Global Stock Network Connected and Resonance Effect Based on Time-Zone VAR Model with LASSO

We propose the LASSO VAR model with the time-zone effect to investigate how stock markets worldwide interact with each other. Our approach overcomes the dimensional problem in the VAR model and characterises connections within financial markets systematically. Using the daily data of 36 national stock indexes from 2006 to 2009, we construct the global stock market network and discover a strong regional structure in which national stock markets behave significantly distinct before, during, and after the 2008 subprime crisis. Besides, the lagged returns in the United States lead a large number of international stock returns