A dynamic analysis of technical inefficiency's generating process with panel data

利用面板数据动态分析技术低效率的产生过程

• 批准号: 16530150
• 负责人: TORII Akio
• 金额: $ 1.28万
• 依托单位: Yokohama National University
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (C)
• 财政年份: 2004
• 资助国家: 日本
• 起止时间: 2004 至 2005
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/en/grant/KAKENHI-PROJECT-16530150/
• 关键词: technical inefficiency; vintage model; exponential power family; 指数べき乗分布

The purpose of this research project is to analyze dynamic changes of technical inefficiency, estimating establishments' inefficiency level and asset vintage. To achieve this purpose, a stochastic process model which explain changes of units' inefficiency level was constructed. Employing the model I investigated some statistical features in distribution of inefficiency levels. Next, employing establishments panel data the distribution of inefficiency level conditioned on inefficiency level at previous period was estimated. I analyzed some statistical character of this distribution and conducted some tests which was lead from theoretical models.On the first step of this research, by theoretical model analysis and empirical analysis employing cross-section data a vintage stochastic process model with exponential power family distribution was chosen to be appropriate. On the second step, I tested this hypothesis.I investigated Japan's regional food supermarket which provides us a large number of homogeneous establishments' samples. I was able to obtain more than 4,000 samples. From this panel data production functions as well as levels of technical inefficiency are estimated. Then differences in estimated inefficiency levels were calculated and employed to test some hypothesis.This procedure provided me almost the same level of a parameter 'p' in stochastic process with those of parameter in exponential power family distribution estimated from cross-section data. Then a hypothesis presented by the author in "Efficiency in Japanese Industries" was supported affirmatively. In Japan's regional food retail market, inefficiencies tend to be improved with higher probability as inefficiency become severer, although the probability increases only in diminishing manner.

本研究旨在分析企业技术无效率的动态变化，估计企业的无效率水平和资产年限。为了达到这一目的，建立了一个解释机组无效率水平变化的随机过程模型。采用该模型，我调查了一些统计特征的分布无效率水平。其次，利用企业的面板数据，估计了以前期无效率水平为条件的无效率水平的分布。本文首先通过理论模型分析和截面数据的实证分析，选择了一个经典的指数幂族分布的随机过程模型，并对该模型进行了检验。第二步，对这一假设进行了检验，考察了日本的地区性食品超市，这为我们提供了大量同质性的场所样本。我收集了4,000多个样本。从这个面板数据生产函数以及技术效率低下的水平进行估计。然后计算估计的无效率水平的差异并用来检验一些假设，这一过程使我获得了几乎与由截面数据估计的指数幂族分布中的参数水平相同的随机过程参数“p”。在此基础上，作者在《日本产业的效率》一书中提出的假设得到了肯定的支持。在日本的地区食品零售市场中，虽然低效率的改善概率逐渐降低，但随着低效率的恶化，低效率的改善概率也逐渐增加。