微观企业全要素生产率测度方法的比较与应用

Comparison of Different TFP Estimation Methods and Its Application Based on Micro Firm Level Data

微观企业TFP通常是不可观测的,它的估计面临内生性、联立性等诸多问题。在系统分析微观生产函数估计所面临主要问题的基础上,经过蒙特卡洛模拟法,比较了微观生产函数最新估计方法的优劣,发现广泛应用的0P、LP和ACF参数估计方法,倾向于高估微观企业的TFP,而De Loecker、GNR和联合估计的TFP测度方法能够得到TFP的稳健估计,增加值类生产函数估计倾向于高估微观企业的TFP,并且高估TFP的离散程度。将前沿TFP测度方法用于中国生产要素错误配置问题的研究,发现要素再配置具有显著的TFP促进效应,但是低于Hsieh和Klenow(2009)政策模拟的估计结果。据此,提出以微观企业TFP开展实证研究的对策建议。

TFP of micro firms is unobservable. It faces lots of identification problems such as endogenous, simultaneity selection bias based on micro dataset. This paper summari- zes all the issues that need to be handled while estimating TFP and uses extensive quasi Monte Carlo simulation method to measure the efficiency as well as consistency for brand- new different TFP estimation methods. The simulation results indicate that the widely used OP, LP and ACF method have the tendency to overestimate TFP while GNR, De Loecker and Joint Method can provide robust TFP estimation. Value-added production function is also has the potential to overestimate TFP and it can also overestimate the inter-quantile distribu- tion of the overall TFP. Based on these brand-new estimation methods, this paper uses the calculated TFP to probe the resource misallocation in China. It demonstrates that although the resource reallocation within an industry has important effect on TFP growth, it is still lower than Hsieh and Peter （2009） simulation results. It concludes with suggestions for car-ry out empirical research based on mico dataset.