资本结构调整速度估计方法比较研究——基于上市公司财务数据的Monte Carlo模拟分析

Comparing the Estimating Methods of the Capital Structure Adjustment Speed:A Monte Carlo Simulation based on the Financial Data of Chinese Listed Firms

资本结构调整速度的衡量是公司财务研究的核心问题之一,衡量方法普遍使用动态面板估计方法,但基于不同理论假设的各种动态面板估计方法在实证研究中的表现大相径庭,导致前期学者估计出的调整速度存在较大差异。本文通过Monte Carlo模拟分析发现,采用不同的估计方法得到的调整速度存在很大的差异,在这些方法中,优化系统GMM估计量的表现最佳,而一阶差分GMM估计量则存在较大的偏误。以此为基础,我们重新估算了中国上市公司的资本结构调整速度。结果表明,在1999-2008年样本区间内,平均调整速度为21.6%,对应的调整半周期为2.85年。

The measure of capital structure adjustment speed is one of the most important issues in corporate finance. The dynamic panel estimation methods are most commonly used. But based on different theoretical assumptions, a variety of dynamic panel estimation methods perform differently in the empirical study, leading to the divergence of capital structure adjustment speed. This paper investigates the performances of various estimators of dynamic adjustment model using Monte Carlo simulation. We find that different estimators can lead to different inference of adjustment speeds. In six estimators investigated, the optimal system GMM estimator performs best, while the first difference GMM estimator suffers serious bias. We then re - estimate the adjustment speed of Chinese listed firms using different estimators, and find that the average adjustment speed is 21.6% with a half - life being 2. 85 years during year 1998 to 2008, when the optimal system GMM estimator is used.