Herding behavior empirical evidence in the vietnamese stock market

MINISTRY OF EDUCATION & TRAINING HO CHI MINH CITY OPEN UNIVERSITY - PHAM THI THANH AN HERDING BEHAVIOR: EMPIRICAL EVIDENCE IN THE VIETNAMESE STOCK MARKET MASTER THESIS FINANCE AND BANKING Ho Chi Minh City, 2018 MINISTRY OF EDUCATION & TRAINING HO CHI MINH CITY OPEN UNIVERSITY - PHAM THI THANH AN HERDING BEHAVIOR: EMPIRICAL EVIDENCE IN THE VIETNAMESE STOCK MARKET Major: Finance and Banking Major code: 60 34 02 01 MASTER THESIS FINANCE AND BANKING Academic Supervisor: Dr Duong Minh Chau Ho Chi Minh City, 2018 DECLARATION I, Pham Thi Thanh An, declare that the research work reported in this dissertation is my own It is submitted to fulfill the requirements for the Masters of Finance and Banking degree at the Open University of Ho Chi Minh City This thesis has not, either in whole or in part, been submitted for a degree or diploma to any other institution or university for a similar qualification except where otherwise indicated and acknowledged There are no other research papers used in this dissertation that are not cited in accordance with regulations This thesis has never been submitted to receive any degree at other universities or training institutions Ho Chi Minh City, 2018 Pham Thi Thanh An i ACKNOWLEDGEMENT Firstly, I would like to express my sincere gratitude to my advisor Prof Duong Minh Chau for the continuous support of my Master thesis, for his patience, motivation, and immense knowledge His guidance helped me in the time of writing research proposal and my thesis I feel lucky thanks to having a devoted support and direction of my advisor for my thesis Beside my advisor, I would like to thank to Ms Ha, Ms.Uyen and all the employees working in Office of Postgraduate Management at Open University Ho Chi Minh City for helping me in procedure of my Master thesis My sincere thanks also goes to my classmates working and helping me in all the time of doing my thesis They are one of the motivations for me to complete my study Thanks for every night, every weekend in three years working with me I appreciate your enthusiasm Last but not the least, I would like to thanks my family: my parents and to my brother for supporting me spiritually and support finance throughout my learning and writing my Master thesis ii ABSTRACT The thesis was conducted to find out the existence of herding behavior and to identify strong herding points and strong influential points on the Vietnamese stock market from 2006 to 2016 Data collected includes non-financial companies on both Ho Chi Minh Stock Exchange and Ha Noi Stock Exchange from 2006 to 2016 The data is divided: the entire period from 2006 to 2016, individual years and sub-periods consisting of before crisis 2006-2007, during crisis 2008-2012 and after the crisis 2013-2016 The model used includes the Cross Sectional Absolute Deviation (CSAD) model of Chang et al (2000) to determine the presence of herding behavior and Weighted Cross- Sectional Variance (WCSV) model of Xie et Al (2015) to find strong herding points and strong influential points on the Vietnamese stock market In addition, the thesis also compares two models: WCSV model based on FamaFrench three factors according to Arbitrage Pricing Theory (APT) and WCSV model based on CAPM one factors through adjusted R2 The purpose of the study is to investigate the impact of herding behavior of individual investors as well as to consider the suitability of the WCSV model when comparing strong herding spots to those of prominent events that influence to the Vietnamese stock market in the period from 2006 to 2016 Regarding the CSAD regression model, the results show that the herding behavior exists in the Vietnamese stock market on both HOSE and HNX in the overall period from 2006 to 2016, individual years (except 2008, 2013, 2015 and 2016) and in three sub-periods: before, during and after the crisis A comparison of the two models including WCSV model based on Fama-French three-factor model and WCSV model based on CAPM one factor shows that the WCSV model according to APT has a higher Adjusted R2 than the WCSV model based on CAPM one factor model Through the WCSV model based on Fama-French three factors as the basic APT, the results show that strong herding points and strong influential points appear in years, periods of relevance to prominent events that strongly influence the Vietnamese stock market In general, strong herding points and strongly influential points appear iii energetically in the periods consisting of before crisis 2006-2007 and during crisis 2008-2012 However, in 2008, the CSAD model does not find the existence of herding behavior, whereas the WCSV model yields strong herding spots at high incidence In addition, there is generally no difference between the results given in the two models as well as there are not many differences between the selected timelines Thus, the above results show that there is a compatible combination of CSAD and WCSV models in determining the existence of herding behavior, eliminating a hypothetical factor of the WCSV model and determining the strong herding points and strong influential points correspond to the time of prominent events on the Vietnamese stock market iv TABLE OF CONTENTS Page Declaration i Acknowledgement ii Abstract iii Table of contents v List of figures vii List of tables viii Abbreviations x CHAPTER INTRODUCTION 1.1 Background 1.2 Research objectives 1.3 Research questions 1.4 Research Subject and Scope 1.5 Contribution .4 1.6 Structure CHAPTER LITERATURE REVIEW AND HYPOTHESES 2.1 Behavioral Finance 2.1.1 Behavior Finance 2.2 Herding Behavior .8 2.2.1 Definitions of Herding Behavior 2.2.2 Types of Herding Behavior 10 2.3 Empirical Evidence 12 2.4 Hypotheses .23 CHAPTER RESEARCH METHODS 24 3.1 Data Description 24 3.2 Research methods 24 v 3.3 Selection of Models .25 3.3.1 The Return Dispersion Model by Chang et al (2000) 25 3.3.2 The Return Dispersion Model by Xie et al (2015) 26 3.3.3 Test for strong herding points and strongly influential points 29 CHAPTER DATA ANALYSIS AND FINDINGS 32 4.1 Descriptive Statistics 32 4.1.1 Descriptive Statistics 32 4.2 Empirical Results 41 4.2.1 Results of testing the existence of herding behavior through CSAD model .41 4.2.2 Empirical results of testing strong herding points and strong influential points 48 CHAPTER CONCLUSIONS AND RECOMMENDATION 58 5.1 Implications 58 5.2 Limitations and Recommendations .61 REFERENCES 63 APPENDIX A 71 APPENDIX B 75 vi LIST OF FIGURES Page Figure 4.1: Time series of the WCSV models during the period from 2006 to 2016 on HOSE 32 Figure 4.2: Time series of the WCSV models during the period from 2006 to 2016 on HNX .33 vii LIST OF TABLES Page Table 3.1: A summarization of strong herding point testing 29 Table 3.2 A detailed summarization of testing strong herding points in the thesis 30 Table 4.1: Summary statistics of dependent variables and independent variables of WCSV model based on Fama-French Three Factor Model for 2006-2016 on HOSE .36 Table 4.2 : Summary statistics of dependent variables and independent variables of WCSV model based on Fama-French Three Factor Model of Year 2006-2016 on HNX .37 Table 4.3: Description of variables in WCSV model in three sub-periods: before, during and after crisis on HOSE .38 Table 4.4: Description of variables in WCSV model in three sub-periods: before, during and after crisis on HNX 39 Table 4.5: The correlation matrix of variables in WCSV model based on FamaFrench three factors on HOSE 40 Table 4.6: The correlation matrix of variables in WCSV model based on Fama French three factors on HNX 40 Table 4.7: Testing multicollinear of variables in WCSV model based on FamaFrench three factors over the period from 2006 to 2016 41 Table 4.8: Testing multicollinear of variables in WCSV model based on FamaFrench three factors in three-sub periods: before, during and after crisis 41 Table 4.9: OLS estimators of CSAD model on HOSE .44 Table 4.10: OLS estimators of (a) WCSV model based on Fama –French three factors as the basic APT (b) WCSV model based on CAPM one factor on HOSE 45 Table 4.11: OLS estimators of CSAD model on HNX .46 Table 4.12: OLS estimators of (a) WCSV model based on Fama –French three factors (b) WCSV model based on CAPM one factor on HNX .47 Table 4.13: Summary of strong herding points and strong influential points with single year data from 2006 to 2016 on HOSE .53 viii 76 10 11 12 10 11 12 2007 2007 2007 2007 2007 2007 2007 2008 2008 2008 2008 2008 2008 2008 2008 2008 2008 2008 2008 2009 2009 2009 2009 2009 2009 2009 2009 6 14 10 12 11 10 0 0 11 12 13 15 12 16 10 16 11 11 1 1 2 0 0 76 0 1 1 0 2 1 0 20 16 17 14 20 21 14 12 7 10 14 12 15 10 12 18 13 12 16 12 15 18 21 21 23 19 20 16 16 10 10 10 10 10 16 16 18 11 12 12 19 20 0 0 0 0 0 0 0 1 0 0 0 0 11 0 0 0 77 10 11 12 10 11 12 10 11 2009 2009 2009 2009 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 2011 2011 2011 2011 2011 2011 2011 2011 2011 2011 2011 6 0 3 5 0 2 1 0 4 2 3 2 2 1 77 2 1 0 0 1 1 2 2 19 16 11 15 12 14 19 17 15 20 22 16 18 19 15 14 18 14 19 14 11 11 17 12 10 18 17 21 20 17 15 17 14 21 19 16 22 22 16 16 19 15 14 18 13 17 17 17 15 18 21 15 19 18 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 78 12 10 11 12 10 11 12 2011 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2013 2013 2013 2013 2013 2013 2013 2013 2013 2013 2013 2013 2014 2014 3 5 7 11 4 3 1 0 0 0 0 0 3 3 4 3 2 78 0 1 1 2 0 0 0 1 15 11 15 14 16 14 16 14 15 10 15 12 12 10 11 16 11 19 19 13 15 17 15 15 13 15 15 16 16 18 19 16 17 18 18 15 18 19 18 20 14 21 18 22 20 21 22 20 23 19 20 17 16 0 0 0 0 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0 79 10 11 12 10 11 12 2014 2014 2014 2014 2014 2014 2014 2014 2014 2014 2015 2015 2015 2015 2015 2015 2015 2015 2015 2015 2015 2015 2016 2016 2016 2016 2016 9 10 2 0 2 2 1 0 0 0 0 0 0 0 0 0 1 2 0 1 2 0 79 1 0 0 0 0 0 0 0 0 0 8 14 17 19 17 19 16 16 18 14 20 19 19 21 21 19 17 21 17 22 13 15 20 17 12 12 13 14 19 23 20 19 22 20 22 20 14 22 19 20 22 23 21 21 22 20 23 17 16 23 20 20 1 0 0 0 0 0 0 0 0 0 0 0 4 0 0 0 0 0 0 0 0 0 0 0 0 80 10 11 12 2016 2016 2016 2016 2016 2016 2016 3 3 0 0 2 3 0 1 1 0 14 17 18 15 17 22 19 21 19 23 20 20 22 21 0 0 0 0 0 0 0 Source: Author’s calculation Table 2: A detailed description of frequency appears strong herding points, weaker herding points, inadequate points and strong influential points examined by WCSV model based on Fama and French three factors as the basic APT with single – year data of Year 2006-2016 Single With –year assumption data Month Year 2006 2006 2006 2006 2006 2006 2006 2006 2006 Strong WCSV HOSE 2 5 WCSV HNX 8 Weaker WCSV HOSE 5 4 80 WCSV HNX 1 2 Inadequate WCSV HOSE 13 14 15 11 13 11 14 11 WCSV HNX 17 16 14 10 15 16 14 18 Strongly influential WCSV HOSE 0 1 0 WCSV HNX 0 81 10 11 12 10 11 12 10 11 12 2006 2006 2006 2007 2007 2007 2007 2007 2007 2007 2007 2007 2007 2007 2007 2008 2008 2008 2008 2008 2008 2008 2008 2008 2008 2008 2008 10 9 3 4 1 0 0 13 14 12 11 6 12 11 10 10 13 9 5 2 3 7 3 5 2 4 81 12 10 13 11 10 10 14 8 13 11 15 10 16 12 13 21 21 21 11 19 11 18 18 17 19 17 11 13 13 10 10 10 10 12 11 13 1 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 1 82 10 11 12 10 11 12 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 2011 2011 2011 11 7 2 10 5 0 1 2 2 0 2 3 5 1 3 6 4 0 1 1 82 10 12 6 11 12 14 15 12 14 13 12 12 20 20 11 16 16 9 17 12 16 16 15 17 12 13 13 15 19 21 20 17 12 15 14 20 19 16 21 22 16 15 19 14 12 16 13 15 0 0 0 0 0 1 0 0 0 0 2 0 0 2 0 0 0 2 83 10 11 12 10 11 12 2011 2011 2011 2011 2011 2011 2011 2011 2011 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2013 2013 2013 2013 2013 2013 12 6 5 5 5 1 2 0 2 2 3 1 1 4 5 2 6 5 2 83 14 14 12 11 12 17 15 14 14 16 13 15 12 13 13 12 10 13 12 17 12 19 17 17 16 14 18 18 14 19 16 16 16 16 16 16 16 18 18 17 15 19 19 20 13 11 14 12 19 19 0 0 1 0 0 1 0 0 0 1 2 0 0 0 0 84 10 11 12 10 11 12 2013 2013 2013 2013 2013 2013 2014 2014 2014 2014 2014 2014 2014 2014 2014 2014 2014 2014 2015 2015 2015 2015 2015 2015 2015 2015 2015 2 3 2 2 2 0 2 0 2 0 1 4 3 2 4 4 5 1 2 2 84 17 15 18 20 14 18 12 17 11 10 13 15 18 20 18 17 14 18 12 13 14 18 17 12 14 19 18 17 22 16 18 17 16 16 14 12 20 21 20 19 20 20 20 16 14 15 16 16 20 20 15 16 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 0 85 10 11 12 10 11 12 2015 2015 2015 2016 2016 2016 2016 2016 2016 2016 2016 2016 2016 2016 2016 5 2 2 0 1 0 1 3 3 2 1 3 0 17 15 16 11 13 19 12 16 15 15 16 18 17 18 18 20 16 22 15 16 23 18 20 17 16 18 20 18 14 20 0 1 2 0 0 0 0 1 1 0 0 0 Source: Author’s calculation Table 3: A detailed description of frequency appears strong herding points, weaker herding points, inadequate points and strong influential points examined by WCSV model based on Fama and French three factors as the basic APT with three subperiods including before crisis: 2006-2007, during crisis 2008-2012 and after crisis 2013-2016 of Year 2006-2016 Three subperiods data Month With assumption Year Strong WCSV HOSE 2006 Weaker Inadequate Strongly influential WCSV WCSV WCSV WCSV WCSV WCSV WCSV HNX HOSE HNX HOSE HNX HOSE HNX 12 17 0 85 86 10 11 12 10 11 12 2006 2006 2006 2006 2006 2006 2006 2006 2006 2006 2006 2007 2007 2007 2007 2007 2007 2007 2007 2007 2007 2007 2007 2008 2008 2008 2008 3 6 14 4 11 14 7 0 2 10 10 10 13 13 11 9 7 10 7 8 1 86 2 3 2 2 2 1 10 6 10 10 13 12 11 11 12 14 12 10 11 9 15 14 11 15 17 16 19 21 22 18 19 11 10 18 18 17 19 17 11 13 13 10 10 10 0 1 0 1 0 1 0 0 0 0 0 0 4 87 10 11 12 10 11 12 2008 2008 2008 2008 2008 2008 2008 2008 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2009 2010 2010 2010 2010 2010 2010 2010 11 11 13 13 8 12 11 3 3 14 10 14 12 8 0 3 0 2 0 0 2 1 1 1 2 2 0 87 0 4 2 0 0 0 12 15 10 17 11 11 15 12 10 14 17 16 16 11 14 11 12 18 14 17 21 21 11 12 11 10 13 16 19 20 13 15 13 19 20 21 21 17 15 17 15 21 19 16 22 22 0 0 0 0 4 0 0 0 0 0 0 0 1 0 0 0 0 0 0 88 10 11 12 10 11 12 10 2010 2010 2010 2010 2010 2011 2011 2011 2011 2011 2011 2011 2011 2011 2011 2011 2011 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 6 3 6 2 2 2 3 1 4 6 3 1 88 3 2 2 1 2 19 17 20 15 17 17 15 17 13 10 15 11 10 17 15 14 15 11 16 14 17 15 13 14 16 15 19 14 14 19 13 17 17 15 16 18 21 15 20 18 18 16 17 18 19 16 17 20 18 16 19 0 0 0 0 1 0 1 0 0 1 0 0 1 0 0 89 11 12 10 11 12 10 11 12 2012 2012 2013 2013 2013 2013 2013 2013 2013 2013 2013 2013 2013 2013 2014 2014 2014 2014 2014 2014 2014 2014 2014 2014 2014 2014 2015 8 3 6 6 5 2 0 1 0 1 1 3 2 2 1 5 4 2 0 89 1 2 0 0 0 2 12 11 10 11 16 18 15 16 15 16 17 10 19 13 17 11 11 12 16 17 18 20 18 17 15 18 19 20 19 12 20 17 21 19 20 22 20 23 20 20 17 16 13 12 12 18 20 20 17 20 20 16 19 2 1 0 1 0 0 0 0 0 0 0 0 0 0 4 0 1 90 10 11 12 10 11 12 2015 2015 2015 2015 2015 2015 2015 2015 2015 2015 2015 2016 2016 2016 2016 2016 2016 2016 2016 2016 2016 2016 2016 2 2 0 3 0 0 0 0 0 0 0 0 0 1 3 1 1 0 0 90 0 0 0 0 0 1 1 1 13 19 16 17 22 18 17 18 22 17 20 13 15 22 17 18 18 18 19 18 20 22 19 14 0 21 0 17 0 19 0 22 0 23 0 18 1 21 0 22 0 20 0 23 0 18 1 16 0 23 0 19 0 20 0 21 0 18 22 0 20 0 20 0 21 0 21 Source: Author’s calculation ... occurrence of herding and the strong influential points in time points Finding the extremely herding pattern helps to examine the influence of herding in each time point to the stock market Therefore,... in this study The aim of this study is finding herding existence in emerging markets and specifically in Chinese and Indian stock markets including testing herding in both stock markets, during... types of herding behaviors and the selection of the thesis in definite the herding and kind of herding in studying for the Vietnamese stock market Besides, previous researches about herding behavior