CAJ | 학술논문

进口和出口的协变性研究对于宏观政策分析和金融决策具有非常重要的意义。本文通过VAR（Vector autoregression）和DCC（Dynamic conditional correlation）建模研究了我国进口和出口之间的静态和时变相依性。主要研究结论如下：13个经济体中，有5个经济体对我国的进口和出口之间的相依性表现出长期的稳定性，其余8个经济体的时变相依性显著，时变路径也趋一致；在汇率变动等外部经济事件的冲击下，影响程度的非对称性和共同影响因素的改变，导致相依结构发生变化；VAR建模显示国际贸易平衡依赖于进口和出口增长率的协调发展。此外，有8个经济体的进口和出口能为彼此提供20％-55％的解释能力。
진구화출구적협변성연구대우굉관정책분석화금융결책구유비상중요적의의。본문통과VAR（Vector autoregression）화DCC（Dynamic conditional correlation）건모연구료아국진구화출구지간적정태화시변상의성。주요연구결론여하：13개경제체중，유5개경제체대아국적진구화출구지간적상의성표현출장기적은정성，기여8개경제체적시변상의성현저，시변로경야추일치；재회솔변동등외부경제사건적충격하，영향정도적비대칭성화공동영향인소적개변，도치상의결구발생변화；VAR건모현시국제무역평형의뢰우진구화출구증장솔적협조발전。차외，유8개경제체적진구화출구능위피차제공20％-55％적해석능력。
The co-movement Behavior of imports and exports contributes to the development of international payment. Moreover, structural change analysis of co-movement plays an important role in evaluating the effect of the exchange rate or some domestic important financial events on the international competitiveness of a country＇s domestic goods （services）. The correlation between imports and exports is a key factor for determining the bilateral trade flows between China, major developed countries and emerging economies. This study may partly explain China＇s import-induced export characteristics in the early stage of economic development. China has been trying to develop long-term import/export relationships with Chinese Taiwan, Chinese Hong Kong, Japan, Korea, the United States and other developed countries. The traditional import-induced export strategy poses many problems, such as the weak capability of independent innovation and the limited space of sustainable growth. The Dynamic Conditional Correlation （DCC） model proposed by Engle is an important and effective method to study bilateral relations between variables. Whereas both domestic and foreign literatures have not introduced the DCC model to study import and export activities, this paper attempts to examine whether there are time-varying correlations between imports and exports （the imports from China to each economy or the exports from each economy to China） of 13 economies including the United States. The time span of the sample data is from January 1993 to April 2010. Data is recorded monthly, and the data has a total of 208 sample points in each sequence. This paper modeled the static and dynamic co-variation between imports and exports and analyzed the features and evolution path of the dynamic co-movement. The driving impact of external events on the structural changes of correlations is studied. The second section compares the advantages and disadvantages of different methods used to model the correlation. The often-used measurement tools of static correlation include linear correlation analysis, regression analysis and the Copula Method. Time-varying correlation coefficient methods include the cross terms used in the regression, Kalman Filters, moving windows （or rolling window） analysis method and DCC modeling. The DCC method has a comparative advantage. The third section describes the models used in this paper. Dynamic Conditional Correlation model is used for dynamic correlation coefficient analysis and the vector auto-regression model is used for static correlation coefficient analysis. The fourth section discussed about data sources and execute some basic statistical analyses, such as summary statistical analysis, seasonal effect test, stationary test, self-correlation test and ARCH effect test. The results of correlation modeling are discussed in the fifth section from the following aspects： the event study, the dynamic correlation analysis based on DCC-GARCH model, and the static correlation of imports and exports. In summary, this paper studies static and time-varying correlations between exports and imports in China. The main results are as followed： Among the 13 economies, five ones have long-term correlations of imports and exports with China and the others have significant dynamic correlations. Under the shocks of external economic events, the structure of correlations will change because of asymmetrical influence and common factors. VAR modeling shows that international trade balance depends on harmonious development of import and export growth. In addition, the exports and imports of 8 economies can represent 22% -55% of 15 economies in this study.