Title

Analyzing Information Transfer Patterns in East Asian Financial Markets

School Name

Governor's School for Science and Mathematics

Grade Level

12th Grade

Presentation Topic

Physics

Presentation Type

Mentored

Abstract

This research uses statistical transfer entropy to map the strength and directionality of connections between major financial companies in South Korea, China, and Japan. A matrix of daily stock prices was compiled from the SSE 250, KOSPI 200, and NIKKEI 225 stock indices, which was then fitted to a logarithmic trend. By modeling the price curves of different companies as connected, stochastic processes and calculating the information-theoretic transfer entropy between them, this study developed a network of the connections between nodes in each country and sector. After the price trends were discretized, a sliding-window plot of transfer entropy was generated for each of the countries in the study. Companies were grouped by financial sector – banking, insurance, investment, real estate – in order to facilitate analysis of information transfer trends between sectors. The research found that compared to both China and Japan, South Korea experienced a significant increase in transfer entropy across all financial sectors. The mapping for Japan shows a spike in mid-2013 which this study attributes to massive bond purchases by the Bank of Japan. Total transfer entropy matrices (TTEs) were also generated for each country in order to compare relationships between individual companies. The most notable result displayed by the aggregate matrices was a consistently lower transfer entropy between companies of the same type (compared to those from other financial sectors).

Location

Neville 306

Start Date

4-14-2018 11:45 AM

Presentation Format

Oral and Written

COinS
 
Apr 14th, 11:45 AM

Analyzing Information Transfer Patterns in East Asian Financial Markets

Neville 306

This research uses statistical transfer entropy to map the strength and directionality of connections between major financial companies in South Korea, China, and Japan. A matrix of daily stock prices was compiled from the SSE 250, KOSPI 200, and NIKKEI 225 stock indices, which was then fitted to a logarithmic trend. By modeling the price curves of different companies as connected, stochastic processes and calculating the information-theoretic transfer entropy between them, this study developed a network of the connections between nodes in each country and sector. After the price trends were discretized, a sliding-window plot of transfer entropy was generated for each of the countries in the study. Companies were grouped by financial sector – banking, insurance, investment, real estate – in order to facilitate analysis of information transfer trends between sectors. The research found that compared to both China and Japan, South Korea experienced a significant increase in transfer entropy across all financial sectors. The mapping for Japan shows a spike in mid-2013 which this study attributes to massive bond purchases by the Bank of Japan. Total transfer entropy matrices (TTEs) were also generated for each country in order to compare relationships between individual companies. The most notable result displayed by the aggregate matrices was a consistently lower transfer entropy between companies of the same type (compared to those from other financial sectors).