CHARACTERISTICS ANALYSIS ON GLOBAL STOCK MARKETS DURING FINANCIAL CRISIS 2008 BY USING RMT-LOUVAIN/LPAM+/COMBO-WAVELET COHERENCE

• Main Author: IQBAL ARRAFI'I (NIM : 10214081), MUHAMMAD
• Format: Final Project
• Language: Indonesia
• Subjects: Fisika
• Online Access: https://digilib.itb.ac.id/gdl/view/29255
• Institution: Institut Teknologi Bandung

Study on stock market closing day prices was done for DAX, FTSE, HSI, Nikkei225, S&P500 and SSE Composite during June 2006-June 2011 which has a purpose to know about stock market characteristics during global financial crisis 2008. Four main steps of this study are data segmentation, data filtering, data clustering, and comovement finding on stock. Data segmentation was done to divide it into three parts which are pra crisis (June 2006-November 2007), crisis (December 2007-June 2009), and pasca crisis (January 2010-June 2011). RMT (Random Matrix Theory) was implemented on the data to eliminate noises and market wide effect within the data. Data clustering was done by using three different networking algorithms which are LPAm+, Louvain, and COMBO which has a purpose to evaluate the formed networks. Comovement finding on the most influential share for each stock indice was done using wavelet Morlet coherence. The result shows that the average value of stock correlation coefficient has the largest value during the crisis (0.226528007) compared with pra crisis (0.190553208) and pasca crisis (0.224701109). Network threshold of stock market correlation network shows the smallest value during the crisis (0.046666667) compared with pra crisis (0.0475) and pasca crisis (0.063333333). The average value of the highest betweenness centrality for six indices shows the smallest value during the crisis (0.030783333) compared with pra crisis (0.071742333) and pasca crisis (0.091245167). The average value of network modularity shows the smallest value during the crisis (0.331859901) compared with pra crisis (0.398361442) and pasca crisis (0.458489167). Comovement finding of the most influential share for each stock indice shows the highest tendency to be positively correlated during the crisis (+7) compared with pra crisis (+6) and pasca crisis (+4).