劳动力的时变产出弹性。实证结果表明时变弹性生产函数具有更好的统计性质。

The income share method isn＇t feasible for calculating the TFP in China due to the lack of national accounting data such as capital price. The output elasticities of capital and labor estimated by traditional Cobb-Douglas production functions, reflecting merely the average level for the whole investigation period, may result in biases or even errors. This paper compares several models with time-varying output elasticities of capital and labor.Our empir- ical results indicate that these models perform better than ordinary models.