(Luận văn) behavioral factors affecting herding bias, the case of ho chi minh stock exchange, vietnam

UNIVERSITY OF ECONOMICS HO CHI MINH CITY t to International School of Business ng hi  ep w n lo ad y th ju Doan Thi Mai Phuong yi pl n ua al BEHAVIORAL FACTORS AFFECTING HERDING BIAS: THE CASE OF HO CHI MINH STOCK EXCHANGE, VIETNAM n va ll fu oi m at nh z z vb k jm ht SUPERVISOR: Prof Nguyen Dong Phong Dr Nguyen Phong Nguyen om l.c gm an Lu n va ey t re HO CHI MINH CITY - 2015 TABLE OF CONTENTS t to CHAPTER 1: INTRODUCTION ng hi 1.1 Background ep Problem statement 1.3 Research question 1.2 w n lo Research scope 1.5 Research methods 1.6 Significance of the research 1.7 Structure of the study ad 1.4 ju y th yi pl al ua CHAPTER 2: LITERATURE REVIEW 10 Theorical background 10 2.2 Review on some behavioral factors and herding bias in stock market 12 n 2.1 n va fu ll 2.2.1 Risk tolerance 12 oi m nh 2.2.2 Over-confidence 13 at 2.2.3 Self-monitoring 15 z z 2.2.4 Gambler’s fallacy 16 vb jm ht 2.2.5 Illusion of control bias 16 k 2.2.6 Herding bias 17 Hypothesis development 20 l.c gm 2.3 CHAPTER 3: RESEARCH METHODOLOGY 25 om 3.1 Measurement scales 25 an Lu 3.1.1 Scales measurement of Risk Tolerance 25 n va 3.1.2 Scales measurement of Over-confidence 25 3.1.5 Scale measurement of Illusion of control 28 ey 3.1.4 Scale measurement of Gambler’s Fallacy 27 t re 3.1.3 Scale measurement of Self-monitoring 26 3.1.6 Herding bias 28 t to 3.2 Sampling 29 ng 3.3 Data collection methods 29 hi ep 3.4 Data analysis methods 30 w 3.4.1 Test of scale measurement reliability 30 n 3.4.2 Exploration factor analysis (EFA) 31 lo ad CHAPTER 4: DATA ANALYSIS AND FINDINGS 33 y th ju 4.1 Descriptive statistics: 33 yi 4.2 Refinement of measurement scales 34 pl ua al 4.2.1 Result of Cronbach’s alpha analysis of formal survey (N=205) 34 n 4.2.2 Factor analysis (EFA) 35 va n 4.3 Testing the assumptions of regression 39 fu ll 4.4 Results of hypothesis testing 40 m oi 4.4.1 Regression model 40 nh at 4.5 Chapter summary 43 z CHAPTER 5: CONCLUSIONS AND IMPLICATIONS 44 z ht vb 5.1 Main Findings 44 k jm 5.2 Managerial Implications 45 gm 5.3 Limitations 47 l.c REFERENCES 48 om APPENDICES 57 an Lu n va ey t re ABSTRACT t to ng hi The stock market is more and more unpredictable which traditional finance theories cannot give reasonable explanation.Behavioral finance, instead, can be helpful in the current situation of the fluctuating market Herding activities among investors have been a popular behavioral explanation for the excess volatility and short term trends observed in financial market However, the number of research focusing on herding and its impacts on the financial market, especially Vietnamese stock market, is limited.This research will identifythe behavioral factors affecting the herding bias existing at the HOSE since it was established (2000) and the impact levels of behavioral elements on the herding bias ep w n lo ad ju y th yi pl Using a data set form a sample of more than 200 investors, we find that investors’ decisions much depends on their own emotion rather than fundermental analysis and technical analysis Moreover, investors generally prefer short-term portfolios (T+3) to long-term ones n ua al n va ll fu In addition,due to the limited level of security understanding, investors usually consider the stock market as a “casino” where luck is the most vital element to win On the other hand, investorswho are too confident about their investment decision usually make mistakes in their short-term investment oi m at nh z z Keywords: Herding behavior, behavioral finance k jm ht vb om l.c gm an Lu n va ey t re CHAPTER 1: INTRODUCTION t to 1.1 Background ng hi Vietnamese stock market was formed in 1998 including Ho Chi Minh City ep Stock Exchange (HOSE) and Hanoi Stock Exchange (HNX) At the very first w stage, the Vietnamese stock market was launched on 28 July 2000 with n lo merely listed companies along with security companies and the growth of ad y th the number of listed companies was quite slow Having passed so many ups- ju and-downs, there are currently 181 security-company members1after yi approximately 14 years However, in comparison to other developed and pl ua al emerging markets, Vietnamese stock market appears to be much smaller in n terms of scale and maturity Although the Ho Chi Minh City stock exchange n va (HOSE) has witnessed considerable developments both in the number of ll fu listed stocks and in transaction value for 14 years, the price movement seems oi m to fluctuate unpredictably over different periods Starting at 100 points in nh July 2000, after year, at June 2001, the VN-Index was fivefold and reached at the peak at 571 points Investors were too excited and dreamed of earning z z money quickly that the demand rose significantly while there were only few vb ht listed stocks in the market (Huy, 2010) However, it was the lack of k jm knowledge and trading experience of investors as well as the insufficient gm support from the authorities which made VN-Index suddenly unceasingly fall l.c until it reached the bottom at 139 points in March 2003 (Huy, 2010) om Investors who joined the market in this period and could not jump out an Lu quickly had to face the financial difficulties because of the huge loss of assets The stock market then seemed to fall in its hibernation status until 2005 and n va eventually woke up in 2006 The boom started in the second half of 2006 and ey t re 1According to the latest available data as at 20 May 2014 on the websites of Ho Chi Minh City and Ha Noi stock markets rocketed up to 1,170 points by March 2007 This fluctuated around 1,000 t to points until October 2007 The Ho Chi Minh City stock market had never been ng “hotter” than that time Nevertheless, VN-Index went to its decline stage after hi ep being pushed up to the peak Ho Chi Minh City stock market experienced a gloomy year in 2008 and VN-Index merely stopped falling in February 2009 w at 235 points (Huy, 2010) The possible explanations for the sharp dip of the n lo ad VN-Index were the impacts of numerous factors, namely, tightening of y th monetary policies, especially lending for stock investment, high deposit ju interest rates, high inflation rate, and a recession of the United State yi pl economy Moreover, lack of timely intervention by the authorities was also a ua al breeding ground of the dramatically decrease of VN-Index (Vo and Pham, n 2008) From 2009 to the first quarter of 2011, VN-Index continued to va n undergo many ups-and-downs, reached another peak at 542 points in May ll fu 2010 and another bottom at 351 points in November 2011 Nonetheless, it oi m seemed to fluctuate between 400 and 500 points with no significant at nh amplitude found in 2012 and finally stood at 505 points on December 31st z 2013, increasing by 23% compared to 2012 (Luyen, 2013).According to z Phu(2010), one of the most important factor resulting in the fluctuation of vb jm ht stock prices is the herding bias Specifically, when the price of stock increases, k a huge number of investors begin buying with the hope that the price will l.c gm continue to rise However, the price only goes up to a specific point and after that will go down When the price decreases, be afraid that the price decline om unceasingly, almost investors will sell with high selling pressure In the other an Lu words, investors seem to takes action followed their psychology and emotion rather than rational mind which does not make the market complied with stock markets; therefore, it should be analyze in depth ey most important factor leading to this phenomenon in the case of Vietnamese t re these fluctuations during the period, I assume that herding bias may be the n va supply-demand rule and may lead to the collapse of the market To explain Regarding the theoretical background on behavioral factors and herding bias, t to if Expected Utility Theory (EUT) may be deemed by a fundamental theory of ng traditional finance, prospect theory is the basic theory of the behavioral hi ep finance Prospect theory concentrates on subjective decision-making affected by the investors’ value system, while EUT focuses on investors’ rational w n expectations (Filbeck, Hatfield & Horvath, 2005) EUT is the normative model lo ad of rational choices and descriptive model of economic behaviors, which y th dominate the analysis of decision making under risk However, this theory is ju criticized for failing to explain why people are appealed to both insurance and yi pl gambling People tend to under-weigh probable outcomes compared to ua al certain ones and they differently response to the similar situation depending n on the context of losses and gains in which they are presented (Kahneman & va n Tversky, 1979) Prospect theory describes some states of mine affecting an fu ll individual’s decision making process including Regret aversion, Loss aversion m oi and Mental accounting (Waweru et al, 2003, p.28) nh Based on the background, the problem statement would state my concern at z about the impact of behavioral factors on herding bias in the case of z 1.2 Problem statement k jm ht vb Vietnamese stock market l.c gm Only recently established in 2000, the Vietnamese stock market is characterized by weak reporting requirements, poor regulations and low om accounting standards (Tran and Truong, 2011) During the formulation and an Lu development process, the Vietnamese stock exchange in general and the fluctuate unpredictably and it seems difficult for investors to make rational Vietnamese stock exchange, particularly dealing with herding issues ey well as the instability of the market, there have been few studies of the t re decisions Despite the dramatic fluctuations of price through the period as n va HOSE in particular went through many different stages Specifically, its prices Although a variety of commentaries based on the conventional financial t to theories have been proposed, they failed to explain what happened at the ng HOSE over the past period Alternatively, behavioral finance can be helpful in hi ep the case of the HOSE since it is based on the psychology to explain why people buy or sell stocks (Waweru et al., 2008) Behavioral factors include w overconfidence, representativeness, availability, loss aversion, regret n lo ad aversion, gamble’s fallacy, over-under reaction, herding and so on (Ritter, y th 2003) Herding activities among investors have been a popular behavioral ju explanation for the excess volatility and short term trends observed in yi pl financial market (Juan Yao et al., 2014) Moreover, many studies carried out ua al in Vietnamese stock market such as Farber et al (2006), and Tran n (2007)identified that herding effect in this market is very strong, especially va n toward positive return of the market Specifically, Farber et al (2006) ll fu conducted a study aboutthe impact of policy on Vietnam stock market and oi m found out the empirical evidence for the well-known herding behavior among at nh investors, by which people suppressed own private information and z expectations to follow the market’s collective action They also stated that z “the trend of herding behavior is stronger toward extreme positive returns of vb k (Farber et al, 2006, p.25) jm ht the market, and in fact, around the consecutive sequence of limit-hits” l.c gm Besides, Chen et al (2003) argued that herding was more likely to happen in emerging markets than developed ones as the government intervention was om high, and the quality of information disclosure was low.Because of this an Lu intervention of the government, the market did not operate following the “demand-supply” rule; therefore, there was no random trend in the market which resulted in the instability of markets Kaminsky and Schmukler (1999) ey information Hence, they always followed the rumor or unofficial information, t re was the fact that individuals in these markets did not have sufficient official n va As regard the low quality of information disclosure in emerging markets, it asserted that during 1997-1998, Asian countries seemed to be driven by t to herding behavior; therefore, it was believed that the herding instinct was ng very strong in Vietnam.Tran and Truong (2011) shown that herding behavior hi ep in Vietnamese stock market was evident Specifically, investors had a tendency to follow the actions of those who were believed to be better w informed They argued that the inadequacy of the regulatory framework in n lo ad Vietnamese market, for example, the lack of transparency or an efficient y th mechanism for reporting information prevented investors from collecting ju accurate and rapid firm-specific information for their own evaluation Hence, yi pl this informational inefficiency together with a relatively high degree of ua al market volatility might induce investors to make decisions based on n consensus, which lead to higher correlations among stock returns The va n dispersion among returns was, therefore, likely to decrease or at least ll fu increase at a decreasing rate As a result, herding was observed Thus, the oi m study aims to analyze herding effect and factors impacting it at nh To understand and give several suitable explanations for the herding bias in z the Vietnamese stock market, it is vital to explore which behavioral factors z affecting the herding bias at the HOSE and how these factors affect the vb jm ht herding bias It would be useful for investors to understand common k investment behaviors, from which justify their decisions for better returns l.c gm Security companies can use this information for the better understanding about investors to forecast more exactly and issue better recommendations om Above all, stock price would reflect its true value and the HOSE would also raise capital for production and expansion n Research question va 1.3 an Lu become the yardstick of the economy prosperity as well as help corporations established (2000) Furthermore, identifying the impact levels of behavioral ey existing at the HOSE over the period surveyed since the HOSE was t re This research identifies the behavioral factors affecting the herding bias elements on the herding bias at the HOSE is another purpose of this research t to For these reasons,this study proposes the following research question: What ng are the behavioral factors affecting investors’ herding bias in the Vietnamese hi ep stock market? 1.4 Research scope w n As mentioned above, the Vietnamese stock market solely includes two stock lo ad exchanges, namely, Ho Chi Minh City Stock Exchange (HOSE) and Hanoi Stock ju y th Exchange (HNX) In this study, the HOSE would be taken into account because yi Ho Chi Minh City is the biggest and the most unceasingly developing city in pl Vietnam Besides, the HOSE has become the yardstick of the economy wealth al ua as well as the most effective channel helping enterprises raise capital for n production and expansion (Luong and Ha, 2011) Above all, selecting the va n HOSE for this study provides a more convenient approach to collect data, fu ll enhance the accuracy of survey information as well as offer more reliable m oi interview data due to the budget and time constraints at Research methods nh 1.5 z This research will employ quantitative methods In particular, the research z ht vb process will include a pilot study and a survey The quantitative research will k gm in the HOSE jm be conducted by sending survey forms to more than 200 investors investing l.c In this study, the hypotheses will be built based upon the existing behavioral om finance theories In particular, the quantitative method will initially be an Lu presented to test these hypotheses To check the reliability of the scales measurement, the research will use SPSS software to test Cronbach’s Alpha hypotheses ey Finally, the paper will apply regression for testing the proposed model and t re measurement validity: factor analysis – EFA (Exploratory Factor Analysis) n va based on standardized items In the next step, this paper will test scales 2.207.1 2.208.1 2.209.1 2.210.1 2.211.1 X43 Some gamblers are just born lucky t to 2.212.1 2.213.1 2.214.1 2.215.1 2.216.1 ng X44 Wins are more likely to occur on a hot machine hi ep X45 A good casino gambler is like a quarterback who knows 2.217.1 2.218.1 2.219.1 2.220.1 2.221.1 winning plays and when to use them w X46 It is good advice to stay with the same pair of dice on a 2.222.1 2.223.1 2.224.1 2.225.1 2.226.1 n lo winning streak ad X47.Show me a gambler with a well-planned system and I would y th show you the winner 2.227.1 2.228.1 2.229.1 2.230.1 2.231.1 ju 2.232.1 2.233.1 2.234.1 2.235.1 2.236.1 yi X48.The longer I've been losing, the more likely I am to win pl X49.The more familiar I am with a casino game, the more likely I al ua am to win 2.237.1 2.238.1 2.239.1 2.240.1 2.241.1 n 2.242.1 2.243.1 2.244.1 2.245.1 2.246.1 va X50 One should pay attention to lottery numbers that often win n ll fu Herding bias would buy shares by following my friend's 2.257.1 2.258.1 2.259.1 2.260.1 2.261.1 z z period at Y52 I would bid the securities whose prices have risen for a nh recommendation 2.252.1 2.253.1 2.254.1 2.255.1 2.256.1 oi I m Y51 2.247.1 2.248.1 2.249.1 2.250.1 2.251.1 Y53 I would bid the same stocks as my friends jm ht vb 2.262.1 2.263.1 2.264.1 2.265.1 2.266.1 k 2.267.1 2.268.1 2.269.1 2.270.1 2.271.1 share price movements, trading volume) 2.272.1 2.273.1 2.274.1 2.275.1 2.276.1 om l.c Y55 I would invest in shares by following technical analysis (e.g gm Y54 I would follow the market information to trade 2.277.1 2.278.1 2.279.1 2.280.1 2.281.1 an Lu Y56 I would invest in shares by following the media 2.282.1 2.283.1 2.284.1 2.285.1 2.286.1 n 2.287.1 2.288.1 2.289.1 2.290.1 2.291.1 ey shares t re Y59 I rely on information from friends and relatives to trade va Y57 I rely on expert's recommendation to trade shares 60 APPENDIX t to ng CRONBACH’S ALPHA; EXPLORATORY FACTOR ANALYSES AND hi MULTIPLES REGRESSION RESULT ep Risk Tolerance w n lo ad ju y th yi pl n ua al n va ll fu oi m at nh z z k jm ht vb om l.c gm an Lu n va ey t re 61 t to ng hi ep w n lo ad ju y th yi pl n ua al Over-confidence n va ll fu oi m at nh z z k jm ht vb om l.c gm an Lu n va ey t re 62 t to ng hi Self-monitoring ep w n lo ad ju y th yi pl n ua al n va ll fu oi m at nh z z k jm ht vb om l.c gm an Lu n va ey t re 63 t to ng hi ep w n lo ad ju y th yi pl n ua al n va Gambler's Fallacy ll fu oi m at nh z z k jm ht vb om l.c gm an Lu n va ey t re 64 t to ng hi ep w n lo ad ju y th yi pl n ua al n va ll fu oi m at nh z z k jm ht vb Illusion of control om l.c gm an Lu n va ey t re 65 t to ng hi ep w n lo ad ju y th yi pl n ua al n va ll fu oi m at nh z z k jm ht vb om l.c gm an Lu n va ey t re 66 t to ng hi ep w n Herding bias lo ad ju y th yi pl n ua al n va ll fu oi m at nh z z k jm ht vb om l.c gm an Lu n va ey t re 67 t to ng hi ep w n lo EFA for all ad ju y th yi pl n ua al n va ll fu oi m at nh z z k jm ht vb om l.c gm an Lu n va ey t re 68 t to ng hi ep w n lo ad ju y th yi pl n ua al n va ll fu oi m at nh z z k jm ht vb om l.c gm an Lu n va ey t re 69 APPENDIX t to EFA and Reliability Test Results – Scales Without Modification ng hi ep Factor loading w Risk Tolerance n lo RIS3 ad RIS4 0.658 0.946 0.598 0.487 0.405 72.77 1.456 0.61 0.853 0.456 0.853 0.456 ua al CONFI7 1.384 pl CONFI6 46.12 45.387 2.269 n Self-monitoring 0.807 SEMO16 0.701 0.537 51.729 Illusion of Control 2.586 jm ht 0.877 0.537 vb GAM3 0.699 z 0.877 1.537 z GAM2 0.478 at 76.852 Gambler's Fallacy 0.611 nh SEMO14 0.367 oi 0.598 m SEMO13 0.378 ll 0.596 fu SEMO11 0.408 n 0.644 0.689 va SEMO7 0.582 ILLUS7 0.631 0.534 ILLUS10 0.649 0.559 ILLUS11 0.652 0.55 62.034 2.481 0.795 0.823 0.654 HERD3 0.834 0.67 HERD8 0.755 0.566 Factor %variance Eigen-value Item-total ey HERD2 t re 0.543 n 0.734 va HERD1 an Lu 0.695 om ILLUS5 l.c 0.451 gm 0.519 0.766 k ILLUS3 Herding Bias Cronbach’s alpha Item-total correlation 0.416 yi Over-confidence Eigen-value 0.501 ju y th RIS5 %variance extracted Cronbach’s 70 loading extracted correlation 46.12 t to Risk Tolerance 1.384 0.658 ng 0.501 0.416 RIS4 0.946 0.598 0.487 0.405 hi RIS3 ep RIS5 72.77 w Over-confidence n CONFI6 lo ad CONFI7 1.456 0.61 0.853 0.456 0.853 0.456 45.387 2.269 0.689 0.408 SEMO11 0.378 SEMO13 0.598 SEMO14 0.807 SEMO16 0.701 SEMO7 ju 0.644 yi y th Self-monitoring 0.596 pl al 0.367 n ua 0.611 va 0.478 76.852 1.537 n Gambler's Fallacy 0.877 m GAM3 0.537 oi 2.586 0.766 0.519 ILLUS5 0.695 ILLUS7 0.631 ILLUS10 0.649 ILLUS11 0.652 0.55 z z 0.582 vb 0.534 k jm 0.559 2.481 0.543 HERD2 0.823 0.654 HERD3 0.834 0.67 HERD8 0.755 0.566 an Lu 0.734 om HERD1 0.795 l.c gm 62.034 Herding Bias 0.451 at ILLUS3 ht nh 51.729 Illusion of Control 0.537 ll 0.877 0.699 fu GAM2 alpha n va ey t re 71 APPENDIX t to ng Items hi ep w Original scale Refind scale Factor Item-total Factor loading correlation loading correlation 0.349 0.288 Eliminated Eliminated 0.304 0.261 Eliminated Eliminated 0.605 0.476 0.501 0.416 0.774 0.55 0.946 0.598 0.527 0.408 0.487 0.405 Item-total Risk Tolerance RIS1 n lo RIS2 ad RIS3 RIS5 ju y th RIS4 yi 29.148 46.12 Eigen value 1.457 1.384 0.643 0.658 pl % variance extracted n ua 0.53 0.262 Eliminated Eliminated CONFI2 0.787 0.197 Eliminated Eliminated CONFI3 0.429 Eliminated Eliminated CONFI4 0.095 0.121 Eliminated Eliminated CONFI5 0.163 0.223 Eliminated Eliminated CONFI6 0.352 0.37 CONFI7 0.35 0.284 % variance extracted 19.53 72.77 Eigen value 1.367 1.456 Cronbach's alpha 0.514 0.61 ll fu CONFI1 gm n va Over-confidence al Cronbach's alpha m oi 0.315 at nh z 0.456 z 0.853 vb 0.456 k jm ht 0.853 om l.c Self-monitoring SEMO2 0.325 0.247 Eliminated Eliminated SEMO3 0.274 0.67 Eliminated Eliminated SEMO4 0.439 0.277 Eliminated Eliminated SEMO5 0.401 0.286 Eliminated Eliminated SEMO6 0.51 0.47 Eliminated Eliminated SEMO7 0.46 0.181 0.644 0.408 SEMO8 0.456 0.304 Eliminated Eliminated ey Eliminated t re Eliminated n 0.17 va 0.45 an Lu SEMO1 72 ng hi ep w 0.438 Eliminated Eliminated SEMO10 0.582 0.577 Eliminated Eliminated SEMO11 0.309 0.314 0.596 0.378 SEMO12 0.43 0.081 Eliminated Eliminated SEMO13 0.411 0.369 0.598 0.367 SEMO14 0.566 0.323 0.807 0.611 SEMO15 0.393 0.173 Eliminated Eliminated SEMO16 0.499 0.353 0.701 0.478 SEMO17 0.4 0.189 Eliminated Eliminated SEMO18 0.567 0.298 Eliminated Eliminated n 0.525 lo t to SEMO9 ad ju y th yi % variance extracted pl Eigen value 45.387 3.358 2.269 0.656 0.689 n va 0.142 n GAM1 ua Gambler's Fallacy al Cronbach's alpha 18.653 fu 0.714 GAM3 0.794 GAM4 0.122 % variance extracted 39.625 Eigen value 1.585 Cronbach's alpha 0.426 Eliminated Eliminated 0.241 0.877 0.537 0.443 0.877 0.537 Eliminated Eliminated ll GAM2 0.161 oi m at nh 0.141 76.852 z 1.537 z Illusion of control 0.028 0.12 Eliminated Eliminated ILLUS2 0.146 0.125 Eliminated ILLUS3 0.586 0.533 0.519 0.451 ILLUS4 0.532 0.4 Eliminated Eliminated ILLUS5 0.664 0.583 0.695 0.582 ILLUS6 0.562 0.428 Eliminated Eliminated ILLUS7 0.612 0.481 0.631 0.534 ILLUS8 0.637 0.5 Eliminated Eliminated ILLUS9 0.341 0.291 Eliminated Eliminated ILLUS10 0.649 0.609 0.649 0.559 k ILLUS1 gm jm ht vb 0.699 Eliminated om l.c an Lu n va ey t re 73 ILLUS11 0.652 t to ng hi % variance extracted 0.569 32.040 0.467 Eigen value 3.524 2.586 Cronbach's alpha 0.753 0.766 0.55 51.729 ep Herding bias w 0.744 0.445 0.734 0.543 HERD2 0.8 0.626 0.823 0.654 HERD3 0.846 0.541 0.834 0.67 HERD4 0.139 0.509 Eliminated Eliminated HERD5 0.026 0.312 Eliminated Eliminated 0.421 0.475 Eliminated Eliminated 0.195 0.58 Eliminated Eliminated 0.631 43.073 0.661 0.755 0.566 n HERD1 lo ad ju y th yi HERD6 pl HERD7 62.034 va 3.446 2.481 n fu Cronbach's alpha n Eigen value ua % variance extracted al HERD8 0.805 0.795 ll oi m at nh z z k jm ht vb om l.c gm an Lu n va ey t re 74