Transfer Entorpy within the Nasdaq Stock Exchange

School Name

Governor's School for Science & Mathematics

Grade Level

12th Grade

Presentation Topic

Consumer Science

Presentation Type

Mentored

Mentor

Mentor: Kyuseong Lim, Korea Advanced Institute of Science and Technology

Written Paper Award

2nd Place

Abstract

Using Transfer Entropy, we attempt to compare information flow between varying industries and the economy as a whole through studying 153 companies in the NASDAQ Stock Exchange and the NASDAQ Composite as an economic indicator. Transfer Entropy provides a model-free approach to detect asymmetrical statistic dependencies and correlations allowing us to calculate both the magnitude and direction of information flow. A delay variable is introduced to determine which industries tend to lead mass movements in the stock market, as well as to characterize the duration of influence leading these movements. Using a time series from 1 January 2000 to 20 July 2016, we perform a time-series analysis between each individual company and the NASDAQ Composite. Our data suggests that the consumer durables and finance industries provide the most information concerning the economy as a whole. Using a one-day delay, we find that introducing a delay variable significantly increases Transfer Entropy of all industries aside from energy, with technology having the most significant increase (103%), compared to the average (28%). After a one-day delay, we find that additional delay decreases Transfer Entropy generally, but not monotonically, suggesting that movements are mostly influenced by two-day trends. The introduction of a delay variable to studying information flow allows us to expand the applications of Transfer Entropy in econophysics, specifically the ability to determine not only magnitude and direction, but also duration. From this, we are able to better understand what leads market movements such as the 2009 global financial crisis.

Location

Wall 210

Start Date

3-25-2017 10:00 AM

Presentation Format

Oral and Written

Group Project

No

COinS
 
Mar 25th, 10:00 AM

Transfer Entorpy within the Nasdaq Stock Exchange

Wall 210

Using Transfer Entropy, we attempt to compare information flow between varying industries and the economy as a whole through studying 153 companies in the NASDAQ Stock Exchange and the NASDAQ Composite as an economic indicator. Transfer Entropy provides a model-free approach to detect asymmetrical statistic dependencies and correlations allowing us to calculate both the magnitude and direction of information flow. A delay variable is introduced to determine which industries tend to lead mass movements in the stock market, as well as to characterize the duration of influence leading these movements. Using a time series from 1 January 2000 to 20 July 2016, we perform a time-series analysis between each individual company and the NASDAQ Composite. Our data suggests that the consumer durables and finance industries provide the most information concerning the economy as a whole. Using a one-day delay, we find that introducing a delay variable significantly increases Transfer Entropy of all industries aside from energy, with technology having the most significant increase (103%), compared to the average (28%). After a one-day delay, we find that additional delay decreases Transfer Entropy generally, but not monotonically, suggesting that movements are mostly influenced by two-day trends. The introduction of a delay variable to studying information flow allows us to expand the applications of Transfer Entropy in econophysics, specifically the ability to determine not only magnitude and direction, but also duration. From this, we are able to better understand what leads market movements such as the 2009 global financial crisis.