中国地方政府融资平台债务的时空演化特征及影响因素

Temporal and Spatial Evolution Characteristics and Influencing Factors of Local Government Financing Platform Debt in China

本文利用中国2007-2018年278个城市地方政府融资平台负债率的面板数据,考察其分布动态与地区差异,揭示其空间相关性特征和集聚模式,并实证考察影响地方政府融资平台负债率变化的主要因素。研究发现:第一,从全国层面和分地区层面看,地方政府融资平台负债率都表现为不同程度的上升态势,且均存在多极分化现象;第二,地方政府融资平台负债率的总体差异、地区内差异和地区间差异均呈不同程度的下降趋势;第三,空间集聚模式主要是以“高-高”型集聚和“低-低”型集聚两种模式为主;第四,地方政府融资平台负债率不仅存在显著的动态效应,也存在显著的空间效应,同时其空间效应也具有显著的动态效应。本文的研究有助于增进政策制定者和社会各界对2007-2018年中国地方政府融资平台债务演化状况的全面认知,为有针对性地解决地方政府隐性债务问题提供实证依据和政策参考。

It is of great practical significance to scientifically and accurately describe the evolution status of local govern⁃ment financing platform debt in China and precisely identify its influencing factors,so as to scientifically formulate government debt policies,thus preventing and resolving the hidden debt risks of local governments.Based on the panel data of debt ratio of local government financing platforms in 278 cities in China from 2007 to 2018,this paper investigates its distribution dynamics and regional differences.It unveils spatial distribution characteristics and ag⁃glomeration mode while empirically investigating the main factors affecting the debt ratio change of local government financing platforms.The results show a generally escalating trend in debt ratios at both national level and sub-regional level,exhibiting a multi-polarization phenomenon.The overall,intra-regional and inter-regional differences in the debt ratio of local government financing platforms experience varying degrees of decline.Spatial agglomeration modes are mainly taking the form of“high-high”and“low-low”clusters.Furthermore,the debt ratio′s impact extends both dynamically and spatially,with both aspects exhibiting significant effects.The research in this paper is helpful for policy makers and all walks of life to fully understand the debt evolution of local government financing platforms in China from 2007 to 2018,and provide empirical basis and policy enlightenment for solving the hidden debt problem of local governments.