分位数回归方法在资本结构影响因素分析中的应用

The Application of Quantile Regression in Study of Determinants of Listed Companies' Capital Structure

传统的OLSE回归只能得到一组系数估计值,因此无法深入分析不同财务杠杆程度下资本结构的影响因素.为此本文尝试着将分位数回归理论引入到回归模型中,通过对我国上市公司的相关数据进行分位数回归分析后发现,不同资产负债率水平下,企业股权结构、企业产生内部资源的能力对资本结构的影响效果不同:低负债率水平下,两者均与债务水平正相关;而高负债率水平下,又均与债务水平负相关.研究还发现,企业的成长性、企业规模以及财务困境成本与债务水平正相关,而盈利能力、非债务税盾则与债务水平负相关。

The traditional OLSE regression can only get a set of estimative coefficient, but can not deeply explain the determinants of capital structure in different financial leverage. Using panel data of listed companies, this paper analyzes it with a quantile regression approach, then draws a conclusion： the share structure and the capability of emerging interior resources effect the capital structure changeably with the different quantiles respectively. That is, for those conditions with low asset-liability ratio, the sign is positive, on the contrary, the sign is negative. The paper also reveals that the growth, the size, and the financial distress cost of company is positively correlated with the capital structure, while the profitability and the non-debt tax shield is negatively correlated with the capital structure.