机构地区: 中山大学岭南学院
出 处: 《经济学(季刊)》 2012年第1期541-558,共18页
摘 要: 当前对于全要素生产率的测算正逐步由宏观层面转向微观企业层面,两者具有完全不同的理论机理。针对企业TFP估计中出现的同时性偏差和样本选择性偏差,一系列的最新修正方案被提出,并形成了很多前沿的估计方法。本文对这些新方法的逻辑进行了梳理,旨在廓清当前实际应用中存在的一些误区。利用1999—2007年中国工业企业数据,本文进一步应用最小二乘法、固定效应方法、OP法和LP法等参数和半参数方法核算了我国主要工业企业的TFP,在横向对比之后,发现半参数方法能够较好地解决传统计量方法中的内生性和样本选择问题。 Measurement of total factor productivity (TFP) is currently shifting from the macro-level to micro-level. The identification of these measures has distinct theoretical logic. Aimed to resolve the simultaneity, selectivity and attrition biases in the traditional method for firm-level TFP estimation, a array of proposals are raised and boiled down to several estimate techniques. This paper provides economic researchers with an up-to-date overview ofissues and relevant solutions associated with the choice of methods. Using a firm-level dataset for 1999—2007, this paper estimates firms' TFP by applying various parametric and semi-parametric methods, such as OLS, FE, OP and LP. We find that the semi-parametric technique can solve the inherent econometric problems in the traditional method.
领 域: [经济管理]