主题：A Spatial Modelling Approach for Large-Scale Mixed Frequency Outcome Interactions                       

时间：2023年9月22日上午9:30

地点：365英国上市网站官网6-210会议室

主讲人简介：韩晓祎，2014年获美国俄亥俄州立大学经济学博士，现为厦门大学王亚南经济研究院与365英国上市网站官网长聘副教授、博士生导师，主要研究领域为计量经济学、应用计量经济学、区域经济学和劳动经济学。多篇论文发表在PNAS、Journal of Business & Economic Statistics、Econometric Theory和Regional Science and Urban Economics、《数量经济技术经济研究》等国内外权威学术期刊上。主持国家自然科学基金面上项目2项、青年项目1项，以及福建省自然科学基金杰青项目。

Abstract: In many scenarios, researchers seek to measure the spatio-temporal interactions among cross-sectional units in different networks. We consider measuring such interaction in a more challenging setting, where the number of cross-sectional units and endogeneous outcomes in the network is large, and outcomes are sampled at different frequencies. We propose a novel spatial modelling approach that employs recent advances in multivariate spatial panel model and mixed frequency VAR to address the modeling and estimation challenges. The proposed model is capable of capturing large-scale high frequency and low frequency outcome interactions along the space and time dimension. We discuss the identification of the model, and develop a computationally tractable Markov Chain Monte Carlo (MCMC) algorithm for estimation and inference. Numerical studies show good finite sample performance and computational tractability. We also apply the new model to study the spatio-temporal effects of global pandemic and unemployment for 68 countries across 6 continents from January 2020 to August 2021, and detect significant spatial-temporal interaction effects for outcomes of different frequencies.