国家外部资本结构的动态演化特征:基于经济发展阶段的分析

External Capital Structure Dynamics of Countries:An Analysis Based on the Stage of Economic Development

本文对经济体在经济发展不同阶段的外部资本结构是否会呈现出系统性演变规律问题进行了研究。研究发现,外部资本结构呈现出一系列与经济发展阶段相关的典型化事实,其中最为重要的是境外股权类融资规模占比与人均GDP之间呈现出先上升后下降的倒U型关系。本文的理论模型为上述发现提供了基于经济体产出风险因素的解释:在经济体从最落后阶段起步而逐步发展至发达阶段的过程中,产出风险逐步降低,外部融资工具的融资成本则依经济体产出风险状况变化而调整,从而依次经历境外股权融资持续替代跨境银行贷款、境外债券融资持续替代境外股权融资两个不同阶段。实证部分验证了前文的典型化事实与理论模型。本文不仅为外部资本结构问题提供了一个新的分析视角,亦有助于认识中国跨境资金流动变动趋势及其管理问题。

The external capital structure of countries,which refers to the functional composition of capital inflows,matters for economic performance and financial stability.Current research focuses on issues related to the driving force of financial development and institutional quality on capital flows.However,less attention has been devoted to understanding how the external capital structure evolves as nations develop from backward to advanced.To fill this gap,we study relationship between the structure of external liabilities and stage of economic development.Using locally weighted scatterplot smoothing procedure,we find that the external capital structure presents many stylized facts related to real GDP per capita.Firstly,there exists an inverted U-shaped relationship between the share of foreign equity and real GDP per capita.And this nonlinear relationship is mainly determined by FDI when distinguishing FDI from equity FPI.Another important finding is that real GDP per capita has different effects on the proportion of OI and the proportion of debt FPI.The proportion of cross-border bank loans decreases with GDP per capita,but the opposite is true for debt FPI.Based on these findings,we propose a new model,in which the level of economic development affects the composition of capital inflows through output risk.A basic fact in macroeconomics lies that the economic growth in more developed countries has lower probability of encountering adverse shocks.We generalize the model of Bolton & Freixas(2000) in two ways.First,introducing social output risk and social cost of default to transform it into a macro model.Second,modifying the contract enforceability and information asymmetry to reflect the particularity of cross border financing.We consider international loans,bond and equity jointly in this new model.The theory highlights that the financing costs of these three types of funds vary with the stage of economic development(as captured by the output risk).Generally,foreign banks have higher ability to help countries through times of financial distress.But such financial flexibility is costly,since bank loans entails additional “intermediation costs”.Therefore,it is desirable to use cross border loans only in countries with the highest probability of output being in a low state.For economies with output risk at the intermediate level,there still be a high risk premium for issuing foreign bonds,and the dilution cost of equity is relatively small,so such economies prefer to use equity for financing.The risk premium caused by bonds is the lowest in developed countries,and thus these countries tap the international bond markets.All these explain why foreign loans have been replaced by foreign equity firstly and then foreign equity have been replaced by international bonds as countries developed from the least development stage to the most developed stage.Our empirical analysis examines the hump-shaped relationship between real GDP per capita and foreign equity share.And the results are still significant when using religious data as instrumental variables.Furthermore,we use the volatility of GDP growth rate to measure output risk and conduct mediation effect tests.The results confirm the impact of output risk on FPI and OI and provide support for theoretical prediction that stage of economic development affects external capital structure through output risk.Chinese external financing has maintained FDI-dominated structure for a long time.How will the external capital structure of China change as per capita income level further increases? This article may provide some insight into these questions.