2024-02-02-亚开行-使用贝叶斯状态空间方法估计区域整合_44页_1mb

Estimating Regional Integration Using the Bayesian State-Space Approach

Economic regional integration is challenging due to incomplete data from missing values and limited time spans. This paper utilizes a dynamic factor model estimated through the Bayesian state-space approach to address these issues. Bilateral Economic Integration (BEI) indexes are estimated across four dimensions: trade, foreign direct investment (FDI), finance, and migration. The Regional Integration Index (RII) is calculated by applying network density to the BEI estimates. The RII for Asia and the Pacific declined slightly in recent years, with greater integration centered around the People’s Republic of China.

The Bayesian approach offers several advantages over traditional methods like Principal Component Analysis (PCA). It integrates handling missing data and considers autoregressive structures in time series, eliminating the need for preliminary data imputation or manipulation. Advanced computational techniques, including Gibbs sampling, are used for Bayesian estimation.

Key findings include:

• The BEI indicates economic linkages between economies, with notable trends such as increased reliance on China in recent years.
• The decline in the RII during the COVID-19 pandemic affected all subregions in Asia.
• Dimensional BEI indexes help identify the primary drivers of integration (e.g., trade, finance).

The study proposes future research directions, including expanding coverage, estimating global integration indexes, enhancing dynamic factor models, and exploring factors influencing integration.