摘要
基于跨境资本流动宏观审慎管理相机抉择机制的理论模型,运用支持向量机回归分析、机器学习算法等方法,规范和实证分析了跨境资本流动宏观审慎管理的相机决策机制及跨境资本流动监测预警的可行性。规范分析得出,在资本和经常账户相对稳定时,利差变化等短期因素是影响跨境资本流动的主要因素;实证分析显示,当前影响中国跨境资金流动的最主要因素仍是国内经济基本面以及国际经济金融大环境,其中,经济增速对跨境资本具有较强引导作用,随着资本账户的持续开放,国际传染等渠道也逐渐成为影响中国跨境资本流动的重要因素;预测分析得出,基于跨境资本流动理论模型的支持向量机回归分析在预测跨境资本基本走势时具有可行性。因此,调节跨境资本流动的趋势性变化,宏观审慎监管应与其他政策有效协调,根本是实施调节经济基本面政策;为了防范跨境资本流动失速,建议构建基于机器学习算法在内的跨境资本流动宏观审慎预警体系。
Based on the theoretical model of discretionary selection mechanism for macroprudential management of cross-border capital flows,this study uses the methods of support vector machine regression analysis and machine learning algorithm,this paper theoretically and empirically analyses discretionary selection mechanism for macroprudential management of cross-border capital flows and the feasibility of monitoring and early-warning of cross-border capital flows.The theoretical analysis shows that short-term factors such as interest margin change are the main factors affecting cross-border capital flows,while capital and current account are relatively stable.The empirical result shows that the main factors influencing China’s cross-border capital flows are domestic economic fundamentals and the international economic and financial environment,thereinto,economic growth has a strong guiding effect on cross-border capital,with the continuous opening of capital account, International Contagion and other channels have gradually become important factors affecting cross-border capital flows in China.The forecasting analysis shows that the support vector machine regression analysis based on the theoretical model of cross-border capital flow is feasible in forecasting the basic trend of cross-border capital.Therefore,in order to regulate the trend of cross-border capital flows,macroprudential supervision should be coordinated with other policies effectively,and the essential thing is toimplement the policies of adjusting the economic fundamentals.For preventing the stall of cross-border capital flows,it is suggested to build a macroprudential early warning system of cross-border capital flows based on machine learning algorithm.
作者
姜哲
王繁荣
吴晓勇
范延岑
Jiang Zhe;Wang Fanrong;Wu Xiaoyong;Fan Yanceng(Wuhan University,Wuhan,Hubei 430072,China;State Administration of Foreign Exchange,Beijing 100048,China;Foreign Exchange Management Department,Yinchuan Central Sub-branch,Peopled Bank of China,Yinchuan,Ningxia 750000,China;Capital Project Management Department,Hangzhou Central Sub-branch,Peopled Bank of China,Hangzhou,Zhejiang 310000,China;Capital Project Management Department,Jinan Branch,People's Bank of China,Jinan,Shandong 250000,China)
出处
《金融经济学研究》
CSSCI
北大核心
2019年第2期25-40,共16页
Financial Economics Research
基金
教育部人文社会科学研究规划基金项目(18YJA790099)
关键词
宏观审慎
相机抉择
因子分析
支持向量机
macroprudential
discretionary selection
factor analysis
support vector machine
