Title: Study on Corporate Governance of Stock Market in Korea: Network Analysis with relationship of Major Shareholders

Authors: Hyejung Moon; JangHo Park; SungKyung Kim

Abstract― The purpose of this paper is to measure the sustainability of corporate through using the relationship of major shareholder using complex system. As a result of case study, we can find two types of meaningful network. First type is the clustering according to connectivity among major shareholders. Second Type is the network according to corporate group. First type shows small world network and characteristics of corporate group such like business style, decision making speed and owner risk. Second type shows scale-free network and investment preference according flowing power law in stock market. Third type is the clustering according to industry sector of stock focusing on Information Technology.

Keywords: Stock Market, Complex System, Corporate Governance, Big Data, Small world network, Scale-free network

I. INTRODUCTION

The purpose of this paper is to measure the sustainability of enterprise through using the relationship of major shareholder without using economic index such like sales volume, profit and stock price. These characteristics show complex society such like unbalance relationship among various shareholders in stock market. We used complex system as a theory to understand enterprise governance following power law (Arthur et al., 1997; Newman, 2000; 2005; Strogatz, 2001; Tse et al., 2010).


Total amount of stock owned by major shareholders is over 556 trillion KRW . National Pension Service owned over 51 trillion KRW of stock. SamSung Insurance owned 19 trillion KRW of stock. LG owned 13 trillion KRW of stock. Top 10 shareholders owned over 35% of total amount.

II. COMPLEX SYSTEM OF STOCK MARKET

A. Complex System

Illustrating the relationship between behaviors like the society of major shareholders of stock market generally use network graph. Existing theory of a small world network has been suitable to describe the society issue with perspective of microeconomic (Levy et al., 2000) such like Six degrees of Kevin Bacon (Watts, 2003), Small world experiment (Newman, 2000; 2005; Watts & Strogatz, 1998). However, the theory of scale-free network is more suitable to explain complex relationship among actors such like countries and companies with perspective of macroeconomics including uncertainty and inequality (Barabasi & Albert, 1999; Wang & Chen, 2003). The cases of scale-free network are Pareto distribution, Long tail law, the collaboration of movie actors in films and WWW (Dorogovtsev & Mendes, 2003).




B. Previous Studies and difference of this paper

Existing papers of stock market using network analysis were the studies for forecasting the stock price using correlation and regression (Bak  et al., 1997; Lux & Marchesi, 1999; 2000). Other papers of major shareholder in stock market were studies about business influences when there is the change of business environment like regulation or structure of major shareholders in enterprises (Ram & Uttam, 2013; Steve  & Igor, 2004).
 Both types of previous studies were lack to explain the phenomenon of stock market according to power laws among major shareholders. I would like to use complex system as a theory to understand business sustainability using network analysis with the relationship between major shareholders in stock market.

III. RESEARCH DESIGN

A. Research Process

This study is a typical inductive research. Durkheim (1897) had spent more than 10 years to find the reason (religion) of suicide after finding statistics of suicide in Germany. However recently in big data era, we can search, gather, visualize and understand huge data in real time. For understanding the governance of enterprises in stock market, I gathered whole data about stock and shareholders at Jan, 17th, 2015 and I conducted inductive analysis using big data tool like network graph at visualization step (Hommes, 2006).


B. Analysis Frmaework

There are three analysis areas in stock market. First area is enterprise value like stock as a node (or vertice) of network. Second area is relationship of major shareholders as a link (or edge) of network. Third area is attribute such like price, number, sector and group of stock. To understand enterprise governance, I analyzed three criteria of network with graph type, features, and centrality. First, there are small world network and scale-free network in network type. Second, there are large hub, density and clustering using Closet-Neman-Moore algorithm. Third, there is centrality with degree, closeness and betweenness.



IV. CASE STUDIES

Before clustering, network graph has very high number of nodes and complex edges. We couldn’t understand stock market with initial graph. We have conducted various clustering with this graph using attributes (sector, group, stock volume such like price and number) and connectivity (degree, closeness and betweenness).  As a result of network analysis, we can find third types of meaningful graph using 39,000 nodes (stock of 2,000 enterprise and over 37,000 major shareholder data). First type is the network according to enterprise group. Second type is the clustering according to connectivity among major shareholders. Third type is the clustering according to industry sector of stock.



A. Small world Network with Erterprise group

 1,300 stocks among about 2,000 companies in stock market are included in subsidiaries of the 35 groups like LG, GS, DongBu, SamSung, HanHwa and etc. When we clustered the relationship among major shareholders in stock market according to enterprise group, each cluster has small world network with similar small degrees of network with short path (Oliver et al., 2008).


There are two singularities as results of clustering according to enterprise group. First, there are high performance of economic growth, decision-making speed, and sustainability in clusters including large nodes like Samsung, LG and HyunDai Car (Bikhchandani & Sharma, 2001; Cont & Bouchaud, 2000). We could find very high trust relationships like cooperation in these clusters (Axelrod,  1984).


Second, there occur separation of management control and owner risk in clusters composed of several similar small nodes as an individual person such like Lotte, CJ and HanJin including complex network without large hubs. There is also high betrayal relationship in these clusters.



B. Scale-free Network of major shareholders

This graph has illustrated by Closet-Neman-Moore algorithm. This algorithm illustrate network according to centrality between nodes. This algorithm can extract meaningful communities from this network, revealing large-scale patterns present in the purchasing habits of customers. Top 10 of total 349 clusters have 35% of nodes(1,351) among 3,918 nodes after clustering according to connectivity in major shareholders. These clusters show the characteristics of investment flowing power law in scale free network. C1 is mainly composed of a large investor as an enterprise. C2 is composed of SamSung and CJ related investors. C3 is composed of GS and HyunDai related investors. C6 is composed of Posco related investors. C8 is composed of Korea Tire related investors. C10 is composed of SongWon related investors. C4 is composed of government investors. Other C5, C7, C9 are composed of Securities companies. This graph shows typically scale-free network with high complexity. There are also large hubs like Korea fund, Samsung, LG and etc.




C. Cluster by Industry Sector

This graph has illustrated by industry sector of major shareholders in stock market. Most activate sector is information and Technology with 452 nodes. Next activate sectors are Industrials (341 nodes), Consumer Discretionary (309 nodes), Materials (260 nodes), Health Care (129 nodes), Consumer Staples (89 nodes), Financials (79 nodes), Utilities (16nodes), Energy (13 nodes) and Telecommunication Services (10 nodes).


The ordinary of economic volume in Korea are energy, Compulsory (education; human health), Distributive trade(including repairs; transport; accommod., food), Real estate activities, Construction, Financial and insurance activities, scientific and technology, Agriculture(including forestry and fishing), Information and communication, Other services by OECD since 1980. However the volume of stock market in Information Technology area is higher than other sectors. This means that the public has high interest in IT area. That means that IT leads whole kinds of industries in Korea. IT might be one of core success factor of business and education in the future.

V. CONCLUSION

As a result of case study, we can find two types of meaningful network. First type is the clustering according to connectivity among major shareholders. Second Type is the network according to enterprise group. First type shows small world network and characteristics of enterprise group such like business style, decision making speed and owner risk. Second type shows scale-free network and investment preference according flowing power law in stock market. Third, sector of Information Technology is most big area of stock market in Korea. An insight of this study is that we can find sustainability using the relationship among major shareholders in stock market including availability of growth in the future. However we couldn't find the reason of different two types and these characteristics, it may be from culture or maturity of democratization in Korea. We would like to find these reasons at next study.



Study on Corporate Governance of Stock Market in Korea: Network Analysis with relationship of Major Shareholders by Hyejung Moon is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.
Based on a work at www.itpolicy.co.kr.