中文摘要：

投入产出模型常被用来刻画最终需求变动对就业的完全影响。传统的投入产出模型以线性就业系数为基础．即假定行业产出的就业弹性等于1。然而大量研究表明。绝大多数行业产出的就业弹性小于1，若使用传统模型就会高估最终需求变动对就业的影响。本文首先通过引入就业弹性提出了一个局部非线性化的投入产出模型。测算最终需求的就业效应；然后，利用固定效应变系数回归模型估算了54个非农行业的就业弹性，并对行业就业弹性差异的影响因素进行了初步探讨：最后．利用所构建的模型和行业就业弹性的估算结果，对国家“四万亿投资”的就业效应进行了实证测算，并与传统模型的测算结果进行对比。结果表明：局部非线性的投入产出就业模型有效修正了传统线性模型对测算结果的高估情况。

英文摘要：

The input-output model is generally used when measuring the employment effects of changes in final demand to calculate its full impact, but the traditional IO model based on the linear employment coefficient, which is equivalent to assuming that each industry employment elasticity of output is 1. Many studies, however, have shown that most industries＇ employment elasticity of output are smaller than 1, so if the employment elasticity is 1, the model will over-estimate the impact of changes in final demand on employment. This paper introduces the employment elasticity into the traditional IO model to get a non-linear input-output model to measure the employment effects of changes in final demand. Then, fixed effects regression model with variable coefficients is used to estimate the employment elasticity of 54 non-agricuhural industries, and the reasonableness of the results is demonstrated by discussing some Influ- encing Factors of industry employment elasticity among different industries. Finally, the non-linear IO model is used in comparison with the results of traditional model to estimate the impact of ＇ RMB four trillion investment＇ on employment. The result by the non-linear IO model fixes the over-estimation caused by linear employment coefficients.