经济学数量模型的选择与科布-道格拉斯生产函数

Matching of Econometric Models with Data and Cobb-Douglas Production Function

科布-道格拉斯生产函数早已成为经济学几乎全部数量方法的可操作工具的基础,然而到今天还存在着几个基础性的理论问题没有解决。此外,它的适用性以数据符合同方差性假设为依据,因此,在选定数量模型之前进行数据检验是必不可少的步骤,以避免有偏的结果。本文采用前沿生产函数模型和随机参数—前沿模型,对采自贵州省湄潭县的农户数据进行对比分析,所得到的结果支持上述关于“模型适用性”的论点。

The Cobb-Douglas production function has become the basis for almost all usable instruments in econometric analysis; however, there are unsolved theoretidal problems that flawed the whole structure. In addition, the assumption of 'homogeneity' on data puts strong restriction on the authentic application of those production function models. It is necessary to do data check before specify the better-fit econometric models and running them. As economists, we should not transplant the outcome from any such models into policy suggestions without further discussion, when working on real world issues, to avoid biased conclusions. Two analytical models are applied in this study on the data collected from farmers' households in Meitan County, Guizhou Province, i.e., the frontier production model, and a random coefficient frontier production model; the diversed outcomes from running these two models support the better-fit-mode argument.